text
stringlengths 87
880k
| pmid
stringlengths 1
8
| accession_id
stringlengths 9
10
| license
stringclasses 2
values | last_updated
stringlengths 19
19
| retracted
stringclasses 2
values | citation
stringlengths 22
94
| decoded_as
stringclasses 2
values | journal
stringlengths 3
48
| year
int32 1.95k
2.02k
| doi
stringlengths 3
61
| oa_subset
stringclasses 1
value |
|---|---|---|---|---|---|---|---|---|---|---|---|
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Elsevier Inc.
S0002-9343(21)00073-5
10.1016/j.amjmed.2021.01.003
Commentary
The Brazilian Scientific Denialism Through The American Journal of Medicine
Machado Silva Heslley Ph.D. ⁎
University Center of Formiga-MG, Formiga, Minas Gerais, Brazil
⁎ Requests for reprints should be addressed to Heslley Machado Silva, University Center of Formiga-MG, Arnaldo de Senna avenue, 329, Formiga, Minas Gerais, Brazil, 35.680-000.
6 2 2021
4 2021
6 2 2021
134 4 415416
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcMisinformation about the coronavirus disease 2019 (COVID-19) pandemic is widespread on social networks in Brazil. Many doctors, journalists, members of the Ministry of Health, and President Jair Bolsonaro spread conspiracy theories on social media that attribute obscure origins of the disease, unproven risks related to vaccines celebrated and desired around the world,1 , 2 and even promoting prescribing medications without scientific proof,3 all without any concern about the consequences of these disclosures.
To my surprise, the name of this journal and the organization it represents have been involved in this kind of antiscientific movement. From various social network sources, I received a video of a Brazilian journalist who vehemently defends the use of drugs without scientific proof, announcing that science has bowed to his and the Brazilian President's prescription, using the name of this journal to validate itself. He says that an article published early in 2021 endorses this thesis. Meanwhile, I received a series of challenging posts from supporters of the Brazilian president and other deniers, exalting that science has bowed to the Bolsonaro speech from a scientific article.
In a new and unpleasant surprise, I looked for data from this article on the official website of the Brazilian Ministry of Health, and there were more forceful references to this supposed publication.* I transcribed in full the words of the page:
“The renowned The American Journal of Medicine , the official journal of the Alliance for Academic Internal Medicine, brings in its first edition of 2021 a study that proves the effectiveness of early treatment in the evolution of COVID-19. The publication states that, through preventive medicine and early treatment, it is possible to avoid the worsening of the patients’ clinical picture and decrease the number of hospital admissions, as well as the evolution of patients to ICU. The article of this Friday (01) reinforces the importance of early treatment, defended by the Federal Government, as a recommendation in the fight against the coronavirus. The instruction published as a scientific article cites the success in combining antivirals and vitamins, including zinc, azithromycin, and hydroxychloroquine, widely used in the Federal Government's protocol to fight the pandemic.”
In fact, the article is there but it was not published on the day indicated; the edition is from January 20214 but was published in August 2020, when the reality about this supposed early treatment was completely different. The reality today is another, and several treatments pointed out in the article proved ineffective. For example, the use of hydroxychloroquine has no effect in the treatment of COVID-19, and its initial and fierce defender, the French doctor and microbiologist Didier Raoult, admitted so in January 2021. Contrary to the Brazilian government's policy and the example of Bolsonaro's attitudes and speech, the article advocates hygiene and social distancing as important ways to confront the pandemic. It is important to highlight that this form of treatment indicated in the article (eg, vitamins, zinc, azithromycin, hydroxychloroquine) was not recommended by the World Health Organization (WHO), and by the country's main health agencies through medical societies, especially those of infectious diseases. Based on this knowledge, no country determined this protocol as principle for the treatment of COVID-19, except Brazil and the followers of Bolsonaro, including the Ministry of Health and some thousands of doctors. The question remains: Are all the other countries wrong and only Brazil has adopted the right path? The numbers show that it is exactly the opposite. Finally, it is important to highlight the seriousness of this kind of misinformation when it comes from official organizations and reaches a country's population, especially when it has a supposed basis in the academic environment. It is necessary to reflect the potential influence in changing preventive behaviors that put at risk the consecrated measures to confront COVID-19.5
The Ministry of Health's website praises the Brazilian government for always recommending this type of treatment to the detriment of all prevention protocols established worldwide by the WHO and science.6 They also boast that, in the words of the website itself: “Brazil is the world leader in the number of patients recovered from COVID-19, and this factor is the result of the actions of the Ministry of Health in response to the pandemic.” Then the Brazilian Minister of Health exalts these Brazilian numbers with his erratic actions and talks about the pandemic, forgetting that Brazil is second in the number of deaths worldwide.7 Ironically, the minister celebrates his performance and that of his government, while his ineptitude regarding the production, purchase, and distribution of vaccines becomes evident, amid a severe worsening in the number of deaths and cases in the pandemic early in 2021.8
The use of an international scientific journal to celebrate the numerous failures occurs amid possibly the most dramatic moment of the COVID-19 pandemic in Brazil. The numbers are so bad, but the Bolsonaro government says the pandemic is at an end and that it does not care about the disease9 or the vaccine, among other unbelievable statements. In a sad and tragic coincidence, in the same week, terrifying images and news in the media detail the tragic situation in the state of Amazonas. In that state capital, Manaus, people with COVID-19 are dying from a lack of oxygen cylinders, demonstrating total incompetence in the Brazilian government's handling of the health crisis.
Maybe I am being unfair. Perhaps it is not incompetence of the Bolsonaro government; I suspect it may be actual competence. Since the beginning of the pandemic, the president and his entourage of fanatical followers suggest that nobody should do social distancing, that people should continue with normal life, that those who do not face the disease are cowards, that we should seek herd immunity,10 that (believe me) the best vaccine for the disease is the virus itself, and that people will really die, but that is life.
Brazil does not have a reasonable plan for beginning vaccinations,11 or for the acquisition of vaccines and syringes, whereas the rest of the world has already started immunization campaigns. It makes sense within this tragic logic. This citation of the scientific article and its disclosure on the government website and its followers’ disclosure, whether journalists or just ideological followers, promotes a false sense of security and pride in the population regarding the disease.
Indeed, there would be no reason to fear for the deniers officially represented by the Brazilian government; there would already be cure and treatment unknown to the rest of the planet. Putting this plan into practice with no regard for ethics and involving a scientific journal means nothing to those who have contempt for science. So everyone can go out without a mask like the president does, spend time with others like he does, and go back to work and school, anticipating the death of the elderly and the weakest, which is “expected,” and, of course, with dark and autocratic objectives achieved.
Funding: None.
Conflicts of Interest: None.
Authorship: The author is solely responsible for the content of this manuscript.
⁎ Available at: https://www.gov.br/saude/pt-br/assuntos/noticias/the-american-journal-of-medicine-defende-tratamento-preventivo-para-covid. Accessed January 15, 2021.
==== Refs
References
1 Yamey G Schäferhoff M Hatchett R Pate M Zhao F McDade KK Ensuring global access to COVID-19 vaccines Lancet 395 10234 2020 1405 1406 10.1016/S0140-6736(20)30763-7 32243778
2 Lurie N Saville M Hatchett R Halton J Developing COVID-19 vaccines at pandemic speed N Engl J Med 382 21 2020 1969 1973 10.1056/NEJMp2005630 32227757
3 Silva HM Medicines and illusions in the fight against COVID-19 in Brazil Ethics Med Public Heal 16 2021 100622 10.1016/j.jemep.2020.100622
4 McCullough PA Kelly RJ Ruocco G Pathophysiological basis and rationale for early outpatient treatment of SARS-CoV-2 (COVID-19) infection Am J Med 134 1 2021 16 22 10.1016/j.amjmed.2020.07.003 32771461
5 Xiao Y Torok ME Taking the right measures to control COVID-19 Lancet Infect Dis 20 5 2020 523 524 10.1016/S1473-3099(20)30152-3 32145766
6 Bedford J Enria D Giesecke J COVID-19: towards controlling of a pandemic Lancet 395 10229 2020 1015 1018 32197103
7 Carvalho TA Boschiero MN Marson FAL COVID-19 in Brazil: 150,000 deaths and the Brazilian underreporting Diagn Microbiol Infect Dis 99 3 2020 115258 10.1016/j.diagmicrobio.2020.115258
8 Cimerman S Chebabo A da Cunha CA Rodríguez-Morales AJ Deep impact of COVID-19 in the healthcare of Latin America: the case of Brazil Braz J Infect Dis 24 2 2020 93 95 10.1016/j.bjid.2020.04.005 32335078
9 COVID-19 in Brazil: “so what?” Lancet 395 10235 2020 1461 10.1016/S0140-6736(20)31095-3 32386576
10 Fontanet A Cauchemez S COVID-19 herd immunity: where are we? Nat Rev Immunol 20 10 2020 583 584 10.1038/s41577-020-00451-5 32908300
11 Ernesto Londoño; Manuela Andreoni; Letícia Casado. Chaotic vaccine plan ‘playing with lives’ in brazil. New York Times. Available at: https://www.nytimes.com/2020/12/14/world/americas/brazil-coronavirus-vaccine.html. Accessed January 15, 2021.
| 33561430 | PMC9745804 | NO-CC CODE | 2022-12-15 00:03:08 | no | Am J Med. 2021 Apr 6; 134(4):415-416 | utf-8 | Am J Med | 2,021 | 10.1016/j.amjmed.2021.01.003 | oa_other |
==== Front
Am Heart J
Am Heart J
American Heart Journal
0002-8703
1097-6744
Mosby
S0002-8703(21)00081-8
10.1016/S0002-8703(21)00081-8
Article
Table of Contents
23 4 2021
5 2021
23 4 2021
235 iii
2019
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745805 | NO-CC CODE | 2022-12-15 00:03:08 | no | Am Heart J. 2021 May 23; 235:i-ii | utf-8 | Am Heart J | 2,021 | 10.1016/S0002-8703(21)00081-8 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier Inc.
S0002-9378(21)00689-X
10.1016/j.ajog.2021.06.055
Letter to the Editor
Reply to “COVID-19 infection just before or during early pregnancy and the possible risk of placenta accreta spectrum or preeclampsia”
Patberg Elizabeth T. MD
Vintzileos Anthony M. MD
Department of Obstetrics and Gynecology, NYU Langone Hospital—Long Island, NYU Long Island School of Medicine, 259 First St, Mineola, NY 11501
17 6 2021
10 2021
17 6 2021
225 4 466466
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcWe would like to thank Dr Al-Lami for the interest in our study and the comments regarding our publication “Coronavirus disease 2019 infection and placental histopathology in women delivering at term.” Dr Al-Lami hypothesized that COVID-19 infection periconceptionally may increase the risk of placenta accreta spectrum (PAS) and/or preeclampsia. The biological mechanism by which this may occur is described by Dr Al-Lami who advocates for further research on the potential connections between them. We are aware of previous studies suggesting that women with COVID-19 infection in the second and third trimesters of pregnancy may be at increased risk for the development of preeclampsia.1, 2, 3 However, reliable data on periconception COVID-19 infection and the subsequent development of preeclampsia and/or PAS are very scarce, if any, at this time.
It should be emphasized that the COVID-19 cases in our study were identified by polymerase chain reaction screening in term patients (>37 weeks gestation) at the time of admission to labor and delivery. Therefore, it is highly unlikely that they would have contracted the virus before pregnancy or even in the first trimester. Because our study was conducted on women delivering at term and we excluded women with preeclampsia or PAS from our control group, we are unable to comment on any potential association. We agree with the author that more research is needed to determine whether there is an association between periconception COVID-19 infection and PAS and/or preeclampsia.
The authors report no conflict of interest.
==== Refs
References
1 Mendoza M. Garcia-Ruiz I. Maiz N. Pre-eclampsia-like syndrome induced by severe COVID-19: a prospective observational study BJOG 127 2020 1374 1380 32479682
2 Antoun L. Taweel N.E. Ahmed I. Patni S. Honest H. Maternal COVID-19 infection, clinical characteristics, pregnancy, and neonatal outcome: a prospective cohort study Eur J Obstet Gynecol Reprod Biol 252 2020 559 562 32732059
3 Di Mascio D. Khalil A. Saccone G. Outcome of coronavirus spectrum infections (SARS, MERS, COVID-19) during pregnancy: a systematic review and meta-analysis Am J Obstet Gynecol MFM 2 2020 100107 32292902
| 34146531 | PMC9745806 | NO-CC CODE | 2022-12-15 00:03:08 | no | Am J Obstet Gynecol. 2021 Oct 17; 225(4):466 | utf-8 | Am J Obstet Gynecol | 2,021 | 10.1016/j.ajog.2021.06.055 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier Inc.
S0002-9378(21)00779-1
10.1016/j.ajog.2021.06.094
Letter to the Editor
The incorporation of telehealth in high-risk pregnancy follow-up needs tailored optimized care scheduled in a strict care protocol
Carbillon Lionel MD, PhD
Benbara Amelie MD
Fermaut Marion MD
Department of Obstetrics and Gynecology, Hôpital Jean Verdier, Assistance Publique—Hôpitaux de Paris, Avenue du 14 Juillet, Bondy 93140, France
Carbillon Lionel MD, PhD
Université Sorbonne Paris Nord, Bobigny, France
1 7 2021
11 2021
1 7 2021
225 5 586586
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcTo the Editors:
Peahl et al1 recently reported their evaluation of patient and provider experiences with a COVID-19 prenatal care model incorporating telehealth and virtual visits. These authors found “perceived improved access to care” through decreased barriers and “perceived high quality of virtual visits” for low-risk patients. However, these authors also reported that across the pre- and postimplementation periods, the average total visit volume (including both in-person and virtual prenatal visit utilization) de facto fell (–16.1%), which is not in accordance with the perception of patients and providers, and can be of concern. Because Peahl et al1 did not define strict inclusion criteria for their prenatal care model and did not address health outcomes, we believe that the conclusion of their study may be misleading.
Indeed, from a meta-analysis of studies pooling data from 198,993 pregnancies before and 168,295 during the pandemic, respectively, Chmielewska et al2 recently evidenced a significant increase (1.37 [1.22–1.53]) in maternal death that was mainly driven by a reduced access to care and not by direct effect of COVID-19 in pregnant women.
In actuality, Peahl et al1 mainly based their prenatal care model on the low-risk schedule “with additional visits and services as appropriate.”
However, only a strict monitoring protocol, depending on the specific risk involved, can meet the needs for high-risk pregnancies, in a rigorous approach specifically tailored for each condition placing patients at higher risk of adverse maternal or neonatal outcomes.3 For the purpose of maintaining close follow-up for high-risk pregnant women during the first wave of the pandemic in New York, Aziz et al3 organized prenatal care in a telehealth framework, allowing to eliminate “approximately one-half of in-person visits for low-risk patients,” but they detailed recommendations scheduled for high-risk pregnancies, pathology by pathology. Indeed, the separation between high and low risk remains challenging: Butler Tobah et al4 randomized low-risk women to an “OB Nest” protocol or usual care (150 in each arm) using a minimization algorithm excluding women with various high-risk conditions or if “obstetrician judgment determined that the pregnancy was at high risk for complications.” Study team clinicians were aware of the assigned arms and used study exclusion criteria if a high risk appeared later. In this strict context, these authors found that maternal and fetal clinical outcomes were similar between groups.
However, Peahl et al1 did not define their inclusion criteria in such a strict manner.
The authors report no conflict of interest.
==== Refs
References
1 Peahl A.F. Powell A. Berlin H. Patient and provider perspectives of a new prenatal care model introduced in response to the coronavirus disease 2019 pandemic Am J Obstet Gynecol 224 2021 384.e1 384.e11
2 Chmielewska B. Barratt I. Townsend R. Effects of the COVID-19 pandemic on maternal and perinatal outcomes: a systematic review and meta-analysis Lancet Glob Health 9 2021 e759 e772 33811827
3 Aziz A. Zork N. Aubey J.J. Telehealth for high-risk pregnancies in the setting of the COVID-19 pandemic Am J Perinatol 37 2020 800 808 32396948
4 Butler Tobah Y.S. LeBlanc A. Branda M.E. Randomized comparison of a reduced-visit prenatal care model enhanced with remote monitoring Am J Obstet Gynecol 221 2019 638.e1 638.e8
| 34217724 | PMC9745807 | NO-CC CODE | 2022-12-15 00:04:02 | no | Am J Obstet Gynecol. 2021 Nov 1; 225(5):586 | utf-8 | Am J Obstet Gynecol | 2,021 | 10.1016/j.ajog.2021.06.094 | oa_other |
==== Front
Am J Cardiol
Am J Cardiol
The American Journal of Cardiology
0002-9149
1879-1913
Elsevier Inc.
S0002-9149(21)00166-1
10.1016/j.amjcard.2021.02.017
Article
COVID-19 and Gender Disparities in Pediatric Cardiologists with Dependent Care Responsibilities
Ferns Sunita J MBBS, MD, MRCPCH (UK), FHRS a⁎
Gautam Shiva PhD b
Hudak Mark L. MD a
a Department of Pediatrics, University of Florida School of Medicine, Jacksonville, FL
b Department of Medicine, University of Florida School of Medicine, Jacksonville, FL
⁎ Corresponding author: Tel.: 1-904-633-4110; fax: 1-904-633-4111.
25 2 2021
15 5 2021
25 2 2021
147 137142
3 1 2021
7 2 2021
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The COVID-19 pandemic disproportionately affects females in the home and workplace. This study aimed to acquire information regarding the gender-specific effects of the COVID-19 lockdown on aspects of professional and personal lives of a subset of pediatric cardiologists. We sent an online multiple-choice survey to a listserv of Pediatric Cardiologists. Data collected included demographics, dependent care details, work hours, leave from work, salary cut, childcare hours before and after the COVID-19 peak lockdown/stay at home mandate and partner involvement. Two hundred forty-two pediatric cardiologists with dependent care responsibilities responded (response rate of 20.2%). A significantly higher proportion of females reported a salary cut (29.1% of females vs 17.6% of males, p = 0.04) and scaled back or discontinued work (14% vs 5.3%; p = 0.03). Prior to the COVID-19 lockdown phase, females provided more hours of dependent care. Females also reported a significantly greater increase in childcare hours overall per week (45 hours post/30 hours pre vs 30 hours post/20 hours pre for men; p < 0.001). Male cardiologists were much more likely to have partners who reduced work hours (67% vs 28%; p < 0.001) and reported that their partners took a salary cut compared with partners of female cardiologists (51% vs 22%; p < 0.001). In conclusion, gender disparity in caregiver responsibilities existed among highly skilled pediatric cardiologists even before the COVID-19 pandemic. The pandemic has disproportionately affected female pediatric cardiologists with respect to dependent care responsibilities, time at work, and financial compensation.
==== Body
pmcThe COVID-19 pandemic continues to affect livelihoods around the world and its economic and social fallouts are disproportionately affecting women in the workplace and at home. Globally, the rate of job loss among women during this pandemic has exceeded 180% of the rate for men.1 Although women comprise 39% of the global work force, they have incurred 54% of all job losses.1 One reason for this disparity is that the virus has significantly increased the family burden of unpaid child and adult care, for which women carry a greater responsibility.1 Recent articles by Reza et al. and Brubaker highlight the need to retain and support female physicians during this pandemic.2 , 3 Although this study was limited to pediatric cardiologists, findings are likely to be similar among other highly skilled medical professionals.
Methods
We obtained approval from the University of Florida Institutional Review Board and conducted an online REDCap-based survey of a subset of current North American Pediatric Cardiologists in practice and training (subscribers to the ‘PediHeartNet’ listserv: PediHeartNet-http://pediheart.net is a private, unmoderated discussion group of predominantly North American Pediatric Cardiologists) who had child or adult dependent care responsibilities. Our 33-item survey (Supplemental File) collected demographic information including age, gender, work experience and details about dependent care. Questions addressed changes in work hours, work leave, and compensation that occurred as a consequence of COVID-19. The survey captured the average amount of time spent by the respondent and by the respondent's partner in providing care or supervision to child and adult dependents before and after many states issued COVID-19 stay-at-home mandates. Finally, the survey provided an opportunity for respondents to relate their most challenging experiences with changes in dependent care consequent to the COVID-19 pandemic. After obtaining permission from the moderators of the listserv, we emailed the survey in July 2020 to approximately 1,200 pediatric cardiologists and requested anonymous responses from pediatric cardiologist physicians in training or in practice whose household included child and/or adult dependents. We emailed a reminder 3, 7, and 14 days after the initial solicitation and closed data collection on day 16. As an incentive, we offered the opportunity to participate in a small online gift card drawing.
We performed descriptive and inferential statistical analyses using SAS v 9.4 and SPSS v 24. Categorical variables were expressed as frequencies and percentages and continuous variables were expressed as mean or median with standard deviation or ranges. The impact of COVID-19 was defined in terms of both categorical and continuous outcome variables. We analyzed categorical continuous variables using logistic regression with gender as the independent variable. Covariates that achieved a 0.15 level of significance were included in the model. We compared continuous variables between genders using t-test or Wilcoxon rank sum test. Significant covariates were included in a linear regression model. If the variable under consideration was not normally distributed then a log linear transformation was performed. p value <0.05 was considered significant.
The principal investigator and 2 other data analysts from the University of Florida with experience in qualitative research methodology (MAB and MF) reviewed all responses to the open-ended question that sought a narrative response to the dependent care challenges posed by COVID-19. Over a 3-week period, each member of the team sorted responses into 10 main topical areas and then all members organized the responses into 5 domains for analysis.
Results
A total of 256 respondents attempted the survey of which 242 respondents were actually eligible (response rate of 20.2%). Of these, 127 (52%) of respondents were female compared with 34% of all board certified pediatric cardiologists in the United States who are female.4 Demographics of respondents are outlined in Table 1 . The largest group of respondents were aged 30 to 39 and had been in practice for less than 5 years. Although there was no gender difference in the median age of surveyed trainees, female cardiologists in practice had a lower median age than male cardiologists. Interventional cardiac subspecialties accounted for 11.6% of overall respondents. Most (86.7%) cardiologists had childcare responsibilities and 12.8% cared for adults. Most single parents or adult caregivers tended to be female (p < 0.0001).Table 1. Study population baseline characteristics
Table 1: All Male Female p Value
Variable
Age Group (years)
20 - 29 5 4 (3.5%) 1 (0.8%) 0.22†
30 - 39 109 45 (39.8%) 64 (50.4%)
40 - 49 85 44 (38.9%) 41 (32.3%)
50 - 59 27 15 (13.3%) 12 (9.4%)
> 60 14 5 (4.4%) 9 (7.1%)
Gender 240 113 (47.1%) 127 (52.9%)
Sexual Orientation
Straight 222 102 (90.3%) 120 (94.5%) 0.10†
Lesbian/Gay 9 6 (5.3%) 3 (2.4%)
Bisexual 6 5 (4.4%) 1 (0.8%)
Race
Asian 67 31 (27.7%) 36 (28.3%) 0.23*
Black/African American 19 11 (9.8%) 8 (6.3%)
White 132 59 (52.7%) 73 (57.5%)
Other 15 9 (8%) 6 (4.8%)
Prefer not to answer 6 2 (1.8%) 4 (3.1%)
Ethnicity
Hispanic/Latino 25 14 (12.7%) 11 (8.8%) 0.61†
Non-Hispanic/Latino 201 92 (83.6%) 109 (87.2%)
Prefer not to answer 9 4 (3.6%) 5 (4%)
Employment:
Fellowship 32 13 (11.5%) 19 (15.1%) 0.42*
Not in fellowship 207 100 (88.5%) 107 (84.9%)
Years of fellowship
1 5 4 (30.8%) 1 (6.7%) 0.17†
2 5 3 (23.1%) 2 (13.3%)
3 9 2 (15.4%) 7 (46.7%)
4 7 4 (30.8%) 3 (20%)
5 2 0 (0%) 2 (13.3%)
Years of practice
≤ 5 86 35 (31.3%) 51 (40.5%) 0.02§
6 - 10 56 23 (20.5%) 33 (26.2%)
11 - 15 47 28 (25%) 19 (15.1%)
16 - 20 28 19 (17%) 9 (7.1%)
≥ 21 21 7 (6.3%) 14 (11.1%)
Sub-specialty
Interventional 28 18 (15.9%) 10 (7.9%)
Non interventional 212 95 (84.1%) 117 (92.1%) 0.05*
Time worked
Full Time 224 109 (96.5%) 115 (91.3%) 0.09*
Percentage Part Time 15 4 (3.5%) 11 (8.7%)
60 - 69 6 1 5
70 - 79 6 2 4
80 - 89 3 1 2
Dependent Care
Children 197 97 100
<7 years Mean (Range) 1.53 (1-4) 1.38 (1-3) 1.67 (1-4) 0.049‡
7-17 years: Mean (Range) 1.63 (1-6) 1.69 (1-6) 1.57 (1-3) 0.87‡
Adult 18 6 12 0.16*
Both Children and Adult 13 6 7 0.78*
Single Parent or Adult Caregiver 20 1 19 <0.0001*
⁎ Chi Square test.
† Fisher's Exact test.
‡ Wilcoxon Rank Sum test.
§ Trend test.
The lockdown phase significantly reduced work hours, with 56% of male and 53% of female cardiologists reporting some decrease in work hours. Most cardiologists reported less than a 20% decrease in work hours. However, despite a similar cut in work hours, a significantly higher proportion of females reported a salary cut and more females reported a salary cut greater than 20% (47% of females vs 25% of males, p = 0.2). More female cardiologists reduced or discontinued work compared with male cardiologists (Table 2 ).Table 2. Impact of lockdown on work distribution by gender
Table 2: All Male Female p Value
Work hours decreased
Yes 130 63 (55.8%) 67 (52.8%) 0.64*
No 110 50 (44.2%) 60 (47.2%)
Percentage decrease in work hours
< 20 104 54 (85.7%) 50 (74.6%) 0.37†
21 - 40 16 5 (7.9%) 11 (16.4%)
41 - 60 6 3 (4.7%) 3 (4.5%)
> 60 4 1(1.6%) 3 (4.5%)
Experienced Salary Cut
Yes 57 20 (17.7%) 37 (29.1%) 0.04*
No 183 93 (82.3%) 90 (70.9%)
% Salary Cut: Median (Range) 12.5 (5-50) 15 (5-50)
Took Leave from Work or Quit Work
Yes 24 6 (5.3%) 18 (14.2%) 0.02*
No 216 107 (94.7%) 109 (85.8%)
Are Genders Affected Disproportionately?
Yes 150 39 (35.5%) 111 (89.5%) <0.0001*
No 70 65 (59.1%) 5 (4.0%)
Don't know 14 6 (5.5%) 8 (6.5%)
Has Caring for Child Impacted Work Life Balance?
Yes 174 79 (79%) 95 (89.6%) 0.035*
No 32 21 (21%) 11 (10.4%)
Has Caring for Adult Impacted Work Life Balance?
Yes 26 9 (81.8%) 17 (89.5%) 0.61†
No 4 2 (18.2%) 2 (10.5%)
⁎ Chi Square test.
† Fisher's Exact test.
COVID-19 had a pronounced gender-specific effect on spousal work hours during the peak lockdown. Male cardiologists were more likely to have partners who cut back on work hours and report that their partners took a salary cut compared with partners of female cardiologists (Table 3 ). Prior to the COVID-19 lockdown, females shouldered a greater burden of dependent care per week than males (30 hours vs 20 hours; p = 0.02). Females also reported spending significantly more time during the regular workweek (M-F, 9 A.M. to 6 P.M.) during which they provided dependent care. During the lockdown, the total number of childcare hours per week and during regular weekday work hours increased significantly for both females and males compared with their pre-COVID-19 baselines. However, when we compared the absolute increase in hours for males and females during lockdown to baseline, females reported a significantly greater increase in childcare hours overall per week (45 hours post/30 hours pre vs 30 hours post/20 hours pre; p < 0.001). Females also were responsible for a significantly greater number of hours of dependent care between 9 A.M. to 6 P.M. (M-F) compared with males (20 hours post/3 hours pre vs 4 hours post/1 hours pre; p < 0.001; Table 4 ).Table 3. Effect of COVID-19 on partner
Table 3:Partner work cut All Male Female p Value
Yes 99 70 (66.7%) 27 (28.4%) <0.0001*
No 103 35 (33.3%) 68 (71.6%)
% Decrease in Work Hours by Partner
< 20 53 38 (54.3%) 15 (55.6%) 0.86†
20 - 40 33 25 (35.7%) 8 (29.6%)
41 - 60 8 5 (7.1%) 3 (11.1%)
> 60 3 2 (2.9%) 1 (3.7%)
Partner Salary Cut
Yes 75 54 (51.4%) 21 (22.3%) <0.0001*
No 124 51 (48.6%) 73 (77.7%)
% Partner Salary Cut: Median (Range) 20 (10-100) 20 (10-100)
⁎ Chi Square test.
† Fisher exact test.
Table 4. Distribution of dependent care responsibilities by gender
Table 4: Male Female
Median (Range) provided for all data Pre COVID-19 During COVID-19 lockdown p Value Pre COVID-19 During COVID-19 Lockdown p Value
Childcare Responsibility – All Hours 20 (0-108) 30 (0-120) <0.0001 30 (0-118) 45 (0-125) <0.0001
Childcare Responsibility – 9 AM – 6 PM 1 (0-18) 9 (0- 43) <0.0001 3 (0-20) 20 (1-45) <0.0001
Partner Childcare Responsibility – All Hours 35 (0-118) 60 (0-128) 20 (0-118) 30 (0-120)
Partner Childcare Responsibility – 9 AM – 6 PM 5 (0-40) 30 (0-45) 2 (0-380) 10 (0-40)
Adult Care Responsibility – All Hours 13.5 (8-35) 20 (15-50) 0.008 20 (1-30) 40 (1-70) 0.002
Adult Care Responsibility – 9 AM – 6 PM 1 (0-5) 5 (0-8) 0.031 1 (0-5) 20 (1-30) 0.008
Values are denoted as median (range).
Wilcoxon sign rank test.
Female cardiologists reported a greater number of hours per week caring for adult dependents than did male cardiologists. During the lockdown phase, the overall number of adult care hours per week between (M-F, 9 A.M. to 6 P.M.) increased significantly for both males and females over their baseline preCOVID-19 hours. However, during the lockdown, females reported a significantly greater increase in adult care hours overall per week (40 hours post/20 hours pre vs 20 hours post/13.5 hours pre; p = 0.04). Females were also responsible for a significantly greater number of absolute hours of adult care between (M-F, 9 A.M. to 6 P.M.) compared with males (20 hours post/1 hours pre vs 5 hours post/1 hours pre; p = 0.01; Table 4). Figure 1 denotes gender distribution for significant binary outcomes. A univariate logistic regression was performed with gender as the independent variable. Covariates selected for the multivariate analysis included years in practice, subspecialty selection, sexual orientation and employment status (Table 5 ). Even when correcting for the imbalance in male and female respondents in the survey (1:1) and the distribution of board certified male and female pediatric cardiologists in practice (2:1) by attaching a weight of 2 for each male and 1 for each female, there were significant discrepancies between males and females with respect to dependent care responsibilities, loss of work and financial compensation. With respect to gender-specific perspectives, females more correctly perceived real disparities in the increased burden of dependent care (89.5% of females vs 35.5% of males; p < 0.0001; (Table 2). Both males and females agreed that caring for a child or adult dependent affected their work-life balance (90% of females vs 79% of males; p = 0.035).Figure 1. Gender distribution for significant binary outcomes.
Figure 1
Table 5. Univariate and multivariate logistic regression
Table 5:Univariate Multivariate
Outcome Variable OR (95 % CI) P OR (95% CI) P
Salary Cut 1.91 (1.03-3.54) 0.04 2.04 (1.09-1.58) 0.03
Leave /Quit work 2.89 (1.10-7.56) 0.03 3.15 (1.19-8.36) 0.02
Work life balance 2.30 (1.04-5.05) 0.04 - -
Perception on gender affection 37.0 (13.89-98.59) <0.0001 44.0(15.65- 123.57) <0.0001
Partner Salary Cut 0.27 (0.15-0.50) <0.0001 - -
Partner Work Cut 0.20 (0.11-0.36) <0.001 - -
Responses to the open-ended question, "What is the most challenging aspect of dependent care you have faced during the COVID-19 lockdown?” were grouped into five domains: (1) Mental Health/Social isolation/Family tensions (2)Work-Home Balance (3) Caretaker (4) Family safety (5) Financial. A few responses crossed over domains, for instance, (1) “We have less money coming in, children who need supervision for school. Trying to work from home, with kids and my husband looking for a job means I am not only trying to do my job but also be a teacher, a cook, a housecleaner with a salary cut and a husband with no job.”(2)“No work life balance, no understanding from colleagues, super stressful, constant guilt and doubt (having to put children back in daycare, no family available etc.), worry about what to do if COVID-19 case at daycare or with online schooling”.
The researchers who designed the questionnaire and analyzed the qualitative responses acknowledged that their personal and work backgrounds might have influenced their qualitative interpretations. SF is a female electrophysiologist with 2 children under 5 whose husband works from home and assumes more than 50% of childcare responsibilities. She kept her 2 young children at home during the lockdown phase and worked from home 1 day a week to help with childcare. MAB is a female research coordinator, wife of an essential health care professional, and mother of 3 young adults who live separately. She and her children worked from home during the various mandated periods in the different states they live in. Her father passed away recently from complications related to hospital-acquired COVID-19. MF is a research administrator and single dad of a 6 year old daughter who along with support from family and friends were able to provide childcare for his daughter.
Discussion
Gender balance in the clinical workplace has been an important topic over the last decade. Pediatric cardiologists, like other healthcare providers, must balance work and family while facing choices that influence their selection of specialty, their choice of academic pursuit, and their employment status. Manifestations of solutions to achieving work-life balance differ between females and males, with female physicians more likely to make accommodations to restrict work time and duties to fulfill responsibilities at home. Family responsibilities may include providing care or supervision of children, aging parents, or both. The COVID-19 pandemic had the potential to exaggerate existing gender differences in the workplace because dually employed couples were no longer able to rely on another individual to care for or supervise their dependents.
Our study results demonstrated that the COVID-19 pandemic did disproportionately increase an already disparate burden of dependent care for both children and adults among female pediatric cardiologists. Our findings echo those of other studies. One study has shown that together, dual working parents devote six hours a day to dependent care with females absorbing two-thirds of this responsibility.5 Similarly, females are much more likely than males to be involved in the care of adult dependents.6 In other fields, both genders have assumed more home responsibilities during the COVID-19 lockdown, but previous disparities persist. Studies of nonmedical professionals have shown that even when males were initially unemployed and their female partners were the ones working full time, males continued to spend less time on household chores.7
Even though the results from this study clearly demonstrate that female cardiologists disproportionately assumed more caregiver hours during the COVID-19 lockdown, most males did not perceive this imbalance. This is consistent with the findings of a British Broadcasting Company survey that showed that perceptions of relative childcare responsibilities differ widely among males and females. Males tended to overestimate their share in domestic chores while females correctly identified their own greater contributions.8
Although the numbers were small, female cardiologists left the workplace at a rate three times greater than their male counterparts. On a global scale, females have been 1.8 times more likely to suffer job loss during the COVID-19 pandemic.1 Women comprise 39% of the global work force but have incurred 54% of all job losses.1 In the United States, for people with children under 12, unemployment increased by 11 absolute percentage points in women compared with 7.3 percentage points in men between February and April 2020.9
During the COVID-19 lockdown, both male and female pediatric cardiologists experienced similar loss of work hours, but females were twice as likely to sustain a decrease in salary. Furthermore, females in medicine have earned 26% less than their male colleagues10 so that the disproportionately larger salary reduction for women has exacerbated pre-existing disparities in compensation. This phenomenon has also played out on the global stage.11
In our study, spouses of male physicians were more likely to take leave from work or sustain a reduction in salary compared with spouses of female physicians. The spousal work differences by gender are in fact much more exaggerated than the reduced work and salary disparity between male and female cardiologists (Figure 1). Other studies during the pre-COVID-19 phase have shown that at baseline, spouses of female physicians on average work more hours and earn significantly more than spouses of male physicians. Therefore, for female physicians, a partner salary cut would impact the overall household income much more. These findings have been thought to place greater pressure on female physicians to trade professional for household responsibilities, highlighting yet another example of gender disparity even among high-earning professionals.12
This voluntary survey was launched using the PediHeartNet listserv and therefore targeted a select group of pediatric cardiologists. As with every survey, those who responded may have been particularly affected by the circumstances under inquiry or have had a particular interest in responding to questions about gender disparities. Our response rate of 20.2% exceeds the average 11% response rate for topical surveys but still does not capture the perspective of all pediatric cardiologists with dependent care responsibilities. Finally, not all respondents answered every question because participants were allowed to proceed to the next question even if they chose not to answer the prior one. The authors acknowledge that there were likely multiple factors influencing salary that were not elicited and controlled for in this study such as private versus academic practice, fixed vs incentive/volume-based salary models or discrepancies with some respondents being in less affected areas or receiving CARES Act funding.
Our study highlights that gender disparity in caregiver duties existed among pediatric cardiologists before the COVID-19 pandemic. The COVID-19 pandemic has disproportionately affected female pediatric cardiologists with respect to dependent care responsibilities, time at work, and financial compensation. Male pediatric cardiologists have not appreciated this gender disparity. The accentuation of gender disparity by the COVID-19 pandemic may continue to have an adverse effect on a woman's professional career long after its more acute effects resolve.
Author Contribution
Dr Ferns conceptualized and designed the study and data collection instruments, collected data, carried out the initial analyses, drafted the initial manuscript, and reviewed and revised the manuscript. Serves as guarantor. Dr Gautam co-designed the data collection instruments, carried out the initial analyses and reviewed and revised the manuscript. Dr Hudak contributed to conceptualization of the study, critically reviewed and revised the manuscript for important intellectual content. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work. The guarantor accepts full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish. The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted.
Acknowledgments
The authors acknowledge the efforts of Maria Abbey Bautista and Mark Fafard from the UF Research Affairs Office for assistance with REDCap survey design and qualitative analysis.
Declaration of interests
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:
Funding: None.
==== Refs
References
1 Madgavkar A White O Krishnan M Mahajan D Azcue X. COVID-19 and Gender Equality: Countering the Regressive Effects. Mc Kinsey Global Institute 2020 https://www.mckinsey.com/featured-insights/future-of-work/covid-19-and-gender-equality-countering-the-regressive-effects. Accessed Oct 20, 2020
2 Reza N DeFilippis EM Michos ED. The cascading effects of COVID-19 on women in cardiology Circulation 2020 10.1161/CIRCULATIONAHA.120.049792 Epub ahead of print. PMID: 33016784
3 Brubaker L. Women physicians and the COVID-19 pandemic JAMA 324 2020 835 836 32735329
4 American Board of Pediatrics Workforce Database 2018 https://www.abp.org/content/workforce Accessed October 20, 2020
5 Miller CC Nearly half of men say they do most of the home schooling. 3 percent of women agree The New York Times 2020 https://www.nytimes.com/2020/05/06/upshot/pandemic-chores-homeschooling-gender.html. Accessed Oct 2020
6 Sharma N Chakrabarti S Grover S. Gender differences in caregiving among family - caregivers of people with mental illnesses World J Psychiatry 6 2016 7 17 27014594
7 Ortiz-Ospina E Tzvetkova S. Women's Employment 2018 OurWorldInData.org https://ourworldindata.org/female-labor-supply Accessed Oct 20, 2020
8 Savage M. How COVID-19 is Changing Women's Lives June 2020 BBC Worklife https://www.bbc.com/worklife/article/20200630-how-covid-19-is-changing-womens-lives. Accessed Oct 20, 2020
9 Gupta S. How COVID-19 worsened gender inequality in the U.S. workforce Science News 2020 https://www.sciencenews.org/article/covid19-worsened-gender-inequality-us-workforce. Accessed Oct 20, 2020
10 Doximity. “Second Annual Physician Compensation Report: March 2018.” https://www.doximity.com/careers/compensation_report?_csrf_ attempted=yes. Accessed Oct 20, 2020.
11 Boniol M McIsaac M Xu L Wuliji T Diallo K Campbell J. Gender Equity in the Health Workforce: Analysis of 104 Countries Health Workforce Working Paper 1 2019 https://www.who.int/hrh/resources/gender_equity-health_workforce_analysis/en/. Accessed Oct 20, 2020
12 Ly DP Seabury SA Jena AB Characteristics of U.S. physician marriages, 2000-2015: an analysis of data from a U.S. census survey Ann Intern Med 168 2018 375 376 29159374
| 33640368 | PMC9745814 | NO-CC CODE | 2022-12-15 00:03:08 | no | Am J Cardiol. 2021 May 15; 147:137-142 | utf-8 | Am J Cardiol | 2,021 | 10.1016/j.amjcard.2021.02.017 | oa_other |
==== Front
Am J Surg
Am J Surg
American Journal of Surgery
0002-9610
1879-1883
Excerpta Medica
S0002-9610(21)00429-3
10.1016/S0002-9610(21)00429-3
Article
Table of Contents (4 pgs)
13 8 2021
9 2021
13 8 2021
222 3 A3A5
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745860 | NO-CC CODE | 2022-12-15 00:03:11 | no | Am J Surg. 2021 Sep 13; 222(3):A3-A5 | utf-8 | Am J Surg | 2,021 | 10.1016/S0002-9610(21)00429-3 | oa_other |
==== Front
Am J Surg
Am J Surg
American Journal of Surgery
0002-9610
1879-1883
Elsevier Inc.
S0002-9610(21)00153-7
10.1016/j.amjsurg.2021.03.018
Invited Commentary
Invited commentary on “the lasting footprint of COVID-19 on surgical education: A resident and attending perspective on the global pandemic”
DeSantis Anthony J.
Rogers Michael P.
Kuo Paul C. ∗
University of South Florida, Morsani College of Medicine, Department of Surgery, 2 Tampa General Circle, Tampa, FL, 33606, USA
∗ Corresponding author. Department of Surgery, University of South Florida, 2 Tampa General Circle, Tampa, FL, 33606, USA.
17 3 2021
9 2021
17 3 2021
222 3 471472
5 3 2021
7 3 2021
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcThere is an often-cited quote (commonly misattributed to Charles Darwin) that reads as follows:“It is not the strongest of species that survives, nor the most intelligent. It is the one that is the most adaptable to change”
While Darwin may have never uttered these exact words himself, the topic of adaptability in the face of change clearly permeates his writings.1 In the years following Darwin’s explorations, many of his ideas and hypotheses have been co-opted and expanded to explain the functioning of societies, cultures, and any number of institutions far beyond their original subjects of study. In keeping with this extension of ideas, we posit that in times of significant upheaval, adaptability to change will be just as important to the training of surgical residents as it was to the Galapagos Finch.
In this issue of AJS, Imai et al. discuss the ways that COVID-19 has impacted the training of surgical residents, and offers perspectives from both faculty and trainees on the severity of these changes and strategies for future adaptation.2 We applaud the authors for providing this timely perspective on how their program has been affected by the recent global pandemic. While a number of recent articles in literature describe the unique challenges faced by surgical residencies of all subtypes, we have yet to determine the best way forward as a cohesive society of surgeons, and unique perspectives from across the surgical discipline will be necessary to develop mechanisms forward that are both comprehensive yet flexible.3, 4, 5, 6
Of particular interest to us was the authors comments regarding the transition to virtual curriculum. Though this transition was precipitated by necessity in the face of COVID-19, we agree that in some regards it may actually result in an improved educational experience for surgical trainees. Our residency program is spread across four different hospital locations in our community, with all formal education and conferences being held at our largest hospital affiliate. Prior to the pandemic, this meant that we had some residents starting their Mondays by travelling out to training sites, rounding on patients and placing initial orders, and returning to their cars to beat the traffic heading into downtown to make it to our weekly conference on time. The inefficiency of this is apparent. In the 80-h workweek era, it is of the highest importance to maximize the training time of residents, and every minute spent in a car on the interstate is not spent in the operating theater or at the patient’s bedside. While it may have taken a global pandemic to push many industries and organizations fully into the digital era, now that we are here we should not forsake some of the advances in efficiency that we have made once the threat of novel coronavirus subsides.
Now, to reverse course on this topic (and hopefully avoid a slew of animated email responses), our above arguments regarding the implementation of virtual curriculum do not mean that we fail to recognize the value and importance of face-to-face interaction. Clearly there is benefit to the surgical resident standing at the podium under the bright lights of the auditorium to present a case for morbidity and mortality conference, or to review recent literature during a grand rounds presentation to the department. These experiences not only serve to prepare trainees for completion of their oral boards, but as a way to interact with their peers and professors on a level not possible by video monitor. We agree that these experiences are important to a surgeon’s education. We simply state that in the age of duty hour restrictions, it is imperative that we maximize the 80 hours a week of in-hospital education that trainees are allowed, and if the use of virtual methods in certain settings can provide residents more time with patients, all the better.
The challenges of the past year have changed both the way we care for patients as well as the way we train the next generation of surgeons. Circumstances have dictated that we adapt or die, and in the grand scheme of things, this is nothing new. However, not all of our adaptations represent negative change, and we would be remiss if we did not identify and expand upon the areas where adaptation has changed us for the better. As vaccinations rise and case numbers fall, the next challenge in surgical training will be determining which processes need to return to the old way, and which should keep moving forward without looking back.
==== Refs
References
1 Darwin C. On the Origin of Species by Means of Natural Selection, or, the Preservation of Favoured Races in the Struggle for Life 1859 London
2 Taryne Imai H.S. Adam Truong Van Chau Amersi Farin The lasting footprint of COVID-19 on surgical education: a resident and attending perspective on the global pandemic Am J Surg 2021 In Press
3 Chick R.C. Clifton G.T. Peace K.M. Using technology to maintain the education of residents during the COVID-19 pandemic J Surg Educ 77 4 2020 729 732 32253133
4 Kogan M. Klein S.E. Hannon C.P. Nolte M.T. Orthopaedic education during the COVID-19 pandemic J Am Acad Orthop Surg 28 11 2020 e456 e464 32282439
5 Rana T. Hackett C. Quezada T. Medicine and surgery residents’ perspectives on the impact of COVID-19 on graduate medical education Med Educ Online 25 1 2020 1818439 32924869
6 Rosen G.H. Murray K.S. Greene K.L. Pruthi R.S. Richstone L. Mirza M. Effect of COVID-19 on urology residency training: a nationwide survey of program directors by the society of academic urologists J Urol 204 5 2020 1039 1045 32463716
| 33752871 | PMC9745861 | NO-CC CODE | 2022-12-15 00:03:11 | no | Am J Surg. 2021 Sep 17; 222(3):471-472 | utf-8 | Am J Surg | 2,021 | 10.1016/j.amjsurg.2021.03.018 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Excerpta Medica
S0002-9343(21)00821-4
10.1016/S0002-9343(21)00821-4
Article
Table of Contents
8 2 2022
2 2022
8 2 2022
135 2 A8A11
2019
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745871 | NO-CC CODE | 2022-12-15 00:03:11 | no | Am J Med. 2022 Feb 8; 135(2):A8-A11 | utf-8 | Am J Med | 2,022 | 10.1016/S0002-9343(21)00821-4 | oa_other |
==== Front
Accid Anal Prev
Accid Anal Prev
Accident; Analysis and Prevention
0001-4575
1879-2057
Elsevier Ltd.
S0001-4575(21)00226-8
10.1016/j.aap.2021.106195
106195
Article
Impact of COVID-19 on motor vehicle injuries and fatalities in older adults in Ontario, Canada
Rapoport Mark J. ab*
Chee Justin N. c
Aljenabi Nadia c
Byrne Patrick A. c
Naglie Gary de
Ilari Frances c
Elzohairy Yoassry c
Vingilis Evelyn fg
Mulsant Benoit H. hi
a Department of Psychiatry, Termerty Faculty of Medicine, University of Toronto, Staff Psychiatrist, Canada
b Sunnybrook Health Sciences Centre, Toronto, Canada
c Ontario Ministry of Transportation, Toronto, Canada
d Division of Geriatric Medicine, Department of Medicine, Temerty Faculty of Medicine, and Institute of Health Policy Management and Evaluation, University of Toronto, Canada
e Department of Medicine and Rotman Research Institute, Baycrest Health Sciences, Toronto, Canada
f Population and Community Health Unit, Canada
g Departments of Family Medicine, and Epidemiology and Biostatistics, Schulich School of Medicine and Dentistry, University of Western Ontario, London, Canada
h Department of Psychiatry, Temerty Faculty of Medicine, University of Toronto, Canada
i Centre for Addiction and Mental Health, Toronto, Canada
⁎ Corresponding author at: Department of Psychiatry, Termerty Faculty of Medicine, University of Toronto, Staff Psychiatrist, Canada.
18 5 2021
7 2021
18 5 2021
157 106195106195
19 7 2020
2 3 2021
13 5 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Background
Older adults constitute the group most vulnerable to COVID-19 mortality. As a result, in North America and elsewhere, older adults have been strongly advised to shelter in place. Older adults also represent the fastest growing segment of licensed drivers.
Objective
We examined the change in injuries and fatalities sustained by younger and older drivers and pedestrians during the first month of the COVID-19 pandemic. We hypothesized that adults ages 80 years and over would have a proportionally larger reduction than the other drivers and pedestrians.
Methods
Using a cohort design, we compared the proportion of drivers and pedestrians involved in injuries and fatalities attributable to individuals aged 80 years and over, as recorded in the Ministry of Transportation of Ontario (Canada) database, between the 30 days prior to shelter-in-place related to the COVID-19 pandemic and the subsequent 30 days. By way of comparison, we conducted a similar comparison for younger age cohorts (16−24 years, 25−34 years, 35−54 years, 55−64 years, and 65−79 years).
Results
Drivers aged 80 years and over represented 21 per 1000 injuries and fatalities in the 30 days prior to March 17, 2020 (95 % CI: 15−29), and 8 per 1000 injuries and fatalities in the 30 days beginning on that date (95 % CI: 2−20), a 64.7 % reduction (exp (β) post 0.353, 95 % CI 0.105−0.892). Drivers in the 35−54 year age range underwent a significant but smaller reduction of 22.9 %; no significant changes were seen for drivers in other age groups, or for pedestrians of any age.
Conclusions and relevance
The physical distancing measures that aimed to reduce the spread of COVID-19 resulted in a marked reduction in driver injuries and fatalities in the oldest old, illustrating the impact of physical distancing recommendations in this population. The excess mortality burden faced by the oldest adults during the COVID-19 pandemic, by direct exposure to the virus, may be indirectly mitigated by the reduction in road-related deaths in this cohort.
Keywords
COVID-19
Aging
Injuries
Fatalities
Driving
Pedestrians
Older drivers
Traffic accident
==== Body
pmc1 Introduction
Older adults are the fastest growing segment of the licensed driver population, with almost a doubling of licensed drivers aged 65 years and over since 1994. They rely on the automobile as their primary mode of transportation (Turcotte, 2012; Vrkljan et al., 2018). Although overall fatality rates due to motor vehicle collisions (MVCs) have declined substantially over recent decades among older adults, their MVC rate and injury rate per miles driven is lower than drivers less than 30 years of age and similar to drivers 30–65 years of age (Tefft, 2017). In contrast, the MVC fatality rate per miles driven of drivers age 80 and older was the highest of any age group(Roberts et al., 2008; Ang et al., 2017; Tefft, 2017), although the increased mortality rate may be in part an artifact of low mileage bias (Langford et al., 2006), increased fragility (Bayam et al., 2005), or different patterns of driving and types of roadways travelled (Keall and Frith, 2004, 2019). Older adults are also over-represented among pedestrian fatalities, accounting for 20–35 % of such fatalities (Transportation Canada, 2011; Centres for Disease Control and Prevention, 2020). In Toronto, Canada’s most populous city, adults ages 55 years and over accounted for 71 % of all pedestrian fatalities between 2016 and 2018 (Toronto Police, 2019).
Older adults are also the group most vulnerable to mortality associated with the COVID-19 pandemic(Wu and McGoogan, 2020). In China, early estimates of the case fatality rates for adults ages 80 years and over was 20.2 %(Onder et al., 2020). At the end of March 2020, the infection fatality ratio outside of China was 1.4 % for those less than age 60 years, and 4.5 % in those 60 years of age and over (Verity et al., 2020). Adults over the age of 80 have a particularly high mortality rate, estimated at 10.1 %, in France(Salje et al., 2020).
Shortly after the World Health Organization (WHO) declared the COVID-19 a worldwide pandemic on March 11, 2020, most jurisdictions in North America and Europe introduced physical distancing measures. On March 12, 2020, Ontario, Canada’s most populous province, closed all schools; on March 17, it declared a state of emergency. The senior driver license renewal program, requiring adults age 80 years and over to pass vision and cognitive screening tests and undergo a review of their driving record, was put on hold, affecting between 10,000 and 12,000 drivers per month. Similarly, requests for specialized on-road testing were put on hold, and mandatory reporting to the Ministry of Transportation of cognitively impaired drivers were considerably decreased as a result of fewer physician visits. Toronto saw a 15–20 % reduction in traffic congestion relative to the previous year (TomTom Traffic Index, 2020). Other direct and indirect impacts of COVID-19 on road safety are less well understood. For example, the Governors Highway Safety Association in the United States documented multiple incidents of speeding and reckless driving during the pandemic (Governors Highway Safety Association (GHSA, 2020), MVCs and fatalities more than doubled in Minnesota, half of which were for speeding, careless, or negligent driving (Governors Highway Safety Association (GHSA, 2020). Similarly, in New York City, speeding tickets on March 27, 2020 doubled compared to a month earlier (Governors Highway Safety Association (GHSA, 2020).
Given uncertainties about the impact of COVID-19 on older drivers, we conducted a study to examine the magnitude of injuries and fatalities sustained by older drivers and pedestrians in the province of Ontario. Our focus was on drivers 80 years and older (the “oldest old”) because of their elevated MVC fatality risk (Tefft, 2017) and high case fatality rates with COVID(Wu and McGoogan, 2020). We hypothesized that a large proportion of the oldest olds would follow physical distancing recommendations, driver and pedestrian injury and fatality rates in this age group would be lower during the pandemic than before it, and they would have a proportionally greater reduction of driver and pedestrian injury and fatalities than younger cohorts.
2 Methods
We used official data provided by the Ontario Ministry of Transportation’s (MTO) Research and Evaluation Office to assess the number and type of injuries and fatalities sustained by drivers and pedestrians for 30 days on and after March 17, 2020, the date of the declaration of the state of emergency in Ontario, and the month prior (i.e., February 17 to March 16, 2020 vs. March 17 to April 16, 2020). When a MVC occurs in the province, police officers are required to submit a standardized collision report to the MTO, which captures basic information about the collision (e.g., date, road jurisdiction) as well as details on drivers (e.g., age, gender, injury severity), vehicles (e.g., class, model, points of impact in the collision), other persons involved (e.g., passengers, pedestrians), and fatalities (e.g., total number resulting from the collision, death date). Accordingly, the dataset contains every MVC that was reportable under the Highway Traffic Act (HTA), which occurred on public roads within all Ontario jurisdictions during the specified period. If any collision-involved individual sustained injury as a result of the collision, then the corresponding police collision report collected by MTO must contain an injury level data element for all individuals involved in that collision. Injuries are defined as minimal (no hospitalization when leaving the scene of the collision, but includes minor abrasions, bruises and complaint of pain), minor (treatment in a hospital emergency room, but not admitted), major (admitted to hospital), and fatal injuries (killed immediately or died within 30 days of the MVC due to collision-related injuries) (Road Safety Research Office, 2016). To assess whether the findings are unique to 2020, we also assessed the number of injuries and fatalities for similar time periods in 2018 and 2019.
Data were extracted from the MTO records by ministry staff on May 4, 2020 and verified for integrity. Fatality data were further assessed for accuracy against files from the Chief Coroner of Ontario, although some coroner investigations may have still been ongoing at the time of data analysis. Since some collision reports can be delayed by several weeks, the numbers available for 2020, and especially those from the more recent March 17-April 16 period, may be an underestimate. Hence, our primary outcome is the proportion of injuries and fatalities sustained by the oldest old group relative to injuries and fatalities in drivers and pedestrians of all ages. If there is an underestimate due to delayed reporting, it should occur randomly in all age groups and the proportion should not be affected.
We compared the time-based changes in the number of injuries and fatalities in the oldest old group and other adult groups by way of comparison: younger young (age 16−24 years); older young (age 25−34 years); younger middle-age (age 35−54 years); older middle age (age 55−64 years); young old (age 65−79 years); and oldest old (age 80 years and over).
Descriptive data are presented. The Agresti-Coull method was used to compute 95 % confidence intervals (CI) for population proportions. A difference-in-differences analysis was used to compare the change in proportion of injuries and fatalities among older senior drivers and pedestrians (i.e., aged 80+) between two points in time and across years (2018–2020). The time points were defined as a pre-period (i.e., February 17 to March 16) and post-period (i.e., March 17 to April 16). We modelled the odds that a driver or pedestrian involved in an injury or fatality collision in the oldest old group, Odds80p, as:log(Odds80p)=β0+βY+βperiod+βY×period
where β0 is the baseline log odds in the pre-period of 2020; βY is the estimated mean difference in the log odds of the oldest old being in an injury or fatality collision between the comparison years (=2018 or Y=2019) and the reference year (Y=2020) during the pre-period (β2020=0) ; βperiod is the expected mean change in the log odds of being 80 or older in an injury or fatality collision from the pre- to post-periods in 2020 (βpre=0). The remaining difference-in-difference estimators, βY×period, represent how the change in log odds of being 80 or older in an injury or fatality collision between the pre- and post-periods in the comparison years differ from what would be expected based on the observed change in the reference year. These parameters are non-zero only for the comparison years and the post-period. Confidence intervals were estimated by “inverting the likelihood ratio”. Statistical significance was determined at ∝ = 0.05, without correction for multiple comparisons. Similar analyses were also conducted for the other age groups.
According to our hypothesis, we would expect expβpost to be less than one, indicating a decrease in odds for the oldest old of being in an injury or fatality collision in the 2020 post-period compared to the 2020 pre-period. We would also expect expβY×post to be greater than one for both comparison years, indicating a lesser decrease in the odds of being 80+ years of age between the pre- to post-periods in 2018 and 2019 compared to 2020. In addition, we estimated the pre- to post-change in odds for 2018 and 2019 as expβpost+β2018×post and expβpost+β2019×post respectively, in order to examine whether drivers and pedestrians in the post-period were more likely to be 80+ years of age than in the pre-period for years prior to 2020.
3 Results
The oldest old group of drivers had a mean age of 83.8 ± 4.9 years and 50 % were male. The mean (± SD) age and gender of the other groups were: 71.0 ± 4.3 years and 68 % male for drivers 65−79 years of age; 59.0 ± 2.7 years and 69 % male for drivers 55−64 years of age; 43.4 ± 6.2 years and 58 % male for drivers 35−54 years of age; 29.2 ± 3.0 years and 67 % male for drivers 25−34 years of age; and 21 ± 2.1 years and 62 % male for drivers 16−24 years of age. Further breakdown (e.g., by geography) is not possible given small numbers.
The number of driver injuries and fatalities across all ages observed in the month before March 17 was 1859 in 2018, 1901 in 2019, and 1620 in 2020. Of those in 2020, 34 were among drivers in the oldest old group, representing 21 per 1000 injuries and fatalities across all ages (95 % CI: 15−29) (Table 1 ). In the month beginning March 17, the number of driver injuries and fatalities across all ages observed was 2065 in 2018, 1870 in 2019, and 532 in 2020. Of those in 2020, 4 were among drivers in the oldest old group, representing 8 per 1000 driver injuries and fatalities across all ages (95 % CI: 2−20).Table 1 Number of Driver and Pedestrian Injuries and Fatalities Per 1000 Injuries and Fatalities by Age Group.
Table 1AGE GROUP YEAR DRIVERS PEDESTRIANS
# per 1000 Driver Injuries and Fatalities Among All Age Groups (95 % Confidence Interval) # per 1000 Pedestrian Injuries and Fatalities Among All Age Groups (95 % Confidence Interval)
Month Before March 17 Month Beginning March 17 Month Before March 17 Month Beginning March 17
Younger Young (Aged 16−24) 2018 174 (157−192) 165 (149−181) 268 (215−327) 227 (181−281)
2019 149 (134−166) 159 (143−177) 199 (148−262) 238 (187−297)
2020 159 (142−177) 179 (148−214) 233 (182−293) 185 (100−312)
Older Young (Aged 25−34) 2018 215 (197−234) 215 (197−233) 189 (145−244) 219 (174−273)
2019 235 (217−255) 204 (187−223) 183 (134−245) 242 (191−302)
2020 226 (206−247) 248 (213−287) 193 (146−250) 278 (173−412)
Younger Middle-Aged (Aged 35−54) 2018 344 (323−366) 350 (329−371) 263 (212−323) 260 (211−316)
2019 354 (333−376) 367 (345−389) 309 (247−378) 229 (180−288)
2020 356 (333−380) 299 (261−339) 287 (231−350) 222 (129−353)
Older Middle-Aged (Aged 55−64) 2018 154 (139−172) 144 (130−160) 136 (98−185) 152 (46−269)
2019 142 (127−158) 142 (127−159) 173 (125−233) 104 (0−236)
2020 140 (124−158) 162 (133−195) 139 (99−191) 185 (0−457)
Young Old (Aged 65−79) 2018 90 (78−104) 96 (84−109) 111 (77−157) 108 (76−151)
2019 94 (82−108) 98 (85−112) 100 (64−151) 139 (99−190)
2020 98 (85−114) 105 (82−134) 121 (84−171) 111 (47−227)
Older Old (Aged 80+) 2018 23 (17−31) 30 (23−38) 33 (16−65) 34 (17−63)
2019 26 (20−34) 29 (23−38) 37 (16−75) 48 (26−84)
2020 21 (15−29) 8 (2−20) 27 (11−59) 19 (0−108)
CI: Confidence Interval, computed using the Agresti-Coull approach.
The number of pedestrian injuries and fatalities across all ages observed in the month before March 17 was 243 in 2018, 191 in 2019, and 223 in 2020. Of those in 2020, 6 were among pedestrians in the oldest old group, representing 27 per 1000 injuries and fatalities across all ages (95 % CI: 11−59) (Table 1). In the month beginning March 17, the number of pedestrian injuries and fatalities across all ages observed was 269 in 2018, 231 in 2019, and 54 in 2020. Of those in 2020, only 1 was in the oldest old group (i.e., an 80-year old male), representing 19 per 1000 pedestrian injuries and fatalities across all ages (95 % CI: CI 0−108). The mean (± SD) age and gender of the other groups were: 68.2 ± 3.2 years and 67 % male for pedestrians 65−79 years of age; 59.7 ± 2.5 years and 40 % male for pedestrians 55−64 years of age; 44.7 ± 6.2 years and 50 % male for pedestrians 35−54 years of age; 29.5 ± 2.4 years and 53 % male for pedestrians 25−34 years of age; and 20.5 ± 2.4 years and 30 % male for pedestrians 16−24 years of age. Similar to drivers, the numbers are insufficient to provide a demographic breakdown by additional variables.
Consistent with our hypothesis, the value of βpost generated by the difference-in-differences model for drivers indicates a 64.7 % lower odds of being an oldest old driver for those involved in an injury or fatality collision during the post-period as compared to the pre-period in 2020 (Table 2a). The values of interaction (difference-in-differences) terms, β2018×post and β2019×post, were both found to be significantly greater than one, indicating the change in the odds of an oldest odd driver being in an injury or fatality collision from pre-to post-periods in 2018 and 2019 were smaller than in 2020. In fact, drivers in the post-period were, if anything, more likely to be oldest old than in the pre-period for 2018 and 2019, though this was not tested statistically. The model for pedestrians revealed a 31.8 % drop in the odds of an oldest old being in an injury or fatality collision in the post- vs. pre-period, but this reduction was not statistically significant (Table 2b).Table 2 Difference-in-Differences Model Parameter Estimates.
Table 2MODEL PARAMETER COEFFICIENT β ODDS RATIO
expβ LOWER 95 % CI UPPER 95 % CI
(a) DRIVERS (Aged 80+) β0 −3.84 0.021 0.015 0.030*
β2018 0.075 1.078 0.684 1.712
β2019 0.210 1.234 0.796 1.935
β2020 0 (reference) 1.000 – –
βpost −1.040 0.353 0.105 0.892*
βpre 0 (reference) 1.000 – –
β2018×post 1.332 3.789 1.375 13.459*
β2019×post 1.176 3.241 1.179 11.489*
(b) PEDESTRIANS (Aged 80+) β0 −3.59 0.028 0.011 0.057
β2018 0.208 1.231 0.422 3.793
β2019 0.319 1.376 0.449 4.342
β2020 0 (reference) 1.000 – –
βpost −0.382 0.682 0.036 4.113
βpre 0 (reference) 1.000 – –
β2018×post 0.399 1.490 0.189 31.909
β2019×post 0.655 1.926 0.246 41.288
(c) DRIVERS (Aged 35−54) β0 −0.592 0.553 0.499 0.612*
β2018 −0.052 0.949 0.825 1.091
β2019 −0.009 0.991 0.863 1.138
β2020 0 (reference) 1.000 – –
βpost −0.261 0.771 0.622 0.951*
βpre 0 (reference) 1.000 – –
β2018×post 0.284 1.329 1.037 1.707*
β2019×post 0.316 1.372 1.070 1.764*
CI: Confidence Interval;
* Significant change in the proportion of injuries and fatalities attributable by age group, using the 95 % profile likelihood confidence interval (p < 0.05).
Post-hoc difference-in-differences analyses were also performed on injury or fatality collisions involving drivers in which being from each of the other age ranges was considered as an outcome instead of being in the 80 and over age range. The odds of a driver being in the 35−54 range underwent a significant but smaller reduction (i.e., 22.9 % in comparison to 64.7 % for the 80+ group) from pre- to the post-periods in 2020, according to the value of βpost in Table 2c. Similar reductions or any other significant changes were not seen for the pedestrians in that age group, or in drivers and pedestrians of the other age groups (data not shown).
4 Discussion
Assessing data from the province of Ontario, Canada, we found a two-third decline in the proportion of driver injuries and fatalities in adults 80 years of age and older during the first month of the physical-distancing measures implemented after the COVID-19 pandemic was declared. We believe that our study is the first to assess the direct effect of a pandemic on injuries and fatalities in oldest old drivers or pedestrians and, indirectly, on the behavior of these older adults. Barnes et al. (2020) recently reported on a similar reduction in proportion of MVCs attributable to adults 65 years and older in the first 50 days of the COVID-19 pandemic in Louisana, but they did not provide data for oldest old adults(Barnes et al., 2020). Preliminary information from the International Transport Forum indicates that almost all countries saw a decline in road fatalities during the initial lockdown associated with COVID (OECD Group, 2020) but again data was not stratified by age. We found that the physical distancing measures that aimed to reduce the spread of COVID-19 and its associated morbidity and mortality appear to have resulted in a marked reduction in driver injuries and fatalities in the oldest old (i.e., those 80 years and older). Although the data available did not allow us to assess this directly, we attribute this marked reduction to the oldest old adhering strictly to the physical distancing measures and staying in their home, i.e., off the roads, to reduce their risk of being infected by COVID-19, and, as a result, reducing their exposure to road risks. Previous literature has shown that older adults drive less at night, perhaps in response to vision-related declines (Charlton et al., 2006). So it is not surprising that they would restrict their driving and mobility in the face of a new significant health threats. There was no similar reduction in the proportion of injuries and fatalities in the other age groups, apart from a smaller reduction in the 35−54 year group. The lack of reduction in the proportion of injuries and fatalities for the 16−24 year old group may reflect differences in telework opportunities. For example, Deng et al., 2020 identified that only 21 % of 15−24 year old in Canada had jobs that allowed for teleworking, while 44 % of 34−54 year old had that opportunity(Deng et al., 2020). Hence, the different requirements related to work and travel of these two younger age groups would be expected to affect their MVC injury and fatality risk differently. Additionally, the lockdown of schools, daycare and after school programs could further keep some parents age 35−54 year at home to care for their children (Statistics Canada, 2017). On the other hand, some young old adults are working or engaged in unpaid occupations, such as caregiving and volunteering. For example, 2015 Canadian data indicate that more than half of men 65 years of age and nearly 30 % of men 70 years of age reported working, particularly as managers in agriculture, retail and wholesale trade, transport truck drivers, janitors, caretakers, or building superintendents (Statistics Canada, 2017). Again, it is possible that young old adults with these jobs had fewer teleworking opportunities and hence their driving exposure was not reduced. Moreover, the young old group may have been more likely than the oldest old to assist their own children with childcare when schools and daycares closed.
It is unclear why there was no similar significant reduction in pedestrian fatalities. This may be due to lack of power. However, it is possible that this relates to the altered forms of community mobility in these early days of the pandemic. For example, people tended to avoid mass transit, taxis, or carpools, but continued to walk for exercise, at times walking in live traffic in order to avoid proximity to other pedestrians on the sidewalk (Baruchman, 2020). A recent report suggests an increase in drugs and alcohol prevalence among road fatalities during the pandemic (Thomas et al., 2020), which in addition to an increase in reckless driving (Governors Highway Safety Association (GHSA, 2020) could make the roads particularly hazardous for pedestrians during this time.
These population-based results are of public health significance. Our early data can provide a benchmark or framework for assessing the impact on the implementation or lifting of physical distancing restrictions as the COVID-19 pandemic waxes and wanes. They dramatically illustrate the impact of the public health physical distancing recommendations for the oldest old. Our results also provide empirical evidence that the excess mortality burden faced by the oldest old during the COVID-19 pandemic, by direct exposure to the virus, may be mitigated indirectly by a reduction in road-related deaths in this age group (Colonna and Intini, 2020). Our data can inform modeling of mortality data; they also suggest that other similar factors may be at play and indirectly affect mortality. For instance, one would expect a decrease in mortality from the suspension of most elective surgeries and an increase in mortality due to decreased physician visits and medical attention or reluctance to present to hospitals for medical emergencies(Gunnell et al., 2020; Ueda et al., 2020; Zhao et al., 2020). Thus, the reduced fatalities among drivers, and other similar indirect factors that affect mortality during the COVID-19 pandemic, need to be taken into account by mortality models. Many of these models are typically limited to integrating observed vs. expected mortality rates(Salje et al., 2020; Wu et al., 2020). The availability of more sophisticated models is important given their use to guide health and public policy decisions in North America.
It is too early to know to what extent the reduction in MVC-related injuries outweighs the potential negative health costs related to the COVID-19 pandemic, although it is very unlikely that the reduction in fatalities approaches the massive death rate associated with the virus in the oldest-old. The COVID-19 pandemic provides an unprecedented naturalistic experiment of the short- and long-term direct and indirect consequences of a marked reduction in the population venturing outdoors, especially among older individuals. The extent to which COVID 19-related stay-at-home measures may have worsened social isolation and pre-existing health disparities in the oldest old is not yet fully appreciated (Hwang et al., 2020; Moore et al., 2020).
There are several limitations to the present study. First, the most recent MTO collision data likely underestimate the absolute number of injuries and fatalities, mainly during the COVID 2020 period, due to the time required for complete reporting. However, our use of proportions and our clear results mitigate this possible limitation. Another limitation is our inability to account for possible unmeasured confounders, including geography, weather, gas prices, gender, driving exposure, or the cognitive or functional status of individuals included in the sample. Finally, our data are relevant to a brief period and we do not know whether the dramatic short-term effects we observed will be maintained, amplified, or will dissipate over time.
Future work will need to examine motor vehicle injuries and fatalities as older adults restart venturing outdoors. Residual physical distancing measures may lead to reduced use of public transportation and a need to promote safer roads for all drivers and pedestrians.(De Vos, 2021) Once physical distancing measures are lifted, some older adults with mild cognitive impairment or dementia whose driving skills are already declining may be at particular risk when they resume driving after a lengthy period without practice. Others may have accommodated to life without driving and may be reluctant to resume driving.
Impact statement
We certify that this work is novel.
Declaration of Competing Interest
Mark Rapoport – no conflicts of interest. Research supported by Canadian Institutes of Health Research, Canadian Consortium on Neurodegeneration in Aging, Canadian Council on Motor Transportation Administration, the Canadian Medical Association/Joule, as well as the Sunnybrook Psychiatry Partnership.
Justin Chee – no conflicts of interest.
Nadia Aljenabi – no conflicts of interest.
Patrick A. Byrne – no conflicts of interest.
Gary Naglie - no conflicts of interest. Research supported by grants from the Canadian Institutes of Health Research and the Canadian Consortium on Neurodegeneration in Aging, as well as by the George, Margaret and Gary Hunt Family Chair in Geriatric Medicine, University of Toronto.
Frances Ilari – no conflicts of interest.
Yoassry Elzohairy – no conflicts of interest.
Benoit Mulsant - no conflicts of interest related to this work. Financial support from the Labatt Family Chair in Biology of Depression in Late-Life Adults at the University of Toronto. Research supported by grants from Brain Canada, the Canadian Institutes of Health Research, the CAMH Foundation, the Patient-Centered Outcomes Research Institute (PCORI), and the US National Institutes of Health, all outside the submitted work; within the past five years, he has also received non-financial research support from Capital Solutions Design LLC, Eli Lilly, Happy-Neuron, and Pfizer, all outside the submitted work.
Evelyn Vingilis – no conflicts of interest. Research supported by grants from Canadian Institutes of Health Research (CIHR); Caskey/Francis Family Award in Clinical Research; Ministry of Transport of Ontario Road Safety Research Partnership Program; Western Research Catalyst Grant 2020.
==== Refs
References
Ang B.H. Global burden of road traffic accidents in older adults: a systematic review and meta-regression analysis Arch. Gerontol. Geriatr. 72 2017 32 38 28527382
Barnes S.R. The Effect of covid-19 Lockdown on Mobility and Traffic Accidents: Evidence From Louisiana. GLO Discussion Paper, No. 616 2020 Global Labor Organization (GLO), Essen.
Baruchman M. Are You Walking More During the Coronavirus Pandemic? 2020 Seattle Times https://www.seattletimes.com/seattle-news/transportation/are-you-walking-more-during-the-coronavirus-pandemic-heres-what-people-are-noticing-and-what-they-say-can-be-improved/
Bayam E. Liebowitz J. Agresti W. Older drivers and accidents: a meta analysis and data mining application on traffic accident data Expert Syst. Appl. 29 2005 598 629
Centres for Disease Control and Prevention Pedestrian Safety 2020 https://www.cdc.gov/motorvehiclesafety/pedestrian_safety/index.html
Charlton J.L. Characteristics of older drivers who adopt self-regulatory driving behaviours Transp. Res. Part F Traffic Psychol. Behav. 9 5 2006 363 372
Colonna P. Intini P. Compensation effect between deaths from covid-19 and crashes: the Italian case Transportation Research Interdisciplinary Perspectives 2020 6 10.1016/j.trip.2020.100170
De Vos, M., In Press. The effect of covid-19 and subsequent social distancing on travel behavior. Transportation Research: Interdisciplinary Perspectives. https://doi.org/10.1016/j.trip.2020.100121.
Deng Z. Morissette R. Messacar D. Running the Economy Remotely: Potential for Working From Home During and After covid-19 2020 Statistics Canada https://www150.statcan.gc.ca/n1/pub/45-28-0001/2020001/article/00026-eng.htm
Governors Highway Safety Association (GHSA) Absent Traffic Jams, Many Drivers Getting More Reckless Available from:https://www.ghsa.org/index.php/resources/news-releases/coronavirus-reckless-driving20(accessed May 12, 2020) 2020
Gunnell D. Suicide risk and prevention during the covid-19 pandemic Lancet Psychiatry 2020 10.1016/S2215-0366(20)30171-1
Hwang T.J. Rabheru K. Peisah C. Reichman W. Ikeda M. Loneliness and social isolation during the covid-19 pandemic Int. Psychogeriatr. 32 10 2020 1217 1220 32450943
Keall M.D. Frith W.J. Older driver crash rates in relation to type and quantity of travel Traffic Inj. Prev. 5 1 2004 26 36 14754672
Langford J. Older drivers do not have a high crash risk--a replication of low mileage bias Accid. Anal. Prev. 38 3 2006 574 578 16426560
Moore J.T. Pilkington W. Kumar D. Diseases with health disparities as drivers of covid-19 outcome J. Cell. Mol. Med. 24 19 2020 11038 11045 32816409
OECD International Traffic Safety Data and Analysis Group Road Safety Annual Report 2020 2020 International Transport Forum https://www.itf-oecd.org/sites/default/files/docs/irtad-road-safety-annual-report-2020_0.pdf
Onder G. Case-fatality rate and characteristics of patients dying in relation to covid-19 in italy JAMA 2020 10.1001/jama.2020.4683
Road Safety Research Office, Ontario road safety annual report Road Safety Research Office, Ontario Ministry of Transportation 2016 Ontario Ministry of Transportation http://www.mto.gov.on.ca/english/publications/pdfs/ontario-road-safety-annual-report-2016.pdf
Roberts S.E. A comparison of self-reported motor vehicle collision injuries compared with official collision data: an analysis of age and sex trends using the canadian national population health survey and transport canada data Accid. Anal. Prev. 40 2 2008 559 566 18329407
Salje H. Estimating the burden of sars-cov-2 in france Science 2020 10.1126/science.abc3517
Statistics Canada Census in Brief: Working Seniors in Canada 2017 https://www12.statcan.gc.ca/census-recensement/2016/as-sa/98-200-x/2016027/98-200-x2016027-eng.cfm
Tefft B.C. Rates of Motor Vehicle Crashes, Injuries and Deaths in Relation to Driver Age, United States, 2014-2015 2017 AAA Foundation for Traffic Safety https://aaafoundation.org/rates-motor-vehicle-crashes-injuries-deaths-relation-driver-age-united-states-2014-2015/
Thomas F.D. Drug and Alcohol Prevalence in Seriously and Fatally Injured Road Users Before and During the covid-19 Public Health Emergency (report No. Dot Hs 813 018) 2020 National Highway Traffic Safety Administration
Tomtom traffic index, 2020. https://www.tomtom.com/en_gb/traffic-index/ (accessed May 5, 2020).
Toronto Police Toronto Collisions Fatalities Dashboard Available from: https://app.powerbi.com/view?r=eyJrIjoiZGNiZmNjZDItZTc4Ny00NDU5LWJmNGYtNGVjN2IxYmE0ZTIwIiwidCI6Ijg1MjljMjI1LWFjNDMtNDc0Yy04ZmI0LTBmNDA5NWFlOGQ1ZCIsImMiOjN9 (Accessed May 18, 2020) 2019
Transportation Canada Road Safety in Canada. Road Safety Canada Consulting 2011 Public Health Agency of Canada https://www.tc.gc.ca/media/documents/roadsafety/tp15145e.pdf
Turcotte M. Profile of seniors’ transportation habits Statistics Canada Canadian Social Trends 2012 (11-008-X)
Ueda M. Managing cancer care during the covid-19 pandemic: agility and collaboration toward a common goal J. Compr. Canc. Netw. 2020 1 4
Verity R. Estimates of the severity of coronavirus disease 2019: a model-based analysis Lancet Infect. Dis. 2020
Vrkljan B. Transportation mobility in the Canadian longitudinal study on aging Raina P. Report on Health and Aging in Canada Findings from Baseline Data Collection 2010-2015 2018 Canadian Longitudinal Study on Aging (accessed May 20, 2020) https://ifa.ngo/wp-content/uploads/2018/12/clsa_report_en_final_web.pdf
Wu Z. Mcgoogan J.M. Characteristics of and important lessons from the coronavirus disease 2019 (covid-19) outbreak in china: summary of a report of 72314 cases from the chinese center for disease control and prevention JAMA 2020 10.1001/jama.2020.2648
Wu J.T. Estimating clinical severity of covid-19 from the transmission dynamics in wuhan, china Nat. Med. 26 4 2020 506 510 32284616
Zhao J. Rudd A. Liu R. Challenges and potential solutions of stroke care during the coronavirus disease 2019 (covid-19) outbreak Stroke 51 5 2020 1356 1357 32228369
| 34020183 | PMC9745872 | NO-CC CODE | 2022-12-15 00:03:12 | no | Accid Anal Prev. 2021 Jul 18; 157:106195 | utf-8 | Accid Anal Prev | 2,021 | 10.1016/j.aap.2021.106195 | oa_other |
==== Front
Am J Surg
Am J Surg
American Journal of Surgery
0002-9610
1879-1883
Elsevier Inc.
S0002-9610(21)00526-2
10.1016/j.amjsurg.2021.09.003
My Thoughts / My Surgical Practice
Finding a place for non-operative management of acute appendicitis: COVID-19 as an example
Emile Sameh Hany ∗
General Surgery Department, Mansoura University Faculty of Medicine, Mansoura, Egypt
∗ Colorectal Surgery Unit, General Surgery Department, Mansoura University Hospitals, Mansoura University, Mansoura City, Egypt
15 9 2021
3 2022
15 9 2021
223 3 605606
22 7 2021
16 8 2021
6 9 2021
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc1 Introduction
Acute appendicitis (AA) is the most common surgical emergency worldwide with an incidence of 1.1 cases per 1000 people per year. The incidence of AA tends to peak in the late teen years and then gradually declines in the elderly. Men are more commonly affected by AA with a male to female ratio of 1.4:1. The diagnosis of AA may be made on clinical basis, yet the use of CT scanning has reduced the rates of negative appendectomy because it has a sensitivity and specificity that exceed 90%. Appendectomy is the gold standard treatment of uncomplicated AA with documented efficacy and safety. However, several clinical trials and meta-analyses1 examined the efficacy and safety of non-operative management (NOM) of AA and reported that NOM is a feasible treatment option with satisfactory outcomes, at least on the short-term.
2 The indications for NOM
Currently, some clinicians are proponents of NOM of AA on the basis that it avoids the possible complications of surgery and preserves the appendix that is not regarded anymore as a vestigial organ since it may have important immunologic functions. On the other end, other clinicians are strict opponents of NOM because they believe appendectomy is the best definitive treatment for uncomplicated AA since NOM is associated with a failure rate of 8.5% and recurrence of symptoms of AA in almost 20% of patients.1
Lying in between the two opposing teams, there exists a grey zone in which NOM of AA does have a place, yet in certain conditions and not as a routine practice. NOM can be a viable treatment option in the settings where access to surgery is restricted as in places where well-equipped operation theaters are not available or when the patient is of ASA class IV or V and is not considered fit for surgery. Recently, another setting where NOM of AA may be a feasible treatment option has emerged, the COVID-19 pandemic.
3 Impact of COVID-19 on surgical practice and management of AA
The COVID-19 pandemic has affected several aspects of life, and medical and surgical practice is not an exception. The impact of the current pandemic on surgical practice has been observed in various domains, importantly elective surgery cancellation has resulted in a significant decrease down to 55% in the volume of procedures performed which has varied among surgical services.2 The impact of the hospital lock down secondary to COVID-19 was assessed in a living systematic review3 that showed a significant negative impact of lock down on patient outcome and procedure volume. Further negative aspects include the increased rates of pulmonary complications and mortality in patients with perioperative SARS-COV-2 infection, and shift from minimally-invasive to open surgery.4
The current pandemic has impacted the presentation and management of AA in various ways. A significant increase in complicated AA along with a simultaneous reduction in uncomplicated AA during the pandemic may imply that patients are not seeking timely surgical care. When compared to the pre-COVID era, patients with AA presented during the pandemic had a 5.5% decrease in uncomplicated AA and a 21.1% and 29% increase in perforated and gangrenous appendicitis, respectively.5
On another hand, the presentation of COVID-19 may mimic the clinical picture of AA, especially in children. Moreover, a recent report assumed a novel association of AA in children infected with SARS-CoV-2. The authors hypothesized that AA in children with SARS-COV-2 infection may represent a postinfectious hyperinflammatory complication of the infection that can occur two weeks after the early manifestation of the disease.6
The impact of COVID-19 pandemic on the management of AA was explored in a population-based analysis in Germany. The analysis found an overall reduction by 12.9% of patients presenting with AA and this reduction was mainly relevant to decreased rates of uncomplicated AA, rather than complicated appendicitis. The treatment methods, rates of extended surgery, complication rates including appendix stump leakage and need for re-operation did not remarkably differ from those before the onset of the pandemic.7
One of the important observed effects of COVID-19 on the management of AA was the increased rate of adoption of NOM of AA as compared to the pre-COVID era. A recent meta-analysis included 14 studies entailing 2140 patients, of whom 45% had a trial of NOM of AA. The application of NOM during COVID-19 was more six times more likely than its application before the pandemic. The weight mean failure rate of NOM was 16.4% and failure was more observed in children and patients with complicated appendicitis. Furthermore, NOM had significantly lower odds for complications than appendectomy. The authors concluded that NOM of AA in the setting of COVID-19 may be a safe, short-term alternative to surgery with acceptably low failure and complication rates.8
4 A new place for NOM of AA
As the number of patients recovering from COVID-19 has been rising, several of these patients may need some kind of surgical intervention and a relevant question would be when would be safe to have surgery after recovery from the infection. The answer was revealed by a recent multicenter study by the CovidSurg collaborative9 that included more than 140,000 patients from 116 countries, 2.2% of whom had a preoperative SARS-CoV-2 diagnosis. When compared to patients without SARS-COV-2 diagnosis, patients having surgery within 0–2 weeks, 3–4 weeks, and 5–6 weeks had significantly higher odds of mortality and complications (odds ratio = 4.1, 3.9, and 3.6). After a seven-week delay in having surgery after infection, the mortality risk dropped to the baseline similar to patients without SARS-COV-2 diagnosis. The study, therefore, recommended to delay elective surgery for at least seven weeks following SARS-CoV-2 infection to decrease the higher odds of pulmonary complications and mortality in these patients to become similar to that of patients without prior COVID-19.
Hence, in light of the recent evidence and given that NOM maybe a safe short-term treatment of uncomplicated AA, another question is raised; can NOM be the first-line treatment for patients with uncomplicated AA who have a history of COVID-19 infection dating less than seven weeks before their current presentation? The answer of this question warrants conducting well-designed trials or at least controlled studies to verify whether delaying surgery in patients with uncomplicated AA who have a recent history of COVID-19 would be effective and manage to avoid the increased risks of postoperative pulmonary complications documented in these patients.
Finally, it is important to note that although NOM can be effective in resolving the acute condition in the majority of these patients, further long-term follow-up is required as it has been estimated that approximately 20% of these patients will need readmission and interval appendectomy for recurrent symptoms.1 , 8
Funding
No funding was received in support of this study.
Declaration of competing interest
No conflict of interest to disclose.
==== Refs
References
1 Podda M. Gerardi C. Cillara N. Antibiotic treatment and appendectomy for uncomplicated acute appendicitis in adults and children: a systematic review and meta-analysis Ann Surg 270 6 2019 1028 1040 10.1097/SLA.0000000000003225 Dec 30720508
2 Utria A.F. Javid P.J. Chen J. Rice-Townsend S.E. Impact of COVID-19 on procedure volume at a tertiary pediatric hospital Am J Surg 221 6 2021 1259 1261 10.1016/j.amjsurg.2021.03.003 Jun 33707079
3 Lee Y. Kirubarajan A. Patro N. Soon M.S. Doumouras A.G. Hong D. Impact of hospital lockdown secondary to COVID-19 and past pandemics on surgical practice: a living rapid systematic review Am J Surg 222 1 2021 67 85 10.1016/j.amjsurg.2020.11.019 Jul 33218675
4 Emile S.H. Should we continue using laparoscopy amid the COVID-19 pandemic? Br J Surg 107 8 2020 e240 e241 10.1002/bjs.11669 Jul 32432344
5 Orthopoulos G. Santone E. Izzo F. Increasing incidence of complicated appendicitis during COVID-19 pandemic Am J Surg 221 5 2021 1056 1060 10.1016/j.amjsurg.2020.09.026 33012500
6 Malhotra A. Sturgill M. Whitley-Williams P. Pediatric COVID-19 and appendicitis: a gut reaction to SARS-CoV-2? Pediatr Infect Dis J 40 2 2021 e49 e55 10.1097/INF.0000000000002998 Feb 1 33298761
7 Köhler Acar L. van den Berg A. Impact of the COVID-19 pandemic on appendicitis treatment in Germany—a population-based analysis Langenbeck's Arch Surg 406 2021 377 383 10.1007/s00423-021-02081-4 33420517
8 Emile S.H. Hamid H.K.S. Khan S.M. Davis G.N. Rate of application and outcome of non-operative management of acute appendicitis in the setting of COVID-19: systematic review and meta-analysis J Gastrointest Surg 2021 1 11 10.1007/s11605-021-04988-1 Mar 26
9 COVIDSurg Collaborative; GlobalSurg Collaborative Timing of surgery following SARS-CoV-2 infection: an international prospective cohort study Anaesthesia 2021 10.1111/anae.15458 Mar 9
| 34538608 | PMC9745884 | NO-CC CODE | 2022-12-15 00:03:12 | no | Am J Surg. 2022 Mar 15; 223(3):605-606 | utf-8 | Am J Surg | 2,021 | 10.1016/j.amjsurg.2021.09.003 | oa_other |
==== Front
Am J Ophthalmol
Am J Ophthalmol
American Journal of Ophthalmology
0002-9394
1879-1891
Elsevier Inc.
S0002-9394(21)00029-5
10.1016/j.ajo.2021.01.009
Editorial
Take a Moment: Reflections on a Pandemic, Suffering, and Humanity
Al-khersan Hasenin ⁎
⁎ Inquiries to Hasenin Al-khersan, Bascom Palmer Eye Institute, University of Miami, 900 NW 17th St, Miami, FL 33130, USA
23 1 2021
7 2021
23 1 2021
227 A1A2
16 12 2020
23 12 2020
9 1 2021
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcWhen reflecting on the COVID-19 pandemic, it is easy to be overwhelmed by the numbers: millions infected, hundreds of thousands losing their lives, and new cases climbing day after day. However, it was through a single patient that I grasped the impact of the pandemic.
On this day in clinic, I was seeing a patient with a branch retinal vein occlusion and resultant macular edema that I had been following for nearly a year in my resident clinic. I probably should not have favorite patients, but she is one of them. She was always funny and energetic during our visits, constantly joking with her husband who accompanied her to each appointment and me. In the midst of busy clinics, she would ask whether I had had time to eat lunch and about how my family was doing.
That day, however, she came alone. Stranger yet, she had missed her last 2 monthly appointments. Her vision had dropped to 20/200 with worsening macular edema in her affected eye after we had previously achieved 20/40 vision. She sat down in the examination chair, lacking her usual charm.
I sensed something was wrong. Glancing at my schedule, I saw the red glare of 6 “waiting for physician” markers. I knew, though, I could not just proceed with business as usual. I finally paused and asked, “What's wrong?”
Suddenly, her face disappeared into her hands as she began to sob. Through muffled cries, I made out that her husband had recently passed away from COVID-19. Truthfully, I didn't know what to say. He, with a spirit to match her own, had been in my clinic only a couple of months ago as enthusiastic as ever. Stunned by her revelation, I put my hand on her shoulder, an honest violation of social distancing. Then we sat together in a moment of silence as she unburdened herself of her grief.
In a practice environment that focuses on volume, it can sometimes be hard to pull away from habits of rapid and efficient patient turnover. The reasons behind the focus on volume are numerous and complex, ranging from the sheer extent of patient need to a compensation system that largely rewards volume.1
The stresses of working in this environment can distract and distance us from the humanity of our patients. How often do we refer to patients by their disease? Do we become upset when a patient asks “too many” questions during a busy clinic day? I am certainly guilty of these offenses. Providers, particularly trainees, may also lose touch with their own humanity, sacrificing their well-being in the setting of impossible external demands.2
For me, the pandemic has served as a sobering reminder of the humbling responsibility that we as physicians have of bearing witness to the suffering of others. Our health care system must of course focus on health, but we also must not forget to care. The simple act of listening, though it may not be charted or billed, not only makes us better doctors but also more compassionate humanitarians.
In addition, as trainees, many of us practice in underserved communities with marginalized populations, many of color, who face unimaginable hardships in their daily lives—all of which have been disproportionately accentuated by the pandemic. We must not allow the callousness with which our society treats these particularly vulnerable patients to bleed into our practice.
I am not naïve to the challenges of carving out even just 5 minutes a visit to listen to patients in the midst of overbooked clinics. Although our personal behavior is of course important in shaping our relationship with patients, we are also constrained by external administrative pressures. It is our duty as physicians, therefore, to advocate for structural changes that will allow us to better serve our patients. We cannot sacrifice the doctor-patient relationship at the altar of clinical volume and stifling bureaucracy.
Since that encounter, my patient has returned every month. Her vision has improved once again, and a smile—punctuated by moments of grief—has returned to her face. Our preinjection ritual has since transformed from telling jokes to recounting memories of her husband. I know that these moments are as important to her as any medical treatment I can offer.
This pandemic has taught me to slow down. To be wholly present in the service of my patients.
To stop.
To take a moment.
Appendix Supplementary materials
Supplemental Material available at AJO.com.
Image, application 1
All authors have completed and submitted the ICMJE form for disclosure of potential conflicts of interest. Funding/Support: No funding support was obtained for the present manuscript. Financial Disclosures: The authors indicate no financial support or conflicts of interest. All authors attest that they meet the current ICMJE criteria for authorship.
Supplemental Material available at AJO.com.
From the Bascom Palmer Eye Institute, University of Miami, Miami, Florida, USA.
==== Refs
REFERENCES
1 Feng PW Ahluwalia A Feng H Adelman RA. National trends in the United States Eye Care Workforce from 1995 to 2017 Am J Ophthalmol 218 2020 128 135 32445703
2 Tran EM Scott IU Clark MA Greenberg PB. Resident wellness in US ophthalmic graduate medical education: the resident perspective JAMA Ophthalmol 136 2018 695 701 29801087
| 33497674 | PMC9745899 | NO-CC CODE | 2022-12-15 00:03:12 | no | Am J Ophthalmol. 2021 Jul 23; 227:A1-A2 | utf-8 | Am J Ophthalmol | 2,021 | 10.1016/j.ajo.2021.01.009 | oa_other |
==== Front
Am J Ophthalmol
Am J Ophthalmol
American Journal of Ophthalmology
0002-9394
1879-1891
Elsevier Inc.
S0002-9394(21)00084-2
10.1016/j.ajo.2021.02.019
Article
Ocular Injury Associated With Prone Positioning in Adult Critical Care: A Systematic Review and Meta-Analysis
Patterson Timothy J. a⁎
Currie Peter a
Williams Michael b
Shevlin Claire a
a From the Department of Acute Critical Care Services, Craigavon Area Hospital, Portadown, Craigavon, United Kingdom
b Department of Ophthalmology, Royal Victoria Hospital, Montreal, Quebec, Canada.
⁎ Inquiries to: Timothy Patterson, Craigavon Area Hospital, 68 Lurgan Road, Portadown, Craigavon BT63 5QQ, United Kingdom
3 3 2021
7 2021
3 3 2021
227 6673
27 10 2020
8 2 2021
16 2 2021
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Purpose
Prone positioning during the COVID-19 pandemic has become increasingly used as an adjunct to increase oxygenation in critical care patients. It is associated with an adverse event profile. This study sought to investigate the occurrence of ocular injuries reported in prone versus supine groups in adult critical care.
Design
Systematic review and meta-analysis.
Methods
A systematic review and meta-analysis were carried out in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines. PubMed, SCOPUS, and the Cochrane Library were searched. The search period was January 1, 1990, to July 1, 2020.
Results
Eleven randomized controlled trials were included, with 2,247 patients. Twenty-eight events were recorded in 3 trials (174 patients) and no events in the other 8 trials (2,073 patients). The rates of eye injury were 5 events in 1,158 patients (1.30%) and 13 events in 1,089 patients (1.19%) in the prone and supine groups, respectively, which were reduced to 2 of 1,158 patients (0.17%) and 2 of 1,089 patients (0.18%), respectively, when reports of eye or eyelid edema were removed. Meta-analysis demonstrated no significant differences between groups with (an OR of 1.40 (95% CI: 0.37–5.27) and without (OR: 0.78; 95% CI: 0.11–5.73) reported edema.
Conclusions
This meta-analysis showed no significant difference in the rate of reported ocular injury between prone and supine critical care groups. These rates remain higher than the incidence reported during general anesthesia. There is a need for studies in critical care settings in which ocular injury is an end-point and which include extended patient follow-up.
==== Body
pmcDuring the first decade of the 21st century, multiple trials demonstrated an improvement in oxygenation in patients with acute respiratory distress syndrome receiving nurse attendance in the prone position.1, 2, 3, 4 In combination with lung protective ventilation, this improvement has subsequently been demonstrated to reduce mortality compared to conventional supine ventilation.5 However, this technique is not without risk. The Faculty of Intensive Care Medicine (FICM) released guidelines for the implementation of this technique in 2019 as there was concern that the increased use of prone positioning was contributing to the rise in critical safety incidents, including pressure injuries to anterior structures.6
One such structure was the eye. Vision loss and impairment due to optic nerve, corneal, and scleral injuries, along with extraocular muscle impingement had been reported previously following spinal and plastic surgery procedures performed with the patient in a prone position,7, 8, 9 with prone positioning having been reported to increase intraocular pressure.10 FICM prone position guidelines aimed to reduce both exposure and pressure, recommending several measures including taping patients’ eyes shut; ensuring there is no direct pressure on patients’ eyes; placing patients in a 30-degrees foot-down position (reverse Trendelenburg) while being nursed in this position (to limit dependent periocular swelling) and rotating patients’ heads from side-to-side at 2-hour intervals.6
Despite the increased use of prone positioning in critical care adults, as highlighted in both popular and scientific press during the recent COVID-19 pandemic, the complication profile of prone positioning is still emerging.11 A previous review of the complications associated with prone positioning revealed only studies in surgical patients.12 This is a different cohort from those in critical care, who should be considered separately. Due to longer cumulative time periods in the prone position, the need for increased repositioning and the easier access for nursing staff (no sterile field to avoid) in critical care patients form a distinct patient group and should be regarded as such. It may also be argued that, as a positive fluid balance may be more likely in critical care patients than in surgical patients, the risks of dependent edema and pressure injury are increased.
THE AIM OF THIS REVIEW
The purpose of this systematic review was to compare the incidence and types of ocular injury reported in randomized controlled trials of prone positioning versus supine positioning in adult critical care and, if possible, perform a meta-analysis. This investigation has not been conducted previously to the authors’ knowledge.
METHODS
This systematic review was conducted in accordance with the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) statement.13 A review protocol was registered with the PROSPERO systematic review database (registration number: CRD42020196917).14
Study Eligibility Criteria
Studies in adult (18 years of age or older at the time of intervention) hospital inpatients in critical care (defined as patients requiring at least Level 2 care, as defined by the National Health Service Critical Care Service Framework) were eligible, regardless of patients’ sedation level, indication for prone positioning (eg, respiratory- or surgical site-related), or geographical location of care.15 Studies including any form of prone positioning adjuncts (where the patient is positioned face down on the bed) were included. Only randomized controlled trials (RCTs) that compared prone to supine positioning were included. Any variation of RCT design was eligible, including cluster or cross-over.
Primary outcome was the incidence of ocular injuries, presented as both crude rates and, after meta-analysis, odds ratios (OR). Only papers published in indexed medical journals (conference abstracts were excluded) were included, with English as the language of publication. The date of publication was restricted to the previous 3 decades (January 1, 1990 to July 1, 2020). This range was defined by the authors because prone positioning in critical care patients became increasingly common following a series of RCTs published during the year 2000. Allowing the search to include studies from between 1990 and 2000 would allow capturing any preceding studies within the same era of critical care.
Information Sources and Search
Four databases, PubMed (MeSH and Advanced), SCOPUS, and Cochrane Library were searched. For PubMed, the search was conducted using both MeSH terms and the advanced search option. MeSH terms (“Position,” “Prone’) were used. An advanced search was conducted using the terms “Prone” AND “Randomised.” Ovid, Embase, and SCOPUS were also searched using the same terms with automatic adjustment for Americanized spelling. The Cochrane Library was searched using the term “Prone.”
Study Selection and Data Collection Process
Two independent reviewers each reviewed all titles retrieved from the initial searches. Duplicates were eliminated, and if possible, using the available abstracts, each reviewer decided its inclusion. If the paper could not be included or excluded with certainty on the basis of the abstract, then the full text was read. Any disagreements between reviewers on papers’ eligibility were resolved by discussion or, if necessary, arbitration by a senior author. If an RCT had been reported by more than one publication, the last publication which reported the trial was used as the reference publication in this review.
Data Items
The following variables were recorded: study information (first author, publication year, and country of origin); participant information (total patients, sex, median/mean age); intervention information where available (prone positioning adjuncts, length of time spent in a prone position, head rotation frequency, eye care provided); and follow-up information (mean and median follow-up duration, planned follow-up period, and how many study participants completed follow-up).
The following outcome data were sought: whether ocular injury data were reported as present or absent; number or proportion of participants in each intervention arm with ocular injuries; injury type; laterality and any patient-reported outcomes available. If further information was required, then individual study authors were contacted with a return period of 30 days.
Risk of Bias in Individual Studies and Across Studies
Two authors independently assessed the potential bias using the Cochrane Collaboration bias tool.16
Summary Measurements and Synthesis of Results
The rates of ocular injury were shown as crude rates and, if appropriate, mean scores, for example, who diagnosed the ocular injury and how and when may vary in relevant ways between studies.
Meta-Analysis
Meta-analysis was performed for rates of injury overall and for each distinct injury type if there were 2 or more RCTs which examined the same injury type. The outcome groups were divided into supine positioning and prone positioning, with subdivisions into specific prone positioning pressure relief adjuncts if sufficient data were available (2 or more homologous studies).
Review Manager 5 software (Nordic Cochrane Center, Copenhagen, Denmark) was used for results synthesis.17 If the outcome data were presented in the form of dichotomous categorical variables, ORs would be reported with corresponding 95% confidence intervals (CI). Statistical heterogeneity between studies was checked and reported using the I2 measure of study heterogeneity. If low heterogeneity (I2: <50%) between studies was reported, then a fixed effect model was used. If higher heterogeneity was displayed, then a random effects model was used. If a meta-analysis displayed a heterogeneity of >75%, it was excluded from the results.18 , 19
The primary summary measurement (rate of ocular injuries) of the meta-analysis was given as an OR with 95% CI. If there was a zero-cell count for any given event, then Review Manager 5 software automatically added a 0.5 continuity correction. A sensitivity analysis was also undertaken. Only studies in which the majority of areas (4 or more) of potential bias were low risk would be analyzed and presented in this analysis.
RESULTS
Study Characteristics
Twelve studies met inclusion criteria. Two studies had included the same participants, resulting in 11 studies with unique populations.1, 2, 3, 4, 5 , 20, 21, 22, 23, 24, 25 The PRISMA flow diagram of search results is presented in Figure 1 .Figure 1 PRISMA flow diagram of study selection. A total of 11 studies were included in the end quantitative synthesis. PRISMA = Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
Figure 1
There were 2,247 randomly assigned patients included in this analysis (1,089 patients underwent supine positioning only, and 1,158 patients were randomized to receive prone positioning). Length of follow-up reporting heterogeneous. Study details are included in Table 1 .TABLE 1 Study and Patient Characteristics of Included Studies
TABLE 1Study (ref), y Location Patient Group Males/Females Age Mean/Median ± SD, y Number Prone Number Supine Mean Length of Follow-Up, Days Supine/Prone
Gattinoni (1), 2001 Italy ALI or ARDS 214.0/90.0 Prone: 57 ± 16
Supine: 59 ± 17 152.0 152.0 10/10
Watanabe (20), 2002 Japan Post-esophagectomy 14.0/2.0 Prone: 66 ± 5.8
Supine: 63 ± 8.1 8.0 8.0 Not available
Beuret (21), 2002 France Patients requiring intubation because of GCS <9.0 36.0/15.0 Prone: 55 ± 19
Supine: 55 ± 20 25.0 26.0 Not available
Guérin (2), 2004 France Acute respiratory failure 593.0/198.0 Prone: 62 ± 15.7
Supine: 62.5 ± 14.7 413.0 378.0 24.5/26.9
Papazian (22), 2005 France ARDS 23.0/16.0 Prone: 51 ± 10.5
Supine: 55 ± 15 26.0 13.0 Not available
Voggenreiter (23), 2005 Germany Trauma patients with ALI or ARDS 33.0/7.0 Prone: 40 ± 14
Supine: 43 ± 10 21.0 19.0 Not available
Mancebo (3), 2006 Spain ARDS 86.0/50.0 Prone: 54 ± 17
Supine: 54 ± 16 76.0 60.0 Not available
Chan (24), 2007 Taiwan CAP 18.0/4.0 Prone: 54.7 ± 21.8
Supine: 69 ± 15.5 11.0 11.0 3.0/3.0
Fernandez (25), 2008 Spain ARDS 25.0/15.0 Prone: 53.9 ± 17.9
Supine: 55.3 ± 14.6 21.0 19.0 17.5/14.7
Taccone (4), 2009 Italy ARDS 246.0/96.0 Population:
60 ± 16 168.0 174.0 28.0/28.0
Guérin (5), 2013 France ARDS 318.0/148.0 Prone: 58 ± 16
Supine: 60 ± 16 237.0 229.0 26.0 (18.0 non-survivors)/
24.0 (21.0 non-survivors)
ALI = acute lung injury; ARDS = acute respiratory distress syndrome; CAP = community-acquired pneumonia; GCS = Glasgow coma scale.
Patient Characteristics
All studies reported patient sex. When the 11 studies were pooled, 641 patients (28.5%) were female patients. The mean ± SD ages of study participants were reported in all studies, ranging from 40 ± 14 to 66 ± 5.8 years in the prone group and 43 ± 10 to 69 ± 15.5 years in the supine group.
Risk of Bias Within Studies
An individual risk of bias analysis for each study is presented in Table 2 . In the case of Beuret and associates,21 although physiological endpoint data assessments were not blinded, as subjective as chest radiographs were, the authors still elected to place this study's blinding of outcome assessment in the negative category. The study by Guérin and associates2 was the only study to report full outcome assessment blinding. Allocation concealment and randomization methodology was unclear for most of the studies, with Chan and associates, 2007, appearing to have been quasi-randomized by the study author.24 Three studies reported adequate randomization and allocation concealment in full.2, 3, 4 The authors found those 3 studies had a minimal risk of bias in most areas so proceeded to a sensitivity analysis including those studies, as reported below.TABLE 2 Cochrane Risk of Biasa
TABLE 2Study (ref), y Random Sequence Generation (Selection Bias) Allocation Concealment (Selection Bias) Blinding of Participants and Personnel (Performance Bias) Blinding of Outcome Assessment (Selection Bias) Incomplete Outcome Data (Attrition Bias) Selective Reporting (Reporting Bias)
Gattinoni (1), 2001 + ? − − + +
Watanabe (20), 2002 ? ? − − + +
Beuret (21), 2002 ? ? − − + +
Guérin (2), 2004 + + − + + +
Papazian (22), 2005 ? ? − − + +
Voggenreiter (23), 2005 + ? − − + +
Mancebo (3), 2006 + + − − + +
Chan (24), 2007 ? ? − − + +
Fernandez (25), 2008 + ? − − + +
Taccone (4), 2009 + ? − − + +
Guérin (5), 2013 + + − − + +
a All studies were found satisfactory in all areas of assessment.
+ = XXX; ? = XXX; − = XXX.
Ocular Injuries
Three of the 11 studies reported ocular injuries, including 1 study reporting subconjunctival hemorrhage and 2 studies reporting eye (chemosis) or eyelid edema. Those occurrences of edema were included in an initial meta-analysis but were excluded from a subgroup analysis because transient dependent facial edema is anecdotally a common feature of prone positioning. Three studies reported ocular findings, with 8 of 11 studies reporting no events in either trial arm. The crude rate of eye injury across all 11 studies consisted of 15 events among 1,158 patients (1.30%) in the prone group and 13 of 1,089 patients (1.19%) in the supine group, which were reduced to 2 of 1,158 patients (0.17%) and 2 of 1,089 patients (0.18%), respectively, when reports of edema were removed. Meta-analysis demonstrated no significant differences between ORs in the supine group and ORs in the prone groups (OR: 1.40; 95% CI: 0.37–5.27) (Figure 2 ). Similarly, no differences were demonstrated when patients with reported edema were removed. An OR of 0.78 (95% CI: 0.11–5.73) (Figure 3 , Table 3 ) was the rate of ocular injuries reported in the included studies.Figure 2 Forest plot analysis of all ocular injuries recorded across the study populations. There were nonsignificant differences shown between groups. OR: 1.02 (95% CI: 0.82–1.26).
Figure 2
Figure 3 Forest plot analysis of all ocular injuries, excluding eye and/or eyelid swelling, recorded across the study populations. There were non-significant difference shown between groups. OR: 0.79 (95% CI: 0.11–5.44).
Figure 3
TABLE 3 Ocular Injuries Recorded in Each of the Included StudiesXXX
TABLE 3Study (ref), y Ocular Injury Recorded Injury Type Supine Prone
Gattinoni (1), 2001 None recorded – – –
Watanabe (20), 2002 Yes Eyelid edema 8 8
Beuret (21), 2002 None recorded – – –
Guérin (2), 2004 None recorded – – –
Papazian (22), 2005 None recorded – – –
Voggenreiter (23), 2005 None recorded – – –
Mancebo (3), 2006 Yes Subconjunctival hemorrhage 2 2
Chan (24), 2007 Yes Chemosis “eye swelling” 3 5
Fernandez (25), 2008 None recorded – – –
Taccone (4), 2009 None recorded – – –
Guérin (5), 2013 None recorded – – –
Following sensitivity analysis, no significant differences were found between the prone group, consisting of 2 of 726 patients (0.275%) or the supine group, consisting of 2 of 667 patients (0.300%; OR: 0.78; 95% CI: 0.11–5.73 for both groups) (Figure 4 ). Only 1 of the 3 studies in the sensitivity analysis reported ocular findings (Table 3).3 Figure 4 Forest plot analysis of all ocular injuries recorded across studies assessed as having a low risk of bias. There were non-significant differences shown between groups. OR: 0.79 (95% CI: 0.11–44).
Figure 4
Each author was contacted and asked for additional information regarding ocular injuries if not reported; authors Guérin and Beuret kindly replied, with no additional information to add.
DISCUSSION
This systematic review and meta-analysis of ocular injury in supine versus prone patients contained 11 unique RCTs. Comparisons were made with all reported ocular injuries and with edema removed, as transient dependent facial edema is anecdotally expected on pronation. When prone versus supine positioning was reported, there was a nonsignificant difference between rates of ocular injuries between the groups. The crude rates of eye injury across all studies, including 15 of 1,158 patients (1.30%) in the prone group and 13 of 1,089 patients (1.19%) in the supine group, which were reduced to 2 of 1,158 patients (0.17%) and 2 of 1,089 patients (0.18%), respectively, when reports of edema were removed. Of note, only 3 studies reported ocular injury, which was reduced to 1 study following removal of eyelid edema and chemosis. The remaining ocular injuries were cases of subconjunctival hemorrhage.
Prior to this review, the risk profile of ocular injury during general anesthesia has been well characterized, with the largest single center study (75,120 non-ocular procedures) reporting a perioperative ocular injury incidence of 0.023% and 0.080% in the prone group, respectively, with patients having been operated on in the prone position at an increased risk of injury (OR: 10.8; 95% CI: 2.4–48.8).26 This poses a discrepancy with the results of the present study, in which incidence rates of 0.17% and 0.18% in prone and supine positions, respectively, were found in critical care patients. Speculatively, a higher incidence in a critical care settings compared to general anesthesia may be because of the prolonged corneal and/or conjunctival exposure time associated with critical care as opposed to procedural general anesthesia.
Previous studies specifically regarding ocular injury following prone positioning in critical care or general anesthesia, however, are limited. Population-based epidemiological studies of the rate of post-operative vision loss following prone spinal surgery showed an incidence of 0.094%, with risk factors, including hypotension, anemia, pre-existing peripheral vascular disease, and pediatric (<18 years old) and geriatric (>84 years old) age for development of vision loss.27 , 28 Loss of vision was not an outcome in this study of the incidence of ocular injury, nor was it reported by any of the included studies.
The present study was constrained by both the lack of ocular assessment for injury as a study endpoint and the limited follow-up period present in these critical care studies, as ocular structural injury may not be evident within the patients’ index hospital admission. Lack of personnel and observer blinding (with only 1 study reporting outcome assessment blinding) were highlighted by the Cochrane Risk of Bias tool; however, the authors acknowledge that pragmatically these would be difficult.2 The randomization process and allocation concealment for several studies was unclear. In addition, only 1 study reported full outcome assessor blinding. This prompted the authors to conduct a sensitivity analysis which included the 3 studies which displayed a minimal risk of bias in most (4 or greater) areas.2 , 3 , 5 This analysis again showed no significant differences between the prone and supine groups, with or without reports of eye or eyelid edema (RR: 0.79; 95% CI: 0.11–5.44) for both groups.
The authors highlight the fact that these results, although conducted to minimize bias, represent data from only 1 study, with the 2 other studies in the sensitivity analysis reporting no data in each study arm. The authors were also prompted to reflect that the study which reported ocular findings had a sample size of 136, whereas the additional 2 studies had a combined sample size of 1,257 participants without reported ocular events. This may indicate the underreporting of ocular injury as an adverse event. Although there is no indication of bias toward supine or prone positioning groups, this observation likely highlights the key deficiency of this study.
To advance this observation, when the studies included in this review were observed as a whole, it is notable that only 3 of 11 studies contained events.3 , 20 , 24 In these studies, the event rates for supine patients (3.33%–100%) and prone patients (3.33%–100%) were greater than those recorded by the review population as a whole. When the authors reviewed the outcome definitions for each of the included studies, they noted that 7 studies specifically mentioned pressure injuries either in the outcome sections of their methodology or results.1, 2, 3 , 20 , 21 Those which did not included the study by Taccone and associates,4 who stated that they would record adverse events relating to remaining in the prone position; Guérin and associates5; Fernandez and associates25; and Papazian and associates,22 who included “complications” as an outcome. None of the included studies included ocular injuries as a specific endpoint in their methodology. This observation highlights the need for ocular injuries as an endpoint in future critical care trials.
Further investigation of the longer-term effects of prone positioning is recommended, with an increased cohort of patients post-prone positioning expected following the COVID-19 pandemic. Clinical recommendations in critical care, beyond the implementation of recommendations of the FICM guidelines, are limited.6 However, the authors recommend that these observation that ocular injuries do occur in critical care, and at a higher rate than in cases of general anesthesia, guides the provision of critical follow-up and post-admission patient guidance.
This meta-analysis of RCTs showed that there was no differences in the rate of ocular injury between prone and supine patient groups in adult critical care. However, the rate of ocular injury was increased compared to general anesthesia, which was displayed. The authors acknowledge that this is a meta-analysis of studies which did not include ocular injury as an endpoint, so the rate of sustained ocular injury may be greater than reported by this study. The authors recommend that this study be used to guide future critical care follow-up services and to prompt further outcome studies and raise awareness of the need for meticulous ocular care in the critical care environment.
ALL AUTHORS HAVE COMPLETED AND SUBMITTED THE ICMJE FORM FOR DISCLOSURE OF POTENTIAL CONFLICTS OF INTEREST and none were reported.
Funding Support: None. Financial Disclosures: The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
(Ocular Injury Associated With Prone Positioning In Adult Critical Care: A Systematic Review And Meta-Analysis [CRD42020196917]; PROSPERO; University of York, York, UK)
==== Refs
References
1 Gattinoni L Tognoni G Pesenti A Effect of prone positioning on the survival of patients with acute respiratory failure N Engl J Med 345 8 2001 568 573 11529210
2 Geurin C Gaillard S Lemasson S. Effects of systematic prone positioning in hypoxaemic acute respiratory failure JAMA 292 19 2004 2379 2387 15547166
3 Mancebo J Fernández R Blanch L. A Multicenter trial of prolonged prone ventilation in severe acute respiratory distress syndrome Am J Respir Crit Care Med 173 11 2006 1233 1239 16556697
4 Taccone P Pesenti A Latini R Prone positioning in patients with moderate and severe acute respiratory distress syndrome: a randomized controlled trial JAMA 302 18 2009 1977 1984 19903918
5 Guérin C. Reignier J. Richard JC for the PROSEVA Study Group. Prone positioning in severe acute respiratory distress syndrome N Engl J Med 368 23 2013 2159 2168 23688302
6 Faculty of Intensive Care Medicine Prone Position in Adult Critical Care 2019 Available at https://www.ficm.ac.uk/sites/default/files/prone_position_in_adult_critical_care_2019.pdf Accessed June 4, 2020
7 Grixti A Sadri M Datta AV. Uncommon ophthalmologic disorders in intensive care unit patients J Crit Care 27 6 2013 746.e9–22
8 Stambough JL Dolan D Werner R Godfrey E. Ophthalmologic complications associated with prone positioning in spine surgery J Am Acad Orthop Surg 15 3 2007 156 165 17341672
9 Woodruff C English M Zaouter C Postoperative visual loss after plastic surgery: case report and a novel continuous real-time video monitoring system for the eyes during prone surgery Br J Anaesth 106 1 2011 149 151 21148645
10 Saran S Gurjar M Kanaujia V Effect of prone positioning on intraocular pressure in patients with acute respiratory distress syndrome Crit Care Med 47 9 2019 e761 e766 31305498
11 Rosenbaum L. Facing Covid-19 in Italy—ethics, logistics, and therapeutics on the epidemic's front line N Engl J Med 382 20 2020 1873 1875 32187459
12 Kwee M Ho Y-H Rozen W. The prone position during surgery and its complications: a systematic review and evidence-based guidelines Int Surg 100 2 2015 292 303 25692433
13 PRISMA Statement and Checklist. 2009. Available at: http://www.prisma-statement.org/. Accessed June 4, 2020.
14 Prospero Systematic Review Database. 2020. Available at: https://www.crd.york.ac.uk/prospero/. Accessed June 4, 2020.
15 Inspection Framework: NHS Acute Hospitals. 2016. Available at: https://www.cqc.org.uk/sites/default/files/20160713_NHS_core_service_inspection_framework_critical_care.pdf. Accessed June 4, 2020.
16 Cochrane Risk of Bias Tool. Cochrane Collaboration. Available at: https://handbook-5-1.cochrane.org/chapter_8/8_assessing_risk_of_bias_in_included_studies.htm. Accessed June 4, 2020.
17 Review Manager 5. Cochrane Collaboration. Available at: https://community.cochrane.org/help/tools-and-software/revman-5. Accessed November 4, 2018.
18 Huedo-Medina TB Sánchez-Meca J Marín-Martínez F Botella J. Assessing heterogeneity in meta-analysis: Q statistic or I2 index? Psychol Methods 11 2 2006 193 206 16784338
19 Demets DL. Methods for combining randomized clinical trials: strengths and limitations Stat Med 3 1987 341 350 3616287
20 Watanabe I Fujihara H Sato K Beneficial effect of a prone position for patients with hypoxemia after transthoracic esophagectomy Crit Care Med 30 8 2002 1799 1802 12163796
21 Beuret P Carton MJ Nourdine K Kaaki M Tramoni G Ducreux JC. Prone position as prevention of lung injury in comatose patients: a prospective, randomized, controlled study Intensive Care Med 28 5 2002 564 569 12029403
22 Papazian L Gainnier M Marin V Comparison of prone positioning and high-frequency oscillatory ventilation in patients with acute respiratory distress syndrome Crit Care Med 33 10 2005 2162 2171 16215365
23 Voggenreiter G Aufmkolk M Stiletto RJ Prone positioning improves oxygenation in post-traumatic lung injury—a prospective randomized trial J Trauma 59 2 2005 333 343 16294072
24 Chan MC Hsu JY Liu HH Effects of prone position on inflammatory markers in patients with ARDS due to community-acquired pneumonia J Formos Med Assoc 106 9 2007 708 716 17908660
25 Fernandez R Trenchs X Klamburg J Prone positioning in acute respiratory distress syndrome: a multicenter randomized clinical trial Intensive Care Med 34 8 2008 1487 1491 18427774
26 Yu HD Chou AH Yang MW Chang CJ. An analysis of perioperative eye injuries after nonocular surgery Acta Anaesthesiol Taiwan 48 3 2010 122 129 20864060
27 Epstein N. Perioperative visual loss following prone spinal surgery: a review Surg Neurol Int 7 13 2016 347 360
28 Patil CG Lad EM Lad SP Ho C Boakye M. Visual loss after spine surgery: a population-based study Spine 33 13 2008 1491 1496 18520945
| 33675753 | PMC9745902 | NO-CC CODE | 2022-12-15 00:03:13 | no | Am J Ophthalmol. 2021 Jul 3; 227:66-73 | utf-8 | Am J Ophthalmol | 2,021 | 10.1016/j.ajo.2021.02.019 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Published by Elsevier Inc.
S0002-9378(21)00538-X
10.1016/j.ajog.2021.04.255
Correction
April 2021 (vol. 224, no. 4, page 382.e2)
11 6 2021
11 6 2021
© 2021 Published by Elsevier Inc.
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcPatberg ET, Adams T, Rekawek P, et al. Coronavirus disease 2019 infection and placental histopathology in women delivering at term. Am J Obstet Gynecol 2021;224:382.e1-18.
In the article cited above, an error appeared in Table 1, “Demographic and clinical patient characteristics by COVID-19 status” (page 382.e2). In row 4, “Birthweight (g),” the finding under column 3, “COVID-19 negative (n=56),” should have been 3358±469 grams, not 3280±402 grams, as in the previous column, “COVID-19 cases (n=77).” The P value shown in column 4 was based on accurate data and is correct.
The difference in mean birthweight was not significant between groups; this variable was not incorporated into any further analyses. This error does not qualitatively change any study findings.
| 34120741 | PMC9745903 | NO-CC CODE | 2022-12-15 00:03:13 | no | Am J Obstet Gynecol. 2021 Jun 11; doi: 10.1016/j.ajog.2021.04.255 | utf-8 | Am J Obstet Gynecol | 2,021 | 10.1016/j.ajog.2021.04.255 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier Inc.
S0002-9378(20)31391-0
10.1016/j.ajog.2020.12.016
Expert Reviews
The evolution of prenatal care delivery guidelines in the United States
Peahl Alex F. MD ab∗
Howell Joel D. MD, PhD bc
a Department of Obstetrics and Gynecology, University of Michigan, Ann Arbor, MI
b Institute for Healthcare Policy and Innovation, University of Michigan, Ann Arbor, MI
c Departments of Internal Medicine and History, University of Michigan, Ann Arbor, MI
∗ Corresponding author: Alex F. Peahl, MD.
13 12 2020
4 2021
13 12 2020
224 4 339347
2 10 2020
24 11 2020
7 12 2020
© 2020 Elsevier Inc. All rights reserved.
2020
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The coronavirus disease 2019 pandemic led to some of the most drastic changes in clinical care delivery ever seen in the United States. Almost overnight, providers of prenatal care adopted virtual visits and reduced visit schedules. These changes stood in stark contrast to the 12 to 14 in-person prenatal visit schedule that had been previously recommended for almost a century. As maternity care providers consider what prenatal care delivery changes we should maintain following the acute pandemic, we may gain insight from understanding the evolution of prenatal care delivery guidelines. In this paper, we start by sketching out the relatively unstructured beginnings of prenatal care in the 19th century. Most medical care fell within the domain of laypeople, and childbirth was a central feature of female domestic culture. We explore how early discoveries about “toxemia” created the groundwork for future prenatal care interventions, including screening of urine and blood pressure—which in turn created a need for routine prenatal care visits. We then discuss the organization of the medical profession, including the field of obstetrics and gynecology. In the early 20th century, new data increasingly revealed high rates of both infant and maternal mortalities, leading to a greater emphasis on prenatal care. These discoveries culminated in the first codification of a prenatal visit schedule in 1930 by the Children’s Bureau. Surprisingly, this schedule remained essentially unchanged for almost a century. Through the founding of the American College of Obstetricians and Gynecologists, significant technological advancements in laboratory testing and ultrasonography, and calls of the National Institutes of Health Task Force for changes in prenatal care delivery in 1989, prenatal care recommendations continued to be the same as they had been in 1930—monthly visits until 28 weeks’ gestation, bimonthly visits until 36 weeks’ gestation, and weekly visits until delivery. However, coronavirus disease 2019 forced us to change, to reconsider both the need for in-person visits and frequency of visits. Currently, as we transition from the acute pandemic, we should consider how to use what we have learned in this unprecedented time to shape future prenatal care. Lessons from a century of prenatal care provide valuable insights to inform the next generation of prenatal care delivery.
Key words
alloimmunization
antenatal care
blood pressure
care delivery
coronavirus disease 2019 pandemic
electronic fetal monitoring
history
preeclampsia
prenatal care
proteinuria
reduced visit schedules
telemedicine
ultrasonography
==== Body
pmcIntroduction
As a chief resident working in an obstetrics clinic serving predominantly low-income women, one of the authors (A.F.P.) noticed a curious pattern in her pregnant patients’ prenatal appointment attendance. Patients would routinely present for their first appointment. They never missed their anatomy scan—the ultrasound where they could (among other things) learn the gender of the baby. They would return between 24 and 28 weeks’ gestation to confirm that their blood count was adequate and they had no signs of gestational diabetes. Subsequently, many would disappear until a few weeks before delivery.
When asked about their absence, patients described the choices they were being forced to make between recommended care and the demands of everyday life: “I couldn’t get a ride”; “I can’t miss work, I gotta put food on the table”; and “I couldn’t get my babies across town.” The reasons were always followed with the reassurance, “but I knew everything was fine.” Women would also express frustration about the frequency and brevity of the appointments (typically no more than 10 minutes) and the lack of fulfillment from their visits. As she thought more critically about the prenatal visit schedule—monthly visits until 28 weeks’ gestation, biweekly visits until 36 weeks’ gestation, and weekly visits until delivery—A.F.P started to question the status quo.
She was surprised to find that the current prenatal care schedule had first been recommended in 1930 (without supporting evidence) and had remained unchanged through the current recommendations published in the “Guidelines for perinatal practice, 8th edition,” in 2017. She learned that the United States has maintained this same one-size-fits-none prenatal care delivery guideline despite drastic changes in technology and population health, evidence to support alternative prenatal care delivery, persistently worse maternity outcomes, and deepening health disparities.1 Moreover, she wondered why.
The coronavirus disease 2019 (COVID-19) pandemic forced us to reconsider prenatal care delivery guidelines in the United States, both to reduce viral exposure during clinic visits and to conserve scarce healthcare resources. As maternity care providers consider whether we should maintain changes, such as reduced visit schedules and telemedicine, understanding prenatal care delivery guidelines over time can provide important insights. Thus, we describe the surprising evolution of prenatal care delivery guidelines over the span of 3 centuries to inform the next generation of prenatal care delivery.
Pregnancy Care in the Early Republic
During the 19th century, medical care in the United States was relatively unstructured. The absence of state licensing laws meant that anyone could claim to be a physician (and many did).2 Moreover, the predominantly rural landscape and difficulty of transportation meant that healthcare advice was often delivered by laypeople, frequently relying on widely read texts that offered advice on all sorts of medical matters, including prenatal care. If one looked, for example, to “Gunn’s domestic medicine,”3 the most important advice for pregnant women was to keep the bowels regular. Tepid baths were also recommended. Another extremely popular text, William Buchan’s “Domestic medicine,”4 suggested that bleeding—a common remedy at the time thought to correct bodily imbalances or remove inflammation through intentional blood loss—be utilized for pregnant women suffering from dependent edema or jaundice.5 Regular visits to a professional provider were neither recommended nor likely to be an option. Most prenatal care remained in the domain of other women in the family who shared an expertise through apprenticeship and lore and for whom the shared experience of pregnancy and childbirth was a central feature of domestic culture.6
By the mid-1800s, European physicians arrived at several insights about what has come to be known as preeclampsia or eclampsia (“toxemia”), which laid the groundwork for future prenatal interventions.7 , 8 It had been recognized since ancient times that pregnancy could be accompanied by headaches and convulsions, and some speculated that the seizures of pregnant women were caused by the uterus. However, it was not always easy to differentiate between convulsions owing to epilepsy and those caused by pregnancy.9 Because of the discovery of new methods for studying components of the blood, the idea that many diseases—including convulsions in pregnancy—were associated with circulating toxins led to the term “toxemia.”10 The association of convulsions in pregnant women with proteinuria was established in 1843 by John Lever,7 working at Guy’s Hospital in London. When Lever decided to examine the urine of every pregnant woman that he saw with convulsions, he found albumen in the urine of all but 1 woman. He suggested that in such cases, rapid delivery was the best course of action. Because he did not continue to find albumen in the urine of these women after delivery, he concluded that the fundamental cause of the convulsions was not an intrinsic disease of the kidney, but was related to pregnancy.11 At about the same time, some physicians noted a hard bounding pulse in pregnant women having convulsions but lacked the technology necessary to measure blood pressure. However, in 1896, the Italian physician Riva-Rocci invented a sphygmomanometer that could easily measure blood pressure, and soon after that, physicians started to assert that hypertension might be an early marker of eclampsia.8 By associating eclampsia with diagnostic findings, such as proteinuria and hypertension, physicians could start to see a rationale for routine examination of pregnant women who were asymptomatic.
During this time, the US medical profession was becoming more organized. States passed licensing laws. The American Medical Association (AMA) was founded in 1846, followed by the formation of specialty groups, including the obstetrical societies in New York and Philadelphia.12 The American Gynecologic Society and American Association of Obstetricians and Gynecologists were founded in 1876 and 1888, respectively, with similar goals of promoting high-quality practice, education, and research.13 Although the societies’ names suggested a national presence, these organizations actually admitted relatively few members—most of whom lived on the East Coast, thus limiting their influence on obstetrical practice.14 However, the formation of these societies created a foundation for further specialization of the field of obstetrics, which would become the dominant platform for prenatal care delivery.
The Early 1900s
Around the turn of the 20th century, some physicians started to advocate for routine prenatal care as a method to reduce maternal and infant mortalities. In 1901, the Scottish practitioner John William Ballantyne pleaded for “promaternity wards”—not only to provide care for women with complications but also to study maternal and neonatal diseases in pregnancy. Later that year, he received funding from the Edinburgh Obstetric Society for the first antenatal bed in the Royal Maternity Hospital; eventually, this promaternity ward grew to over 23 beds.15 , 16 Ballantyne studied pregnancy using the latest technology of the day—in this case, X-rays and pulse measurement. He also emphasized the value of having patients seen by specialized practitioners who were skilled in obstetrical practice rather than general practitioners. However, this single hospital ward could obviously help only a limited number of women.
Ballantyne’s message was amplified in the United States through the American Journal of Obstetrics in 1901.17, 18, 19 Prominent leaders, such as Johns Hopkins obstetrician John Whitridge Williams, recognized the precariousness of health in pregnancy, stating “it is apparent that the border-line between health and disease is less clearly marked during gestation, and derangements… may readily give risk to pathological conditions which seriously threaten the life of mother, child, or both.”20
One of the first broad-based attempts at intervention came in 1901 in Boston, when public health nurses from the District Nursing Association began trying to reduce infant mortality by conducting home prenatal care visits with the Boston Lying-in Hospital for childbirth.21 , 22 New York City public health nurses followed in 1907.23 National attention became increasingly focused on high infant mortality rates. In 1909, the White House held a Conference on the Care of Dependent Children.24 Hoping that a new federal agency could improve children’s health throughout the country, US President William Howard Taft formed the Federal Children’s Bureau in 1912.25
A year later, in 1913, the Children’s Bureau released a slim booklet offering advice on prenatal care. It provided information on common symptoms and complications of pregnancy, preparation for childbirth, and hints for a smooth postpartum recovery. Of note, women were encouraged to consult with the doctor from the beginning of pregnancy, although the booklet noted that “he [sic] may have very little to do beyond giving advice and making the routine examinations of the urine [for protein].”26 The booklet did not offer advice on how often pregnant women needed to see their physician.
Infant deaths increasingly came to the attention of the medical profession. In his 1914 presidential address to the American Association for Study and Prevention of Infant Mortality, John Whitridge Williams presented a massive study of 10,000 consecutive admissions of pregnant women, with 705 fetal deaths. He concluded that 40% of infant deaths could be prevented with prenatal care. Williams outlined the ideal prenatal care plan: all women would present for an early prenatal visit and receive a full physical examination and Wassermann test (for syphilis). He suggested that a nurse visit every woman in her home to assess her “social situation” and that women return 1 month before delivery to assess for proper delivery location (home vs hospital). Of note, the author stressed the much worse outcomes for African American mothers.27 Williams was one of the most influential obstetricians of his time. He was the founding author of the dominant reference text, “Williams obstetrics,” which went through 17 editions from 1903 to 1985. Williams’ work not only raised concerns about prenatal deaths but also offered a systematic approach to improving outcomes through prenatal care.28
To systematically keep track of births throughout the country, the Census created the national birth-registration area in 1915, which provided national data to study the connection between prenatal care and infant and maternal deaths.29 If such data could demonstrate a connection between increasing care and better outcomes, it offered a means to improve health and an opportunity for physicians to strengthen their own position in the marketplace: “As the knowledge grows that the attendance of pregnancy and the guarding of young infant life are a great and important scientific function, the market will be created for good obstetric care.”30 Thus, prenatal care became not only a preventive service but also a reason for routine physician services.31
Women’s groups were becoming a political force in the Progressive Era. They worked to pass the 19th Amendment to the US Constitution in 1919, which enfranchised women. They went on to push for passage of the Sheppard-Towner Bill in 1921, which provided federal funding for 2987 prenatal care centers and public health nurses and community distribution of educational materials.32 However, funding was discontinued in 1929 following lobbying by the AMA that this was a “step toward socialized medicine.”22
Perhaps in response to the growing awareness of prenatal care’s ability to influence both infant and maternal outcomes, the Children’s Bureau published a new booklet on prenatal care in 1930. Unlike earlier publications, this guideline detailed a specific schedule for prenatal physician visits: monthly visits until 28 weeks’ gestation, biweekly visits until 36 weeks’ gestation, and weekly visits until delivery. In other words, depending on precisely how early a pregnancy was diagnosed, this was a recommendation for 12 to 14 visits during pregnancy. The booklet did not reference any evidence supporting the recommended visit schedule, nor did it specify how or if the schedule should be modified for patients with additional risk factors. The number of recommended visits remained remarkably unchanged over the years.
Subsequent editions of the booklet did reflect changing ideas and knowledge. For example, in 1942, updated booklets added recommendations for a public health or private nurse to help patients achieve recommended care.33 In 1949, the revised booklet acknowledged the role of the father in the pregnancy and birth process and the importance of social and emotional health.34 By the 1962 revision, mothers were admonished to seek a doctor with training and experience in delivering prenatal care, such as a specialist obstetrician. However, despite the many changes occurring in medical practice, new editions of the booklet continued to recommend the same schedule of 12 to 14 prenatal visits.35
During this period, prenatal care was not the only obstetrical service increasingly delivered by physicians. Birth moved from the home to the hospital as physicians continued to campaign for the medicalization of childbirth. In addition, the increasingly popular method of “twilight sleep delivery” (use of anesthetics during delivery to allow women to experience a pain-free childbirth) required physician supervision in a hospital setting, further medicalizing birth and prenatal care.36, 37, 38 By 1938, only about half of all births remained in women’s homes.39
Midcentury
Systematizing birth within the hospital supported the argument that births ought to be attended by specialists in obstetrical care—an argument consistent with a general trend toward the importance of specialization within medicine. Specialization came to be marked by certificates that were provided by private organizations (specialty boards), and in 1930, the incorporation of the American Board of Obstetricians and Gynecologists provided a formal mechanism by which physicians could legitimately claim particular expertise in caring for pregnant women.14 , 40
In 1951, the American Academy of Obstetrics and Gynecology (AAOG) was formed to serve the “average obstetrician gynecologist” by promoting high standards of practice, education, and research; promoting positive relationships with the public; and contributing to the scientific literature. In 1957, the name was changed to the American College of Obstetricians and Gynecologists (ACOG).12 Around the same time, Certified Nurse Midwifery became increasingly organized with the founding of the American College of Nurse Midwives in 1955.39
In 1959, the ACOG released their first “Manual of standards in obstetric-gynecologic practice” intended for a wide audience.41 The authors stressed that clinical practice was rapidly developing and that changes in their recommendations were to be expected. They upheld many of the recommendations of the previous Children’s Bureau pamphlets. A section went over fees and suggested a single bill that would include any needed operative procedures. The section on lay education did not contemplate any parental pairing other than the traditional husband and wife. In a nod to how care may have changed since previous generations, a separate section on discussion with the patient’s “mother and mother in law” suggested that the physician point out “differences in modern practice.” However, most significant for this paper, the ACOG saw no reason to reconsider the same 12 to 14 visit schedules that had first been articulated some 3 decades ago or to provide additional specifications for patients with varying levels of medical or social risk.
Just as earlier technological discoveries, such as the X-ray machine and the sphygmomanometer, had been used to improve prenatal care, the next few decades saw the introduction of several more technological innovations. The 1959 guidelines emphasized Rh testing, and the first clinical trial documenting the efficacy of Rh immunoglobulin for preventing alloimmunization was published in 1968.42 In the 1970s, radioimmunoassay detection of human chorionic growth hormone laid the foundation for earlier discovery of pregnancy and home pregnancy tests,43 whereas use of ultrasound and electronic fetal heart monitoring became routine in the late 1970s.44, 45, 46, 47 Genetic screening through amniocentesis and alpha fetal protein was introduced in the 1970s, with widespread adoption by the 1980s—predominantly for high-risk populations, including women of advanced maternal age—giving pregnant patients access to earlier diagnosis of genetic disorders and congenital anomalies.48 , 49 Additional changes included the first use of the Kessner Index (a composite measure of the timing of prenatal care initiation and total visit number completed) to assess the adequacy of prenatal care in 1970.50 Simultaneously, increasing ability of digital access to data enabled a detailed analysis of the impact of low birthweight as one of many racial disparities in the United States.51, 52, 53
In 1980, the US Surgeon General declared that a major national health objective was reduction of low birthweight infants.54 In 1982, the Institute of Medicine convened the Committee to Study the Prevention of Low Birthweight to investigate the most promising strategies for improving infant outcomes. Findings were published in 1985. The committee concluded that evidence supported the causal relationship between prenatal care and reduction of infants with low birthweight,53 , 55 estimating that $3.38 could be saved for every preventive dollar spent on prenatal care. Following the conference, several federal and state initiatives attempted to improve prenatal care access—particularly for low-income women—through Medicaid expansion and increased funding for prenatal care programs.56
The committee also called for a revision of prenatal care to “encourage the provision of improved, more flexible prenatal care services,” including use of medical and social assessments to determine appropriate care.53 Therefore, the Department of Health and Human Services commissioned the Public Health Service Expert Panel on the Content of Prenatal Care in 1989 to review the “effectiveness and efficiency of current prenatal care.”57 , 58 As they reviewed existing evidence, it became clear, as the panel’s chair concluded, that “the amazing and humbling message... was how little we knew.” Although data were insufficient to guide recommendations for a specific frequency of prenatal appointments, the committee felt comfortable recommending a flexible schedule of prenatal visits based on patients’ medical and social risk factors. Their proposed schedule included 7 visits for low-risk multiparous patients and 9 visits for low-risk nulliparous patients, with additional visits added as needed for high-risk patients based on medical and social risk factors. Interestingly, they suggested a phone visit for multiparous patients at 10 weeks’ gestation, perhaps a first step toward what we now see as telemedicine for prenatal care. In addition, the document advocated for preconception care, postpartum care extending through the first year after delivery, and a variety of social and mental health services designed to support the pregnant patient. The director of the National Institute of Child Health and Human Development, Dr Duane Alexander, anticipated controversy surrounding the new guidelines for the number of visits for low-risk patients, foreshadowing to 1 reporter “these changes will be fought by a lot of people.”59
As Alexander anticipated, this high-profile advice to cut down on prenatal visits attracted quick attention from the national media, including a front page article in the New York Times. It also drew attention from leading obstetrics and gynecology physicians.59 In 1990, the ACOG Executive Committee discussed the new recommendations. Even though ACOG members had been involved in the panel, the committee found the rationale for some changes to be unconvincing, reporting the panel’s “objectives were very broad and not always supported by data.” Perhaps unsurprisingly, they focused on the new visit schedule. Although existing historic data do not allow a detailed analysis of the committee’s discussions, they did note that “the data recommends reducing the number of prenatal visits for low-risk women on the assumption that this will produce more resources for those at risk of delivering prematurely. However, the organization of healthcare delivery services does not make such a direct transfer of resources possible” (October 1990 Executive Committee minutes, retrieved from the ACOG archives). Thus, the committee doubted (and was probably correct) whether saving money on fewer visits for low-risk patients would lead to more money for high-risk patients. To match prenatal services to patients’ needs, the new recommendations were apparently rejected for insufficient evidence. Of note, rejecting these recommendations implied maintaining an existing visit structure that was also not evidence based. In a 1991 commentary, 3 prominent ACOG members publicly questioned the new advice on visit timing, noting concern that lay press coverage might lead pregnant women to make fewer visits to their obstetrician.58 However, for low-risk women, that was, of course, precisely the point.
Although not mentioned in the brief comments recorded in the ACOG Committee minutes, payment incentives may have played a role in the deliberations. Most births in the 1980s and early 1990s were covered by commercial insurance60 and largely financed through a fee-for-service structure, which meant higher physician reimbursement for more prenatal visits.60, 61, 62 Although states implemented global fee structures for physician services within Medicaid as early as 1983, over 40% of private physicians refused to take patients with Medicaid.63 , 64 Therefore, at the time of the task force’s recommendations, physicians may have had significant financial motivation to maintain more intensive visit schedules. Although private payers may have had financial incentives to advocate for the new guidelines, they may have not pushed for changes because (1) prenatal care is relatively inexpensive, (2) reduced visit schedules were not widely supported by providers or specialty leadership, and (3) they wished to avoid covering other expanded services that the panel recommended, such as education and nutrition (Milton Kotelchuck, PhD, MPH, e-mail communication, September 27, 2020). It was not until managed care became more common later in the 1990s that global provider payments became ubiquitous, removing one of the incentives for more prenatal visits.65 It is also possible that patient and provider preferences drove resistance to the new visit schedule. Morton Lebow, an ACOG spokesperson, reflected that prenatal care was “based on experience, and that experience has been very good.”59
Some elements of the Public Health Service Expert Panel’s work were adopted by the ACOG in their guidelines, such as the emphasis on preconception visits, care tailoring, and psychosocial support.66 Simultaneously, the question as to how many visits were needed was studied more intensively. During the 1990s and early 2000s, clinical trials studied reduced visits for low-risk women and more intensive services—often known as “enhanced prenatal care”—for women at higher risk of preterm birth and low birthweight.67 , 68 A meta-analysis of more than 5000 patients from the United States and other high-income countries demonstrated equivalent maternal and neonatal outcomes when antenatal visits were reduced from 12 to 14 visits to 9 visits for low-risk patients.69 The World Health Organization has recommended an 8-visit schedule, with the use of women-held case notes, community-based interventions, and task-shifting components of prenatal care to community-based health workers to improve access and patient experience, particularly in low-resource settings.70 Although peer countries adopted reduced visit schedules for low-risk patients with no clear harmful effect, most major US maternity care organizations maintained the same visit schedule originally proposed in 1930.71 Attempting to reduce rates of preterm birth and low birthweight, public health researchers studied numerous other models of enhanced prenatal care, including increased case management, prenatal education, and better integration of social services.72 Most trials showed equivocal results, with large investments in prenatal care delivery not yielding significant changes in outcomes.73 , 74
Over the past decades, the United States has seen the introduction of still more innovative prenatal care delivery models. Group prenatal care, which includes enhanced education and relationship building, started in the 1990s and has recently enjoyed greater popularity.75 Some studies have documented improved patient outcomes, particularly for medically and socially complex patients.76 , 77 Starting in 2014, the University of Utah and the Mayo Clinic introduced new approaches to prenatal care, including virtual visits and leveraging nurse care managers. Preliminary evidence from these new telemedicine models demonstrates equivalent maternal and neonatal outcomes, high patient satisfaction, and even lower healthcare costs.78 , 79 However, further data are needed as results are from highly controlled trial settings; include largely homogenous, high-income patient populations; and are focused on low-risk patients. In recent years, significant innovation has been driven by the private sector, with startups, such as Babyscripts and Maven, offering consumers new, flexible ways to engage in prenatal care, through home monitoring, digital educational platforms, and telemedicine visits.80
However, despite all the new technologies that has been developed over the past century and despite all the new sciences and innovative ideas and techniques, the same 12 to 14 in-person prenatal visit schedule first advocated in 1930 has remained stubbornly and firmly in place until the COVID-19 pandemic.
The Coronavirus Disease 2019 Pandemic and Beyond
In March 2020, the world changed. Patients and providers became increasingly concerned about viral exposure in healthcare settings. In-person prenatal visits no longer seemed so benign.81 Practices across the country rapidly adopted reduced prenatal visit schedules, telemedicine, or hybrid care models. The ACOG endorsed the use of virtual care. In addition, for the first time since 1930, ACOG also endorsed reducing the number of prenatal visits.82 Early data on the feasibility and acceptability of both reduced visit schedules and telemedicine during this time are promising; however, data on outcomes, patient experience, and equity across diverse settings are still pending.81 , 83, 84, 85 The Table summarizes the key events that have shaped prenatal care delivery from the 1800s to today.Table Key events in the evolution of prenatal care delivery guidelines
Period Key events in prenatal care delivery
Early 1800s Prenatal care relatively unstructured and delivered by laypeople.
Mid-1800s Recognition of association among blood pressure, proteinuria, and preeclampsia or eclampsia.
Late 1800s Increasing organization of the medical profession.
1901 John Ballantyne (Edinburgh General Hospital) introduced “promaternity wards”; first home prenatal visits conducted by the Boston Lying-in Hospital.
1909 White House Conference on the Care of Dependent Children.
1912 The Children’s Bureau was formed.
1913 The Children’s Bureau released the first prenatal care booklet, recommending consultation with a physician early in pregnancy.
1914 John Whitridge Williams (Johns Hopkins Hospital) presented data suggesting that prenatal care can reduce infant mortality.
1915 The national birth-registration area was formed, providing national data on maternal and infant deaths.
1921 The Sheppard-Towner Bill was passed, providing federal funding for prenatal care.
1930 The Children’s Bureau released a second prenatal care booklet, with specific recommendations for physician visit schedule; the American Board of Obstetricians and Gynecologists first provides specialty certification.
1951 AAOG was formed.
1955 The ACNM was founded.
1957 The AAOG changed its name to the ACOG.
1959 The ACOG released the first “Manual of standards in obstetric-gynecologic practice,” which maintains the original prenatal visit schedule.
1970 The Kessner Index was introduced to assess the adequacy of prenatal care.
1985 Findings from the Institute of Medicine Committee to Study the Prevention of Low Birthweight were released, supporting the causal relationship between prenatal care and reduction of low birthweight infants.
1989 NIH Public Health Service Expert Panel on the Content of Prenatal Care that recommended a schedule of prenatal visits based on medical and social risk factors; Medicaid expansion occurred to improve prenatal care access.
1990s Clinical trials demonstrated the safety of reduced visit schedules for low-risk patients; group prenatal care was first introduced.
2019 The first RCT of the prenatal care model integrating telemedicine was published.
2020 Outbreak of COVID-19, which resulted in a pandemic.
AAOG, American Academy of Obstetrics and Gynecology; ACNM, American College of Nurse Midwives; ACOG, American College of Obstetricians and Gynecologists; COVID-19, coronavirus disease 2019; NIH, National Institutes of Health; RCT, randomized clinical trial.
Peahl. Evolution of prenatal care guidelines. Am J Obstet Gynecol 2021.
What comes next? As we transition out of the acute pandemic into our “new normal,” what can be learned from a century of prenatal care history? First, we should continue to be humbled by how little we know about appropriate prenatal care delivery. Although we now know more about what services are important for improving maternal and neonatal outcomes, we still lack key information on how to deliver them, and how often. The right visit number, frequency, and modality—in person, telemedicine, group care, etc.—remain elusive. Similarly, we continue to struggle with how best to tailor services to patients’ medical and social needs. However, after a century, we seem to be ready to seriously reconsider the prenatal visit schedule originally proposed in 1930.
Studying the history of prenatal care delivery guidelines reveals a recurrent flaw in our design of prenatal care delivery—namely, that we have ignored it. Therefore, we have treated visit frequency and modality as fixed boxes, into which we must fit an ever-changing set of care recommendations. The 1989 National Institutes of Health panel reconsidered this idea, recommending a prenatal visit schedule anchored around the delivery of key services that could be individualized to patients’ medical and social risk factors. More recently, at our institution, we have redesigned prenatal care based on 2 fundamental principles: designing care delivery around essential services and creating flexible services to address the needs of specific patients.81 , 86 It is important to note that this does not mean a universal reduction in the number of visits; medically high-risk patients may benefit from additional healthcare contacts, as would low-risk patients with psychosocial risk factors (eg, intimate partner violence, low support). Some of these additional services may be better delivered outside of routine in-person prenatal visits with physicians, through programs, such as home visiting programs,87 peer support,88 nutritional interventions,89 and numerous others. We do not have data to support a specific prenatal visit schedule, recommended number of telemedicine visits, or specifications of additional services, and we never have. However, one thing is clear: we are long overdue for new prenatal care delivery guidelines in the United States. The Figure provides an overview of how prenatal visit schedules have changed over time and what they may look like in the future.Figure Low-risk prenatal care guidelines from 1930 to 2020
ACOG, American College of Obstetricians and Gynecologists; COVID-19, coronavirus disease 2019; NIH, National Institutes of Health.
Peahl. Evolution of prenatal care guidelines. Am J Obstet Gynecol 2021.
Over 100 years after Ballantyne proposed “promaternity care,” his optimism for the future of prenatal care still rings true. Thinking back to those who called progress in prenatal care to be “fantastic, imaginary, and impossible,” he asked “who shall dare, in full remembrance of what has been accomplished in the past century, to set limits to the progress to be achieved in the present.”16 The COVID-19 pandemic has provided an opportunity for us to reflect on over a century of prenatal care delivery, incorporate what evidence has been gained, and strive to generate new knowledge to inform the next century of care for pregnant patients.
Acknowledgments
The authors would like to thank Dr Milton Kotelchuck for his historic insights and contribution to the manuscript and Mary Hyde, former senior director of the American College of Obstetricians and Gynecologists Resource Center, for her contributions to the historic research for this project. They would also like to thank Sarah Block for her assistance with the preparation of this manuscript. Ms Block is employed by the University of Michigan. Dr Kotelchuck, Ms Hyde, and Ms Block did not receive compensation for their contributions.
The authors report no conflict of interest.
==== Refs
References
1 Severe maternal morbidity in the United States. Centers for Disease Control and Prevention Available at: www.cdc.gov/reproductivehealth/maternalinfanthealth/severematernalmorbidity.html 2017
2 Shryock R.H. Medical licensing in America, 1650-1965 1967 Johns Hopkins University Press Baltimore, MD
3 Gunn J.C. Gunn’s domestic medicine. Knoxville, TN: printed under the immediate supervision of the author 1830 a physician of Knoxville
4 Buchan W. Domestic medicine 1848 Otis Broaders, and Company Boston, MA
5 Davis A.B. Appel T.A. Bloodletting instruments in the National Museum of History and Technology 1979 Smithsonian Institution Press Washington, DC
6 Merkatz I.R. Thompson J.E. New perspectives on prenatal care 1990 Elsevier New York, NY
7 Lever J.C.W. Cases of puerperal convulsions, with remarks Guy’s Hosp Rep 1 1843 495 517
8 Cook H.W. Briggs J.B. Clinical observations on blood pressure Johns Hopkins Hosp Rep 11 1903 451 534
9 Temkin O. The falling sickness: a history of epilepsy from the Greeks to the beginnings of modern neurology 2nd ed. 1971 Johns Hopkins University Press Baltimore, MD
10 Bell M.J. A historical overview of preeclampsia-eclampsia J Obstet Gynecol Neonatal Nurs 39 2010 510 518
11 Purkerson M.L. Vekerdy L. A history of eclampsia, toxemia and the kidney in pregnancy Am J Nephrol 19 1999 313 319 10213834
12 Mengert W.F. History of the American College of Obstetricians and Gynecologists 1950-1970 1971 American College of Obstetricians and Gynecologists Washington, DC
13 About the American Association of Obstetricians and Gynecologists Foundation 2020 The American Association of Obstetricians and Gynecologists Foundation Available at: https://www.aaogf.org/about.asp
14 American College of Obstetricians and Gynecologists History of the American College of Obstetricians and Gynecologists: 1951-2017 2019 American College of Obstetricians and Gynecologists Washington, DC
15 Ballantyne J.W. A plea for a pro-maternity hospital Br Med J 1 1901 813 814 20759529
16 Ballantyne J.W. Manual of antenatal pathology and hygiene 1902 William Green & Sons Edinburgh, Scotland
17 Ballantyne J.W. Visits to the wards of the pro-maternity hospital, a vision of the twentieth century Am J Obstet Dis Women Child 43 1901 596
18 Alexander G.R. Kotelchuck M. Assessing the role and effectiveness of prenatal care: history, challenges, and directions for future research Public Health Rep 116 2001 306 316 12037259
19 Romero R. Prenatal medicine: the child is the father of the man. 1996 J Matern Fetal Neonatal Med 22 2009 636 639 19736614
20 The management of normal pregnancy Williams J.W. Obstetrics: a text-book for the use of students and practitioners 2nd ed. 1910 D. Appleton and Company New York, NY 199 204
21 Worcester A. The Boston Lying-in Hospital fifty years ago N Engl J Med 209 1933 1109 1112
22 Thompson J.E. Walsh L.V. Merkatz I.R. The history of prenatal care: cultural, social, and medical contexts Merkatz I.R. Thompson J.E. New perspectives on parental care 1990 Elsevier New York, NY 9 30
23 Heaton C. Fifty years of progress in obstetrics and gynecology N Y State J Med 51 1951 83 85 14796874
24 Yarrow A.L. History of U.S. Children’s Policy 1900-present Available at: https://firstfocus.org/wp-content/uploads/2014/06/Childrens-Policy-History.pdf 2009
25 Centennial series: the creation of the Children’s Bureau. Children’s Bureau Express Available at: https://cbexpress.acf.hhs.gov/index.cfm?event=website.viewArticles&issueid=134§ionid=24&articleid=3463 2012
26 Prenatal care. United States Department of Labor, Children’s Bureau Available at: https://www.mchlibrary.org/history/chbu/2265-1930.PDF 1930
27 Williams J.W. The limitations and possibilities of prenatal care: based on the study of 705 fetal deaths occurring in 10,000 consecutive admissions to the obstetrical department of the Johns Hopkins Hospital J Am Med Assoc 64 1915 95 101
28 Williams J.W. ed. Obstetrics: a text-book for the use of students and practitioners. 1st ed 1905 D. Appleton and Company New York, NY
29 Shapiro S. Development of birth registration and birth statistics in the United States Popul Stud (Camb) 4 1950 86 111
30 Hill I.L. Extending the care of pregnancy JAMA 71 1918 885 887
31 Hsiao F. Controlling pregnancy: Fred Lyman Adair and the influence of eugenics on the development of prenatal care Available at: https://elischolar.library.yale.edu/ymtdl/3504 2019 Accessed May 10, 2020
32 Kobrin F.E. The American midwife controversy: a crisis of professsionalization Bull Hist Med 40 1966 350 363 5949832
33 Prenatal care United States Department of Labor, Children’s Bureau Available at: https://www.mchlibrary.org/history/chbu/2265-1942.PDF 1942
34 Prenatal care Federal Security Agency, Social Security Administration, Children’s Bureau Available at: https://www.mchlibrary.org/history/chbu/2265-1949.PDF 1949
35 Prenatal care United States Department of Health, Education, and Welfare, Office of Child Development, Children’s Bureau Available at: https://www.mchlibrary.org/history/chbu/2265-1962.PDF 1962
36 Litoff J.B. American midwives: 1860 to the present 1978 Greenwood Press Westport, CT
37 Wertz R.W. Wertz D.C. Lying-in: a history of childbirth in America 1989 Yale University Press New Haven, CT
38 McCool W.F. Simeone S.A. Birth in the United States: an overview of trends past and present Nurs Clin North Am 37 2002 735 746 12587371
39 Leavitt J.W. Brought to bed: childbearing in America, 1750-1950 1986 Oxford University Press New York, NY
40 Everett H.S. Taylor E.S. The history of the American Gynecological Society and the scientific contributions of its Fellows Am J Obstet Gynecol 126 1976 908 919 793392
41 American College of Obstetricians and Gynecologists Manual of standards in obstetric-gynecologic practice 1959 American College of Obstetricians and Gynecologists Chicago, IL
42 Pollack W. Gorman J.G. Freda V.J. Ascari W.Q. Allen A.E. Baker W.J. Results of clinical trials of RhoGAM in women Transfusion 8 1968 151 153 4173363
43 Vaitukaitis J.L. Braunstein G.D. Ross G.T. A radioimmunoassay which specifically measures human chorionic gonadotropin in the presence of human luteinizing hormone Am J Obstet Gynecol 113 1972 751 758 4673805
44 Goldberg B.B. Isard H.J. Gershon-Cohen J. Ostrum B.J. Ultrasonic fetal cephalometry Radiology 87 1966 328–32 passim
45 McLeary R.D. The radiologist’s role in obstetric ultrasound Radiology 137 1980 565 566 7433694
46 Kubli F.W. Hon E.H. Khazin A.F. Takemura H. Observations on heart rate and pH in the human fetus during labor Am J Obstet Gynecol 104 1969 1190 1206 5807980
47 Myers R.E. Mueller-Heubach E. Adamsons K. Predictability of the state of fetal oxygenation from a quantitative analysis of the components of late deceleration Am J Obstet Gynecol 115 1973 1083 1094 4632417
48 Burke W. Tarini B. Press N.A. Evans J.P. Genetic screening Epidemiol Rev 33 2011 148 164 21709145
49 Prenatal care United States Department of Health and Human Services Available at: https://www.mchlibrary.org/history/chbu/2265-1983.PDF 1983
50 Institute of Medicine. Infant Death: An Analysis by Maternal Risk and Health Care. Kessner DM, ed. Vol. 1. Washington, DC: National Academy of Sciences; 1973.
51 Eisner V. Brazie J.V. Pratt M.W. Hexter A.C. The risk of low birthweight Am J Public Health 69 1979 887 893 474844
52 United Nations, Department of Economic and Social Affairs, Statistics Division Demographic and Social Statistics Branch. United Nations demographic yearbook, 1981 1983 United Nations New York, NY
53 Committee to Study the Prevention of Low Birthweight Division of Health Promotion and Disease Prevention, Institute of Medicine. Preventing low birthweight 1985 National Academies Press Washington, DC
54 Brown S.S. Can low birth weight be prevented? Fam Plann Perspect 17 1985 112 118 3842660
55 Krans E.E. Davis M.M. Preventing low birthweight: 25 years, prenatal risk, and the failure to reinvent prenatal care Am J Obstet Gynecol 206 2012 398 403 21889122
56 Omnibus Budget Reconciliation Act of 1989, HR 3299, 101st Cong, Pub L No. 101–239 (1989-1990). Congress.gov Available at: https://www.congress.gov/bill/101st-congress/house-bill/3299
57 Caring for our future: the content of prenatal care A report of the Public Health Service Expert Panel on the Content of Prenatal Care. National Institutes of Health Available at: https://eric.ed.gov/?id=ED334018 1989
58 Rosen M.G. Merkatz I.R. Hill J.G. Caring for our future: a report by the expert panel on the content of prenatal care Obstet Gynecol 77 1991 782 787 2014096
59 Kolata G. Less prenatal care urged for most healthy women. New York Times. October 4, 1989:A1 Available at: https://www.nytimes.com/1989/10/04/us/less-prenatal-care-urged-for-most-healthy-women.html Accessed Oct. 12, 2020
60 Gold R.B. Kenney A.M. Singh S. Paying for maternity care in the United States Fam Plann Perspect 19 1987 190 206 3319681
61 Barriers to the use of prenatal care Brown S.S. Prenatal care: reaching mothers, reaching infants 1988 National Academy Press Washington, DC 54 87
62 Centers for Disease Control and Prevention (CDC) Adequacy of prenatal-care utilization--California, 1989-1994 MMWR Morb Mortal Wkly Rep 45 1996 653 656 8769473
63 Howell E.M. Brown G.A. Prenatal, delivery, and infant care under Medicaid in three states Health Care Financ Rev 10 1989 1 15
64 Orr M.T. Forrest J.D. The availability of reproductive health services from U.S. private physicians Fam Plann Perspect 17 1985 63 69 3842655
65 Schwalberg R. Mathis S.A. Giffen M. Mohamadi L. Zimmerman B. Sines E. Medicaid coverage of perinatal services: results of a national survey. The Henry J. Kaiser Family Foundation Available at: https://www.kff.org/wp-content/uploads/2013/01/medicaid-coverage-of-perinatal-services-results-of-a-national-survey-report.pdf 2013
66 American College of Obstetricians and Gynecologists Standards for obstetric gynecologic services 7th ed. 1989 American College of Obstetricians and Gynecologists Washington, DC
67 Medicaid prenatal care: states improve access and enhance services, but face new challenges. United States General Accounting Office Available at: https://www.gao.gov/assets/80/78856.pdf 1994
68 Alexander G.R. Korenbrot C.C. The role of prenatal care in preventing low birth weight Future Child 5 1995 103 120 7633858
69 Dowswell T. Carroli G. Duley L. Alternative versus standard packages of antenatal care for low-risk pregnancy Cochrane Database Syst Rev 7 2015 CD000934
70 WHO recommendations on antenatal care for a positive pregnancy experience. World Health Organization Available at: https://www.who.int/reproductivehealth/publications/maternal_perinatal_health/anc-positive-pregnancy-experience/en/ 2016
71 Friedman Peahl A. Heisler M. Essenmacher L.K. A comparison of international prenatal care guidelines for low-risk women to inform high-value care Am J Obstet Gynecol 222 2020 505 507 31962108
72 Hill I. Dubay L. Courtot B. Strong start for mothers and newborns evaluation: year 5 project synthesis. Urban Institute Available at: https://downloads.cms.gov/files/cmmi/strongstart-prenatal-finalevalrpt-v1.pdf 2018
73 Anum E.A. Retchin S.M. Strauss J.F. 3rd Medicaid and preterm birth and low birth weight: the last two decades J Womens Health (Larchmt) 19 2010 443 451 20141370
74 Krans E.E. Davis M.M. Strong start for mothers and newborns: implications for prenatal care delivery Curr Opin Obstet Gynecol 26 2014 511 515 25379768
75 Rising S.S. Centering pregnancy. An interdisciplinary model of empowerment J Nurse Midwif 43 1998 46 54
76 Byerley B.M. Haas D.M. A systematic overview of the literature regarding group prenatal care for high-risk pregnant women BMC Pregnancy Childbirth 17 2017 329 28962601
77 Carter E.B. Temming L.A. Akin J. Group prenatal care compared with traditional prenatal care: a systematic review and meta-analysis Obstet Gynecol 128 2016 551 561 27500348
78 Butler Tobah Y.S. LeBlanc A. Branda M.E. Randomized comparison of a reduced-visit prenatal care model enhanced with remote monitoring Am J Obstet Gynecol 221 2019 638.e1 638.e8
79 Virtual prenatal care. University of Utah Health Available at: https://healthcare.utah.edu/virtual-care/virtual-prenatal-care/ 2020
80 Marko K.I. Krapf J.M. Meltzer A.C. Testing the feasibility of remote patient monitoring in prenatal care using a mobile app and connected devices: a prospective observational trial JMIR Res Protoc 5 2016 e200 27864167
81 Peahl A.F. Smith R.D. Moniz M.H. Prenatal care redesign: creating flexible maternity care models through virtual care Am J Obstet Gynecol 223 2020 389.e1 389.e10
82 COVID-19 FAQs for obstetrician-gynecologists, obstetrics. American College of Obstetricians and Gynecologists Available at: https://www.acog.org/clinical-information/physician-faqs/covid-19-faqs-for-ob-gyns-obstetrics 2020
83 Peahl A.F. Powell A. Berlin H. Patient and provider perspectives of a new prenatal care model introduced in response to the coronavirus disease 2019 pandemic Am J Obstet Gynecol 2020 [Epub ahead of print]
84 Fryer K. Delgado A. Foti T. Reid C.N. Marshall J. Implementation of obstetric telehealth during COVID-19 and beyond Matern Child Health J 24 2020 1104 1110 32564248
85 Aziz A. Zork N. Aubey J.J. Telehealth for high-risk pregnancies in the setting of the COVID-19 pandemic Am J Perinatol 37 2020 800 808 32396948
86 Peahl A.F. Gourevitch R.A. Luo E.M. Right-sizing prenatal care to meet patients’ needs and improve maternity care value Obstet Gynecol 135 2020 1027 1037 32282594
87 Issel L.M. Forrestal S.G. Slaughter J. Wiencrot A. Handler A. A review of prenatal home-visiting effectiveness for improving birth outcomes J Obstet Gynecol Neonatal Nurs 40 2011 157 165
88 Dennis C.L. Dowswell T. Psychosocial and psychological interventions for preventing postpartum depression Cochrane Database Syst Rev 2 2013 CD001134
89 Soneji S. Beltrán-Sánchez H. Association of special supplemental nutrition program for women, infants, and children with preterm birth and infant mortality JAMA Netw Open 2 2019 e1916722
| 33316276 | PMC9745905 | NO-CC CODE | 2022-12-15 00:03:13 | no | Am J Obstet Gynecol. 2021 Apr 13; 224(4):339-347 | utf-8 | Am J Obstet Gynecol | 2,020 | 10.1016/j.ajog.2020.12.016 | oa_other |
==== Front
Am Heart J
Am Heart J
American Heart Journal
0002-8703
1097-6744
Mosby
S0002-8703(21)00155-1
10.1016/S0002-8703(21)00155-1
Article
Table of Contents
20 6 2021
8 2021
20 6 2021
238 iii
2019
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745917 | NO-CC CODE | 2022-12-15 00:03:15 | no | Am Heart J. 2021 Aug 20; 238:i-ii | utf-8 | Am Heart J | 2,021 | 10.1016/S0002-8703(21)00155-1 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Excerpta Medica
S0002-9343(22)00157-7
10.1016/S0002-9343(22)00157-7
Article
Table of Contents
4 4 2022
4 2022
4 4 2022
135 4 A8A10
2019
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745918 | NO-CC CODE | 2022-12-15 00:03:15 | no | Am J Med. 2022 Apr 4; 135(4):A8-A10 | utf-8 | Am J Med | 2,022 | 10.1016/S0002-9343(22)00157-7 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier
S0002-9378(21)01126-1
10.1016/S0002-9378(21)01126-1
Article
Table of Contents
27 11 2021
12 2021
27 11 2021
225 6 A5A14
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745919 | NO-CC CODE | 2022-12-15 00:03:15 | no | Am J Obstet Gynecol. 2021 Dec 27; 225(6):A5-A14 | utf-8 | Am J Obstet Gynecol | 2,021 | 10.1016/S0002-9378(21)01126-1 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier
S0002-9378(20)30678-5
10.1016/S0002-9378(20)30678-5
Article
Table of Contents
28 7 2020
8 2020
28 7 2020
223 2 A3A11
2020
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745920 | NO-CC CODE | 2022-12-15 00:03:15 | no | Am J Obstet Gynecol. 2020 Aug 28; 223(2):A3-A11 | utf-8 | Am J Obstet Gynecol | 2,020 | 10.1016/S0002-9378(20)30678-5 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Elsevier Inc.
S0002-9343(21)00080-2
10.1016/j.amjmed.2021.01.007
Commentary
Cardiac Rehab in the COVID-19 Pandemic
Pecci Cristina DO a
Ajmal Muhammad MD b⁎
a Cardiology Fellow, College of Medicine, University of Arizona, Phoenix
b Cardiology Fellow, College of Medicine, University of Arizona, Tucson
⁎ Requests for reprints should be addressed to Muhammad Ajmal, MD, Cardiology Fellow University of Arizona, School of Medicine, Tucson, AZ, 85719.
10 2 2021
5 2021
10 2 2021
134 5 559560
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcTraditional cardiac rehabilitation is a medically supervised program that has played a pivotal role in reducing future cardiac events and improving overall quality of life in patients with cardiac diseases. Unfortunately, cardiac rehab has been underused due to a multitude of factors, and evidence shows that only 30% of eligible patients participate in it. Between 2007 and 2011, only 16.3% of patients with Medicare and 10.3% of veterans participated in cardiac rehab.1 Participation was even lower in women, older adults, and individuals from underserved populations.1 , 2 With the evolution of the coronavirus disease 2019 (COVID-19) crisis, most cardiac rehab programs have shut down completely to maintain social distancing. According to statistics from Transcatheter Cardiovascular Therapeutics Data, 2685 cardiac rehab programs that provide services to hundreds of thousands of patients have closed.
However, although the concept of traditional rehab involves a face-to-face encounter between physician and patient, home rehab may provide a sound alternative. According to Dr Randal Thomas, preventative cardiologist and medical director of Mayo Clinic's Cardiac Rehabilitation Program, home-based cardiac rehabilitation is pivotal in keeping patients out of the hospital, enforcing social distancing among high-risk patients, promoting healthier eating at home, improving mental health, and encouraging patients to quit smoking. The American Heart Association and the American Association of Cardiovascular and Pulmonary Rehabilitation published a recent meta-analysis showing that all-cause mortality data for up to 12 months demonstrated no statistically significant differences between center-based and home-based rehab.1
Some studies demonstrate that the expansion of home-based rehab might improve participation rates. One of the first studies conducted to assess the feasibility and effectiveness of cardiac rehab at home showed that remote telephone-based delivery of rehab is a viable alternative to center-based rehab.3 , 4 In this study, half of the eligible patients refused to participate in rehab at all due to lack of interest. Of those that were offered to choose between home-based rehab and center-based rehab, 77% chose home-based rehab. Participants in remote rehab were highly satisfied with their care and completion rates approached 89% compared with 73% of those doing face-to-face rehab.3 , 4 Furthermore, costs for each program were similar,3 , 4
With the current COVID-19 crisis, telemedicine has become a necessity in providing health care to patients. Unfortunately, the Centers for Medicare and Medicaid Services (CMS) is not reimbursing cardiac rehab telehealth services. According to Randal J. Thomas and Laurence Sperling, a request for emergency, temporary coverage of home-based cardiac rehab has been issued to CMS.5 The American Association of Cardiovascular and Pulmonary Rehabilitation (AACVPR) has been at the forefront of this battle. This organization has established the Innovative Delivery Model Collaborative (IDMC), a multidisciplinary coalition of volunteers and professionals that work together to foster the successful implementation of home-based programs that meet the traditional standard of care.
So, what exactly does home-based cardiac rehab look like? According to Jonathan Whiteson, MD, medical director of cardiac and pulmonary rehabilitation at New York University, Langone Rusk Rehab, home-based rehab appears unique in different settings. Some patients are being observed in real time, whereas others are given instructions to do on their own with periodic follow-up. Aside from a structured exercise regimen, patients can be monitored with devices and even receive virtual visits from nutritionists and psychologists.6 , 7
The first meta-analysis on telehealth exercised-based cardiac rehab was published in 2016. The main findings of the 11 randomized trials included were that telehealth cardiac rehab appeared to be at least as effective and, in some cases, more effective for improving cardiovascular risk factors, enhancing physical activity levels, improving diastolic blood pressure, reducing cholesterol, and increasing functional capacity.7
A recent randomized controlled noninferiority trial study compared the effectiveness of real-time remote rehab to traditional rehab over 12 weeks. The remote rehab arm comprised a smart phone and a chest-worn wearable sensor. During exercise training, participants’ heart and respiratory rates, single-lead electrocardiogram (ECG), and geopositional data were displayed in the smartphone app for self-monitoring, streamed to a web server, and visualized in the web app for specialist review. Outcomes assessed at baseline and 12 weeks included maximal oxygen uptake (VO2 max), modifiable cardiovascular risk factors, exercise adherence, motivation, quality of life and program delivery, and medication costs.8 In this study involving 162 participants, maximal oxygen uptake was comparable in both groups at 12 weeks and remote rehab was noninferior to center-based rehab.8
It is important to note that free resources remain available and include public webcasts, YouTube videos, online education, and reading material. The AACVPR is continuously providing information on upcoming webcasts. One such webcast was given by the Canadian association of Cardiovascular Prevention and Rehabilitation on April 8, 2020. In this free webcast, the panel discussed protocols and strategies to transition on-site rehabs to virtual or home-based models. YouTube has become another essential resource. The Henry Ford Health System in Detroit, Michigan, has offered a free, extensive home cardiac rehab program that offers educational materials to any patient around the world on YouTube. The University of Michigan provides a free online step-based aerobic program for patients at home.
Traditional education materials remain options for patients without Internet access. The United Kingdom Heart manual is the most extensively studied self-management book for patients recovering from acute coronary syndromes or open-heart surgery.1 , 2 The American Heart Association, University Health Network Toronto Rehabilitation Institute, National Heart Foundation of Australia, and the Veterans Health Administration offer similar textbooks. Each book is tailored to a patient's specific needs and offers secondary prevention topics.
Although the options for home-based care continue to evolve, it is important that all home rehab programs have the content and structure of traditional and center-based care. The AACVPR stresses that a physician medical director should lead all programs and that all patients should undergo a baseline assessment and receive an individualized treatment plan. This plan should consist of nutrition education, weight counseling, mental health evaluation, risk factor control, tobacco cessation, and exercise training.
In conclusion, although many center-based traditional rehab centers have shut down during the COVID-19 crisis, remote cardiac rehab remains a viable alternative. Virtual telehealth allows patients to exercise in their own homes or other fitness centers while being supervised by a physician or other health care professional by live, 2-way video teleconferencing. Not only does this improve the quality of life for patients during the pandemic, but it also serves to overcome barriers such as transportation and social issues that prohibited patients from normally attending center-based rehab. As Dr Whiteson notes, “It's much better for patients to be engaged with us and to have some guidance than to be left at home on their own and not know what to do.”
Funding: None.
Conflicts of Interest: None.
Authorship: Both authors had access to the data and a role in writing this manuscript.
==== Refs
References
1 Thomas RJ Beatty AL Beckie TM Home-based cardiac rehabilitation: a scientific statement from the American Association of Cardiovascular and Pulmonary Rehabilitation, the American Heart Association, and the American College of Cardiology J Am Coll Cardiol 140 2019 e69 e89
2 Heran BS Chen JM Ebrahim S Exercise-based cardiac rehabilitation for coronary heart disease Cochrane Database Syst Rev. 7 2011 CD001800
3 Wakefield B Drwal K Scherubel M Klobucar T Johnson S Kaboli P. Feasibility and Effectiveness of Remote, Telephone-Based Delivery of Cardiac Rehabilitation Telemed J E Health 20 2014 32 38 24161003
4 Williams MA Ades PA Hamm LF Clinical evidence for a health benefit from cardiac rehabilitation: An update Am Heart J 152 2006 835 841 17070142
5 McKeown LA. Cardiac rehab during COVID-19: telehealth, unpaid heroes step up to help at home. Available at:https://www.tctmd.com/news/cardiac-rehab-during-covid-19-telehealth-unpaid-heroes-step-help-home. Accessed on December 14, 2020.
6 Maddison R Rawstorn JC Stewart RAH Effects and costs of real-time cardiac telerehabilitation: randomised controlled non-inferiority trial Heart 105 2019 122 129 30150328
7 Rawstorn JC Gant N Direito A Beckmann C Maddison R Telehealth exercise-based cardiac rehabilitation: a systematic review and meta-analysis Heart 102 2016 1183 1192 10.1136/heartjnl-2015-308966 26936337
8 Maddison R Rawstorn JC Rolleston A The remote exercise monitoring trial for exercise-based cardiac rehabilitation (REMOTE-CR): a randomised controlled trial protocol BMC Public Health 14 2014 1236 25432467
| 33577752 | PMC9745923 | NO-CC CODE | 2022-12-15 00:03:15 | no | Am J Med. 2021 May 10; 134(5):559-560 | utf-8 | Am J Med | 2,021 | 10.1016/j.amjmed.2021.01.007 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Published by Elsevier Inc.
S0002-9343(21)00709-9
10.1016/j.amjmed.2021.10.011
Commentary
Physician Burnout: Fix the Doctor or Fix the System?
Greep Nancy C. MD a⁎
Woolhandler Steffie MD, MPH bc
Himmelstein David MD bc
a Private Practice, Santa Barbara, Calif
b School of Public Urban Health, City University of New York at Hunter College, New York, NY
c Lecturer, Harvard Medical School, Boston, Mass
⁎ Requests for reprints should be addressed to Nancy C. Greep, MD, 542 Litchfield Lane, Santa Barbara, CA, 93109.
1 11 2021
4 2022
1 11 2021
135 4 416417
© 2021 Published by Elsevier Inc.
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcBurnout—emotional exhaustion and depersonalization (ie, treating patients like objects)—affects almost half of American physicians, even before the coronavirus disease 2019 (COVID-19) pandemic.1 Burnout diminishes the quality of physician's lives2 and undermines patient care by leading to more medical errors and physicians leaving practice.3
Physicians report that administrative burdens (eg, filling out forms, dealing with multiple formularies, prior authorization and network restrictions, and documenting quality metrics) are important contributors to burnout. Other factors include cumbersome electronic health records (EHRs), productivity pressures, long work hours, poor work-life balance, and loss of autonomy.4
Two types of remedies for burnout have been suggested: personal interventions to decrease stress and improving the work environment. The former category includes everything from resilience training, mindfulness, yoga, battle buddies and psychotherapy to doctors’ dining rooms. However, these approaches, many of which put the onus of fixing burnout on already overburdened and time-pressured physicians, have mostly had little or no enduring effect.4
Suggestions for improving the work environment have included improving patient flow, using scribes, enlarging workspaces, having a social worker readily available, flexible scheduling, more vacations, and an executive wellness officer.5 Although many such changes have merit, they generally require protected time for the physician, additional personnel or space, and expense. Moreover, they would not address the underlying cause of burnout.
As with most maladies, finding a cure requires understanding the cause. In the case of burnout, the main cause is not nonresilient physicians or suboptimal work environments. Rather it is our failed health care system, a system characterized by impaired access to care, high and uncontrolled costs, and suboptimal quality care for many. These shortcomings stem from a dysfunctional insurance system and the invasion of the profit motive into what should be a public service.
The insurance system's dysfunction is attributable to the multiplicity of health insurers, their pursuit of profit, and the tethering of private insurance to employment. A multipayer system leads to billing nightmares. The profit motive means insurers seek to decrease costs by restricting care through prior authorizations and narrow networks. Because private insurance is tied to employment, patients lose their coverage, and often their doctors, when they lose their jobs, or their employer switches insurers. As a result, physicians are deprived of the gratification derived from long-term relationships with their patients.
For-profit enterprises have procured many hospitals, physician practices, dialysis units, nursing homes, home health agencies, and other health care entities.6 They prioritize their bottom line, a priority that often conflicts with physicians’ obligation to prioritize optimal care.7 The drive for profit, which has infected even many nonprofit health care organizations, has reduced physicians’ clinical autonomy and changed their role from leaders to expendable, albeit it, highly paid workers.8
To address burnout, we advocate moving to a single-payer, nonprofit insurance system, eliminating patients’ out-of-pocket costs, and removing the profit motive from patient care. We are not alone: A single-payer system has been endorsed by the American College of Physicians and the Society of General Internal Medicine. How might such reform mitigate physician burnout?
Access: Because everyone would be equally and well insured, there would be no need to check whether a patient had insurance, a drug was on their insurer's formulary, or a specialist to whom one wants to refer a patient was in the patient's network. No time would be wasted searching for charity care or workarounds for uninsured or underinsured patients. Prior authorization requirements would be largely eliminated, as they have in many nations. Billing would be streamlined, and billing personnel redeployed to facilitate better care. Coding would be simplified and documentation for quality metrics decreased. EHRs could be redesigned for patient care rather than billing, and current commercially driven impediments to interoperability reduced. With administrative tasks decreased, and EHRs made more user friendly, physicians could spend more time facing patients and less time interacting with a computer—both during and after visits.
Cost: Physicians’ offices would not have to collect copays or determine if a patient had met their annual deductible. Physicians could prescribe medications, imaging studies, and ancillary tests without worrying whether a patient could afford them.
Quality: Freeing clinicians from administrative and EHR burdens would allow them to devote more time to their patients and to thinking and learning. Eliminating insurance-mandated changes in providers would facilitate long-term doctor-patient relationships, a better understanding of patients’ priorities and concerns, and the avoidance of redundant workups that often result from breaks in continuity. Patients would no longer have to skip medications because of cost. The pressure to meet productivity targets would be reduced, decreasing one source of errors. Care would be directed by clinicians and not by bureaucrats and businessmen.
Converting to a nonprofit single-payer system will not be easy. For-profit medical enterprises will need to convert to nonprofit status and our multipayer system of insurance replaced by a single insurer, the federal government. Although a majority of the public and of physicians already support such reform,9 congresspeople, many of whom receive large donations from the corporate medical industrial complex, will need to vote for it.
It is unrealistic to expect that burnout would completely disappear with single payer reform. Care of sick patients will remain emotionally and physically taxing in any health care system. Stress from maintenance of certification would continue. Although a single-payer system would help alleviate health disparities, caring for disadvantaged patients who face challenges the physician cannot fix, like homelessness, would continue to cause provider distress.
We recognize that our view is somewhat speculative as little hard data is available on the effects of single payer on burnout. However, a recent Canadian Medical Association survey offers optimism. More than 80% of Canadian physicians said that their emotional and psychologic well-being was high, and only 30% met criteria for burnout.10 To verify whether the level of burnout is lower in a nonprofit single-payer system such as Canada, a study using the same questionnaire in both countries at the same time would be informative.
Funding: None.
Conflicts of Interest: None.
Authorship: All authors had access to the data and a role in writing this manuscript.
==== Refs
References
1 Shanafelt T Hasan O Dyrbye L. Changes in burnout and satisfaction with work-life balance in physicians and the general US working population between 2011 and 2014 Mayo Clin Proc 90 12 2015 1600 1613 10.1016/j.mayocp.2015.08.023 26653297
2 Dyrbye LN Varkey P Boone SL Satele DV Sloan JA Shanafelt TD. Physician satisfaction and burnout at different career stages Mayo Clin Proc 88 12 2013 1358 1367 10.1016/j.mayocp.2013.07.016 24290109
3 Tawfik DS Profit J Morgenthaler TI Physician burnout, well-being, and work unit safety grades in relationship to reported medical errors Mayo Clin Proc 93 11 2018 1571 1580 10.1016/j.mayocp.2018.05.014 30001832
4 West CP Dyrbye LN Shanafelt TD. Physician burnout: contributors, consequences and solutions J Intern Med 283 6 2018 516 529 10.1111/joim.12752 29505159
5 Gergen Barnett KA. In Pursuit of the Fourth Aim in Health Care Med Clin North Am 101 5 2017 1031 1040 10.1016/j.mcna.2017.04.014 28802466
6 Physicians for a National Health Program. The case against privatization. Available at: https://pnhp.org/2018/12/03/the-case-against-privatization-of-u-s-health-care/. Accessed September13, 2021.
7 Relman AS. The problem of commercialism in medicine Camb Q Healthc Ethics 16 04 2007 375 10.1017/S0963180107070466 18018916
8 Relman AS. The new medical-industrial complex N Engl J Med 303 17 1980 963 970 10.1056/nejm198010233031703 7412851
9 Pew Research Center. More Americans now favor single payer health coverage than in 2019. Available at:https://www.pewresearch.org/fact-tank/2020/09/29/increasing-share-of-americans-favor-a-single-government-program-to-provide-health-care-coverage/. Accessed September 20, 2021.
10 Canadian Medical Association. National Physician Health Survey 2018. Available at:https://www.cma.ca/physician-wellness-hub/resources/measurements-outcomes/2018-national-physician-health-survey/. Accessed June 17, 2021.
| 34736881 | PMC9745939 | NO-CC CODE | 2022-12-15 00:03:17 | no | Am J Med. 2022 Apr 1; 135(4):416-417 | utf-8 | Am J Med | 2,021 | 10.1016/j.amjmed.2021.10.011 | oa_other |
==== Front
Am J Surg
Am J Surg
American Journal of Surgery
0002-9610
1879-1883
Excerpta Medica
S0002-9610(21)00357-3
10.1016/S0002-9610(21)00357-3
Article
Table of Contents (3 pgs)
26 7 2021
8 2021
26 7 2021
222 2 A3A6
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745940 | NO-CC CODE | 2022-12-15 00:03:17 | no | Am J Surg. 2021 Aug 26; 222(2):A3-A6 | utf-8 | Am J Surg | 2,021 | 10.1016/S0002-9610(21)00357-3 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Excerpta Medica
S0002-9343(22)00410-7
10.1016/S0002-9343(22)00410-7
Article
Table of Contents
16 6 2022
7 2022
16 6 2022
135 7 A6A10
2019
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745941 | NO-CC CODE | 2022-12-15 00:03:17 | no | Am J Med. 2022 Jul 16; 135(7):A6-A10 | utf-8 | Am J Med | 2,022 | 10.1016/S0002-9343(22)00410-7 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier
S0002-9378(20)30476-2
10.1016/S0002-9378(20)30476-2
Article
Table of Contents
28 5 2020
6 2020
28 5 2020
222 6 A3A8
2020
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9745944 | NO-CC CODE | 2022-12-15 00:03:17 | no | Am J Obstet Gynecol. 2020 Jun 28; 222(6):A3-A8 | utf-8 | Am J Obstet Gynecol | 2,020 | 10.1016/S0002-9378(20)30476-2 | oa_other |
==== Front
Am J Surg
Am J Surg
American Journal of Surgery
0002-9610
1879-1883
Elsevier Inc.
S0002-9610(22)00165-9
10.1016/j.amjsurg.2022.03.017
Editorial
More than a moment: Why surgical societies must continuously strive for diversification and creating a culture of equity and inclusivity in academic surgery
Clarke Callisia N.
Division of Surgical Oncology, Medical College of Wisconsin, 8701 W. Watertown Plank Road, Milwaukee, WI, 53226, United States
17 3 2022
7 2022
17 3 2022
224 1 292293
15 3 2022
15 3 2022
© 2022 Elsevier Inc. All rights reserved.
2022
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcThe year 2020 arrived with a bang as a pandemic of epic proportions rattled our healthcare system to its core. We all tuned in to news broadcasts and radio shows to keep us updated on emerging details about COVID-19. That, along with the fact that our usual distractors: sports, movies, travel and even schools, were all paused, resulted in a captive audience that most desperately clung to the media for any bit of news regarding the path forward. With so few distractors, the entire country was also forced to bear witness to some ugly truths. A pandemic killing brown and black Americans disproportionately, the murders of George Floyd and others, the rise in Anti-Asian violence, and the mistreatment of Hispanic immigrants at the border were all on display. The combination created an awakening in many Americans who were previously either unaware or unbothered by ongoing racial and ethnic prejudices that exist within our society.
Academic surgery was not a stranger to this awakening. More than ever, institutions and societies acknowledged the impact of bias and microaggressions on diversity and inclusivity in our profession. Many quickly pivoted to at least acknowledge, and at times address, these shortcomings with statements and pledges to address inequities. Perhaps for the first time, we as a profession were unified in the desire to make necessary change as we strive to achieve true diversity, justice, equity and inclusion in academic surgery and surgical leadership. But these efforts must extend far beyond the moment. Two years into the pandemic, not much has changed in academia, and the necessary work to diversify our surgeon workforce is just beginning.
The representation of ethnic and racial minority groups in the United States population is increasing at a pace much faster than earlier predicted; by 2045, it is anticipated that White Americans will no longer be a majority.1 As the demographic shifts in our population continue, the potential for widening the already concerning health disparities by race, ethnicity, sex, gender identity and socioeconomic status is of major concern and has been identified as a top public health priority.2 Racial and ethnic minorities are more likely to provide care for medically underserved communities, and patients of color are more likely to seek out physicians of color to provide their care.3, 4, 5 Concordance in physician race and gender has been found to positively influence preventative services and patient satisfaction.6 , 7 As such, efforts to eliminate health disparities must focus on creating a physician workforce that reflects our patient population.
In 2019, 60.1% of the US population identified as White, 18.5% Hispanic, 12.5% Black, 5.8% Asian, 2.2% other, and 0.7% American Indian/Alaska Native.1 During the same year, the AAMC US Physician Workforce Data demonstrated that Hispanics, Blacks and Native Americans remained grossly underrepresented in medicine, accounting for only 5.8%, 5.0% and 0.3% of practicing physicians respectively.8 Surgery in particular faces a steep challenge in attaining the necessary physician diversity to meet our patient needs. Low racial and ethnic diversity in surgery is persistent; only 13.6% of students matriculating into medical school are underrepresented in medicine (URM), with Blacks accounting for only 7.1%, and Hispanics accounting for 6.2% of matriculating medical students in 2018.8 More disappointingly, there has been no significant change in these numbers over the past 30 years, despite efforts to increase the pipeline for URMs into medicine and other STEM careers. Even fewer URMs enter general surgery residencies. In 2020, only 5.3% of surgical trainees identified as Black, 8.1% as Hispanic/Latino, 0.8% as American Indian and 0.3% as Pacific Islander.9
The number of individuals who identify as members of the lesbian, gay, bisexual, transgender, queer, and others community (LGBTQ+) continues to increase with 5.4% of US adults identifying as LGBTQ+.10 Approximately 5–11% of surgical residents identify as LGBTQ+, but a significant proportion report harassment and mistreatment throughout their training—likely impacting retention in academia after residency completion.11 , 12 The paucity of data on the LGBTQ + surgical workforce highlights the lag in coordinated efforts to specifically address the needs of our LGBTQ + patients and colleagues.
Lack of diversity within surgery, however, cannot be not simply attributed to problems in the pipeline. While almost half of students matriculating into medical school are women, fewer women in general are entering surgical residencies. After completing surgical training, women and minorities are less likely to enter academic practices, and those that do are faced with limited opportunities for advancement and promotion when compared to their white male colleagues.13 , 14 There has been an approximately 3-fold increase in the number of female surgical faculty over the last three decades. Still, women only account for a quarter of assistant professors, 20% of associate professors, 10% of full professors and 15% of surgical chairs despite embodying 40% of surgical trainees.13 , 15 , 16 When controlled for rank, female surgical faculty were often compensated significantly less than male counterparts.17, 18, 19 The gender gap in leadership of surgical societies persists despite increased female representation in the field. A recent analysis of gender parity in general surgical societies noted 83% male representation amongst society leaders when compared to only 17% female.20 Other studies show that these disparities are most pronounced at the highest echelons of leadership, with the largest gender gap seen at the level of society president.21
Asian Americans in medicine are a heterogenous racial and ethnic group largely consisting of individuals who identify as being of Indian, Chinese or Korean heritage. Asian Americans constitute 21% of medical students but only 12% of surgical trainees.22 While well represented in surgery, Asian Americans are underrepresented in surgical leadership, accounting for 12% surgical chairs of departments and only 2.3% of governing boards of surgical societies.16 , 22
So how do we combat these barriers to diversity and inclusivity in surgery? Academic institutions and surgical societies carry the brunt of the responsibility in working to correct these issues. After all, we construct the systems that shape the experiences and careers of minoritized students, trainees and junior faculty, often serving as the gate keepers that decide who gets the ‘privilege’ of entry into medical school, residencies, and fellowships. If we are to really address the issues of lack of diversity and inclusivity in surgery, we must start here first.
Fortunately, organizations such as the Association of Women Surgeons (AWS), the Latino Surgical Society (LSS), the Society of Asian Academic Surgeons (SAAS), the Society of Black Academic Surgeons (SBAS) and the Society of Out Surgeons and Allies (SOSA) exist in part to support the professional development of their respective members. They may serve as a significant source of mentorship and sponsorship for those who may not be able to identify mentors and sponsors at their home institutions. The Association for Academic Surgery (AAS) and the Society of University Surgeons (SUS) are committed to improving diversity, equity and inclusion in academic surgery by supporting our sister societies and championing the diversity of our members through targeted courses, programs and antiracist educational content. These efforts require intentionality and financial investment, and they must continue for the long term—a succession of moments—to achieve to true equity and justice. In this focused issue of the American Journal of Surgery, we will discuss the needs of minoritized populations in academic surgery and targeted efforts to address them based on the “Together We Rise” sessions from the Academic Surgical Congress 2021 Virtual Meeting.
Declaration of competing interest
The authors report no relationships that could be construed as a conflict of interest.
==== Refs
References
1 >U.S. Census Bureau Population Data Retrieved from 2020 2020 https://usafacts.org/data/topics/people-society/population-and-demographics/population-data/population/#chart-114185-1
2 2/2/2022]; Available from: https://www.healthypeople.gov/2020/About-Healthy-People/Development-Healthy-People-2030.
3 Xu G. The relationship between the race/ethnicity of generalist physicians and their care for underserved populations Am J Publ Health 87 5 1997 817 822
4 Komaromy M. The role of black and Hispanic physicians in providing health care for underserved populations N Engl J Med 334 20 1996 1305 1310 8609949
5 Saha S. Patient-physician racial concordance and the perceived quality and use of health care Arch Intern Med 159 9 1999 997 1004 10326942
6 Henderson J.T. Weisman C.S. Physician gender effects on preventive screening and counseling: an analysis of male and female patients' health care experiences Med Care 39 12 2001 1281 1292 11717570
7 Schmittdiel J. Effect of physician and patient gender concordance on patient satisfaction and preventive care practices J Gen Intern Med 15 11 2000 761 769 11119167
8 Association of American Medical Colleges Diversity in Medicine: Facts and Figures 2019 Retrieved from 2019 2019 https://www.aamc.org/data-reports/workforce/interactive-data/figure-18-percentage-all-active-physicians-race/ethnicity-2018
9 Association of American Medical Colleges Report on Residents Retrieved from 2020 2020 https://www.aamc.org/data-reports/students-residents/interactive-data/report-residents/2020/table-b5-md-residents-race-ethnicity-and-specialty
10 Jones J. LGBT Identification Rises to 5.6% in Latest U.S Estimate GALLUP® Feb 24, 2021 2021 September 5, 2021; Available from: https://news.gallup.com/poll/329708/lgbt-identification-rises-latest-estimate.aspx
11 Heiderscheit E.A. Experiences of LGBTQ+ residents in US general surgery training programs JAMA Surg 157 1 2022 23 32 34668969
12 Lee K.P. Attitude and perceptions of the other underrepresented minority in surgery J Surg Educ 71 6 2014 e47 52 24974336
13 Abelson J.S. The climb to break the glass ceiling in surgery: trends in women progressing from medical school to surgical training and academic leadership from 1994 to 2015 Am J Surg 212 4 2016 566 572 e1 27649976
14 Yeo H.L. Association of time to attrition in surgical residency with individual resident and programmatic factors JAMA Surg 153 6 2018 511 517 29466536
15 Hu Y.Y. Discrimination, abuse, harassment, and burnout in surgical residency training N Engl J Med 381 18 2019 1741 1752 31657887
16 Kassam A.F. Gender and ethnic diversity in academic general surgery department leadership Am J Surg 221 2 2021 363 368 33261852
17 Hoops H.E. Analysis of gender-based differences in surgery faculty compensation, promotion, and retention: establishing equity Ann Surg 268 3 2018 479 487 30063494
18 Freund K.M. Inequities in academic compensation by gender: a follow-up to the national faculty survey cohort study Acad Med 91 8 2016 1068 1073 27276007
19 Jena A.B. Olenski A.R. Blumenthal D.M. Sex differences in physician salary in US public medical schools JAMA Intern Med 176 9 2016 1294 1304 27400435
20 Wu B. Gender disparity in leadership positions of general surgical societies in North America, Europe, and Oceania Cureus 11 12 2019
21 Lyons N.B. Gender disparity in surgery: an evaluation of surgical societies Surg Infect 20 5 2019 406 410
22 Kuo L.E. Asian American and Pacific Islander experiences and challenges in academic surgery Am J Surg 223 1 2022 211 213 34353621
| 35337646 | PMC9745945 | NO-CC CODE | 2022-12-15 00:03:17 | no | Am J Surg. 2022 Jul 17; 224(1):292-293 | utf-8 | Am J Surg | 2,022 | 10.1016/j.amjsurg.2022.03.017 | oa_other |
==== Front
Am J Cardiol
Am J Cardiol
The American Journal of Cardiology
0002-9149
1879-1913
Excerpta Medica
S0002-9149(21)00403-3
10.1016/j.amjcard.2021.04.027
Reader's Comments
Substance Use and Premature Atherosclerotic Cardiovascular Disease (From the CDC Behavioral Risk Factor Surveillance System [BRFSS] Survey)
Daher Marilyne MD
Al Rifai Mahmoud MD, MPH
Mahtta Dhruv DO, MBA
Krittanawong Chayakrit MD
Berman Jeffrey MD
Ullah Waqas MD
Alam Mahboob MD
Virani Salim S MD, PhD ⁎
Michael E. DeBakey VA Medical Center, Baylor College of Medicine, Houston, Texas
⁎ Corresponding author: S.S. Virani, Tel: (713) 794-8517; fax: (713) 748-7359.
6 6 2021
1 8 2021
6 6 2021
152 177178
19 4 2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcSubstance use (SU) affects a significant number of individuals globally. More recently, national statistics showed that 13.3% of the U.S. population recently initiated or increased SU to cope with stress during the COVID-19 pandemic, with young adults being at higher risk (24.7% and 19.5% of those aged 18-24 and 25-44 years, respectively.).1 The association of SU and atherosclerotic cardiovascular disease (ASCVD) is well-described in the literature.2 Given the prevalence of SU in the young population, we examined the association of SU and premature ASCVD.
We used data from the Behavioral Risk Factor Surveillance System survey (2016 to 2019), a nationwide telephone-based survey of a random sample of U.S. adults. Variables were self-reported and weighted based on sampling methodology. Premature ASCVD was defined as that occurring in individuals <55 years. SU was defined as current use of e-cigarettes or cigarettes, or use of smokeless tobacco (ST), heavy alcohol (>14 drinks in men and >7 drinks in women per week), or marijuana; high-risk behavior was defined as intravenous drug use or high-risk sexual behavior. Analyses were performed via Stata version 16.1 (Stata Corp, College Station, Texas).
The study population consisted of 1,122,765 individuals younger than 55 years, 50% of whom were women, 59% white, 13% black, 19% Hispanic, and 5% with ASCVD (n = 72,233). Those with premature ASCVD (Table 1 ) had higher prevalence of e-cigarette, cigarette, and ST use (p < 0.001 for all). In multivariable analysis, e-cigarette (Odds Ratio 1.45; 95% Confidence Interval 1.15,1.81), cigarette (1.44; 1.34,1.54), ST (1.23; 1.07,1.41), marijuana use (1.25; 1.01,1.55), and high-risk behavior (1.26; 1.09,1.44) were independently associated with premature ASCVD. There was a significant interaction between gender and ST: among those with ST use, prevalence of ASCVD was 7.7% among men vs. 2.0% among women (p for interaction <0.05). There was no significant interaction between race and any SU.Table 1 Prevalence and odds ratios (95% confidence interval) of substance use and premature ASCVD
Table 1 Presence of Atherosclerotic Cardiovascular Disease Odds Ratio (95% Confidence Interval)
No Yes Unadjusted Adjusted*
E-cigarette Smoking 6.90% 8.30% 1.21 (1.11,1.33) 1.45 (1.15,1.81)
Cigarette Smoking 16.90% 30.60% 2.16 (2.09,2.24) 1.44 (1.34,1.54)
Smokeless Tobacco 4.00% 5.30% 1.33 (1.23,1.43) 1.23 (1.07,1.41)
Heavy Alcohol 6.90% 5.90% 0.84 (0.77,0.91) 0.95 (0.79,1.14)
Marijuana 12.50% 11.90% 0.95 (0.84,1.06) 1.25 (1.01,1.55)
High-Risk Behavior† 8.00% 5.90% 0.73 (0.68,0.78) 1.26 (1.09,1.44)
⁎ Results adjusted for age, gender, race/ethnicity, education, employment, incomes, presence of comorbidities (hypertension, hyperlipidemia, diabetes mellitus, chronic kidney disease, chronic obstructive pulmonary disorder, and cancer).
† High-risk behavior was defined as intravenous drug use or high-risk sexual behavior.
Our results highlight the significance of addressing SU as a risk factor for premature ASCVD in young individuals, especially given increased SU noted during the COVID-19 pandemic. Importantly, the association between SU and premature ASCVD in our study was independent of other cardiovascular risk factors. The decline in rates of cigarette use has been partially attributed to the increased use of alternative products, such as e-cigarettes and ST. Data from the National Health Interview Survey showed that e-cigarette users had higher rates of myocardial infarction and stroke compared to non-users.3 Although there is mixed data on ST and ASCVD, ST may be associated with increased rates of fatal MI.4 Marijuana use has been shown to be associated with myocardial infarction in meta-analyses, with a temporal association.5 Data from the VITAL (Veterans Affairs Healthcare database and theVeterans wIth premaTure AtheroscLerosis) registry showed the use of alcohol, cocaine, and amphetamine were each independently associated with premature ASCVD (OR 1.50, 2.44, and 2.74, respectively).2 Interestingly in our results, high-risk behavior was only significant after adjusting for multiple variables, indicating the importance of further examination of the role of gender, education, income, and other factors that may drive this association.
Our results should be interpreted in the context of potential limitations, including the self-reporting of SU, which can generate recall bias. Furthermore, we did not assess the association between the frequency of SU and premature ASCVD.
In conclusion, the use of e-cigarette, cigarette, ST, marijuana, and high-risk behavior are associated with higher rates of premature ASCVD and need to be specifically addressed alongside traditional metabolic risk factors in order to mitigate the rates of ASCVD in this patient population.
Disclosures
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Salim S. Virani reports a relationship with Department of Veterans Affairs, World Heart Federation, Tahir and Jooma Family that includes: funding grants.
==== Refs
References
1 Czeisler MÉ LR, Petrosky E. Mental health, substance use, and suicidal ideation during the COVID-19 pandemic — United States, June 24–30, 2020. MMWR Morb Mortal Wkly Rep2020;69:1049–1057 http://dxdoiorg/1015585/mmwrmm6932a1.
2 Mahtta D Ramsey D Krittanawong C Al Rifai M Khurram N Samad Z Recreational substance use among patients with premature atherosclerotic cardiovascular disease Heart 107 2020 650 656
3 Vindhyal MR Ndunda P Munguti C Vindhyal S Okut H. Impact on cardiovascular outcomes among e-cigarette users: a review from national health interview surveys J Am Coll Cardiol 73 9_Supplement_2 2019 11–11
4 Piano MR Benowitz NL Fitzgerald GA Corbridge S Heath J Hahn E Impact of smokeless tobacco products on cardiovascular disease: implications for policy, prevention, and treatment: a policy statement from the American Heart Association Circulation 122 2010 1520 1544 20837898
5 DeFilippis EM Bajaj NS Singh A Malloy R Givertz MM Blankstein R Marijuana Use in patients with cardiovascular disease: JACC review topic of the week J Am Coll Cardiol 75 2020 320 332 31976871
| 34103157 | PMC9745946 | NO-CC CODE | 2022-12-15 00:03:17 | no | Am J Cardiol. 2021 Aug 1; 152:177-178 | utf-8 | Am J Cardiol | 2,021 | 10.1016/j.amjcard.2021.04.027 | oa_other |
==== Front
Bull Acad Natl Med
Bull Acad Natl Med
Bulletin De L'Academie Nationale De Medecine
0001-4079
2271-4820
Published by Elsevier Masson SAS on behalf of l'Académie nationale de médecine.
S0001-4079(22)00377-6
10.1016/j.banm.2022.11.013
Article Original
Impact de la première année de la pandémie de COVID-19 sur l'épidémiologie des infections invasives (bactériémies) dans les hôpitaux de l’Assistance Publique – Hôpitaux de Paris**
Impact of the first year of the COVID-19 pandemic on the epidemiology of invasive infections (bacteremia) in the hospitals of the Assistance Publique - Hôpitaux de ParisAmarsy Rishma 1
Robert Jérôme 2
Jarlier Vincent 3⁎
1 Groupe hospitalo-universitaire APHP Nord-Université de Paris, Site Lariboisière et Fernand Widal, équipe Infection-Prévention-Contrôle et CIMI-Paris, Inserm U1135, Sorbonne Université, Paris, France
2 Groupe hospitalo-universitaire APHP. Sorbonne Université, Site Pitié-Salpêtrière, Laboratoire de Bactériologie-Hygiène et CIMI-Paris, Inserm U1135, Sorbonne Université, Paris, France
3 Membre correspondant de l’Académie Nationale de Médecine. Service de Bactériologie-Hygiène et CIMI-Paris, Inserm U1135, Sorbonne Université, Paris, France
⁎ Auteur correspondant:
13 12 2022
13 12 2022
16 10 2022
14 11 2022
© 2022 Published by Elsevier Masson SAS on behalf of l'Académie nationale de médecine.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Introduction : La pandémie de COVID-19 a entrainé un afflux massif de patients atteints de formes sévères dans les hôpitaux, nécessitant souvent des soins intensifs (cathéters vasculaires, ventilation, etc.) qui exposent à des risques élevés d'infections nosocomiales en particulier les infections invasives (bactériémies).
Méthode : L'impact de la pandémie de COVID-19 sur l'épidémiologie des bactériémies en 2020 a été analysé dans 25 hôpitaux de l’Assistance Publique – Hôpitaux de Paris (environ 14.000 lits, couvrant la région Ile de France). Jusqu'à un quart des patients admis à l’AP-HP durant la période mars-avril (pic de la première vague) étaient infectés par la COVID-19. L’incidence/100 admissions des bactériémies a globalement augmenté par rapport aux années précédentes : de 24 % en mars 2020 et de 115 % en avril.
Résultats :
L’évolution de l’incidence des bactériémies n’a pas été la même pour 2 groupes de microorganismes d’écologies bien différentes.
- Pour les microorganismes de type « hospitalier » classiquement responsables d’infections nosocomiales, l’incidence a beaucoup augmenté en mars-avril 2020 : Klebsiella pneumoniae (x2,3), Pseudomonas aeruginosa (x2,4), Staphylococcus aureus (x2,4), entérocoques (x3,4), levures (x2,7). Les 2/3 des bactériémies causées par ces microorganismes ont été considérées comme acquises lors de l’hospitalisation. Fait important, il s’est aussi produit une forte augmentation de l’incidence des bactériémies causées par des souches résistantes aux antibiotiques. Les antibiotiques utilisés comme indicateurs ont été les céphalosporines de 3e génération (3GC), antibiotiques majeurs du traitement des infections grave et utilisés pour la surveillance des résistances bactériennes en Europe. À titre d’exemple, l’incidence des bactériémies à souches résistantes aux C3G a été multipliée par 3 en avril 2020 pour K. pneumoniae et S.aureus (résistance croisée aux 3GC et à la méticilline chez cette espèce). Durant la même période, la consommation de 3GC a fortement augmenté dans les mêmes hôpitaux (+131% en mars et + 148% en avril).
- Pour Streptococcus pneumoniae (pneumocoque) et Streptococcus pyogenes (streptocoque hémolytique du groupe A), deux pathogènes responsables d’infections essentiellement communautaires et de transmission respiratoire, la pandémie a eu l’effet inverse. Il s’est produit une diminution de l’incidence en 2020 de 34% et 28% respectivement, en particulier au printemps au moment des mesures strictes de confinement, de distanciation physique et de port du masque. Une légère réémergence des infections à ces deux espèces s’est produite l’été 2020 après l'assouplissement des mesures de prévention. Par contraste avec ce qui a été vu plus haut, les 4/5 des bactériémies causées par ces deux espèces ont été considérées comme communautaires.
Conclusion : La pandémie de COVID-19 qui a eu un fort impact sur la gestion hospitalière et sur l’organisation sociale dans la population générale, et a eu des impacts opposés sur l’incidence des bactériémies selon les pathogènes et leur mode de transmission.
The COVID-19 pandemic has led to a massive influx of patients suffering from severe forms of the disease into hospitals, often requiring intensive care (vascular catheters, ventilation, etc.) which exposes them to high risks of nosocomial infections, particularly invasive infections (bacteremia). The impact of the COVID-19 pandemic on the epidemiology of bacteremia in 2020 was analysed in 25 hospitals of the Assistance Publique - Hôpitaux de Paris (APHP,approximately 14,000 beds, covering the Ile de France region). Up to a quarter of patients admitted to AP-HP during the March-April period (peak of the 1st wave) were infected with COVID-19. The incidence/100 admissions of bacteraemia increased overall compared to previous years: by 24% in March 2020 and by 115% in April.
The evolution of the incidence of bacteremia was not the same for 2 groups of microorganisms with very different ecologies.
- For the "hospital" type microorganisms classically responsible for nosocomial infections, the incidence increased significantly in March-April 2020: Klebsiella pneumoniae (x2.3), Pseudomonas aeruginosa (x2.4), Staphylococcus aureus (x2.4), enterococci (x3.4), yeasts (x2.7). Two thirds of the bacteremias caused by these microorganisms were considered as acquired during hospitalization. Importantly, there was also a sharp increase in the incidence of bacteremia caused by antibiotic-resistant strains. The antibiotics used as indicators were the 3rd generation cephalosporins (3GCs), major antibiotics in the treatment of serious infections used for monitoring bacterial resistance in Europe. For example, the incidence of bacteremia with C3G-resistant strains increased threefold in April 2020 for K. pneumoniae and S. aureus (cross-resistance to 3GC and meticillin in this species). During the same period, the consumption of 3GC increased sharply in the same hospitals (+131% in March and + 148% in April).
- For Streptococcus pneumoniae (pneumococcus) and Streptococcus pyogenes (group A hemolytic streptococcus), two pathogens responsible for mainly community and respiratory-transmittedinfections, the pandemic had the opposite effect. There was a decrease in incidence in 2020 of 34% and 28% respectively, particularly in the spring when strict containment, physical distancing and mask-wearing measures were in place. A slight re-emergence of infections with these two species occurred in the summer of 2020 after the relaxation of prevention measures. In contrast to what was seen above, 4/5 of the bacteremias caused by these two species were considered community-acquired.
The COVID-19 pandemic had a strong impact on hospital management and social organization in the general population, and had opposite impacts on the incidence of bacteremia depending on the pathogens and their mode of transmission.
Mots-clés
Bactériémie
Infections bactériennes
Infections fongiques invasives
Virus du SRAS
Keywords
Bacteremia
Bacterial infections
Invasive Fungal Infections
SARS Virus
==== Body
pmc* Séance du 15/11/2022
| 0 | PMC9745959 | NO-CC CODE | 2022-12-15 00:03:19 | no | Bull Acad Natl Med. 2022 Dec 13; doi: 10.1016/j.banm.2022.11.013 | utf-8 | Bull Acad Natl Med | 2,022 | 10.1016/j.banm.2022.11.013 | oa_other |
==== Front
J Virus Erad
J Virus Erad
Journal of Virus Eradication
2055-6640
2055-6659
The Authors. Published by Elsevier Ltd.
S2055-6640(22)00246-1
10.1016/j.jve.2022.100308
100308
Original Research
Evolution of anti-SARS-CoV-2 spike protein titers after two-dose COVID-19 vaccination among people living with HIV
Liu Wang-Da ab
Pang Man Wai c
Wang Jann-Tay ad
Sun Hsin-Yun a
Huang Yu-Shan a
Lin Kuan-Yin a
Wu Un-In ab
Li Guei-Chi a
Liu Wen-Chun a
Su Yi-Ching a
He Pu-Chi ae
Lin Chia-Yi f
Yeh Chih-Yu f
Cheng Yu-Chen c
Yao Yi c
Chen Yi-Ting g
Wu Pei-Ying g
Chen Ling-Ya g
Luo Yu-Zhen g
Chang Hsi-Yen g
Sheng Wang-Huei aeh
Hsieh Szu-Min a
Chang Sui-Yuan ci∗∗
Hung Chien-Ching ajklm∗
Chang Shan-Chwen ah
a Department of Internal Medicine, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan
b Department of Medicine, National Taiwan University Cancer Center, Taipei, Taiwan
c Department of Clinical Laboratory Sciences and Medical Biotechnology, College of Medicine, National Taiwan University, Taipei, Taiwan
d Institute of Infectious Diseases and Vaccinology, National Health Research Institutes, Zhunan, Taiwan
e Department of Medical Education, National Taiwan University, Taipei, Taiwan
f Department of Nursing, National Taiwan University Hospital, Taipei, Taiwan
g Center of Infection Control, National Taiwan University Hospital, Taipei, Taiwan
h School of Medicine, National Taiwan University College of Medicine, Taipei, Taiwan
i Department of Laboratory Medicine, National Taiwan University Hospital, National Taiwan University College of Medicine, Taipei, Taiwan
j Department of Tropical Medicine and Parasitology, National Taiwan University College of Medicine, Taipei, Taiwan
k Department of Internal Medicine, National Taiwan University Hospital Yun-Lin Branch, Yun-Lin County, Taiwan
l Department of Medical Research, China Medical University Hospital, Taichung, Taiwan
m China Medical University, Taichung, Taiwan
∗ Corresponding author. Department of Internal Medicine, National Taiwan University Hospital, 7 Chung-Shan South Rd., Taipei City, 10002, Taiwan.
∗∗ Corresponding author. Department of Laboratory Medicine, National Taiwan University Hospital, 7 Chung-Shan South Rd., Taipei City, 10002, Taiwan.
13 12 2022
13 12 2022
10030811 8 2022
9 12 2022
© 2022 The Authors. Published by Elsevier Ltd.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Background
A community COVID-19 outbreak caused by the B.1.1.7 SARS-CoV-2 variant occurred in Taiwan in May 2021. High-risk populations such as people living with HIV (PLWH) were recommended to receive two doses of COVID-19 vaccines. While SARS-CoV-2 vaccines have demonstrated promising results in general population, real-world information on the serological responses remains limited among PLWH.
Methods
PLWH receiving the first dose of SARS-CoV-2 vaccine from 2020 to 2021 were enrolled. Determinations of anti-SARS-CoV-2 spike IgG titers were performed every one to three months, until PLWH received the third dose of SARS-CoV-2 vaccine or had confirmed SARS-CoV-2 infection. All serum samples were tested for anti-nucleocapsid antibody and those tested positive were excluded from analysis.
Results
A total of 1189 PLWH were enrolled: 829 (69.7%) receiving two doses of AZD1222 vaccine, 232 (19.5%) mRNA-1273 vaccine, and 128 (10.8%) BNT162b2 vaccine. Of all time points, PLWH receiving two doses of mRNA vaccines had consistently higher antibody levels than those receiving AZD1222 vaccine (p < 0.001 for all time-point comparisons). Of all PLWH, factors associated with failure to achieve an anti-spike IgG titer >141 BAU/mL within 12 weeks included type 2 diabetes mellitus (DM) (adjusted odds ratio [aOR], 2.24; 95% CI, 1.25–4), CD4 count <200 cells/mm3 upon receipt of the first dose of vaccination (aOR, 3.43; 95% CI, 1.31–9) and two homologous AZD1222 vaccination (aOR, 16.85; 95%CI, 10.13–28). For those receiving two doses of mRNA vaccines, factors associated with failure to achieve an anti-spike IgG titer >899 BAU/mL within 12 weeks were CD4 count <200 cells/mm3 on first-dose vaccination (aOR, 3.95; 95% CI, 1.08–14.42) and dual BNT162b2 vaccination (aOR, 4.21; 95% CI, 2.57–6.89).
Conclusions
Two doses of homologous mRNA vaccination achieved significantly higher serological responses than vaccination with AZD1222 among PLWH and PLWH with CD4 counts <200 cells/mm3 and DM had consistently lower serological responses.
Keywords
Antibody
Humoral immunity
ChAdOx1 nCoV-19 (AZD1222) vaccine
mRNA-1273 vaccine
BNT162b2 vaccine
==== Body
pmc1 Introduction
The severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection (COVID-19) had become pandemic since 2020 and caused a high number of morbidities and mortalities, especially among immunocompromised patients.1 Previous studies had demonstrated that people living with HIV (PLWH) of older age or with comorbidities such as diabetes mellitus (DM) or cardiovascular diseases might have poor outcomes when infected with SARS-CoV-2 compared with those without HIV infection.2 , 3 Further studies indicated that a worse outcome was associated with a lower CD4 count, even in PLWH receiving stable antiretroviral therapy (ART) with virological suppression.3, 4, 5, 6, 7 Therefore, current guidelines suggested PWLH who have mild to moderate COVID-19 were eligible to receive antiviral agents or monoclonal antibody treatment to prevent from severe or critical disease, especially those with advanced HIV infection, which was defined as having CD4 counts less than 200 cells/mm3 or ever having AIDS-defining illnesses.8
Compared with other countries, the epidemic of COVID-19 in Taiwan was relatively mild before 2022. There were only a few indigenous cases of COVID-19 before May 2021 under strict non-pharmaceutical interventions such as border control, contact tracing and isolation, social distancing and a nationwide campaign for using personal protective equipments.9 A community outbreak caused by the B.1.1.7 (alpha) variant took place in May 2021, right after the implementation of SARS-CoV-2 vaccination program. Overall, a total of 14,308 patients with indigenous cases of COVID-19 were reported from May to August 2021; of those, only 192 (1.3%) PLWH were found to be infected by SARS-CoV-2.10 There was a consensus that all PLWH should receive the full series of two doses of COVID-19 vaccine, regardless of CD4 counts or plasma HIV RNA load,8 though the safety and efficacy data of COVID-19 vaccine for PLWH were limited. Previous studies have demonstrated that certain vaccines such as influenza or HBV vaccine induce suboptimal serological responses in PLWH.11 , 12 Data from most clinical trials of COVID-19 vaccines conducted in PLWH were scarce and safety and short-term immunogenicity remained limited9 , 13, 14, 15, 16, 17, 18; moreover, the durability of antibody responses and the correlation between antibody titers and the risk of SARS-CoV-2 infection remained unclear. In this prospective observational study, we aimed to investigated the serological responses in PLWH who had been on stable ART and received two homologous doses of COVID-19 vaccines, including the AZD1222, mRNA-1273 and BNT162b2 vaccine.
2 Methods
2.1 Study population and setting
This observational study was conducted at the National Taiwan University Hospital (NTUH) to include PLWH aged 20 years or older who had been receiving HIV care as outpatients. Those receiving the first dose of SARS-CoV-2 vaccine between January 2020 and December 2021 were enrolled. A second homologous vaccine including the AZD1222, mRNA-1273 and BNT162b2 vaccine was given according the guidelines by Taiwan Centers of Disease Control, with at least an interval of 4–10 weeks between the two doses. Determinations of anti-SARS-CoV-2 spike IgG titers were performed every one to three months, depending on the clinical appointments made for HIV care, until the participants received the third dose of SARS-CoV-2 vaccine, the diagnosis of COVID-19, loss to follow-up or death, whichever took place first. Participants who had a history of confirmed SARS-CoV-2 infection were excluded. All serum samples were also tested for anti-SARS-CoV-2 nucleocapsid antibody and those tested positive at baseline and during follow-up were excluded from analysis. The medical records of the included PLWH were reviewed and the information on the demographic and clinical characteristics were reviewed, including age, body-mass index (BMI), date of vaccination, CD4 counts and plasma HIV RNA load (PVL) on vaccination, ART, and underlying diseases that might influence immune responses such as type 2 DM, chronic kidney disease (CKD) of stage 3–5 (defined as an estimated glomerular filtration rate less than 60 mL/min/1.73 m2), malignancies, autoimmune disease and viral hepatitis. The study was approved by the Research Ethics Committee of NTUH (NTUH 202106149RIND) and all participating PLWH gave written informed consent.
2.2 Laboratory investigations
All serum samples were stored centrifuged at −20 °C before testing. Anti-SARS-CoV-2 spike IgG in serum samples was determined using SARS-CoV-2 IgG II Quant assay (Abbott, Abbott Park, Illinois, U.S.A.) according to the manufacturer's instructions. An anti-SARS-CoV-2 spike IgG level higher than 50 arbitrary units per milliliter (AU/mL) was considered positive. The mathematical relationship of the Abbott AU/mL unit to WHO unit (binding antibody unit per mL [BAU/mL]) would follow the equation: BAU/mL = 0.142∗AU/mL.
In addition, anti-SARS-CoV-2 nucleocapsid antibody IgG was determined using Elecsys® Anti-SARS-CoV-2 assay (Roche, U.S.A), while an anti-SARS-CoV-2 nucleocapsid IgG level higher than 1.0 cutoff index (COI) was considered reactive.
2.3 Outcome assessment
The primary endpoints were serologic responses at weeks 1–24 after the second dose of SARS-CoV-2 vaccination. Acquisition of COVID-19 included symptomatic infection and asymptomatic infection; the history of symptomatic infection was retrieved from the National Notification System for Infectious Diseases, while asymptomatic infection was defined as a positive result of anti-SARS-CoV-2 nucleocapsid antibody in the absence of clinical symptoms. In order to speculate the potential vaccine effectiveness through antibody measurements, two cut-off values including 141 and 899 BAU/mL of anti-spike IgG were used. Dimeglio C et al. recently demonstrated that an anti-spike antibody titer greater than 141 BAU/mL was correlated with the presence of neutralizing antibodies through the evaluation of 8758 vaccinated and unvaccinated healthcare workers.19 Prediction of clinical efficacy performed by Feng S et al. suggested that an anti-spike antibody titer of 899 BAU/mL predicted a 90% vaccine efficacy of AZD1222 vaccine against symptomatic infection by the B.1.1.7 variant.20
2.4 Statistical analysis
Categorical variables, such as gender, underlying diseases and ART were compared between different vaccination groups using Fisher's exact test and Pearson's chi-squared test. Continuous variables, such as age and laboratory results at vaccination were analyzed using Mann-Whitney U test. The geometric mean titers (GMTs) of SARS-CoV-2 anti-spike IgG were calculated in Ln-transformed data for statistics. Antibody titers in Ln form of different vaccines were compared using Student-T tests of all time points. We applied the logistic regression model to estimate the adjusted odds ratios (aORs) for those with relatively low anti-spike IgG (<141 or <899 BAU/mL) within 12 weeks, and for those whose anti-spike IgG titers rapidly declined within 12–24 weeks after receiving the second dose of COVID-19 vaccine. A backward stepwise regression with removal threshold of p = 0.2 to select among covariates to be included into the multivariable model. A two-tailed p value less than 0.05 was considered statistically significant. All analyses were performed using Stata/SE software, Version 17.0 (https://www.stata.com).
3 Results
Between January 2020 and December 2021, 1252 PLWH who were followed at the HIV clinics of NTUH received the first dose of COVID-19 vaccines and were included in this study. After excluding those with confirmed COVID-19, positive anti-SARS-CoV-2 nucleocapsid antibody at baseline or during follow-up, and loss to follow-up, 1189 PLWH who completed 2-dose vaccination were included for further analysis, including 829 (69.7%) who received two homologous doses of AZD1222 vaccine, 232 (19.5%) mRNA-1273 vaccine and 128 (10.8%) BNT162b2 vaccine (Fig. 1 ).Fig. 1 Study population.
Fig. 1
Table 1 shows the characteristics of included PLWH. There were mostly male (97.6%) with a median age of 40 years and 98.6% were virologically suppressed with ART with a median baseline CD4 counts of 632 cells/mm3. Overall, 1145 (96.3%) PLWH were receiving integrase strand-transfer inhibitor-based antiretroviral regimens before vaccination. Those who received two doses of mRNA vaccines tended to be older compared with those who received two doses of AZD1222 vaccines (41 vs 39 years, p < 0.001), while PLWH who received mRNA vaccines were more likely to have type 2 DM, CKD stage 3–5 and a previous history of lymphoma. The intervals between the two homologous doses of AZD1222 vaccination and mRNA vaccination were 13 (interquartile range, 13–14) and 13 (6–15) weeks, respectively.Table 1 Baseline characteristics of the included participants. Group 1 represents the participants receiving two doses of AZD1222 vaccines, while Group 2 were those had two doses of mRNA vaccines, including mRNA-1273 and BNT162b2 vaccines.
Table 1 Total (N = 1189) Group 1 (N = 829) Group 2 (N = 360) p value
Age (IQR), years 40 (33–48) 39 (33–46) 41 (35–50) < 0.001
Male gender 1160 (97.6) 815 (98.3) 345 (95.8) 0.01
BMI >30 kg/m2 110 (9.3) 76 (9.2) 34 (9.4) 0.88
Interval between the two doses, weeks 13 (12–14) 13 (13–14) 13 (6–15) 0.002
HIV status
CD4 (IQR), cells/mm3 632 (474–812) 641 (496–810) 612 (436–830) 0.11
CD8 (IQR), cells/mm3 809 (624–1062) 798 (624–1057) 845 (618–1076) 0.58
CD4 <200 cells/mm3 25 (2.1) 14 (1.7) 11 (3.1) 0.13
CD4 <350 cells/mm3 125 (10.5) 74 (8.9) 51 (14.2) 0.007
CD4 <500 cells/mm3 336 (28.3) 211 (25.5) 125 (34.7) 0.001
CD4/CD8 0.77 (0.57–1.03) 0.77 (0.59–1.03) 0.78 (0.51–1.03) 0.18
Median PVL, (IQR), log copies/mL 0 (0–0) 0 (0–0) 0 (0–0) 0.6
PVL >20 copies/mL 192 (16.2) 131 (15.8) 61 (16.9) 0.62
PVL >200 copies/mL 17 (1.43) 11 (1.3) 6 (1.7) 0.65
cART
INSTI-based 1145 (96.3) 794 (95.8) 351 (97.5) 0.15
NNRTI-based 42 (3.5) 35 (4.2) 7 (1.9) 0.051
PI-based 6 (0.5) 3 (0.4) 3 (0.8) 0.38
Comorbidities
Type 2 DM 74 (6.22) 42 (5.1) 32 (8.9) 0.01
CKD stage III–V 49 (4.12) 24 (2.9) 25 (6.9) 0.001
Solid-organ cancer 25 (2.1) 14 (1.7) 11 (3.1) 0.13
Lymphoma 24 (2) 13 (1.6) 11 (3.1) 0.09
Autoimmune disease 15 (1.3) 11 (1.3) 4 (1.1) 1
Concurrent systemic steroid use 7 (0.6) 4 (0.5) 3 (0.8) 0.44
Chronic hepatitis B 138 (11.7) 97 (11.8) 41 (11.5) 0.88
Anti-HCV positivity 170 (14.3) 116 (14) 54 (15) 0.65
Abbreviations: BMI, body mass index; cART, combination antiretroviral therapy; CKD, chronic kidney disease; DM, diabetes mellitus; HCV, hepatitis C virus; HIV, human immunodeficiency virus; INSTI, integrase strand transfer inhibitor; IQR, interquartile range; NNRTI, non-nucleoside reverse transcriptase inhibitor; PI, protease inhibitor; PVL, plasma HIV RNA load.
Among the PLWH who received two homologous doses of AZD1222 vaccine and mRNA vaccines, anti-spike IgG titer reached the peak on the third and fourth weeks and subsequently declined. At all time points, PLWH receiving two homologous doses of mRNA vaccines had consistently higher antibody levels than those who received two homologous doses of AZD1222 vaccine (p < 0.001 for all time-point comparisons).
Of 1136 PLWH who underwent anti-spike IgG testing within 12 weeks after the second dose of COVID-19, 385 (33.9%) failed to achieve anti-spike IgG titers >141 BAU/mL, including 372 (46.7%) of PLWH receiving AZD1222, 8 (3.6%) of PLWH receiving mRNA-1273 and 10 (8.5%) of PLWH receiving BNT162b2 vaccine (Fig. 2 ). Associated factors with failure to achieve anti-spike IgG titers >141 BAU/mL in the multivariate logistic regression model including PLWH with CD4 counts <200 cells/mm3 before vaccination (aOR, 3.43; 95% confidence interval [CI], 1.31–9), presence of type 2 DM (aOR, 2.24; 95% CI, 1.25–4), and those who received two homologous doses of AZD1222 vaccine (aOR, 16.85; 95% CI, 10.13–28) (Table 2 ) (Fig. 3 ). Moreover, 733 (72.9%) of 1005 PLWH who underwent anti-spike antibody testing 4–12 weeks after the second dose of vaccination failed to achieve anti-spike IgG titers >899 BAU/mL. PLWH receiving 2 consecutive doses of AZD1222 vaccine were more likely to fail to achieve anti-spike IgG titers >899 BAU/mL compared with those receiving two homologous doses of mRNA vaccines (OR, 45.91; 95% CI, 30.7–68.65) (data not shown).Fig. 2 Serologic responses after the second dose of COVID-19 vaccination at different follow-up intervals. (A). Evolution of anti-spike IgG of PWLH receiving two AZD1222 or mRNA vaccines. (B). Evolution of anti-spike IgG of PWLH receiving two AZD1222 vaccines. (C). Evolution of anti-spike IgG of PWLH receiving two mRNA vaccines. (D). Evolution of anti-spike IgG of PWLH receiving two mRNA-1273 (M) or BNT162b2 (B) vaccines. The number in the table below the X-axis in each figure represented the number of PLWH undergoing antibody testing and the GMT of anti-spike IgG in each period.
Fig. 2
Table 2 Factors associated with low titers of anti-spike antibody (<141 BAU/mL) in PLWH within 12 weeks after receiving the second dose of COVID-19 vaccine.
Table 2 Univariable Multivariable
HR (95% CI) p value aOR (95% CI) p value
Age, per 1-year increase 1 (0.99–1.01) 0.71 – –
Male gender 3.14 (1.08–9.13) 0.04 2.41 (0.77–7.59) 0.13
BMI >30 kg/m2 1 (0.66–1.53) 0.99 – –
Type 2 DM 1.43 (0.88–2.32) 0.15 2.24 (1.25–4) 0.006
CKD stage III–V 0.94 (0.51–1.74) 0.85 – –
Solid-organ cancer 0.64 (0.25–1.64) 0.36 – –
Lymphoma 0.91 (0.37–2.25) 0.84 – –
Autoimmune disease 0.71 (0.22–2.23) 0.55 – –
Concurrent systemic steroid use 0.39 (0.05–3.34) 0.39 – –
Chronic hepatitis B 1.39 (0.96–2.01) 0.08 1.4 (0.92–2.12) 0.12
Anti-HCV positivity 1.03 (0.73–1.46) 0.85 – –
CD4 <200 cells/mm3 1.84 (0.86–3.95) 0.12 3.43 (1.31–9) 0.01
CD4 <350 cells/mm3 1.12 (0.75–1.66) 0.59 – –
CD4 <500 cells/mm3 0.9 (0.69–1.19) 0.48 – –
PVL >20 copies/mL 0.91 (0.64–1.27) 0.57 – –
PVL >200 copies/mL 1.72 (0.62–4.78) 0.3 – –
PVL >100,000 copies/mL 1.96 (0.27–13.94) 0.5 – –
Two doses of AZD1222 vaccine 15.3 (9.33–25.1) < 0.001 16.85 (10.13–28) < 0.001
Interval between the two doses >8 weeks 4.89 (2.76–8.63) < 0.001 – –
Abbreviations: BMI, body mass index; CKD, chronic kidney disease; DM, diabetes mellitus; HCV, hepatitis C virus; PVL, plasma HIV RNA load.
Fig. 3 Risk associated with low anti-spike IgG response within the first 12 weeks after PLWH received two COVID-19 vaccines in the multivariate logistic regression model. (A) All PLWH with antibody titers failing to reach 141 BAU/mL. (B) PLWH with two AZD1222 vaccines, with antibody titers failing to reach 141 BAU/mL. (C) PLWH with two mRNA vaccines, with antibody titers failing to reach 899 BAU/mL.
Fig. 3
Of the 796 participants who had anti-spike IgG titers determined within the first 12 weeks after receiving the second dose of AZD1222 vaccine, 367 (46.1%) and 759 (95.4%) failed to achieve anti-spike IgG titers >141 and 899 BAU/mL, respectively. The major independent factor associated with failure to achieve anti-spike IgG titers >141 BAU/mL was type 2 DM (aOR, 2.66; 95% CI, 1.35–5.23) (Supplementary Table 1). In addition, among 340 PLWH receiving two homologous mRNA vaccines, 18 (5.3%) and 105 (30.9%) failed to achieve anti-spike IgG titers >141 and > 899 BAU/mL, respectively. In the multivariate analysis, factors associated with an anti-spike IgG <899 BAU/mL were CD4 count <200 cells/mm3 (aOR, 3.95; 95% CI, 1.08–14.42) and two doses of BNT162b2 vaccine (aOR, 4.2; 95% CI, 2.57–6.89) (Supplementary Table 2) (Fig. 3).
Among the 228 PLWH who underwent anti-spike IgG testing within 12–24 weeks after receiving the second dose of COVID-19, 111 (48.7%) had rapid declines of anti-spike IgG titer, which was defined as a decline of anti-spike IgG titers to <141 BAU/mL, including 108 (70.6%) PLWH who received AZD1222, 1 (1.8%) who received mRNA-1273, and 2 (10%) who received BNT162b2 vaccine (Fig. 2). Receipt of two homologous doses of AZD1222 vaccine was the major factor associated with the rapid declines of the anti-spike antibody (aOR, 57.62; 95% CI, 16.9–196.6) (Supplementary Table 3). Furthermore, 151 (98.7%) PLWH who received two doses of AZD-122 vaccine, 42 (76.4%) who received mRNA-1273 and 19 (95%) who received BNT162b2 vaccine had a decline of anti-spike IgG titer to <899 BAU/mL 12–24 weeks after the second dose of vaccination.
4 Discussion
In this study investigating the serologic responses of COVID-19 vaccination among PLWH, we demonstrated the longitudinal follow-up of antibody responses in PLWH who had been mostly well-controlled with ART and received different types of 2-dose homologous vaccine in the real-word setting. PLWH who underwent two homologous mRNA vaccines were more likely to achieve higher and more sustained antibody responses compared with those who received 2-dose homologous AZD1222 vaccine.
To date, several studies have demonstrated that well-controlled PLWH had similar serological responses to those without HIV infection within 2–4 weeks after the second dose of AZD1222 or BNT162b2 vaccine.12 , 16 However, the titers of anti-spike IgG in our cohort were slightly lower when compared with the titers observed in the studies mentioned above .12 , 16 In a retrospective study of 100 PLWH and 152 matched HIV-negative control participants, HIV was not significantly associated with the magnitude of any humoral response. However, two homologous doses of AZD1222 vaccination was significantly associated with lower antibody responses.21 When compared to the antibody responses in a study that assessed the evolution of antibody titers in young, HIV-negative participants of similar age who received homologous or heterologous SARS-CoV-2 vaccines,22 we found lower antibody responses in our study consisting only PLWH. In addition to HIV infection, types of vaccines used, and vaccination strategies (homologous vs heterologous), the discrepancy may be attributed to the intervals between the two doses of vaccines administered. The included PLWH in our study mostly received the second doses of vaccines after a longer interval (>12 weeks) due to limited availability of vaccines in Taiwan in 2021 compared with those of other studies in which participants received the second dose at an interval of 4–8 weeks.12 , 16 , 22 In our study, we found that an interval >8 weeks between two doses of vaccination had a 4.9-fold higher risk associated with lower antibody responses. However, such correlation was not found in the subgroup analysis in PLWH receiving AZD1222 or mRNA vaccine, respectively. Previous meta-analysis by Voysey M et al. has demonstrated an interval of 8–12 weeks for AZD1222 vaccine led to higher immunogenicity while an extended interval of 14–17 weeks showed a similar trend in antibody responses.23 , 24 Nevertheless, there were scarce data to support this finding among PLWH. In addition, the storage of the serum sample at −20 °C before testing might also contribute the lower antibody responses observed in our study, although previous studies have demonstrated that multiple freeze--thaw cycles had no effect on the ability of the ELISA assay to detect SARS-CoV-2 IgG antibodies.25 , 26
Similar to the findings in studies among healthy adults, our study demonstrated a waning of anti-spike IgG.27 , 28 Lapointe HR et al. pointed out that virologically suppressed PLWH showed typical antibody durability after two doses of COVID-19 vaccination regardless of types of vaccine administered.14 However, in the context of communicating variants and evidence of correlation between anti-spike IgG and neutralizing ability, investigations to examine the clinical efficacy of a higher standard of antibody titer is warranted. Seventy percent of PLWH who received AZD1222 vaccine in this study failed to achieve sustained antibody responses by maintaining anti-spike IgG titers >141 BAU/mL, an indicator of presence of neutralizing antibodies. Given the peaks of anti-spike IgG titers might be lower compared with those without HIV infection, the durability of maintaining neutralizing antibodies might be overestimated. Therefore, our findings support that additional doses of booster vaccination are needed in PLWH during the ongoing SARS-CoV pandemic.
Even with only a small number in our cohort, PLWH with a CD4 count <200 cells/mm3 was shown to be associated with poor serological responses to COVID-19 vaccination. PLWH with CD4 counts <200 cells/mm3 in this study continued to achieve inferior antibody responses even if they received two homologous doses of mRNA vaccines. Such findings were consistent with those of previous studies.29, 30, 31, 32, 33 Moreover, our study illustrated that, in addition to CD4 counts and vaccine type, traditional risk factors including CKD stage 3–5 and type 2 DM were associated with poor serological responses.34, 35, 36 As a result, PLWH with comorbidities that may affect immune responses should be prioritized for a second booster dose of SARS-CoV-2 vaccine.
Strengths of this study include a relatively large number of PLWH and a longer observational period for the evaluation of durability of antibody responses. However, the observational nature of this study results in several limitations. First, anti-spike IgG measurements were performed at intervals according to clinical care practices or convenient time points for PLWH but not at fixed time points; while the trends were similar to those observed in other studies, the serological responses should be interpreted with caution. Second, only a few PLWH underwent antibody testing 12–24 weeks after the second dose of vaccination, the power to understand the extent to which the serological responses waned might be dampened. Third, the low-level viral transmission rate in Taiwan at the time the study was conducted hindered the evaluation of vaccine effectiveness. Nevertheless, our previous study during the early phase of implementation of COVID-19 vaccination program in Taiwan had demonstrated that COVID-19 vaccination was still clinically effective among PLWH during the outbreak setting where non-pharmaceutical interventions were strictly implemented.9 , 10 Fourth, no HIV-negative participants were included as a control group, which might preclude us from understanding if virologically-suppressed PLWH with improved immunity might mount similar immune responses to COVID-19 vaccination to HIV-negative people of the same age groups. Last, we did not assess the neutralizing antibody or T-cell related immunity. Instead, two cut-off values derived from immune-bridging model were used in our study. The performance of using the two cut-off values in predicting clinical vaccine effectiveness warrants more investigations among PLWH.
In conclusion, we found that PLWH receiving two homologous SARS-CoV-2 vaccine mounted antibody responses with similar trends of antibody responses over time to those observed in clinical trial settings. Two doses of homologous mRNA vaccination had significantly higher immunogenicity than AZD1222 among PLWH. PLWH with CD4 counts less than 200 cells/mm3 had consistently lower antibody responses to either mRNA or non-mRNA vaccination. Traditional risk factor such as DM also contributed to lower anti-spike IgG titers.
Author contributions
Wang-Da Liu – manuscript drafting, data analysis.
Man Wai Pang – laboratory assistant for antibody test.
Sui-Yuan Chang – laboratory chief, handling the antibody measurement.
Jann-Tay Wang – statistics, data analysis.
Hsin-Yun Sun – participants enrolling.
Yu-Shan Huang - participants enrolling.
Kuan-Yin Lin - participants enrolling;
Un-In Wu - participants enrolling;
Guei-Chi Li – data collection.
Wen-Chun Liu - participants enrolling, laboratory assistant.
Yi-Ching Su - laboratory assistant.
Pu-Chi He - participants enrolling, data collection.
Chia-Yi Lin – data collection.
Chih-Yu Yeh – data collection.
Yu-Chen Cheng - laboratory assistant for antibody test.
Yi Yao - laboratory assistant for antibody test.
Yi-Ting Chen - participants enrolling.
Yu-Zhen Luo - participants enrolling.
Pei-Ying Wu - participants enrolling.
Ling-Ya Chen - participants enrolling.
Hsi-Yen Chang - participants enrolling.
Wang-Huei Sheng - participants enrolling;
Szu-Min Hsieh - participants enrolling.
Chien-Ching Hung – concept of the study, manuscript drafting.
Shan-Chwen Chang – concept of the study.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supplementary data
The following is the Supplementary data to this article:Multimedia component 1
Multimedia component 1
Acknowledgment
This study is supported by Taiwan Center of Disease Control (MOHW111-CDC-C-114-000104), 10.13039/501100004663 MOST (NO. 111-2321-B-002-017), and 10.13039/501100005762 National Taiwan University Hospital (NO. MM022-1).
Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.jve.2022.100308.
==== Refs
References
1 Williamson E.J. Walker A.J. Bhaskaran K. Factors associated with COVID-19-related death using OpenSAFELY Nature 584 2020 430 436 32640463
2 Bhaskaran K. Rentsch C.T. MacKenna B. HIV infection and COVID-19 death: a population-based cohort analysis of UK primary care data and linked national death registrations within the OpenSAFELY platform Lancet HIV 8 2021 e24 e32 33316211
3 Dandachi D. Geiger G. Montgomery M.W. Characteristics, comorbidities, and outcomes in a multicenter registry of patients with human immunodeficiency virus and coronavirus disease 2019 Clin Infect Dis 73 2021 e1964 e1972 32905581
4 Braunstein S.L. Lazar R. Wahnich A. Daskalakis D.C. Blackstock O.J. Coronavirus disease 2019 (COVID-19) infection among people with human immunodeficiency virus in New York City: a population-level analysis of linked surveillance data Clin Infect Dis 72 2021 e1021 e1029 33252620
5 Jassat W. Abdool Karim S.S. Mudara C. Clinical severity of COVID-19 in patients admitted to hospital during the omicron wave in South Africa: a retrospective observational study Lancet Global Health 10 2022 e961 e969 35597249
6 Durstenfeld M.S. Sun K. Ma Y. Association of HIV infection with outcomes among adults hospitalized with COVID-19 AIDS 36 2022 391 398 34750295
7 Hoffmann C. Casado J.L. Härter G. Immune deficiency is a risk factor for severe COVID-19 in people living with HIV HIV Med 22 2021 372 378 33368966
8 Department of Health and Human ServicesNational Institutes of Health Guidance for COVID-19 and people with HIV Available at: https://clinicalinfo.hiv.gov/en/guidelines/guidance-covid-19-and-people-hiv/guidance-covid-19-and-people-hiv
9 Lin K.Y. Wu P.Y. Liu W.D. Effectiveness of COVID-19 vaccination among people living with HIV during a COVID-19 outbreak J Microbiol Immunol Infect 55 2022 535 539 35577735
10 Huang H.Y. Chan P.C. Huang Y.C. An outbreak of SARS-CoV-2 infections among people living with HIV and its successful containment-Taiwan, May to August 2021 J Formos Med Assoc 2022 S0929-6646(22)00180-00182. [Epub ahead of print]
11 Tebas P. Frank I. Lewis M. Poor immunogenicity of the H1N1 2009 vaccine in well controlled HIV-infected individuals AIDS 24 2010 2187 2192 20616698
12 Mena G. García-Basteiro A.L. Bayas J.M. Hepatitis B and A vaccination in HIV-infected adults: a review Hum Vaccines Immunother 11 2015 2582 2598
13 Madhi S.A. Koen A.L. Izu A. Safety and immunogenicity of the ChAdOx1 nCoV-19 (AZD1222) vaccine against SARS-CoV-2 in people living with and without HIV in South Africa: an interim analysis of a randomised, double-blind, placebo-controlled, phase 1B/2A trial Lancet HIV 8 2021 e568 e580 34416193
14 Frater J. Ewer K.J. Ogbe A. Safety and immunogenicity of the ChAdOx1 nCoV-19 (AZD1222) vaccine against SARS-CoV-2 in HIV infection: a single-arm substudy of a phase 2/3 clinical trial Lancet HIV 8 2021 e474 e485 34153264
15 Lapointe H.R. Mwimanzi F. Cheung P.K. People with HIV receiving suppressive antiretroviral therapy show typical antibody durability after dual COVID-19 vaccination, and strong third dose responses J Infect Dis 2022 jiac229 ([epub ahead print])
16 Jedicke N. Stankov M.V. Cossmann A. Humoral immune response following prime and boost BNT162b2 vaccination in people living with HIV on antiretroviral therapy HIV Med 23 2022 558 563 34725907
17 Woldemeskel B.A. Karaba A.H. Garliss C.C. The BNT162b2 mRNA vaccine elicits robust humoral and cellular immune responses in people living with human immunodeficiency virus (HIV) Clin Infect Dis 7 2022 1268 1270
18 González de Aledo M. Cañizares A. Vázquez-Rodríguez P. Safety and Immunogenicity of SARS-CoV-2 vaccines in people with HIV AIDS 36 2022 691 695 34999608
19 Dimeglio C. Herin F. Martin-Blondel G. Miedougé M. Izopet J. Antibody titers and protection against a SARS-CoV-2 infection J Infect 84 2022 248 288
20 Feng S. Phillips D.J. White T. Correlates of protection against symptomatic and asymptomatic SARS-CoV-2 infection Nat Med 27 2021 2032 2040 34588689
21 Brumme Z.L. Mwimanzi F. Lapointe H.R. Humoral immune responses to COVID-19 vaccination in people living with HIV receiving suppressive antiretroviral therapy NPJ Vaccines 7 2022 28 35228535
22 Sheng W.H. Chang S.Y. Lin P.H. Immune response and safety of heterologous ChAdOx1-nCoV-19/mRNA-1273 vaccination compared with homologous ChAdOx1-nCoV-19 or homologous mRNA-1273 vaccination J Formos Med Assoc 121 2022 766 777 35305895
23 Voysey M. Costa Clemens S.A. Madhi S.A. Single-dose administration and the influence of the timing of the booster dose on immunogenicity and efficacy of ChAdOx1 nCoV-19 (AZD1222) vaccine: a pooled analysis of four randomised trials Lancet 397 2021 881 891 33617777
24 Grunau B. Goldfarb D.M. Asamoah-Boaheng M. Immunogenicity of extended mRNA SARS-CoV-2 vaccine dosing intervals JAMA 327 2022 279 281 34860253
25 Kanji J.N. Bailey A. Fenton J. Stability of SARS-CoV-2 IgG in multiple laboratory conditions and blood sample types J Clin Virol 142 2021 104933
26 Shurrab F.M. Al-Sadeq D.W. Amanullah F. Effect of multiple freeze-thaw cycles on the detection of anti-SARS-CoV-2 IgG antibodies J Med Microbiol 70 2021 001402
27 Mishra S.K. Pradhan S.K. Pati S. Sahu S. Nanda R.K. Waning of anti-spike antibodies in AZD1222 (ChAdOx1) vaccinated healthcare providers: a prospective longitudinal study Cureus 13 2021 e19879
28 Levin E.G. Lustig Y. Cohen C. Waning immune humoral response to BNT162b2 COVID-19 vaccine over 6 months N Engl J Med 385 2021 e84 34614326
29 Corma-Gómez A. Fernández-Fuertes M. García E. Severe immunosuppression is related to poorer immunogenicity to SARS-CoV-2 vaccines among people living with HIV Clin Microbiol Infect 2022 S1198-743X(22)00276-2. [Epub ahead of print]
30 Nault L. Marchitto L. Goyette G. Covid-19 vaccine immunogenicity in people living with HIV-1 Vaccine 40 2022 3633 3637 35568588
31 Heftdal L.D. Knudsen A.D. Hamm S.R. Humoral response to two doses of BNT162b2 vaccination in people with HIV J Intern Med 291 2022 513 518 34841629
32 Touizer E. Alrubayyi A. Rees-Spear C. Failure to seroconvert after two doses of BNT162b2 SARS-CoV-2 vaccine in a patient with uncontrolled HIV Lancet HIV 8 2021 e317 e318 34087093
33 Antinori A. Cicalini S. Meschi S. Humoral and cellular immune response elicited by mRNA vaccination against SARS-CoV-2 in people living with HIV (PLWH) receiving antiretroviral therapy (ART) according with current CD4 T-lymphocyte count Clin Infect Dis 2022 ciac238 ([Epub ahead of print])
34 Zimmermann P. Curtis N. Factors that influence the immune response to vaccination Clin Microbiol Rev 32 2019 e00084-18
35 Sanders J.F. Bemelman F.J. Messchendorp A.L. The RECOVAC immune-response study: the immunogenicity, tolerability, and safety of COVID-19 vaccination in patients with chronic kidney disease, on dialysis, or living with a kidney transplant Transplantation 106 2022 821 834 34753894
36 Soetedjo N.N.M. Iryaningrum M.R. Lawrensia S. Permana H. Antibody response following SARS-CoV-2 vaccination among patients with type 2 diabetes mellitus: a systematic review Diabetes Metabol Syndr 16 2022 102406
| 0 | PMC9745965 | NO-CC CODE | 2022-12-16 23:21:38 | no | J Virus Erad. 2022 Dec 13; 8(4):100308 | utf-8 | J Virus Erad | 2,022 | 10.1016/j.jve.2022.100308 | oa_other |
==== Front
J Infect Chemother
J Infect Chemother
Journal of Infection and Chemotherapy
1341-321X
1437-7780
Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd.
S1341-321X(22)00326-9
10.1016/j.jiac.2022.12.003
Original Article
Comparison of the clinical characteristics and outcomes of Japanese patients with COVID-19 treated in primary, secondary, and tertiary care facilities
Tomidokoro Daiki a
Asai Yusuke b
Hayakawa Kayoko c
Kutsuna Satoshi e
Terada Mari cd
Sugiura Wataru d
Ohmagari Norio c
Hiroi Yukio a∗
a Department of Cardiology, National Center for Global Health and Medicine, Tokyo, Japan
b AMR Clinical Reference Center, National Center for Global Health and Medicine, Tokyo, Japan
c Disease Control and Prevention Center, National Center for Global Health and Medicine, Tokyo, Japan
d Center for Clinical Sciences, National Center for Global Health and Medicine, Tokyo, Japan
e Department of Infection Control and Prevention, Graduate School of Medicine, Faculty of Medicine, Osaka University, Osaka, Japan
∗ Corresponding author. Department of Cardiology, National Center for Global Health and Medicine, 1-21-1 Toyama, Shinjuku-ku, Tokyo, 162-8655, Japan.
13 12 2022
13 12 2022
23 10 2022
25 11 2022
7 12 2022
© 2022 Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases. Published by Elsevier Ltd.
2022
Japanese Society of Chemotherapy and The Japanese Association for Infectious Diseases
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Aim
To compare the characteristics and clinical course of patients with coronavirus disease (COVID-19) according to the healthcare level of the admitted hospital, to provide an insight into determining the appropriate level of care for each patient.
Methods
This retrospective, observational study utilized data from the COVID-19 Registry Japan (COVIREGI-JP), the largest Japanese registry of hospitalized patients with COVID-19. Datasets were obtained from reports filed as of May 31, 2022.
Results
A total of 59,707 patients (2004 in the primary care group, 41,420 in the secondary care group, and 16,283 in the tertiary care group) from 585 facilities were included in the analysis. Patients with established risk factors for severe disease, such as old age and the presence of comorbidities, were treated at higher care facilities and had poorer initial conditions and in-hospital clinical course, as well as higher mortality. Analysis of the fatality rates for each complication suggested that patients with complications requiring procedures (e.g. pleural effusions, myocardial ischemia, and arrhythmia) may have better survival rates in facilities with specialist availability. The number of deaths and severe COVID-19 cases in this study were notably less than those reported overseas.
Conclusion
Our results showed that more difficult COVID-19 cases with poor outcomes were treated at higher care level facilities in Japan. Attending to possible complications may be useful for selecting an appropriate treatment hospital. Healthcare providers need to maintain a broad perspective on the distribution of medical resources.
Keywords
COVID-19
Medical resources
Healthcare policy
Level of care
Japan
Registry
==== Body
pmc1 Introduction
The surges of the coronavirus disease (COVID-19) pandemic continue to burden the Japanese healthcare system. Data suggests that newer strains of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2; the causative agent of COVID-19), such as the Omicron variant, cause less severe illness and death, but are more transmissible, leading to a sudden surge in cases. Under such circumstances, it is important to consider the allocation of limited medical resources. Additionally, even when the respiratory infection itself is mild, significant risks of poor outcomes remain owing to worsening comorbidities or the development of new complications. With different degrees of accessibility to specialized treatment in each healthcare facility, it is often difficult for healthcare providers to determine the appropriate level of medical service for the patient. The “COVID-19 Registry Japan” (COVIREGI-JP), a nationwide COVID-19 inpatient registry, provides a comprehensive analysis of Japanese COVID-19 cases among various healthcare facilities [1]. However, as no analysis has been based on the facility's level of care, the inter-hospital differences in patient characteristics and their clinical course remain unclear. Here, we compared the characteristics and outcomes of infected patients among primary, secondary, and tertiary level care facilities with the aim to examine whether the level of care they received was appropriate.
2 Material and methods
2.1 Study design and data sources
This retrospective observational study utilized data from the COVIREGI-JP, a Japanese registry of more than 70,000 patients with COVID-19, hospitalized in 695 facilities [1]. Additional details of the registry including patient inclusion criteria and data collection method have been described elsewhere [1,2]. Each research collaborator completed a report form that was developed specifically for the registry. Datasets were obtained from reports filed as of May 31, 2022.
2.2 Healthcare level of facility
Participating facilities in the registry were classified as primary, secondary, or tertiary care hospitals. To determine the level of care, we referred to the emergency care category of the facility, which was designated based on the Medical Care Act. Japan has three categories of emergency hospitals: a “primary” emergency center, dealing with patients who can be managed as outpatients; a “secondary” emergency center, dealing with patients who can be managed as inpatients on a general medical floor; and a “tertiary” emergency center, dealing with patients who require specialized or intensive care. Here in our study, a facility is classified as a primary, secondary, and tertiary care hospital if it is designated as a primary, secondary, and tertiary emergency center, respectively. Smaller facilities that did not participate in emergency care (e.g. psychiatric hospitals or clinics with minimal beds) were also included in the primary care category. Patients treated at primary, secondary, and tertiary care hospitals were defined as primary, secondary, and tertiary care groups, respectively.
2.3 Baseline characteristics
Data on patients' demographics, including age, sex, body mass index (BMI), smoking and alcohol habits, and comorbidities were collected. Comorbidities were pre-specified in the report form and included myocardial infarction, congestive heart failure, peripheral vascular disease, cerebrovascular disease, dementia, chronic obstructive pulmonary disease (COPD), chronic lung disease, bronchial asthma, liver disease, peptic ulcer, hypertension, hyperlipidemia, diabetes with and without complications, obesity, kidney disease, hemodialysis, solid tumor, leukemia, lymphoma, immunosuppression, and collagen disease. The vital signs (temperature, heart rate, respiratory rate, and blood pressures), COVID-19 severity, oxygen support and route, and findings on imaging (X-ray and/or CT) on admission were also recorded. COVID-19 severity on admission was defined as either severe or non-severe, based on previous studies [1,3]. Severe cases were those that met at least one of the following criteria: (1) requiring invasive or non-invasive mechanical ventilation, (2) requiring supplemental oxygen, (3) SpO2 ≤ 94% in room air, or (4) tachypnea with a respiratory rate ≥24 breaths per minute. Those who did not meet any of these criteria were classified as non-severe.
2.4 Clinical course and complications
The clinical course for each group was evaluated by intensive care unit (ICU) admissions and the use of supportive care, including oxygen support, prone positioning, tracheotomy, nitric oxide (NO) inhalation, vasopressor/inotropic administration, renal replacement therapy/hemodialysis, and blood transfusion. Complications specified in the report form were collected, which included pneumonia coinfections, respiratory conditions (pneumothorax, pleural effusion, bloody sputum, acute respiratory distress syndrome [ARDS]), bacteremia, cardiovascular conditions (myocarditis/pericarditis, arrhythmia, myocardial ischemia, deep vein thrombosis [DVT]), pulmonary thromboembolism [PE]), neurological conditions (meningitis, seizure, stroke), and gastrointestinal bleeding.
2.5 Outcomes
We collected data on the outcome (death or discharge) and the length of hospitalization. Discharge destination was classified as one of the following: discharge to home, transfer to a non-medical facility, transfer to a rehabilitation/long-term care facility, transfer to a medical facility for intensive care, death, and others. Mortality rates by complications were also calculated to determine which complications had a fatal course and whether they differed at the institutional level.
2.6 Statistical analysis
Continuous variables are expressed as medians and interquartile ranges and categorical variables as number of cases and percentages. Continuous variables are compared between the groups using the Kruskal-Wallis test and categorical variables using the chi-square test. All statistical analysis were performed using R version 4.1.3 (R Foundation for Statistical Computing, Vienna, Austria).
2.7 Ethics
This study was approved by the National Center for Global Health and Medicine Ethics Review Board, (NCGM-G-004037-02) which waived the requirement for informed consent due to the retrospective nature of the study. The protocol was conducted in accordance with the Declaration of Helsinki.
3 Results
Data of 59,707 cases from 585 facilities were included in the analysis, comprising 2,004, 41,420, and 16,283 patients in the primary, secondary, and tertiary care groups, respectively. The healthcare levels of the included facilities were categorized as 20, 421, and 144 primary, secondary, and tertiary care hospitals, respectively.
3.1 Patient characteristics and condition on admission
The patient demographics and comorbidities are summarized in Table 1 . The median age of the patients was 54, 57, and 58 years in the primary, secondary, and tertiary care groups, respectively. The sex and BMI were similar in the primary and tertiary care groups, with 60% males and a median BMI of 23.6. The secondary care group consisted of 55.6% males and had a lower median BMI of 23.2. Regarding alcohol consumption and drinking habits, patients in higher-level facilities were more likely to consume alcohol daily and have a history of smoking. Those who had at least one comorbidity constituted 51.8%, 55.5%, and 60.6% of the patients in the primary, secondary, and tertiary care groups, respectively. The tertiary care group had the highest prevalence for all listed comorbidities, with the only exception of dementia (9.5%, 8.6%, and 6.8% in the primary, secondary, and tertiary care groups, respectively).Table 1 Baseline characteristics of patients with COVID-19 by care level of the admitted hospital. Data are presented as n (%) or median [IQR]. Numbers do not always add up due to missing values.
Table 1 Total (n = 59,707) Primary (n = 2004) Secondary (n = 41,420) Tertiary (n = 16,283) p-value
Age (years) 57 [39, 74] 54 [37, 74] 57 [39, 74] 58 [41, 74] <0.001
Sex Male 33990 (56.9) 1194 (59.6) 23026 (55.6) 9770 (60.0) <0.001
Female 25698 (43.0) 807 (40.3) 18381 (44.4) 6510 (40.0)
BMI (kg/m^2) 23.4 [20.7, 26.5] 23.6 [20.8, 26.9] 23.2 [20.6, 26.3] 23.6 [20.9, 26.9] <0.001
Smoking history Current smoking 8936 (15.0) 347 (17.3) 6142 (14.8) 2447 (15.0) <0.001
Past smoking 12391 (20.8) 380 (19.0) 8136 (19.6) 3875 (23.8)
Never 29023 (48.6) 1028 (51.3) 20533 (49.6) 7462 (45.8)
Unknown 9250 (15.5) 244 (12.2) 6554 (15.8) 2452 (15.1)
Alcohol intake Daily 3309 (5.5) 77 (3.8) 2201 (5.3) 1031 (6.3) <0.001
Occasional 18185 (30.5) 372 (18.6) 12623 (30.5) 5190 (31.9)
Never 23365 (39.1) 580 (28.9) 16742 (40.4) 6043 (37.1)
Unknown 14151 (23.7) 965 (48.2) 9397 (22.7) 3789 (23.3)
Comorbidities Any comorbidity 33892 (56.8) 1039 (51.8) 22988 (55.5) 9865 (60.6) <0.001
Myocardial infarction 1175 (2.0) 42 (2.1) 733 (1.8) 400 (2.5) <0.001
Congestive heart failure 1979 (3.3) 48 (2.4) 1344 (3.2) 587 (3.6) 0.008
Peripheral vascular disease 1002 (1.7) 20 (1.0) 626 (1.5) 356 (2.2) <0.001
Cerebrovascular disease 4057 (6.8) 103 (5.1) 2827 (6.8) 1127 (6.9) 0.009
Dementia 4851 (8.1) 190 (9.5) 3548 (8.6) 1113 (6.8) <0.001
COPD 1480 (2.5) 29 (1.4) 958 (2.3) 493 (3.0) <0.001
Chronic lung disease (excluding COPD) 940 (1.6) 34 (1.7) 602 (1.5) 304 (1.9) 0.003
Bronchial asthma 3277 (5.5) 97 (4.8) 2284 (5.5) 896 (5.5) 0.46
Mild liver disease 1418 (2.4) 31 (1.5) 773 (1.9) 614 (3.8) <0.001
Moderate to severe liver disease 228 (0.4) 7 (0.3) 150 (0.4) 71 (0.4) 0.42
Peptic ulcer 494 (0.8) 12 (0.6) 317 (0.8) 165 (1.0) 0.011
Hypertension 18259 (30.6) 508 (25.3) 12422 (30.0) 5329 (32.7) <0.001
Hyperlipidemia 8679 (14.5) 191 (9.5) 5822 (14.1) 2666 (16.4) <0.001
Diabetes without complications 8886 (14.9) 243 (12.1) 6035 (14.6) 2608 (16.0) <0.001
Diabetes with complications 1311 (2.2) 24 (1.2) 737 (1.8) 550 (3.4) <0.001
Obesity 4190 (7.0) 91 (4.5) 2770 (6.7) 1329 (8.2) <0.001
Moderate to severe kidney disease 1080 (1.8) 20 (1.0) 603 (1.5) 457 (2.8) <0.001
Hemodialysis 611 (1.0) 6 (0.3) 351 (0.8) 254 (1.6) <0.001
Solid tumor 2148 (3.6) 70 (3.5) 1401 (3.4) 677 (4.2) <0.001
Metastatic solid tumor 555 (0.9) 10 (0.5) 375 (0.9) 170 (1.0) 0.03
Leukemia 176 (0.3) 5 (0.2) 109 (0.3) 62 (0.4) 0.06
Lymphoma 342 (0.6) 8 (0.4) 237 (0.6) 97 (0.6) 0.52
Immunosuppression 1508 (2.5) 48 (2.4) 957 (2.3) 503 (3.1) <0.001
Collagen disease 817 (1.4) 17 (0.8) 554 (1.3) 246 (1.5) 0.02
BMI, body mass index; COPD, chronic obstructive pulmonary disease.
Table 2 shows the patients' medical conditions on hospital admission. The percentage of patients with severe COVID-19 was 26.2%, 29.3%, and 44.9% in the primary, secondary, and tertiary care groups, respectively. Advanced care facilities treated more patients with oxygen demand on admission (10.4% in primary. vs. 13.4% in secondary vs. 13.4% in tertiary care group). Regarding the administration route for patients on oxygen support, 76.1% of patients with oxygen support in the primary care group and 71.3% in the secondary care group required only nasal cannula. In contrast, in the tertiary care group, nasal cannula was sufficient for only 46.3% of those requiring oxygen support, while the remainder of patients required more intensive respiratory support. The percentage of patients with pneumonia findings on imaging in the primary and secondary care groups was similar at 52.7% for X-ray and 71% for CT, while it reached 63.3% for X-ray and 79.8% for CT in the tertiary care group.Table 2 Initial condition of patients with COVID-19 by care level of the admitted hospital. Data are presented as n (%) or median [IQR]. Numbers do not always add up due to missing values.
Table 2 Total (n = 59,707) Primary (n = 2004) Secondary (n = 41,420) Tertiary (n = 16,283) p-value
Vital signs Temperature (°C) 37 [36.6, 37.8] 37 [36.6, 37.7] 37 [36.6, 37.7] 37.1 [36.6, 37.8] <0.001
Heart rate (beats/minute) 86 [76, 98] 87 [76, 98] 86 [75, 98] 87 [76, 99] <0.001
Respiratory rate (breaths/minute) 18 [16,21] 18 [16,20] 18 [16,21] 19 [16,23] <0.001
Systolic blood pressure (mmHg) 127 [114, 142] 125 [113, 138] 127 [114, 142] 127 [114, 142] <0.001
Diastolic blood pressure (mmHg) 79 [70, 88] 80 [70, 88] 79 [70, 88] 78 [69, 88] <0.001
Severity of COVID-19 Severe 19975 (33.5) 525 (26.2) 12142 (29.3) 7308 (44.9) <0.001
Non-severe 39732 (66.5) 1479 (73.8) 29278 (70.7) 8975 (55.1)
Oxygen support Yes 10401 (17.4) 209 (10.4) 5541 (13.4) 4651 (28.6) <0.001
Route of oxygen support Nasal cannula 6265 (60.2) 159 (76.1) 3953 (71.3) 2153 (46.3)
Face mask 1885 (18.1) 33 (15.8) 863 (15.6) 989 (21.3)
Reservoir mask 1326 (12.7) 16 (7.7) 502 (9.1) 808 (17.4)
High-flow oxygen device 252 (2.4) 0 (0) 108 (1.9) 144 (3.1)
Non-invasive ventilation 46 (0.4) 0 (0) 15 (0.3) 31 (0.7)
Invasive mechanical ventilation 623 (6.0) 1 (0.5) 100 (1.8) 522 (11.2)
ECMO 4 (0) 0 (0) 0 (0) 4 (0.1)
None (room air) 48974 (82) 1795 (89.6) 35657 (86.1) 11522 (70.8)
Imaging findings
Chest X-ray Data available 0 0 0 0
No abnormality 13244 (42.2) 397 (46.4) 9320 (45.6) 3527 (35.1) <0.001
Pneumonia 17596 (56.1) 451 (52.7) 10781 (52.7) 6364 (63.3)
Abnormality (excluding pneumonia) 520 (1.7) 7 (0.8) 354 (1.7) 159 (1.6)
CT Data available 0 0 0 0
No abnormality 10128 (23.8) 437 (25.8) 7769 (26.1) 1922 (17.3) <0.001
Pneumonia 31244 (73.4) 1212 (71.5) 21166 (71.1) 8866 (79.8)
Abnormality (excluding pneumonia) 1214 (2.9) 46 (2.7) 844 (2.8) 324 (2.9)
ECMO, extracorporeal membrane oxygenation; CT, computed tomography.
3.2 Clinical course, complications, and outcomes
Table 3 shows the clinical indicators of the in-hospital course for each group. ICU admission occurred in 0.4%, 3.8%, and 18.1% of patients in the primary, secondary, and tertiary care groups, respectively. Various types of intensive treatment (e.g. oxygen support, vasopressor/inotropic administration, and renal replacement therapy), were consistently used more frequently in higher care facilities. With the exception of nasal cannula and high-flow oxygen device, such intensive treatment was used in less than ten patients in the primary care group. ECMO was used in a total of 208 patients, which consisted of 0, 32, and 176 patients in the primary, secondary, and tertiary care groups, respectively.Table 3 In-hospital clinical course of patients with COVID-19 by care level of the admitted hospital. Data are presented as n (%). Numbers do not always add up due to missing values.
Table 3 Total (n = 59,707) Primary (n = 2004) Secondary (n = 41,420) Tertiary (n = 16,283) p-value
ICU admission 4554 (7.6) 9 (0.4) 1593 (3.8) 2952 (18.1) <0.001
Oxygen support 22785 (38.2) 631 (31.5) 14174 (34.2) 7980 (49.0) <0.001
Route of oxygen support Nasal cannula or face/resorvoir mask 22526 (37.7) 629 (31.4) 14069 (34.0) 7828 (48.1)
High-flow oxygen device 3085 (5.2) 52 (2.6) 1515 (3.7) 1518 (9.3)
Non-invasive ventilation 804 (1.3) 4 (0.2) 252 (0.6) 548 (3.4)
Invasive mechanical ventilation 2664 (4.5) 8 (0.4) 945 (2.3) 1711 (10.5)
ECMO 208 (0.3) 0 (0) 32 (0.1) 176 (1.1)
Prone positioning 2896 (4.9) 2 (0.1) 1201 (2.9) 1693 (10.4) <0.001
Tracheotomy 500 (0.8) 2 (0.1) 120 (0.3) 378 (2.3) <0.001
Nitric oxide inhalation 42 (0.1) 0 (0) 13 (0) 29 (0.2) <0.001
Vasopressor/Inotropic administration 1520 (2.5) 1 (0) 507 (1.2) 1012 (6.2) <0.001
RRT or hemodialysis 806 (1.3) 7 (0.3) 352 (0.8) 447 (2.7) <0.001
Blood transfusion 1126 (1.9) 7 (0.3) 410 (1.0) 709 (4.4) <0.001
ICU, intensive care unit; ECMO, extracorporeal membrane oxygenation; RRT; renal replacement therapy.
As indicated in Table 4 , complication rates tended to be higher in the tertiary care group. In hospitals providing a higher level of care, non-COVID-19 pneumonia coinfection and respiratory-related complications occurred more frequently than in hospitals that provided a lower level of care. Non-COVID-19 pneumonia coinfection cases were infrequent in the primary care group; there were only 5 cases of viral pneumonia, 67 cases of bacterial pneumonia, 3 cases of methicillin-resistant Staphylococcus aureus (MRSA) pneumonia, and 2 cases of Pseudomonas aeruginosa pneumonia. Patients developed ARDS in 1.0%, 3.2%, and 8.5% of the primary, secondary, and tertiary care groups, respectively. The incidence rates for almost all the other complications were lowest in the primary care group and highest in the tertiary care group.Table 4 Complications in patients with COVID-19 by care level of the admitted hospital. Data are presented as n (%).
Table 4 Total (n = 59,707) Primary (n = 2004) Secondary (n = 41,420) Tertiary (n = 16,283) p-value
Pneumonia coinfection
Viral pneumonia (excluding COVID-19) 511 (0.9) 5 (0.2) 308 (0.7) 198 (1.2) <0.001
Bacterial pneumonia (including HAP/VAP) 3226 (5.4) 67 (3.3) 1703 (4.1) 1456 (8.9) <0.001
MRSA pneumonia 281 (0.5) 3 (0.1) 119 (0.3) 159 (1.0) <0.001
Pseudomonas aeruginosa pneumonia 190 (0.3) 2 (0.1) 85 (0.2) 103 (0.6) <0.001
MDRP pneumonia 30 (0.1) 0 (0) 8 (0) 22 (0.1) <0.001
Respiratory
Pneumothorax 249 (0.4) 0 (0) 95 (0.2) 154 (0.9) <0.001
Pleural effusion 1751 (2.9) 31 (1.5) 957 (2.3) 763 (4.7) <0.001
Bloody sputum 459 (0.8) 13 (0.6) 236 (0.6) 210 (1.3) <0.001
ARDS 2747 (4.6) 20 (1.0) 1335 (3.2) 1392 (8.5) <0.001
Infectious
Bacteremia 595 (1.0) 5 (0.2) 249 (0.6) 341 (2.1) <0.001
Cardiovascular
Endocarditis 18 (0) 0 (0) 11 (0) 7 (0) <0.001
Myocarditis/Pericarditis 58 (0.1) 0 (0) 27 (0.1) 31 (0.2) <0.001
Arrhythmia 262 (0.4) 2 (0.1) 150 (0.4) 110 (0.7) <0.001
Cardiac ischemia 104 (0.2) 7 (0.3) 47 (0.1) 50 (0.3) <0.001
Deep vein thrombosis 450 (0.8) 4 (0.2) 248 (0.6) 198 (1.2) <0.001
Pulmonary thromboembolism 218 (0.4) 1 (0) 134 (0.3) 83 (0.5) <0.001
Neurological
Meningitis 36 (0.1) 0 (0) 21 (0.1) 15 (0.1) <0.001
Seizure 152 (0.3) 1 (0) 80 (0.2) 71 (0.4) <0.001
Stroke 212 (0.4) 2 (0.1) 119 (0.3) 91 (0.6) <0.001
Gastrointestinal
Gastrointestinal bleeding 363 (0.6) 4 (0.2) 184 (0.4) 175 (1.1) <0.001
HAP, hospital-acquired (or nosocomial) pneumonia; VAP, ventilator-associated pneumonia; MRSA, Methicillin-resistant Staphylococcus aureus; MDRP, Multidrug-resistant Pseudomonas aeruginosa; ARDS, acute respiratory distress syndrome.
The outcome and length of hospital stay among the three groups are shown in Table 5 . More deaths and fewer home discharges were observed in patients treated at higher care facilities. The median length of stay was 11 days in the tertiary care group and 10 days in the primary and secondary care groups.Table 5 Outcomes of patients with COVID-19 by care level of the admitted hospital. Data are presented as n (%). Numbers do not always add up due to missing values.
Table 5 Total (n = 59,707) Primary (n = 2004) Secondary (n = 41,420) Tertiary (n = 16,283) p-value
Outcome Death 2816 (4.7) 52 (2.6) 1672 (4.0) 1092 (6.7) <0.001
Discharge 56865 (95.2) 1952 (97.4) 39723 (95.9) 15190 (93.3)
Discharge destination Home 44580 (78.4) 1627 (83.4) 31347 (78.9) 11606 (76.4)
Non-medical facility 2714 (4.8) 36 (1.8) 2023 (5.1) 655 (4.3)
Rehabilitation/long-term care facility 9353 (16.4) 276 (14.1) 6162 (15.5) 2915 (19.2)
Medical facility for intensive care 199 (0.3) 13 (0.7) 178 (0.4) 8 (0.1)
Others 19 (0) 0 (0) 13 (0) 6 (0)
Length of hospital stay (days) 10 [8,15] 10 [7,14] 10 [8,14] 11 [8,17] <0.001
3.3 Outcomes for each complication
Table 6 shows the analysis of in-hospital mortality by complication. The complications with the highest mortality rates overall were MRSA pneumonia (50.2%), pneumothorax (47.8%), gastrointestinal bleeding (44.6%), bacteremia (44.2%), Pseudomonas aeruginosa pneumonia (42.1%, and 43.3% for multiple drug resistant strains), and myocardial ischemia (37.5%).Table 6 In-hospital mortality of patients with COVID-19 with various complications by care level of the admitted hospital. Data are presented as n (%). “-” in primary care group indicates that there were no reported cases of the complication.
Table 6 Total (n = 59,707) Primary (n = 2004) Secondary (n = 41,420) Tertiary (n = 16,283) p-value (primary vs secondary) p-value (secondary vs tertiary)
Cases Deaths (mortality rate) Cases Deaths Cases Deaths Cases Deaths
Pneumonia coinfection
Viral pneumonia (excluding COVID-19) 511 91 (17.8) 5 0 (0) 308 44 (14.3) 198 47 (23.7) 0.36 0.006
Bacterial pneumonia (including HAP/VAP) 3226 913 (28.3) 67 22 (32.8) 1703 448 (26.3) 1456 443 (30.4) 0.23 0.01
MRSA pneumonia 281 141 (50.2) 3 2 (66.7) 119 61 (51.3) 159 78 (49.1) 0.59 0.71
Pseudomonas aeruginosa pneumonia 190 80 (42.1) 2 1 (50.0) 85 34 (40.0) 103 45 (43.7) 0.77 0.61
MDRP pneumonia 30 13 (43.3) 0 0 (−) 8 3 (37.5) 22 10 (45.5) N/A 0.69
Respiratory
Pneumothorax 249 119 (47.8) 0 0 (−) 95 44 (46.3) 154 75 (48.7) N/A 0.71
Pleural effusion 1751 544 (31.1) 31 13 (41.9) 957 316 (33.0) 763 215 (28.2) 0.29 0.03
Bloody sputum 459 130 (28.3) 13 0 (0) 236 68 (28.8) 210 62 (29.5) 0.02 0.86
ARDS 2747 864 (31.5) 20 10 (50) 1335 419 (31.4) 1392 435 (31.3) 0.07 0.93
Infectious
Bacteremia 595 263 (44.2) 5 1 (20.0) 249 119 (47.8) 341 143 (41.9) 0.21 0.15
Cardiovascular
Endocarditis 18 1 (5.6) 0 0 (−) 11 1 (9.1) 7 0 (0) N/A 0.41
Myocarditis/Pericarditis 58 13 (22.4) 0 0 (−) 27 5 (18.5) 31 8 (25.8) N/A 0.5
Arrhythmia 262 96 (36.6) 2 1 (50.0) 150 62 (41.3) 110 33 (30.0) 0.8 0.06
Myocardial ischemia 104 39 (37.5) 7 6 (85.7) 47 17 (36.2) 50 16 (32.0) 0.01 0.66
Deep vein thrombosis 450 87 (19.3) 4 0 (0) 248 39 (15.7) 198 48 (24.2) 0.38 0.02
Pulmonary thromboembolism 218 34 (15.6) 1 0 (0) 134 14 (10.4) 83 20 (24.1) 0.73 0.007
Neurological
Meningitis 36 10 (27.8) 0 0 (−) 21 7 (33.3) 15 3 (20.0) N/A 0.37
Seizure 152 37 (24.3) 1 1 (100.0) 80 19 (23.8) 71 17 (23.9) 0.07 0.97
Stroke 212 70 (33.0) 2 0 (0) 119 34 (28.6) 91 36 (39.6) 0.37 0.09
Gastrointestinal
Gastrointestinal bleeding 363 162 (44.6) 4 1 (25.0) 184 85 (46.2) 175 76 (43.4) 0.39 0.59
HAP, hospital-acquired (or nosocomial) pneumonia; VAP, ventilator-associated pneumonia; MRSA, Methicillin-resistant Staphylococcus aureus; MDRP, Multidrug-resistant Pseudomonas aeruginosa; ARDS, acute respiratory distress syndrome.
Mortality rates varied substantially by the type of complication. ARDS and non-COVID-19 mixed pneumonia were rare in primary care facilities, but when they did occur, the mortality rate exceeded 30%, which was much higher than those in secondary and tertiary care settings. Myocardial ischemia also had a high fatality rate of 85.7% in the primary care group, which was significantly higher than the 36.2% and 32.0% in the secondary and tertiary care groups, respectively.
Compared between the primary group and the secondary care groups, mortality rate of myocardial ischemia was significantly lower in the secondary group (36.2% vs. 85.7%, p = 0.01), whereas for bloody sputum it was higher in secondary group (28.8% vs. 0%, p = 0.02). The fatality rate was higher in the tertiary care group than in the secondary care group when complicated by viral pneumonia other than COVID-19 (23.7% vs. 14.3%, p = 0.006), bacterial pneumonia (30.4% vs. 26.3%, p = 0.01), deep vein thrombosis (24.2% vs. 15.7%, p = 0.02), and pulmonary embolism (24.1% vs. 10.4%, p = 0.007). Conversely, the fatality rate was higher in the secondary care group than that in the tertiary care group when complicated by pleural effusion (33.0% vs. 28.2%, p = 0.03), with a similar trend observed for arrhythmia (41.3% vs. 30.0%, p = 0.06).
4 Discussion and conclusions
This is the first report in Japan to compare the characteristics and clinical course of patients with COVID-19 according to the health care level of the admitted hospital. Previous reports from Japan have focused on analyses classified according to patient factors; however, to improve actual practice, assessments based on factors regarding healthcare providers are equally important. In Japan, during the stable phases of the pandemic, there have been no major restrictions on inpatient or outpatient practice. However, even 3 years after the start of the pandemic, hospitals remain short of beds during surges of infections. Determining the appropriate level of care for patients with COVID-19 remains unresolved. Here, we revealed that patients treated at higher-level medical facilities had a higher risk of severe disease, more comorbidities, and a poorer clinical course.
Differences in outcomes among healthcare facilities have previously been reported outside Japan. In a study in the USA, during the first 6 months of the pandemic, the mortality rates were strongly associated with the prevalence of COVID-19 in the hospital's surrounding communities, but neither the number of intensive care unit beds nor academic status was relevant [4]. Another USA report suggested that treatment at a hospital dedicated to COVID-19 care performed better in terms of the survival rates of patients with severe COVID-19 [5]. In that study, mortality rates were higher in specialized hospitals, but multivariate adjustment verified that their outcomes were better when patients' severity was accounted for. Socioeconomic determinants, such as poor prognosis in hospitals with poor finances in disadvantaged areas, have also been shown to influence outcomes [6].
We showed that patients at higher risk of severe disease were treated at higher care institutions in Japan. Age, smoking, and the presence of comorbidities are established risk factors for severe COVID-19, and these characteristics are more prevalent in patients in higher care facilities [1,7]. Surprisingly, the opposite trend was observed for dementia. Dementia and end-of-life care are major issues in Japan's aging society [8]. Patients with COVID-19 with dementia may have been sent to a lower-level care hospital because they had previously decided to avoid intensive treatment. Male sex and high BMI are other risk factors for severe disease but were less common in the secondary care group. Japan is a hyper-aged society, with a large proportion of frail elderly women [9], and frailty is also associated with poor prognosis in patients with COVID-19 [10]. A subgroup of elderly women with low BMI may have been treated at higher-care-level facilities, resulting in a unique distribution in terms of sex and BMI among the three groups.
Despite there being no clear difference among the three groups in terms of vital signs alone, severe cases, as defined by respiratory status, were transported to higher care facilities. This was supported by imaging findings and oxygen administration status on admission. Unsurprisingly, patients admitted to higher-level facilities had a poor clinical course after admission. Intensive interventions were most common in tertiary care hospitals, and the tertiary group experienced more complications and had poor outcomes. Conversely, patients treated in primary care settings rarely required intensive care or experienced complications during their course. Although the present study was not adjusted with multivariate analysis, the less favorable clinical course in higher-level care facilities is reasonable, given the percentage of high-risk patients with COVID-19.
The fatality rates classified by each complication provide clues for better management of COVID-19. First, our data also revealed that when complicated by viral pneumonia, bacterial pneumonia, deep vein thrombosis, and pulmonary embolism, tertiary groups had lower rates of survival. This probably reflects the fact that the majority of very severe cases of COVID-19 were treated in tertiary care hospitals. Indeed, severe COVID-19 infection often triggers a process known as immunothrombosis, resulting in DVT and/or PE. Second, for some complications, lower-level care facilities performed worse despite the patients having less severe pneumonia. Specifically, the mortality rate of cardiac ischemia was significantly higher in the primary group than the secondary group, and the mortality rate of pleural effusion and arrhythmia was higher in the secondary group than the tertiary group. This may be attributed to the fact that treatments for these complications are usually procedure-based, whereas other complications in the list are often followed up with pharmacologic or conservative treatment. It is plausible that in patients at higher-level care facilities, who are better prepared by specialists in each field, undergo such procedures quickly and effectively, resulting in better outcomes. Now that the in-hospital survival rate has plateaued because of the development of novel COVID-19 medications [11], further improvement in outcomes depends on the handling of complicated cases. The development of methods for predicting the occurrence of each complication remains a subject for future research.
We also revealed data concerning the healthcare delivery system in Japan. The mortality rate in this study was 2.6%, 4%, and 6.7% in primary, secondary, and tertiary care facilities, respectively, all of which were lower than those reported overseas [[12], [13], [14], [15], [16], [17], [18]]. It is possible that the majority of hospitalized patients in Japan have exceptionally mild disease. Indeed, only 33.5% of the patients in our study were classified as having severe COVID-19 on admission, which is lower than the proportion of severely ill patients reported in several international studies. In a USA study that analyzed 466,677 admissions from April 2020 to April 2021, more than 50% of hospitalized patients presented with acute respiratory failure throughout the study period [19]. Especially with the development of effective oral antiviral medications, clinicians may set a higher bar for admitting non-severe COVID-19 patients. Furthermore, in the present study, the median length of hospital stay was 10–11 days, which is much longer than the 7 days reported in the USA [20]. In fact, a considerable proportion of hospital stay in Japan is due to social reasons (i.e. no caregiver at home) rather than medical indications. The COVID-19 pandemic has had a significant negative impact on the management of other diseases [21,22]. It is crucial to develop a healthcare resource allocation strategy that takes a view of the entire picture as opposed to focusing on COVID-19. At the same time, the government needs to make more efforts to promote smooth transition from acute care centers to recovery/long-term beds.
This study has several limitations. First, selection bias may have occurred due to the nature of this registry study, as described previously [23]. Also, if a patient had a history of multiple COVID-19 hospitalizations or had been transferred, the data may have been duplicated. Furthermore, only hospitalized patients were included in this study. Now that outpatient treatment has become mainstream for mild cases [24], both inpatient and outpatient cases need to be examined to establish a rational healthcare policy.
In conclusion, our study provides evidence that more severely ill patients with COVID-19 with poorer prognoses are treated at higher care facilities in Japan. In addition to the severity of pneumonia, attending to possible complications may assist with selecting an appropriate medical facility for treatment. Furthermore, our findings serve as a reminder for healthcare providers that they may be allocating too many resources toward COVID-19 care at the cost of managing other diseases.
Funding sources
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
Authorship statement
DT and YH contributed to the concept, data interpretation, and writing of the original manuscript. YA was responsible for the data acquisition and analysis. KH, SK, MT, WS, and NO provided supervision and were responsible for the organization and management of the trial. All authors have approved the final version of the manuscript for publication.
Declaration of competing interest
None.
==== Refs
References
1 Matsunaga N. Hayakawa K. Terada M. Clinical epidemiology of hospitalized patients with coronavirus disease 2019 (COVID-19) in Japan: report of the COVID-19 registry Japan Clin Infect Dis 73 2021 10.1093/cid/ciaa1470 e3677–89–e3689
2 Terada M. Ohtsu H. Saito S. Hayakawa K. Tsuzuki S. Asai Y. Risk factors for severity on admission and the disease progression during hospitalisation in a large cohort of patients with COVID-19 in Japan BMJ Open 11 2021 e047007 10.1136/bmjopen-2020-047007
3 Beigel J.H. Tomashek K.M. Dodd L.E. Mehta A.K. Zingman B.S. Kalil A.C. Remdesivir for the treatment of COVID-19 – final Report. N Engl J Med. 383 2020 1813 1826 10.1056/NEJMoa2007764 32445440
4 Asch D.A. Sheils N.E. Islam M.N. Chen Y. Werner R.M. Buresh J. Variation in US hospital mortality rates for patients admitted with COVID-19 during the first 6 months of the pandemic JAMA Intern Med 181 2021 471 478 10.1001/jamainternmed.2020.8193 33351068
5 Bergman Z.R. Usher M. Olson A. Chipman J.G. Brunsvold M.E. Beilman G. Comparison of outcomes and process of care for patients treated at hospitals dedicated for COVID-19 care vs other hospitals JAMA Netw Open 5 2022 e220873 10.1001/jamanetworkopen.2022.0873
6 Asch D.A. Islam M.N. Sheils N.E. Chen Y. Doshi J.A. Buresh J. Patient and hospital factors associated with differences in mortality rates among black and white US Medicare beneficiaries hospitalized with COVID-19 Infection JAMA Netw Open 4 2021 e2112842 10.1001/jamanetworkopen.2021.12842
7 Clift A.K. von Ende A. Tan P.S. Sallis H.M. Lindson N. Coupland C.A.C. Smoking and COVID-19 outcomes: an observational and Mendelian randomisation study using the UK biobank cohort Thorax 77 2022 65 73 10.1136/thoraxjnl-2021-217080 34580193
8 Nakanishi M. Honda T. Processes of decision making and end-of-life care for patients with dementia in group homes in Japan Arch Gerontol Geriatr 48 2009 296 299 10.1016/j.archger.2008.02.009 18359112
9 Kojima G. Iliffe S. Taniguchi Y. Shimada H. Rakugi H. Walters K. Prevalence of frailty in Japan: a systematic review and meta-analysis J Epidemiol 27 2017 347 353 10.1016/j.je.2016.09.008 28142044
10 Zhang X.M. Jiao J. Cao J. Huo X.P. Zhu C. Wu X.J. Frailty as a predictor of mortality among patients with COVID-19: a systematic review and meta-analysis BMC Geriatr 21 2021 186 10.1186/s12877-021-02138-5 33731018
11 Moon R.C. Mackey R.H. Cao Z. Emot S. Schott L.L. Gayle J.A. Is coronavirus disease 2019 (COVID-19) less deadly now? Trends in in-hospital mortality among hospitalized COVID-19 patients in the United States Clin Infect Dis 74 2022 2238 2242 34534276
12 Mozaffari E. Chandak A. Zhang Z. Liang S. Gayle J. Thrun M. Clinical management of hospitalized coronavirus disease 2019 patients in the United States Open Forum Infect Dis 9 2022 ofab498 10.1093/ofid/ofab498
13 Finelli L. Gupta V. Petigara T. Yu K. Bauer K.A. Puzniak L.A. Mortality among US patients hospitalized with SARS-CoV-2 infection in 2020 JAMA Netw Open 4 2021 e216556 10.1001/jamanetworkopen.2021.6556
14 Nguyen N.T. Chinn J. Kirby K. Hohmann S.F. Amin A. Outcomes of COVID-19 adults managed in an outpatient versus hospital setting PLoS One 17 2022 e0263813 10.1371/journal.pone.0263813
15 Macedo A. Gonçalves N. Febra C. COVID-19 fatality rates in hospitalized patients: systematic review and meta-analysis Ann Epidemiol 57 2021 14 21 10.1016/j.annepidem.2021.02.012 33662494
16 Roso-Llorach A. Serra-Picamal X. Cos F.X. Pallejà-Millán M. Mateu L. Rosell A. Evolving mortality and clinical outcomes of hospitalized subjects during successive COVID-19 waves in Catalonia, Spain Glob Epidemiol 4 2022 100071 10.1016/j.gloepi.2022.100071
17 Comparative epidemiology of four waves of COVID-19 Mexico 2020 - April 2022 10.21203/rs.3.rs-1875498/v1
18 Drake T.M. Riad A.M. Fairfield C.J. Egan C. Knight S.R. Pius R. Characterisation of in-hospital complications associated with COVID-19 using the ISARIC WHO Clinical Characterisation Protocol UK: a prospective, multicentre cohort study Lancet 398 2021 223 237 10.1016/S0140-6736(21)00799-6 34274064
19 Whitfield G.P. Harris A.M. Kadri S.S. Warner S. Bamrah Morris S. Giovanni J.E. Trends in clinical severity of hospitalized patients with coronavirus disease 2019 – premier hospital dataset, April 2020 – April 2021 vol. 9 2022 Open Forum Infect Dis ofab599 10.1093/ofid/ofab599
20 Mozaffari E. Chandak A. Zhang Z. Liang S. Gayle J. Thrun M. Clinical management of hospitalized coronavirus disease 2019 patients in the United States Open Forum Infect Dis 9 2022 ofab498 10.1093/ofid/ofab498
21 Yamaguchi S. Okada A. Sunaga S. Ikeda Kurakawa K. Yamauchi T. Nangaku M. Impact of COVID-19 pandemic on healthcare service use for non-COVID-19 patients in Japan: retrospective cohort study BMJ Open 12 2022 e060390 10.1136/bmjopen-2021-060390
22 Dang A. Thakker R. Li S. Hommel E. Mehta H.B. Goodwin J.S. Hospitalizations and mortality from non–SARS-CoV-2 causes among Medicare beneficiaries at US hospitals during the SARS-CoV-2 pandemic JAMA Netw Open 5 2022 e221754 10.1001/jamanetworkopen.2022.1754
23 Hayakawa K. Asai Y. Matsunaga N. Tsuzuki S. Terada M. Suzuki S. Evaluation of the representativeness of data in the COVID-19 Registry Japan during the first six waves of the epidemic Glob Health Med 4 2022 2022 10.35772/ghm.2022.01033 20201033:204–9
24 Perico N. Cortinovis M. Suter F. Remuzzi G. Home as the new frontier for the treatment of COVID-19: the case for anti-inflammatory agents Lancet Infect Dis 2022 10.1016/S1473-3099(22)00433-9
| 0 | PMC9745966 | NO-CC CODE | 2022-12-15 23:17:55 | no | J Infect Chemother. 2022 Dec 13; doi: 10.1016/j.jiac.2022.12.003 | utf-8 | J Infect Chemother | 2,022 | 10.1016/j.jiac.2022.12.003 | oa_other |
==== Front
Appl Nurs Res
Appl Nurs Res
Applied Nursing Research
0897-1897
1532-8201
Elsevier Inc.
S0897-1897(22)00101-X
10.1016/j.apnr.2022.151659
151659
Article
Factors associated with changes in nurses' emotional distress during the COVID-19 pandemic
Brown Robin a⁎
Da Rosa Patricia b
Pravecek Brandi c
Carson Paula d
a College of Nursing, South Dakota State University, Wagner Hall 311, Box 2275, Brookings, SD 57007, United States of America
b Public Health Research Associate, College of Nursing, South Dakota State University, Brookings, SD 57007, United States of America
c College of Nursing, South Dakota State University, Sioux Falls, SD, United States of America
d College of Nursing, South Dakota State University, Brookings, SD, United States of America
⁎ Correspondence author.
13 12 2022
13 12 2022
15165930 5 2022
28 8 2022
6 12 2022
© 2022 Elsevier Inc. All rights reserved.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Purpose
The purpose of this study was twofold: to assess if nurses experienced changes in emotional distress (stress, depression, and anxiety) as the number of patients infected with coronavirus disease 2019 (COVID-19) increased and if there were any sociodemographic, psychosocial, and work environmental influence on the change.
Methods
Using a repeated cross-sectional study design, we collected survey data among 198 South Dakota (SD) nurses. Data were collected in two waves, during the first 12 months of the COVID-19 pandemic in the United States (July and December 2020). Participants completed two online surveys: (a) The Depression, Anxiety, and Stress Scale (DASS-21); and (b) Change Fatigue Scale. Predictive factors were divided into three groups: sociodemographic, psychosocial, and work environment variables. Multiple linear regression models were run to estimate the factors associated with the change in DASS-21 subscale score.
Results
Total DASS-21 score and scores for all subscales significantly increased from Survey 1 to Survey 2. Significant positive associations were found between change fatigue and workplace barriers with change in depression, anxiety, and stress scores. A linear relationship was identified between self-worry about COVID-19 risk and depression and stress and being male and young were associated with changes in depression.
Conclusions
Increase in emotional distress of nurses as the pandemic progresses is consistent with other studies. It is vital for healthcare organizations to recognize the factors associated with the changes in emotional distress and their role in decreasing the stress levels of nurses.
Keywords
COVID-19
Nurses
Emotional distress
Change fatigue
==== Body
pmc1 Introduction
The novel coronavirus disease 2019 (COVID-19) pandemic has led to many unexpected changes not only in nurses' daily lives but also in the way healthcare facilities function. During the pandemic, organizations were forced into rapid and radical transformations, which led to a shift in how people value work (Amis, 2020). Healthcare is well-known for being in a state of constant change, but the pandemic has led to additional organizational changes, which has led to acute stress (Shahrout & Dardas, 2020). In addition, excessive organizational change can cause change fatigue, decrease in job satisfaction, increase in turnover rates (Bernerth et al., 2016; Brown et al., 2018; Camilleri et al., 2018), and affect the overall relations with the organization (Li et al., 2021). Lack of adequate time and preparation with organizational changes may form barriers to change and pose threats to the organization (Li et al., 2021). The current pandemic forced healthcare facilities to suddenly implement organizational changes with minimal preparation in order to meet the needs of the COVID-19 patients.
Change fatigue is the overwhelming feeling of stress, exhaustion, and burnout associated with rapid and continuous change in the workplace (McMillan & Perron, 2013). Change fatigue evolved from the discipline of management and the psychological effects of organizational change have been fundamentally under-researched in nursing. Qualitative research by McMillan and Perron (2020) revealed that the intensification of the work required of nurses and nurses' perception of repeated and ongoing self-sacrifice were core themes in nurses' experience of change fatigue. Even though organizational change and demands on nurses are at an all-time high due to the pandemic, there is currently no research focusing on nurses and change fatigue during the pandemic. It is imperative to the success of organizations to understand the effects of these changes and how nurses cope with these unexpected changes (Li et al., 2021).
Current research demonstrates the extent to which nurses are experiencing acute stress and psychological distress during the pandemic. In a recent study by Nursing Standard (2021), eight out of ten nurses reported their mental health has been affected by the pandemic. According to Shahrout and Dardas (2020), 64 % of Jordanian nurses are experiencing acute stress disorder (ASD) and 41 % are also suffering significant psychologic distress during the COVID-19 pandemic. A cross-sectional survey study carried out in February of 2020 revealed that 25.1 % of Chinese frontline nurses experienced significant psychological distress due to the COVID-19 pandemic, as evidenced by their 12-item General Health Questionnaire (GHQ-12) scores (Nie et al., 2020). These findings demonstrated a psychological distress level nearly two times as high as the general population of China. Cai et al. (2020) found similar increases in psychological distress, notably depression, anxiety, and insomnia, among nurses in the epicenter of the pandemic in Wuhan, China. Firew et al. (2020) conducted a cross-sectional survey of healthcare workers in the United States (U.S.), including nurses, with results demonstrating higher levels of psychological stress during the pandemic due to a variety of factors, including isolation and having to send cohabitants away during the height of COVID infections. In addition, Chen et al. (2020) reported a moderate degree of emotional exhaustion among nurses caring for COVID patients.
Minimal longitudinal research has been conducted on the changes of stress and anxiety of nurses as the pandemic continues. A recent survey conducted by the American Nurses Foundation (2021) reported nurses continue to be stressed (75 %), frustrated (69 %), and overwhelmed (62 %), and over 34 % rated their emotional health as unhealthy. One longitudinal study conducted by Cai et al. (2020) in China found depression, anxiety, and posttraumatic stress disorder (PTSD) was statistically higher in the outbreak period (Survey 1) compared to the stable periods (Survey 2). Researchers in Belgium evaluated the mental health impact of COVID-19 from April 1 to June 30, 2020. Data was collected at three different points of time within this period, initially following the onset of the pandemic, at 4 weeks, and at 8 weeks. The study reported that while rates of depression, anxiety and somatization did not increase, the rate of emotional distress was consistently high despite a decrease in COVID-19 patient admissions over time (Van Steenkiste et al., 2021).
The purpose of this study was twofold. The first purpose was to assess if nurses experienced changes in emotional distress (stress, depression, and anxiety) as the number of patients infected with COVID-19 increased. The seven-day moving average COVID-19 cases in South Dakota (SD) increased from 60.7 (July 1st) to 817 (December 1, 2020) (South Dakota Department of Health, n.d.). Secondly, researchers evaluated if there were any sociodemographic, psychosocial, and work environmental variables associated with the change in emotional distress during the COVID-19 pandemic.
2 Methods
Using a repeated cross-sectional study design, we collected survey data among SD nurses. Data were collected in two waves during the first 12 months of the COVID-19 pandemic in the U.S. (July and December 2020). The follow up survey was used to examine any change in stress, depression, and anxiety over time as the positivity rate of the COVID-19 pandemic increased. To recruit nurses to the first phase of the study, a link to the survey was distributed to nurses registered with the South Dakota Board of Nursing (N = 19,249). In addition, information about the study was circulated through the College of Nursing's Facebook. Detailed information regarding the methodology of the first survey referred in this study is published in a previous manuscript in Applied Nursing Research (Da Rosa et al., 2021). Nurses who consented to participate in a follow up survey were contacted via email. The study was approved by the university's Human Research Ethics Committee (IRB-2006012-EXM), and online informed consent was obtained for all participants.
2.1 Instruments
Data were collected in two waves to investigate changes in emotional distress levels. Instruments were linked using a common participant ID. The initial survey (Survey 1) included the following sections: emotional distress (anxiety, depression, and stress); socio-demographic characteristics; work environment variables; and concerns based on COVID-19 risks. The follow up survey (Survey 2) included the following sections: emotional distress; change fatigue; work environment variables and barriers at work; and concerns based on COVID-19 risks. Change fatigue was added to Survey 2 to assess the effects of multiple changes being made in facilities due to the COVID-19 pandemic.
The Depression, Anxiety, and Stress Scale (DASS-21) is a self-report scale and was used to assess emotional distress. Each of the three DASS-21 scales contains seven items. Scores for depression, anxiety, and stress are calculated by summing the scores for the relevant items (Lovibond & Lovibond, 1995a). Each of these constructs is interrelated and the combined score (sum of the three subscales) can be used to screen for general psychological distress. The DASS-21 has demonstrated excellent internal consistency in a variety of populations, with a Cronbach's ranging from 0.9 to 0.97 (Henry & Crawford, 2005; Lovibond & Lovibond, 1995b). Similarly, alpha for this study (n = 198) was 0.94. To assess change in DASS-21 subscales, absolute difference in the subscale scores between Survey 1 and 2 were calculated. Job satisfaction, number of COVID patients, and two questions on worry based on COVID-19 risk were asked in both surveys. The change in these variables was calculated as the difference of responses. Then, the new variable, representing the change in DASS-21 from Survey 1 to Survey 2, was categorized as decrease, no change, or increase.
The Change Fatigue Scale measures well-being, organizational commitment, and turnover intentions in employees experiencing multiple organizational changes. The scale is a 7-item Likert scale and has shown good reliability and internal consistency. Cronbach alpha with non-nurses is 0.85 (Bernerth et al., 2011) and with nurses 0.94 (Brown et al., 2018). In our study, the Cronbach alpha for the study population (n = 198) was 0.94. Higher total score reflects higher change fatigue.
The race/ethnicity question allowed multiracial responses; the variable was recoded to White and Non-White because 96 % of the sample was White. The Work Environment Survey was developed by a team of interdisciplinary research experts from results of a review of literature of previous COVID-19 pandemic studies. Work environment variables included current nursing degree, years practicing as a nurse, primary work setting, job satisfaction, suspected cases of COVID-19 with direct contact, level of preparation to provide direct care, number of extra hours at work, and change fatigue. Workplace barriers score included personal protection equipment (PPE) availability, number of COVID-19 patients, lack of staffing, lack of emotional support, communication between leadership/team, and hours scheduled. These barriers were rated from not a barrier to extreme barrier. Psychosocial variables included concerns with COVID-19 exposure and mental health and worry about COVID-19 risk for self and others.
2.2 Data analysis
The outcome variables of interest in this study were change in the DASS-21 subscale scores (depression, anxiety, and stress) from Survey 1 to Survey 2 and the objective of the analysis was to determine significant associations with nurse characteristics; in particular, associations with sociodemographic and psychosocial factors, and work environment characteristics. Multiple linear regression was used to estimate associations of factors with the change in DASS-21 subscale score. Given the sample size relative to the number of factors to be evaluated, a hierarchical approach to regression model development was used. To adjust for regression towards the mean for the change in DASS-21 subscale scores, the baseline DASS subscale scores were included in all regression models. The factors were divided into three groups and separate models were run using block sequential backwards elimination. Group A included sociodemographic variables; group B included psychosocial variables (e.g., concerns with COVID-19 risks) and group C included work environment variables. At each hierarchical stage, backwards elimination was conducted by removing the factor with the highest (2-sided) p-value >0.2 until all remaining factors at that hierarchical stage had an association with the dependent variable with p < 0.2; these factors were retained in all models for subsequent hierarchical stage backward elimination procedures. For ordinal factors, linear trend and quadratic trend (when justified) were evaluated. For interpretation of final models, statistical significance for outcome variable – factor associations were set at p < 0.05. Questionnaire data were exported from QuestionPro and analyzed using Stata version 15.1 (StataCorp, College Station, TX).
3 Results
From 1599 participants with complete data on the first survey (Survey 1), 715 respondents agreed to participate in the following up survey (Survey 2). Among those, 220 respondents participated in both surveys and 198 provided complete data available for the regression analysis. Table 1A shows the descriptive characteristics of the study sample. The majority were female, white, married, and working as registered nurses. Over half were older than 40 years old, satisfied with their jobs, and felt prepared in providing direct care. Table 1B depicts the characteristics of the work environment (e.g., number of extra hours worked per week). One third reported working >10 extra hours during the first months of the pandemic, and over a third percent reported seeing >20 cases.Table 1A Participants' characteristics.
Table 1ACharacteristic Counts (%)
Total 198
Gender
Male 22 (11.1)
Female 176 (88.9)
Age in years
20–29 years 25 (12.6)
30–39 years 62 (31.3)
40–49 years 39 (19.7)
50–59 years 46 (23.2)
60+ years 26 (13.1)
Marital status
Single, never married 34 (17.2)
Married 137 (69.2)
Single, previously married 27 (13.6)
Two-level race
White 190 (96.0)
Non-White 8 (4.0)
Highest degree
Licensed Practical Nurse (LPN) 15 (7.6)
Associate degree (ADN)/diploma 39 (19.7)
Bachelor's degree (BSN) 101 (51.0)
Advanced practice degree 43 (21.7)
Household Income (Annual)
<$50,000 21 (10.6)
$50,000 to $74,999 57 (28.8)
$75,000 to $99,999 43 (21.7)
$100,000+ 77 (38.9)
Any children living with you?
Yes 96 (48.5)
No 102 (51.5)
Current nursing degree
LPN 18 (9.1)
RN 152 (76.8)
CNP/CRNA/CNM/CNS 28 (14.1)
Years practicing as nurse
<6 years 36 (18.2)
6–10 years 42 (21.2)
11–20 years 47 (23.7)
>20 years 73 (36.9)
Primary work setting
Health center/hospital 93 (47.0)
Clinic 44 (22.2)
Ambulatory care 10 (5.1)
Long-term/assisted care 22 (11.1)
Other 29 (14.6)
Satisfied with job (1)
Very unsatisfied 4 (2.0)
Unsatisfied 10 (5.1)
Satisfied 80 (40.4)
Very satisfied 104 (52.5)
Suspected cases of COVID with direct contact (1)
None 68 (34.3)
1–10 87 (43.9)
11–20 17 (8.6)
>20 26 (13.1)
Level of preparation to provide direct care
Completely unprepared 9 (4.5)
Somewhat unprepared 32 (16.2)
Somewhat prepared 106 (53.5)
Very prepared 51 (25.8)
Concerned for COVID exposure and mental health
Yes 44 (22.2)
No 84 (42.4)
No mental health condition 70 (35.4)
Worry about COVID risk for self (1)
Extremely worried 22 (11.1)
Generally worried 97 (49.0)
Generally not worried 63 (31.8)
Not worried at all 16 (8.1)
Worry about COVID risk for others (1)
Extremely worried 44 (22.2)
Generally worried 96 (48.5)
Generally not worried 48 (24.2)
Not worried at all 10 (5.1)
Note: Household income is in U.S. dollar. Based on the study population with complete data for all variables from both Survey 1 (1) and Survey 2 that are used in the regression models.
Table 1B Psychosocial variables and work environment characteristics (Survey 2).
Table 1BCharacteristic Counts (%)
Total 198
Extra hours worked per week since July 2020
0 h 57 (28.8)
1–10 h 70 (35.4)
>10 h 71 (35.9)
Suspected direct contact with COVID patients (2)
Zero 15 (7.6)
1–10 75 (37.9)
11–20 38 (19.2)
>20 70 (35.4)
Worry about COVID risk for self (2)
Extremely worried 14 (7.1)
Generally worried 70 (35.4)
Generally not worried 96 (48.5)
Not worried at all 18 (9.1)
Worry about COVID risk for others (2)
Extremely worried 40 (20.2)
Generally worried 86 (43.4)
Generally not worried 60 (30.3)
Not worried at all 12 (6.1)
Satisfied with job (2)
Completely dissatisfied 12 (6.1)
Somewhat dissatisfied 31 (15.7)
Somewhat satisfied 86 (43.4)
Very satisfied 69 (34.8)
Changed facilities since July 2020
No 180 (90.9)
Yes 18 (9.1)
Changed nursing units since July 2020
No 178 (89.9)
Yes 20 (10.1)
Based on the study population with complete data for all variables from both Survey 1 and Survey 2 that are used in the regression models.
As shown on Table 2 , the total DASS-21 score and the scores for all subscales significantly increased from Survey 1 to Survey 2. The total average score (DASS-21) changed from 16.5 (SD = 17.4) to 24.31 (SD = 17.8), an increase of 7.8 units (p-value <0.001). A similar trend was observed for all subscales.Table 2 Descriptive statistics (mean, standard deviation) for numerical variables on Surveys 1 and 2.
Table 2Variable Survey 1 Survey 2 Change P-value
Mean ± SD
DASS score total 16.51 ± 17.40 24.31 ± 17.83 7.81 ± 17.38 0.001
DASS anxiety subscale 3.67 ± 5.67 6.03 ± 6.42 2.36 ± 5.80 0.001
DASS depression subscale 4.77 ± 5.89 6.82 ± 6.77 2.05 ± 6.63 0.001
DASS stress subscale 8.07 ± 7.87 11.46 ± 7.30 3.39 ± 7.79
Change fatigue total score 25.27 ± 8.59
Workplace barriers total score 12.12 ± 4.15
N = 198; based on the study population with complete data for all variables from both Survey 1 and Survey 2 that are used in the regression models. The first number in parentheses is the p-value for the 2-sided t-test that the mean difference is zero, and the second number is the 2-sided p-value for the non-parametric Wilcoxon signed-rank test.
For the linear regression, results for change in DASS-21 total score, anxiety, depression, and stress score are shown in Table 3 . Significant positive associations were found between change fatigue and workplace barriers with change in depression, anxiety, and stress scores. For instance, a one unit increase in the change fatigue score was associated with a 0.18 unit increase in the mean change in stress score, and 0.11 increase in the mean change in depression score. For workplace barriers, an increase in the workplace barrier score was associated with increased mean change in stress (Beta = 0.394, p-value = 0.001), depression (Beta = 0.306, p-value = 0.012), and anxiety (Beta = 0.259, p-value = 0.010) scores. Another significant factor was concern for COVID-19 and worsening of a previous mental health condition. The mean change for anxiety concern was 2.77 higher (p-value = 0.034) than the mean change for nurses that reported No concern for anxiety. A similar result was found for the mean change in depression score (Beta = 0.880, p-value = 0.031). In addition, a significant linear trend association was found with mean change in depression and stress with self-worry about COVID-19 risk, meaning that as self-worry increases the mean change in depression and stress score increases. Gender and age were also associated with mean change in depression. The mean change in depression score for males was 2.63 higher than the mean change in depression score for females. Conversely, older nurses (50–59 years) reported a decrease in the mean change (Beta = −4.94, p-value = 0.018) in depression compared to nurses ages 20–29. Race, income, years practicing, preparedness, job satisfaction and primary setting were not associated with change in anxiety, depression, and stress scores.Table 3 Associations with change in DASS-21 subscale scores.
Table 3Characteristic Depression Anxiety Stress
Beta P-value Beta P-value Beta P-value
DASS depression subscale, initial −0.510 <0.001 −0.321 0.001 −0.105 0.345
DASS anxiety subscale, initial −0.097 0.372 −0.552 <0.001 −0.087 0.433
DASS stress subscale, initial −0.006 0.950 0.182 0.029 −0.668 <0.001
Change fatigue total score 0.109 0.046 0.096 0.045 0.180 0.001
Workplace barriers total score 0.306 0.012 0.259 0.010 0.394 0.001
Changed nursing units since July 2020 n/m n/m
No R 0.187(e)
Yes −1.623
Concerned for COVID and mental health
Yes 0.880 0.031(e) 2.773 0.034(e) 2.042 0.094(e)
No R R R
No mental health condition −1.898 0.257 −0.704
Worry about COVID risk for self (Survey 1)
Extremely worried R 0.026(e) R 0.225(e) R 0.181(e)
Generally worried −0.300 0.035(l) −0.464 0.069(l) −2.093 0.040(l)
Generally not worried −3.092 0.760(q) −0.824 0.218(q) −2.124 0.762(q)
Not worried at all −4.032 −3.624 −4.904
Change in self worry
Decrease −0.375 0.021(e) −1.059 0.020(e) −0.068 0.232(e)
No change R 0.011(l) R 0.005(l) R 0.138(l)
Increase 3.408 0.073(q) 2.599 0.312(q) 2.231 0.218(q)
Gender n/m n/m
Male 2.634 0.048(e)
Female R
Age in years
20–29 years R 0.018(e) R 0.669(e) R 0.338(e)
30–39 years −4.316 0.036(l) −1.590 0.965(l) 0.246 0.525(l)
40–49 years −3.009 0.039(q) −0.589 0.407(q) 2.550 0.672(q)
50–59 years −4.935 −0.919 0.350
60+ years −3.529 −0.409 1.896
Highest degree n/m n/m
Licensed Practical Nurse (LPN) 1.447 0.192(e)
Associate degree (ADN)/diploma R 0.045(l)
Bachelors (BS) −0.206 0.970(q)
Advanced degree −1.586
Regression models used change in subscales scores as dependent variables and block sequential backwards elimination to determine factors (characteristics) to retain in models. Characteristics listed include only those retained in at least one of the subscale models; n/m indicates the variable was not retained in the model; R indicates the reference category. Variable also entered in the model: race, household income, years practicing, primary setting, preparedness, and job satisfaction.
For categorical variable p-values, (e) is for the overall effect; (l) is for a linear trend; (q) is for a quadratic trend.
4 Discussion
The purpose of this study was to determine if nurses experienced changes in emotional distress (stress, depression, and anxiety) as the COVID-19 pandemic progressed and the number of infections increased. The impact of sociodemographic, psychosocial, and work environment factors on changes in emotional distress was also evaluated. Nurses have experienced significant psychological distress throughout the COVID-19 pandemic. Our findings demonstrate a statistical increase in depression, anxiety, and stress in nurses caring for patients as COVID-19 cases increased. Early recognition of the impact of COVID-19 on the mental health of nurses may be a key factor in preventing nurse burnout and lessening the nursing shortage.
These results support recent research findings demonstrating increased emotional distress in nurses during the pandemic (Chen et al., 2020; Nie et al., 2020; Shahrout & Dardas, 2020). In contrast, Cai et al. (2020) found in a longitudinal study that nurses in Wuhan, China had significantly higher depression, anxiety, and posttraumatic stress disorder (PTSD) at the outbreak period (Survey 1) compared to the stable period (Survey 2). The variation in the timing of the most significant mental health impact of COVID suggests that nurses are at risk for emotional distress throughout all phases of the pandemic. This highlights the importance of continual administrative and facility support of nurses, regardless of the current pandemic phase. In our study, gender and age were associated with change in depression, with males having higher change in depression, compared to females. In contrast, with a longitudinal study by Cai et al. (2020), age and gender were not significant factors for depression, anxiety, and PTSD with nurses in Wuhan, China. Chen et al. (2020) found female nurses had significantly higher emotional exhaustion during the pandemic. The COVID-19 pandemic has led to increased emotional stress for nurses, but there are conflicting results related to individual characteristics. The reason for this is unclear but demonstrates the importance of implementing supportive measures for both male and female nurses as both may have other significant outside life stresses that compound the impact of COVID-19. More research is needed to study individual characteristics and the relationship to emotional distress.
Workplace barriers total score, including PPE availability, number of COVID patients, lack of staffing, lack of emotional support, communication between leadership/team, and hours scheduled, were found to have a significant impact on the stress scores of nurses. Nie et al. (2020) found similar results with the lack of PPE affecting the mental health of nurses in China. Understanding the workplace barriers that have the potential to increase nurses' emotional distress is essential when designing strategies and solutions to mitigate the burden of mental health problems during the pandemic. Knowledge of these workplace barriers may also inform future surge planning efforts in the event of another pandemic.
Change fatigue was also found to be a significant factor of changes in stress and depression scores of nurses. Healthcare facilities have been in a constant state of change since the onset of the pandemic and the impact of these continual and rapid changes has undoubtedly been experienced by nurses as vital frontline employees. To our knowledge, there is no current research with change fatigue and the pandemic. One study prior to the pandemic reported nurses' job satisfaction had a statistically significant negative association with change fatigue (Brown et al., 2018). Change fatigue caused changes in stress and depression, but there is minimal research on this important topic. More research is needed to assess the effects of change fatigue during the pandemic and whether it contributes to nurses leaving the profession, especially now given an increased focus on the nursing shortage and the development of solutions to combat its rising prevalence.
In this study, a concern for COVID-19 and worsening of a previous mental health condition was a strong factor. In addition, a linear trend was found with change in depression and stress and self-worry about COVID-19 risk. As self-worry increased, depression and stress increased. Given the negative impact high levels of stress can have on mental health and well-being, it is vital for healthcare organizations to recognize these workplace barriers and their role in increasing stress levels of nurses. Ensuring the availability of adequate PPE, establishing safe and acceptable nurse to patient care ratios, and fostering effective communication between leadership and the healthcare team may support the emotional health of nurses. These efforts to reduce the impact of workplace barriers and subsequently decrease nurses' stress levels may in turn result in decreased nurse burnout and turnover (Mosadeghrad, 2013). In addition, implementing interventions to lessen the psychological impact of COVID-19 on nurses and removing barriers to access to mental health support resources during and following the pandemic are essential. Cai et al. (2020) found nurses in China benefited from online psychological consultation. Increasing the availability of telehealth mental health services for nurses may be beneficial in both urban and rural settings.
4.1 Limitations
There are important methodological limitations to this study. First, this study was carried out among a convenience sample that comprised mostly white females, more aware of their mental health condition, and from a rural state in which a small percent consented to the second survey. Thus, the study population may not be representative of nurses in general. Third, emotional distress was measured using a self-reported questionnaire and not an objective assessment of a health professional. However, the scales used in this study demonstrated good reliability and have been successfully used in several other studies (Bernerth et al., 2011; Brown et al., 2018; Crawford, et al., 2008). An additional follow-up survey of this study would allow us to assess whether the levels of emotional distress differ over time and investigate whether the associated factors remained the same. Finally, similar studies carried out among a more representative sample would be warranted.
5 Conclusion
In conclusion, this repeated cross-sectional study among nurses found that workplace environmental factors, including change fatigue, concerns for COVID-19, worsening of a previous mental health condition, and self-worry about COVID-19 exposure were risk factors for changes in emotional distress. This study further indicated that being male and young were also associated with changes in depression. Strategies to reduce workplace stressors and provide mental wellbeing among nurses are crucial to ensure quality of care and nurses' ability to provide healthcare services during the COVID-19 pandemic. Given the uncertain trajectory of the pandemic, developing, and implementing interventions to lessen the psychological impact of COVID on nurses in a timely manner is essential.
Declaration of competing interest
None.
Acknowledgements
The authors would like to thank Howard Wey, PhD, MS for his statistical analysis support.
==== Refs
References
American Nurses Foundation Pulse on the Nation’s Nurses COVID-19 Survey Series: Mental Health and Wellness Survey 3 September 2021 https://www.nursingworld.org/practice-policy/work-environment/health-safety/disaster-preparedness/coronavirus/what-you-need-to-know/pulse-on-the-nations-nurses-covid-19-survey-series-mental-health-and-wellness-survey-3-september-2021/
Amis J.M. Organizational change in a (post-) pandemic world: Rediscovering interests and values Journal of Management Studies 58 2 2020 582 586 10.111/joms.12663
Bernerth J. Walker H. Harris S. Change fatigue: Development and initial validation of a new measure Work & Stress 25 4 2011 321 337
Brown R. Wey H. Foland K. The relationship among change fatigue, resilience, and job satisfaction of hospital staff nurses Journal of Nursing Scholarship 50 3 2018 306 313 doi: 10.111/jnu.12373 29517141
Cai Z. Cui Q. Liu Z. Li J. Gong X. Liu J. Wan Z. Yuan X. Li X. Chen C. Wang G. Nurses endured high risks of psychological problems under the epidemic of COVID-19 in a longitudinal study in Wuhan China Journal of Psychiatric Research 131 2020 132 137 10.1016/j.jpsychires.2020.09.007 32971356
Camilleri J. Cope V. Murray M. Change fatigue: The frontline nursing experience of large-scale organizational change and the influence of teamwork Journal of Nursing Management 27 3 2018 655 660 30354000
Chen R. Sun C. Chen J. Jen H. King X.L. Kao C. Chou K. A large-scale survey on trauma, burnout, and posttraumatic growth among nurses during the COVID-19 pandemic International Journal of Mental Health Nursing 30 1 2020 102 116 https://doi-org.excelsior.sdstate.edu/10.1111/inm.12796 33107677
Da Rosa P. Brown R. Pravecek R. Carotta C. Garcia A.S. Carson P. Callies D. Vukovich M. Factors associated with nurses’ emotional distress during the COVID-19 pandemic Applied Nursing Research 2021 62 10.1016/j.apnr.2021.151502
Firew T. Sano E.D. Lee J.W. Flores S. Lang K. Salman K. Greene M.C. Chang B.P. Protecting the front line: A cross-sectional survey analysis of the occupational factors contributing to healthcare workers’ infection and psychological distress during the COVID-19 pandemic in the USA BMJ Open 10 10 2020 10.1136/bmjopen-2020-042752
Henry J.D. Crawford J.R. The short-form version of the depression anxiety stress scales (DASS-21): Construct validity and normative data in a large non-clinical sample British Journal of Clinical Psychology 44 2 2005 227 239 16004657
Li J.Y. Sun R. Weiting T. Yeunjare L. Employee coping with organizational change in the face of a pandemic: The role of transparent internal communication Public Relations Review 47 1 2021 101984
Lovibond P.F. Lovibond S.H. The structure of negative emotional states: Comparison of the depression anxiety stress scales (DASS) with the Beck depression and anxiety inventories Behaviour Research and Therapy 33 3 1995 335 343 10.1016/0005-7967(94)00075-u 7726811
Lovibond S.H. Lovibond P.F. Manual for the depression anxiety and stress scales 2nd ed. 1995 Psychology Foundation Sydney
McMillan K. Perron A. Nurses amidst change: The concept of change fatigue offers an alternative perspective on organizational change Policy, Politics & Nursing Practice 1 2013 26 32 10.1177/1527154413481811
McMillan K. Perron A. Change fatigue in nurses: A qualitative study Journal of Advanced Nursing 76 10 2020 2627 2636 32542846
Mosadeghrad A.M. Occupational stress and turnover intention: Implications for nursing management International Journal of Health Policy & Management 1 2 2013 169 176 10.15171/ijhpm.2013.30 24596858
Nie A. Su X. Zhang S. Guan W. Li J. Psychological impact of COVID-19 on frontline nurses: A cross-sectional survey study Journal of Clinical Nursing 29 21/22 2020 4217 4226 32786150
Nursing Standard How COVID-19 is affecting nurses mental health and what to do about it January 2021 https://rcni.com/nursing-standard/features/how-covid-19-affecting-nurses-mental-health-and-what-to-do-about-it-159456
Shahrout G. Dardas L.A. Acute stress disorder, coping self-efficacy and subsequent psychological distress among nurses amid COVID-19 Journal of Nursing Management 28 7 2020 1686 1695 32767827
South Dakota Department of Health, n.d.South Dakota Department of Health (n.d.). South Dakota COVID-19 dashboard. https://doh.sd.gov/COVID/Dashboard.aspx
Van Steenkiste E. Schoofs J. Gilis S. Messiaen P. Mental health impact of COVID-19 in frontline healthcare workers in a Belgian tertiary care hospital: A prospective longitudinal study Acta Clinica Belgica 2021 10.1080/17843286.2021.1903660
| 0 | PMC9745970 | NO-CC CODE | 2022-12-15 23:17:44 | no | Appl Nurs Res. 2022 Dec 13;:151659 | utf-8 | Appl Nurs Res | 2,022 | 10.1016/j.apnr.2022.151659 | oa_other |
==== Front
Int J Afr Nurs Sci
Int J Afr Nurs Sci
International Journal of Africa Nursing Sciences
2214-1391
The Author(s). Published by Elsevier Ltd.
S2214-1391(22)00125-1
10.1016/j.ijans.2022.100518
100518
Article
Clinical nursing care protocol for convalescent plasma transfusion in patients with COVID-19
Maiara Ferreira Barreto Pires Bruna a⁎
Marcia Peres Ellen b
Marcos Tosoli Gomes Antonio b
Valéria Dantas de Oliveira Souza Norma b
Carneiro Carvalho Eloá c
Cristina da Silva Thiengo de Andrade Priscila b
Mayerhofer Kubota Thais d
Faria Cristiene d
Carvalho Leite Dayana d
Conceição das Merces Magno e
Teti Toledo Thelma b
Sant'anna Nunes Alessandra b
Cabral Pereira da Costa Carolina b
Fajin de Mello dos Santos Lívia b
Guimarães Assad Luciana f
Maria de Sá Basílio Lins Silvia f
Gustavo Torres Dias da Cruz Luiz d
Paula Oliveira Motta Ana d
Almeida de Oliveira Juliana d
dos Santos Fernandes Eveline d
Olinda Ferreira de Sousa Maria d
de Sousa Chami Ariana d
Oliveira Duarte Martins Mônica d
de Aguiar Ciríaco Alexandrina d
Britto Ribeiro de Jesus Patrícia f
da Silva Pires Ariane b
Perez Fuentes Junior Eugênio b
Ferraz Gomes Helena b
a Departamento de Fundamentos e Administração de Enfermagem, Universidade Federal Fluminense (UFF) – Niterói (RJ), Brazil
b Departamento de Enfermagem Médico-Cirúrgica, Universidade do Estado do Rio de Janeiro (UERJ) – Rio de Janeiro (RJ), Brazil
c Departamento de Enfermagem em Saúde Pública, UERJ - Rio de Janeiro (RJ), Brazil
d Hospital Universitário Pedro Ernesto (HUPE) - Rio de Janeiro (RJ), Brazil
e Departamento de Ciências da Vida, Universidade Estadual da Bahia (UESB) –Salvador (BA), Brazil
f Departamento de Fundamentos de Enfermagem, UERJ - Rio de Janeiro (RJ), Brazil
⁎ Corresponding author.
13 12 2022
13 12 2022
1005184 10 2021
10 8 2022
12 12 2022
© 2022 The Author(s). Published by Elsevier Ltd.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Introduction
The treatment of COVID-19 is still challenge. So convalescent plasma can be an important alternative of treatment. Protocols with nursing care during infusion is very important to guide an effective and safety care. Objective: to analyze the evidence in the literature on the action of convalescent plasma, of the use of protocols with nursing care to use convalescent plasma and build a nursing care protocol for transfusion in patients with COVID-19. Methods: Methodological study carried out in two stages: scoping review. The search was done using the descriptors: convalescent plasma transfusion, convalescent plasma, and acute respiratory syndromes or COVID-19, to found protocols and effectiveness of convalescent plasm. Beside was done a specialist panel to build the protocol. Results: Low-evidence studies have shown improvement in the clinical signs of COVID-19 using Convalescent Plasma, reduction or elimination of viral load, benefits in the production of lymphocytes, decreases C-reactive protein, increases titers of anti-SARS-CoV-2 antibodies, positive evolution in lung involvement identified by X-rays, decrease in hospitalization. No studies were found in the databases on the protocol for clinical nursing care in plasma transfusion. Therefore, a protocol was developed with the description of clinical nursing care to be performed before, during and after the transfusion by plasma: checking of vital signs and indicative signs of transfusion reaction, measurement of oxygen saturation, assessment of venous access and checking of the level of consciousness. Conclusion: There are no evidence studies to support the use of plasma, nor anything related to bundles.
Keywords
Coronavirus Infections
COVID-19
Nursing
Plasma
Protocols
Therapy
==== Body
pmc1 Introduction
COVID-19 usually manifests with flu-like symptoms such as fever, body pain, headache, loss of smell, taste and dyspnea, which can worsen and progress to Severe Acute Respiratory Syndrome (SARS) (Brazil, 2020).
SARS is defined by the presence of dyspnea or the following signs of severity: Peripheral Capillary Oxygen Saturation (SpO2) less than 95% in room air, signs of respiratory distress or increase in respiratory rate assessed according to age, worsening in clinical conditions of the disease and hypotension regarding to usual blood pressure; individual of any age with acute respiratory failure during seasonal period (Brazil, 2020).
Currently, alternatives have been sought for dealing with this disease, especially in patients whose clinical condition worsens. In this context, administration of plasma or immunoglobulin has been shown to be a potential treatment by reducing the mortality rate in patients with Systemic Respiratory Distress Syndrome (Soo et al., 2004; Gheng et al., 2005).
Studies testing convalescent plasma in patients with SARS and influenza A showed that in both cases there was a relative reduction in the mortality rate, as well as a decrease in viral load, interleukin 6, interleukin 10 levels and tumor necrosis factor. Plasma has been shown to be an important clinical strategy and an alternative treatment in patients with COVID-19 (Hung et al., 2011, Junior et al., 2020).
Thus, it is imperative to search for scientific evidence on the use of convalescent plasma in clinical management of patients with COVID-19, as well as on clinical nursing care protocols referring to it, in the context of the nursing care practice.
Based on the Population, Concept, and Context (PCC) Strategy of scope review, the guiding question was the following: what is the evidence on convalescent plasma transfusion and clinical protocols in the context of patient care with COVID-19?
To answer it, the objectives were to analyze the evidence found in the literature on the action of convalescent plasma in patients with COVID-19 and to build a protocol for clinical nursing care during convalescent plasma transfusion.
2 Method
Methodological study carried out in two stages, scoping review and a panel of experts to build the protocol.
The review was carried out in June 2020 and was based on the electronic bibliographic survey in MEDLINE (Medical Literature Analysis and Retrieval System Online) databases via Pubmed and LILACS (Latin American and Caribbean Literature in Health Sciences).Gray literature was accessed on academic Google and medical society websites to search for guidelines. The review followed the PRISMA protocol.
Primary search strategy carried out at MEDLINE:((((“plasma”[MeSH Terms] OR “plasma”[All Fields]) OR “plasmas”[All Fields]) OR “plasma s”[All Fields]) AND (((((((“covid 19”[All Fields] OR “covid 2019”[All Fields]) OR “severe acute respiratory syndrome coronavirus 2”[Supplementary Concept]) OR “severe acute respiratory syndrome coronavirus 2”[All Fields]) OR “2019 ncov”[All Fields]) OR “sarscov 2”[All Fields]) OR “2019ncov”[All Fields]) OR ((“wuhan”[All Fields] AND (“coronavirus”[MeSH Terms] OR “coronavirus”[All Fields])) AND (2019/12/1:2019/12/31[Date - Publication] OR 2020/1/1:2020/12/31[Date - Publication])))) AND ((((((“blood transfusion”[MeSH Terms] OR (“blood”[All Fields] AND “transfusion”[All Fields])) OR “blood transfusion”[All Fields]) OR ((“blood”[All Fields] AND “component”[All Fields]) AND “transfusion”[All Fields])) OR “blood component transfusion”[All Fields]) OR “blood component transfusion”[MeSH Terms]) OR ((“blood”[All Fields] AND “component”[All Fields]) AND “transfusion”[All Fields])).
The same search strategy was adapted to the LILACS database. All search strategies can be obtained by contacting the authors via email. The records were imported into the reference manager Endnote Basic (Clarivate Analytics), and duplication was removed.
Through the association of descriptive terms, a Boolean search (AND) was performed corresponding to the conceptual blocks to retrieve studies on the clinical protocol for convalescent plasma transfusion, convalescent plasma, and acute respiratory syndromes or COVID-19. The descriptors were determined from the controlled vocabularies MeSH (Medical Subject Headings Section), DeCS (Health Sciences Descriptors).
Inclusion criteria: studies and/or protocols on clinical care for convalescent plasma transfusion in adult and elderly patients with COVID-19 in available full texts. Exclusion criteria: studies containing only clinical trial records and/or summaries of integrative reviews; animal studies, duplicate studies keeping only one.
In MEDLINE via Pubmed, 552 articles were found, and 39 were fully read. In LILACS, 9 articles were found and only one fully read.
From the bibliographic data, a clinical care protocol for convalescent plasma transfusion was developed and then submitted to the panel of experts with national and international members. The panel was consisted of health experts who met at least one of the following criteria: the professional must have at least three years of experience in clinical nursing practice; must be assisting patients with a confirmed or suspected diagnosis of COVID-19 and a diagnosis of SARS.
According to the literature, there is no consensus on the number of evaluators to validate the items of an instrument. However, the number of experts will depend on the available sample with which the researcher can be in contact(7). The sample was non-probabilistic, intentional and composed of 7 judges: 3 professionals with less than 10 years of professional training and experience in clinical practice and 4 professionals with more than 10 years of professional training and experience in clinical practice. Regarding the highest degree, 40% had a PhD and 60% specialization. As for professional performance, 40% was a hematologist or oncologist nurse, 40% a university professor and 20% a nurse working in COVID-19 clinical units.
The specialists received a Likert scale to evaluate all the items of the nursing care protocol in convalescent plasma transfusion, with the options as follows:1- adequate, 2- adequate with changes, 3-inadequate, and a space for observations, in order to suggest necessary changes.
When any of the members disagreed with the text, suggestions were analyzed by the researchers in this study and changed in order to achieve the best version of the instrument. The changes were approved when 100% of the Committee members (panel and researchers) agreed with the proposal. At the end of the evaluation, the authors analyzed the suggestions and made the final version of the protocol, which was reassessed and approved.
3 Results
The flowchart (figure 1 ) details the search steps for selecting the articles included in the review.Figure 1 Flowchart of search and selection of articles.
In the analysis, 27 studies were included. Figure 2 shows the data extraction after analytical reading of the selected studies.Figure 2 Clinical studies on the use of convalescent plasma in the treatment of COVID-19.
From the analysis of the material, it is possible to state that studies on the use of convalescent plasma in the treatment of COVID-19 are incipient, and in nursing care studies there is no material available in national and international literature in databases searched. However, based on the transfusion scientific evidence generally submitted to expertanalysis of specialists who care patients with COVID-19, it was possible to produce a Nursing Care Protocol (Figure 3 ).Figure 3 Clinical nursing care protocol for convalescent plasma transfusion.
4 Discussion
The treatment of COVID-19 nowadays is a challenge. There are still no specific therapeutic agents or vaccines in the clinical treatment of patients affected by COVID-19. Antivirals and antibacterials are used, however, in the need for randomized controlled studies to determine the effectiveness of these drugs against this infectious agent (Wu and McGoogan, 2020, Lu, 2020). Studies have been pointing out to the possibility of using convalescent plasma as an adjuvant therapy for treating COVID-19 (Zeng et al., 2020; Barone and DeSimone, 2020, Knudson and Jackson, 2020).
The use of convalescent plasma was recommended as an empirical treatment during the Ebola virus pandemic in 2014 and as a protocol for the treatment of Midwestern Coronavirus Syndrome in 2015 (FDA, 2020).In addition, a study on the use of convalescent plasma during the influenza A pandemic demonstrated a reduction in viral load in the respiratory tract and mortality (Hung et al., 2011).
On March 24, 2020, FDA (Food and Drug Administration) addressed how patients should be selected and when to use convalescent plasma, namely: identifying subjects and confirming molecularly in case of COVID-19 after being asymptomatic for 14 days, then testing them and confirm that there is no sign of infection by COVID-19 (nasopharyngeal CRP). Therefore, these are the patients eligible to donate convalescent plasma (FDA, 2020; Knudson & Jackson et al., 2020). It is also suggested that convalescent plasma has at least 1: 160 titers of SARS-CoV-2 neutralizing antibodies. If this is not possible, FDA considers the 1:80 titration as acceptable.
Low-evidence studies indicate that the use of convalescent plasma for COVID-19 reduces mortality rates, improves the clinical respiratory signs of the disease, decreases C-reactive protein, increases titers of anti-SARS-CoV-2 antibodies, reduces and/or eliminates viral load, decreases infiltration and lung involvement and shortens hospital stay (Valk et al., 2020; Kong et al., 2019; Rajendran et al., 2020, da Silva, 2020).
Despite these results, a randomized clinical trial conducted in China did not show significantly greater improvement in patients treated with plasma when compared to untreated patients, although plasma therapy demonstrated a reduction in viral load more quickly (72 hours) (Li et al., 2020). Corroborating this study previously referred, a quick review following the Cochrane methodology and a systematic review highlighted that there is no robust evidence to confirm the effectiveness and safety on the use of plasma in patients with COVID-19 (Devasenapathy et al., 2020, Valk et al., 2020). In addition, a case series study conducted in China highlights that the use of plasma in patients with severe COVID-19 does not reduce mortality (Zeng & Xiang, 2020).
The use of convalescent plasma needs to be evaluated for its effectiveness and efficacy, although, in the absence of specific treatment, plasma can be considered a promising technology in the treatment of COVID-19 since important adverse effects related to its use were not identified (Duan et al., 2020, Sahu et al., 2020, Devasenapathy et al., 2020).
Within the scope of their clinical approach, it is up to nurses to adopt qualified and scientific intervention in the face of convalescent plasma transfusion since this technology has been used as an alternative in the absence of specific promising treatment for COVID -19.
No studies addressing the clinical care protocol for plasma transfusion to patients with COVID-19 were found. Thus, based on clinical experience and studies (INS, 2020; Wallis et al., 2014) that address infusion care in general, the aspects/actions that the nurse must follow in the context of convalescent plasma transfusion are highlighted below (Figure 4 ).Figure 4 Nursing care to perform convalescent plasma transfusion.
5 Conclusion
In this review, the authors did not find high-quality studies to prove that convalescent plasma is effective in the treatment of COVID-19, nor studies on clinical protocol with nursing care for transfusion of convalescent plasma in patients with COVID-19. Therefore, a protocol was developed in accordance with infusion studies in generaland the clinical experience of the authors and the examiners. This protocol established the main nursing care in the context of plasma transfusion as follows: assessment of vital signs, level of oxygen saturation, characteristics of venous access, indicative signs of transfusion reactions such as changes in heart and respiratory rate and level of consciousness (if possible).
Among the study limitations, the lack of high-evidence studies with robust methodologies to guarantee efficiency and safety of the use of convalescent plasma in the treatment of COVID-19 stands out.
Further studies are suggestedwith clinical protocols developed in nursing care settings with patients affected by COVID-19.
6 Relevance to clinical practice:
This study build the first protocol of clinical care to the transfusion of Plasm convalescent based on literature and specialist panel, subsidizing clinical practice.
Uncited references
Ahn et al., 2020, Anderson et al., 2020, Bloch et al., 2020, Chen et al., 2020, Cheng et al., 2005, Food and Drug Administration, 2020, Gniadek and Donnersberger, 2020, Honore et al., 2020, Infusion Nurses Society Brasil (BR). Diretrizes práticas para a terapia infusional. 3rd ed. São Paulo: Casa Verde;, 2018, Keith et al., 2020, Kesici et al., 2020, Kong et al., 2020, Ministério da Saúde (BR)., 2020a, Ministério da Saúde (BR)., 2020b, Ministério da Saúde (BR), 2020 [Acesso 11 Jul 2020]., Oliveira et al., 2008, Pawar et al., 2020, Roback and Guarner, 2020, Salazar et al., 2020, Shen et al., 2020, Sullivan and Roback, 2020, Syal, 2020, Tiberghien et al., 2020, Ye et al., 2020, xxxx.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
==== Refs
References
Ahn J.Y. Sohn Y. Lee S.H. Cho Y. Hyun J.H. Baek Y.J. …Choi J.Y. Use of Convalescent Plasma Therapy in Two COVID-19 Patients with Acute Respiratory Distress Syndrome in Korea J Korean Med Sci 35 14 2020 e149 32281317
Anderson J. Schauer J. Bryant S. Graves C.R. The use of convalescent plasma therapy and remdesivir in the successful management of a critically ill obstetric patient with novel coronavirus 2019 infection: A case report. Case Rep Womens Health 27 e00221 2020 10.1016/j.crwh.2020.e00221
Barone P. DeSimone R.A. Convalescent plasma to treat coronavirus disease 2019 (COVID-19): Considerations for clinical trial design Transfusion 60 6 2020 1123 1127 10.1111/trf.15843 32374891
Bloch E.M. Shoham S. Casadevall A. Sachais B.S. Shaz B. Winters J.L. …Tobian A.A. Deployment of convalescent plasma for the prevention and treatment of COVID-19 J Clin Invest 130 6 2020 2757 2765 10.1172/JCI138745 32254064
Chen, L., Xiong, J., Bao, L., & Shi, Y. (2020). Convalescent plasma as a potentialtherapy for COVID-19. Lancet Infect Dis, S1473-3099(20), 30141-9. 10.1016/
Cheng Y. Wong R. Soo Y.O. Wong W.S. Lee C.K. Ng M.H. …Cheng G. Use of convalescent plasma therapy in SARS patients in Hong Kong Eur J Clin Microbiol Infect Dis 24 1 2005 44 46 10.1007/s10096-004-1271-9 15616839
da Silva J.A.T. Convalescent plasma: A possible treatment of COVID-19 in India Med J Armed Forces India. 76 2 2020 236 237 10.1016/j.mjafi.2020.04.006 32296259
Devasenapathy N. Ye Z. Loeb M. Fang F. Najafabadi B.T. Xiao Y. …Guyatt G. Efficacy and safety of convalescent plasma for severe COVID-19 based on evidence in other severe respiratory viral infections: A systematic review and meta-analysis CMAJ 192 2020 E745 E755 10.1503/cmaj.200642 32444482
Duan K. Liu B. Li C. Zhang H. Yu T. Qu J. …Yang X. Effectiveness of convalescent plasma therapy in severe COVID-19 patients Proc Natl Acad Sci 117 17 2020 9490 9496 10.1073/pnas.2004168117 32253318
Food and Drug Administration Investigational-covid-19-convalescent-plasma-emergency-inds. US Food and Drug Administration 2020 Food and Drug Administration USA [cited Jul 8 2020]. Available from:
Gniadek T.J. Donnersberger D. COVID-19 convalescent plasma donor recruitment: Beware the Faustian bargains Transfusion 2020 2020 10.1111/trf.15871
Honore P.M. Mugisha A. Kugener L. Redant S. Attou R. Gallerani A. …De Bels D. Therapeutic plasma exchange as a routine therapy in septic shock and as anexperimental treatment for COVID-19: We are not sure. Version 2 Crit Care 24 1 2020 226 10.1186/s13054-020-02943-1 32414417
Hung I.F. To K.K. Lee C.K. Lee K.L. Chan K. Yan W.W. Yuen K.Y. Convalescent plasma treatment reduced mortality in patients with severe pandemic influenza A (H1N1) 2009 virus infection Clin Infect Dis. 52 2011 447 456 10.1093/cid/ciq106 21248066
Infusion Nurses Society Brasil (BR). Diretrizes práticas para a terapia infusional. 3rd ed. São Paulo: Casa Verde; 2018.
Junior D.M.L. Santis G.C. Bordin J.o. COVID-19 convalescent plasma transfusion Hematology, transfusion and cell therapy 42 2 2020 113 115 32313872
Keith P. Day M. Perkins L. Moyer L. Hewitt K. Wells A. A novel treatment approach to the novel coronavirus: An argument for the use of therapeutic plasmaexchange for fulminant COVID-19. Version 2 Crit Care. 24 1 2020 128 10.1186/s13054-020-2836-4 32241301
Kesici S. Yavuz S. Bayrakci B. Get rid of the bad first: Therapeutic plasma exchange with convalescent plasma for severe COVID-19 Proc Natl Acad Sci 23 2020 12526 12527 10.1073/pnas.2006691117
Knudson C.M. Jackson J.B. COVID-19 convalescent plasma: Phase 2 Transfusion 60 6 2020 1123 1127 10.1111/trf.15842 32374891
Kong Y. Cai C. Ling L. Zeng L. Wu M. Wu Y. …Liu Z. Successful treatment of a centenarian with coronavirus disease 2019 (COVID-19) using convalescent plasma Transfus Apher Sci 102820 2020 10.1016/j.transci.2020.102820
Li L. Zhang W. Hu Y. Tong X. Zheng S. Yang J. …Liu Z. Effect of Convalescent Plasma Therapy on Time to Clinical Improvement in Patients With Severe and Life-threatening COVID-19: A Randomized Clinical Trial JAMA 2020 10.1001/jama.2020.10044
Lu H. Drug treatment options for the2019-newcoronavirus (2019-nCoV) Biosci Trends 14 1 2020 69 71 10.5582/bst.2020.01020 31996494
Ministério da Saúde (BR). (2020). Diretrizes para diagnóstico e tratamento da COVID-19. Brasília: Ministério da Saúde; 2020 [Access 10 jul 2020]. Available from: http://portalarquivos.saude.gov.br/images/pdf/2020/April/10/Diretrizes-covid-V2-9.4.pdf
Ministério da Saúde (BR). (2020). Protocolo de Manejo Clínico da Covid-19 na atenção especializada. Brasília: Ministério da Saúde; 2020 [Access 10 jul 2020]. Available from: http://bvsms.saude.gov.br/bvs/publicacoes/manejo_clinico_covid-19_atencao_especializada.pdf
Ministério da Saúde (BR) Secretaria de Ciência, Tecnologia e Insumos Estratégicos em Saúde. Utilização de plasma convalescente para tratamento de COVID-19 2020 [Acesso 11 Jul 2020]. Ministério da Saúde Brasília Available from:
Oliveira M.S. Fernandes A.F.C. Sawada N.O. Educational handbook for selfcare in women with mastectomies: A validation study Texto & Contexto Enferm 17 1 2008 115 123
Pawar A.Y. Hiray A.P. Sonawane D.D. Bhambar R.S. Derle D.V. Ahire Y.S. Convalescent plasma: A possible treatment protocol for COVID- 19 patients suffering from diabetes or underlying liver diseases Diabetes Metab Syndr 14 4 2020 665 669 10.1016/j.dsx.2020.05.023 32438330
Rajendran K. Krishnasamy N. Rangarajan J. Rathinam J. Natarajan M. Ramachandran A. Convalescent plasma transfusion for the treatment of COVID-19: Systematic review J Med Virol 10 2020 10.1002/jmv.25961
Roback, J.D., & Guarner, J. (2020). Convalescent Plasma to Treat COVID-19: Possibilities and Challenges. JAMA, 323(16), 1561-62. 10.1001/jama.2020.4940 S1473-3099(20)30141-9
Sahu K.K. Jindal V.S. Ahmad D. Cerny J. Gerber J.M. Convalescent Plasma Therapy: A Passive Therapy for An Aggressive COVID-19 J Med Virol 2020 10.1002/jmv.26047
Salazar, E., Perez, K.K., Ashraf, M., Chen, J., Castillo, B., Christensen, P.A., … Musser, J.M. (2020). Treatment of COVID-19 Patients with Convalescent Plasma. Am J Pathol, 27(20), 30257-1. 10.1016/j.ajpath.2020.05.014
Shen C. Wang Z. Zhao F. Yang Y. Li J. Yuan J. …Liu L. Treatment of 5 Critically Ill Patients With COVID-19 With Convalescent Plasma JAMA 323 16 2020 1582 1589 10.1001/jama.2020.4783 32219428
Soo Y.O. Cheng Y. Wong R. Hui D.S. Lee C.K. Tsang K.K. …Sung J.J.Y. Retrospective comparison of convalescent plasma with continuing high-dose methylprednisolone treatment in SARS patients Clin Microbiol Infect 10 2004 676 678 10.1111/j.1469-0691.2004.00956.x 15214887
Sullivan H.C. Roback J.D. Convalescent Plasma: Therapeutic Hope or Hopeless Strategy in the SARS-CoV-2 Pandemic Transfus Med Rev S0887–7963 20 2020 30025 30030 10.1016/j.tmrv.2020.04.001
Syal K. COVID-19: Herm Immunity and Covalescent Plasma Transfer Therapy J Med Virol 2020 2020 1 3 10.1002/jmv.25870
Tiberghien P. de Lamballerie X. Morel P. Gallian P. Lacombe K. Yazdanpanah Y. Collecting and evaluating convalescent plasma for COVID-19 treatment: Why and how? Vox Sang 2020 2020 10.1111/vox.12926
Valk S.J. Piechotta V. Chai K.L. Doree C. Monsef I. Wood E.M. …Skoetz N. Convalescent plasma or hyperimmune immunoglobulin for people with COVID-19: A rapid review Cochrane Database Syst Rev 5 2020 CD013600 10.1002/14651858 32406927
Wallis M.C. McGrail M. Webster J. Marsh N. Gowardman J. Playford E.G. Rickard C.M. Risk factors for peripheral intravenous catheter failure: Amultivariate analysis of data from a randomized controlled trial Infect Control 35 1 2014 63 68 10.1086/674398
Wu Z. McGoogan J.M. Characteristics of and important lessons from the coronavirus disease 2019 (COVID-19) outbreak in China: Summary of a report of 72314 cases from the Chinese Center for Disease Control and Prevention JAMA 323 13 2020 1239 1242 10.1001/jama.2020.2648 32091533
Ye M. Fu D. Ren Y. Wang F. Wang D. Zhang F. …Tangfeng L. Treatment with convalescent plasma for COVID-19 patients in Wuhan China. J Med Virol 1–12 2020 10.1002/jmv.25882
Zeng F.C. Xiang D. Convalescent plasma for patients with COVID-19 Proc Natl Acad Sci 117 23 2020 12528 10.1073/pnas.2006961117 32398379
Zeng, Q.L., Yu, Z.J., Gou, J.J., Li, G.M., Ma, S.H., Zhang, G.F., … Liu, Z. Effect of Convalescent Plasma Therapy on Viral Shedding and Survival in COVID-19 Patients. J Infect Dis, 16(1), 38-43. 10.1093/infdis/jiaa228
| 0 | PMC9745971 | NO-CC CODE | 2022-12-15 00:03:20 | no | Int J Afr Nurs Sci. 2023 Dec 13; 18:100518 | utf-8 | Int J Afr Nurs Sci | 2,022 | 10.1016/j.ijans.2022.100518 | oa_other |
==== Front
J Drug Deliv Sci Technol
J Drug Deliv Sci Technol
Journal of Drug Delivery Science and Technology
1773-2247
2588-8943
Elsevier B.V.
S1773-2247(22)00993-5
10.1016/j.jddst.2022.104082
104082
Article
Development of favipiravir loaded PLGA nanoparticles entrapped in in-situ gel for treatment of Covid-19 via nasal route
Gattani Vaishnavi a
Dawre Shilpa ab∗
a Department of Pharmaceutics, School of Pharmacy & Technology Management, SVKMS, NMIMS, Babulde Banks of Tapi River, MPTP Park, Mumbai-Agra Road, Shirpur, Maharashtra, 425405, India
b Department of Pharmaceutics, School of Pharmacy, Vishwakarma University, Laxmi Nagar, Kondhwa, Pune, Maharashtra, 411048, India
∗ Corresponding author. Department of Pharmaceutics, School of Pharmacy & Technology Management, SVKMS, NMIMS, Babulde Banks of Tapi River, MPTP Park, Mumbai-Agra Road, Shirpur, Maharashtra, 425405, India.
13 12 2022
13 12 2022
1040827 10 2022
8 12 2022
10 12 2022
© 2022 Elsevier B.V. All rights reserved.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
In 2019 the emergence of SARS-COV-2 caused pandemic situation worldwide and claimed ∼6.4 M lives (WHO 2022). Favipiravir (FAV) is recommended as a therapy for Covid-19 which belongs to BCS class III with a short half-life of 2–5.5h. Thus, the objective of current study was the development of favipiravir loaded PLGA nanoparticles (NPs) by box-behnken design. Moreover, these NPs were entrapped in thermosensitive gel to increase the permeation through nasal route. The nanoparticles exhibit particle size of 175.6 ± 2 nm with >70 ± 0.5 %EE. NPs showed PDI (0.130) and zeta potential (−17.1 mV) suggesting homogeneity and stability of NPs. DSC, XRD, and FTIR studies concluded absence of any interaction of FAV and the excipients. SEM and AFM studies demonstrated spherical morphology of NPs with smooth surface. The NPs entrapped in-situ gel showed clarity and pH 5.5–6.1. The gelation temperature of NPs dispersed in-situ gel was found in the range of 35 °C −37 °C. The gel has viscosity in range of 34592–4568 cps. The texture analysis profile of gel showed good gelling properties. Dissolution study suggested a sustained release of FAV from NPs (24h) and NPs dispersed gel (32h) as compared to FAV solution (4h). The gel showed good mucoadhesion properties (9373.9 dyne/cm2). Ex-vivo permeation through nasal mucosa of goat elucidated NPs dispersed gel demonstrated significantly higher permeation than solution and NPs. Therefore, it would be a prospective formulation to combat Covid-19 infection with high patient compliance.
Graphical abstract
Image 1
Keywords
Favipiravir
Mucoadhesion
PLGA Nanoparticles
Thermoreversible in-situ gel
Nasal ex vivo permeation
Box-behnken design
==== Body
pmcAbbreviations
FAV Favipiravir
SARS-COV-2 Severe-acute respiratory syndrome corona virus-2
DOE Design of Experimentation
PLGA Poly-Latic-Glycolic acid
HPMC Hydroxy-Propyl Methyl Cellulose
BBD Box-Behnken, design
%EE % Entrapment Efficiency
PDI Poly dispersity index
FTIR Fourier Transforms Infrared spectroscopy
DSC Differential-Scanning-Calorimetry
SEM Scanning Electron Microscopy
AFM Atomic force microscopy
XRD X-ray diffraction
PBS Phosphate Buffer Saline
NPs Nanoparticles
ANOVA Analysis of variance
1 Introduction
The severe acute respiratory syndrome coronavirus-2 (SARS COV-2) is a new-fangled microorganism and origin of Covid-19 pandemic situation globally [1]. Current therapies, research & development groups, and medical facilities of all countries were trying to find solutions and therapies of the Covid-19 infection that has claimed ∼6.4 M lives to date [2]. This viral strain is extremely infectious and contagious. It causes flu-like symptoms causing severe respiratory syndrome leading to death due to respiratory failure [3]. Many antivirals were repurposed that could be useful against this unique coronavirus, due to the lack of new drugs to combat the condition. In addition, the occurrence of multidrug resistance and low antiviral concentration at the site of infection posed challenges in treatment [4]. Thus, there is a need of new formulation strategies and modalities that could assist in tackling this Covid-19 infection.
Numerous novel formulations such as solid lipid nanoparticles [5,6], liposomes, NLCs [7], nanoparticles (NPs), etc. have been studied for delivery of drugs via nasal route. Moreover, these novel formulations were explored as a therapy of Covid-19 infection [8] and amongst these nanoparticles were successfully applied for development of vaccine [9]. PLGA is a U.S. FDA approved biodegradable polymer with great efficiency of drug loading. It has been previously explored for development of nanoparticles loaded with variety of drugs. Surnar B. et al. developed ivermectin encapsulated nanoparticles using PLGA polymer modified with PEG and maleimide (PLGA-b-PEG-Mal). Furthermore, they had decorated nanoparticles by Fc immunoglobulin fragment to cross the epithelial layer of gut which allowed drug to reach into bloodstream. In vitro cytotoxicity and uptake studies in HEK293T cells showed better uptake of NPs than free drug. Moreover, oral bioavailability of drug was found higher than drug solution [10]. Ucar et al. designed oseltamivir-phosphate loaded PLGA nanoparticles decorated by means peptide binder of SARS-COV-2 spike protein for ACE-2 receptor targeting. Nanoparticles demonstrated a sustained release of drug for two months [11]. In another study Wu et al. developed lisinopril tagged nanoparticles loaded with remdesivir by molecular docking approach to target on angiotensin-converting enzyme −1 receptor [12].
Favipiravir (FAV) is a molecule approved in Japan for treatment of influenza virus infection and recently approved as a therapy of SARS-COV-2. It specifically inhibits viral RNA-dependent RNA polymerase (RdRp). Earlier studies revealed that favipiravir showed antiviral activity for large range of RNA viruses than other antiviral agents [13]. The available marketed formulation of favipiravir is oral tablet (Avigan®, Fabiflu®). The dosing frequency of FabiFlu® is 18 tablets on day 1 (200 mg) (nine in the morning and nine in the evening), trailed by eight tablets each day subsequently for a maximum of 14 days. It is highly patient incompliance with high toxicity and adverse effects [14]. Furthermore, FAV belongs to the BCS class III with a short half-life of 2–5.5h. Therefore, to decrease this dose size and dosing frequency, there is a need of novel formulation. The intranasal route could be suitable route for Covid-19 treatment that can deliver high drug concentration at the target site and circumvent undesirable adverse effects as lower doses are required [15]. Moreover, it avoids first pass metabolism and related side effects. Gilead Sciences, USA also conducting clinical trials of remdesivir in inhaled form for Covid-19 [16].
In the present study we have prepared and optimized FAV loaded PLGA nanoparticles (NPs) by using box-behkhen design for sustained release action. Furthermore, these nanoparticles were entrapped in thermosensitive in situ-gel for enhancement of nasal permeation. The developed formulation showed sol to gel transition gel in nasal area which could reduce mucociliary clearance and can allow NPs to diffuse through nasal mucosa. Moreover, intranasal route enhances patient compliance and reduce dose & dosing frequency. FAV-NPs and FAV-NPs dispersed in-situ gel were assessed for physicochemical characteristics, dissolution pattern, and permeation through nasal mucosa of goat.
2 Materials and methods
2.1 Materials
Favipiravir was generously presented by Viruj Pharmaceuticals (Hyderabad, India). The polymer Poly lactide-co-glycolide (PLGA, 50:50) was donated by Evonik India Pvt. Ltd. (Mumbai, India). Other polymers Hydroxy-Propyl Methyl Cellulose (HPMC) and Kolliphor P-407 & Kolliphor P-188 were bought from Rankem Labs, Mumbai, India and BASF, India, respectively. Other chemicals like hydrogen chloride, sodium hydroxide, potassium dihydrogen orthophosphate phosphate, acetonitrile, acetone, and analytical grades solvents were procured from SD Fine Chem, (India).
2.2 Methods
2.2.1 Development of FAV nanoparticles
The nanoprecipitation method was used to prepare nanoparticles with little modification [17]. The organic phase was produced by solubilizing PLGA (5 mg & 20 mg) and favipiravir (30 mg & 200 mg) in acetone (10 mL). Poloxamer 407 (10 mg & 20 mg) was solubilized in water (20 mL) to make an aqueous phase. After that it was placed on water bath for complete dissolution and cooled for 10–15 min. Then an aqueous phase was mixed gradually into an organic phase. The acetone was then permitted to evaporate for 3h by stirring emulsion via magnetic stirrer. Later nanoparticles were centrifuged at 14,000 rpm for 20 min. The nanoparticles were collected and rinsed three times with distilled water.
2.2.2 Nanoparticles optimization by design of experimentation
Optimization of the favipiravir loaded polymeric nanoparticles was done with box-behnken design (BBD) using three parameters at two levels. The selected optimization parameters were favipiravir concentration (X1), the PLGA amount (X2), and Poloxamer 407 amount (X3) were considered as an independent parameter and the two dependent factors, particles size (Y1) and percentage entrapment efficiency (Y2) were considered (Table 1 ). Table 2 illustrates trials suggested by software with five center locations. To elude biasness, the trials were conducted randomly.Table 1 Independent parameters for design of nanoparticles by BBD.
Table 1No. Parameters Low (mg) High (mg)
1 FAV amount 30 200
2 PLGA amount 5 20
3 Poloxamer 407 amount 10 20
Table 2 Outcome of independent parameters on response factors.
Table 2Run Parameter 1 Parameter 2 Parameter 3 Response 1 Response 2
X1: Favipiravir amount X2: PLGA amount X3: Poloxamer amount Y1: Particle size Y2: Entrapment efficiency
mg Mg mg nm %
1 115 20 15 208.2 73.72
2 115 11.5 10 175.6 78.24
3 115 11.5 10 175.6 78.24
4 200 11.5 5 180.5 74.23
5 115 3 5 171.5 69.63
6 30 20 10 196.1 65.59
7 115 20 5 179.9 72.35
8 30 11.5 15 168.9 68.89
9 200 11.5 15 188.8 72.56
10 115 11.5 10 175.6 78.24
11 200 20 10 212.24 75.54
12 115 3 15 148.8 65.23
13 200 3 10 168.8 68.23
14 30 11.5 5 188.5 67.35
15 30 3 10 174.5 62.54
16 115 11.5 10 175.6 78.24
17 115 11.5 10 175.6 78.24
The DOE generated the non-linear quadratic design, which was found to beY = b0 + b1X1 + b2X2 + b3X3 +b4X1X2 + b5X1X3 + b6X2X3 + b7 X21 + b8X22 + b9X23
In the above equation Y represents the response due to the interaction of all factors in combination. The intercept is b0, and the regression coefficients are b1-b9, were obtained from the experimental outcomes. The independent factors are X1, X2, and X3. The additional parameters (mixing time, non-aqueous and aqueous phase amount, etc.) were retained constant. The statistical data from the linear regression as well as the 3D surface response graphs were examined using Design-Expert® Version 11.
2.2.3 Development of FAV-NPs entrapped gel
The in-situ gel was developed by using previously reported cold method [18]. Poloxamer 407 (2 gm) and Poloxamer 188 (0.5 gm) were accurately weighed and added slowly in nanoparticle suspension at 4 °C temperature. The blend was agitated using a magnetic stirrer at 800 rpm (Remi Lab, India) till polymers were solubilized. After that hydroxyl propyl methyl cellulose (1% w/v) solution which is a mucoadhesive polymer was mixed and solubilized in above mixture. The developed gel was characterized for gel properties.
2.2.4 Evaluation of FAV nanoparticles
2.2.4.1 Entrapment-efficiency (%EE)
The amount of favipiravir encapsulated in nanoparticle was determined by using UV spectroscopy. Nanoparticles dispersion was rotated at 14000 rpm at 4 °C for 20mins by using centrifuge. After this the supernatant was accumulated and diluted with ethanol. Then analyzed for FAV by UV spectrophotometer at ƛmax 237 nm. The formula; % EE = quantity of FAV in nanoparticles/theoretical amount of drug * 100 was utilized. In which the theoretical amount signifies the amount of drug initially loaded in NPs.
2.2.4.2 ζ-potential, polydispersity index, & particle size
The size of FAV NPs was calculated by utilizing dynamic light scattering technology (Malvern particle size analyser 2000 series, UK). The nanoparticle dispersion (0.5 mL) was prepared in purified water (10 mL) using ultrasonic homogenizer (DP120, Mumbai, India) operated at 150 voltages with on-off cycle of 10s for 5mins. The particle size analysis was conducted at scattering angle of 90° at 25 °C. Then this dispersion was analyzed for ζ-potential, polydispersity index, and particle size.
2.2.4.3 FTIR study
FTIR scanning of FAV, Poloxamer-407, PLGA, and excipients with FAV was carried out by using Shimadzu Affinity- 1S. All the samples were in powder from, small quantity of this sample was carefully kept into the empty space and later were examined at resolution of 4/cm −1 in the IR area (400–4000/cm −1).
2.2.4.4 DSC study
Favipiravir, PLGA, and FAV-loaded nanoparticles were analyzed for DSC at warming rate of 10 °C (DSC 60 Shimadzu). Samples were analyzed at 30 °C – 400 °C. The movement of heat is from the primary and secondary warming gradients used to determine the melting temperature (Tm).
2.2.4.5 XRD study
X-ray diffraction study was done by using Philips Analytical Xperto Pro PW-1710, with a resolution of 0.001A. The diffraction pattern of samples was recorded with 2 θ range of 5–60° in the steps of 0.01o/sec. Cu Kα radiation operated at 40 kV and 30 mA was used as an X-ray source.
2.2.4.6 Scanning electron microscopy (SEM)
The superficial structure of nanoparticles was studied by scanning electron microscope (ZEISS, SUPRA 55VP). The nanoparticle suspension was fixed upon a slide (1*1 cm) using a tape. Then sputtered with gold and observed using microscope at a magnifying power of 13830. Polymeric nanoparticles shape and size were recorded.
2.2.4.7 Atomic force microscopy (AFM)
Atomic force microscopy was assessed utilizing dimension XR. The nanoparticles were diluted with distilled water and dried on mica coated glass slide. The images were taken with a silicon-type probe in non-contact operating mode [19]. Then the slide was analyzed under 100 μ m scanner and canteliver of 25 N/m and of radius 10 nm.
2.2.4.8 Stability studies
The stability NPs was checked in accordance with International Council of Harmonization guidelines Q1(R2). The nanoparticles were placed for six months at 70 ± 5% RH and 40 ± 2 °C. The aliquots were taken and assessed for their %EE, particle size, and physical appearance in every one month.
2.2.5 Evaluation of FAV-NPs entrapped gel
2.2.5.1 Transparency and pH
The gel was cautiously inspected under brightened background for any sign of foreign particles or turbidity. The pH of the gel was calculated by an electronic pH meter (LABINDIA Pico+ design).
2.2.5.2 Viscosity measurement
The viscosity of gel was performed utilizing 10 ml of gel in beaker under Brookfield digital DV-II Model. The viscous property of gel was tested at different speed (10–100). The generated % torque was recorded.
2.2.5.3 Gelation temperature
The temperature at which gel formation occurs was determined by stirring 20 ml gel at 37 °C and speed of 50 rpm. The degree centigrade condition at which the magnetic bar become steady was recorded.
2.2.5.4 Texture profile
The gel texture profile was assessed using an instrument (TA-XT texture analyser). The gel (10 mL) was kept in a beaker and adjusted under analyser at 25 mm height at 37 ± 2 °C. After that a probe was permitted to insert at least 10 mm inside the gel at a speed of 2 mm/s. Then the probe was raised slowly which allows the detachment of gel from the layer. The power requisite for separation provides the gel strength.
2.2.6 Dissolution studies
The dissolution profile of favipiravir was conducted using Franz diffusion cell [20]. The FAV release pattern was assessed form solution, FAV-NPs, and FAV-NPs distributed gel. The dissolution media used was Phosphate Buffer (PBS) at 5.5 pH. The cellophane membrane (capacity 2.41 mL/cm and diameter 17.5 mm) was measured as per requirement and activated in PBS 5.5. After 24h the cellophane membrane was washed with purified water to eradicate any impurity, then placed in Franz diffusion cell apparatus. Weighed quantity of FAV solution, FAV-NPs dispersion, and FAV-NPs dispersed gel were taken in equivalence to the drug concentration (3 ml = 17 mg) and kept into donor cell. The receptor cell was filled with dissolution media (20 ml). The apparatus was then placed at 37 °C onto a stirrer (300 rpm). The sample of 1 ml was withdrawn at various time points (30mins, 1, 1.5, 2, 2.5, 3, 4, 6, 8, 12, 24, 28, & 32h) and was replenished with PBS for each aliquot. The aliquots were evaluated by the UV spectrophotometer for FAV at 237 nm.
2.2.7 Mucoadhesion study
The nasal mucosal adhesion of the FAV-NPs dispersed gel was performed by finding out the power needed to remove formulation adhered to mucosal sheath. The nasal mucosa of goat was procured from a local butcher shop and adhere on to the vial (0.785 cm2 surface area). The upper probe attached with the mucosal sheath by using cyanoacrylate adhesive. However, gel was attached with the bottom side of probe and allowed to maintained at 37 °C. Force of 0.1 N was utilized to assure close contact between the mucosa and gel. An equilibrium tool was utilized for the upward movement of upper and lower probes at a constant rate (0.15 nm/s). The formula, mucoadhesive strength = Fmax *g/A was applied for calculation of mucoadhesive strength of gel in dyne/cm2. Wherein, Fmax = maximum force required for separation (grams), g = acceleration due to gravity [980 cm/s2], and A = area of mucosa layer attached.
2.2.8 Ex vivo permeation study
The permeation of FAV solution, FAV-NPs, and FAV-NPs dispersed gel were conducted using nasal mucosa of goat [19]. It was obtained from local slaughter shop. Then mucosa was placed in Franz diffusion cell (area 2.0 cm2 & thickness of 0.3 mm). Two minutes were spent after adhering the nasal mucosa so that the drug solution or gel should not get leaked. The FAV suspension in phosphate buffer, FAV-NPs dispersion, and FAV-NPs dispersed gel 100 μL (1.4 mg/ml) were placed separately in different cells. The samples were taken at interval of 5, 15, 30 min, 1h, 2h, 4h, & 6h which was replaced by fresh phosphate buffer at every point of collection of samples. The aliquots were examined for FAV by validated RP-HPLC method at 237 nm. Aliquots were individually substituted by an equivalent amount media. The graph was plotted between cumulative amount of FAV permeated from nasal mucosa vs time to calculate flux. The below mentioned formula was utilized to calculate permeability coefficient (P):P = dQ/dt CoA
where Co indicates the initial concentration in the donor cell, dQ/dt signifies the rate of permeability or flux [mg/h], and A epitomizes the active area of the nasal mucosa.
2.2.9 Statistical study
The standard deviation and mean were calculated for entire data set. Dunnett's test and one-way ANOVA on data was conducted by GraphPad Prism 5, USA software, in which p < 0.05 indicates a considerable difference.
3 Results and discussion
3.1 Development of nanoparticles
Favipiravir encapsulated polymeric nanoparticles were successfully prepared using nanoprecipitation technique with high drug loading (>70%). Nanoprecipitation technique provide nanoparticles with high encapsulation efficiency [21]. Furthermore, these NPs were successfully entrapped in thermosensitive in-situ gel for enhancement of nasal permeation.
3.2 Nanoparticles optimization by design of experimentation
The BBD design was utilized for development of NPs. The impact of independent factors (X1) such as FAV amount, (X2) PLGA amount, and (X3) Poloxamer 407 amount on responses (NPs size and %EE) studied by response surface methodology (RSM). The software suggested 17 runs with 5-center points and formulations were designed (Table 2). Moreover, software provided quadratic equation that explains the interaction of all variables and their effects. The significance level (5%) of the quadratic equations was measured by ANOVA. Table 3, Table 4 illustrates the coefficients of quadratic equation and p-values. The desirability value of optimized formulation was found 0.85 which is close to an ideal value 1. The desirability value equal to 0 was considered not satisfactory, however a value near to 1 signified an acceptable value [22].Table 3 Summarization of model properties for responses Y1 and Y2.
Table 3Models R2 Adjusted R2 Predicted R2 SD Adequate precision % CV Remarks
Particle size (Y1)
Quadratic 0.9899 0.9770 0.8388 2.30 37.31 1.27 Significant
% Entrapment efficiency (Y2)
Quadratic 0.9901 0.9774 0.8417 0.79 25.004 1.10 Significant
Table 4 Resultant p-values of dependent factors.
Table 4Variables p-value of NPs particle size p-value of %EE
X1 0.0108 <0.0001
X2 <0.0001 <0.0001
X3 0.4093 0.2007
X1X2 0.0021 0.0310
X1X3 0.0005 0.0820
X2X3 <0.0001 0.0082
X12 0.0001 <0.0001
X22 0.0106 <0.0001
X32 0.0722 0.0003
3.2.1 Influence on particle size
The following equation illustrates the influence of several independent variables on the particle size (Y1):Particlesize(y)=175.60+2.79x1+16.61x2−0.7125x3+5.46x1x2+6.98x1x3+12.75x2x3+8.44x12+3.87x22−2.37x32
The F-value of the model was 76.41, concluded that the design was significant (p < 0.001). The correlation between the experimental and estimated values was indicated by the R2 of 0.9899. The “Predicted R2” (0.8388) and “Adjusted R2” of 0.9770, suggested that the design was found to be accurate to predict the outcomes. The S/N ratio was used to compute “adequate precision.” The resultant ratio of 10.94 recommended that the signal was within acceptable range. A percent CV of 1.27 reflected the model's precision and reliability.
The p-value was utilized to predict the effects of coefficients. The lower the p-value, the greater the significance of the associated coefficient. The model terms are significant when “Prob > F" (<0.05). Table 4 elucidates that the all-design variables X1, X2, X3, X1X2, X2X3, and X1X3 were significant. A positive symbol signifies a synergistic effect; however, negative symbol denotes an antagonistic result.
The RSM plot illustrates the influence of independent parameters on nanoparticles size (Fig. 1 ). It was detected that with increase in the PLGA amount, there was increase in NPs size occurred. This effect was observed due to increase in polymer concentration that resulted into enhancement in viscosity and created problem in preparation of minute globules at steady mixing speed [23]. Thus, high particle size was obtained owing to high concentration of PLGA. These results were in accordance with study done by Huang et al. They had studied effect of varied PLGA concentrations on particle size and concluded that concentration of PLGA is directly proportional with the NPs size [24].Fig. 1 3D surface Plots of A) showing effects of PLGA concentration (X2) and Favipiravir Concentration (X1); B) showing effects of Poloxamer 407 concentration (X3) and Favipiravir Concentration (X1) C) PLGA concentration (X2) and Poloxamer concentration (X3) on response particle size (Y1).
Fig. 1
Additionally, the concentration of surfactant is an important risk analysis parameter in development of nanoparticles and it is a primary influencer on nanoparticle size [25]. Generally, it has been observed that concentration of poloxamer 407 is inversely proportional with the nanoparticle size. The surfactant plays a crucial role in reduction of surface tension and provide stability to nanoparticles [26]. In our study it was observed that enhancement of the poloxamer 407 concentration reduces NPs size, however, after some point, increase in particle size was observed. The probable reason could be as initially the amount of poloxamer 407 could be sufficient for reduction of particle size. Nevertheless, further increase in surfactant concentration caused increase in particle size due to adsorption of poloxamer 407 on NPs surface resulted formation of thick coating [27]. Furthermore, with increase in surfactant amount viscosity of aqueous phase also rises resulting large particle size [28].
The above-mentioned results were in accordance with previous studies [27,29,30]. Sakhi and co-authors developed polymeric nanoparticles loaded with paclitaxel for treatment of breast cancer. They observed similar results as with increase in poloxamer concentration PLGA, NPs size increases [27]. Thus, it could be concluded that concentrations of polymer and surfactant both significantly effect NPs size.
3.2.2 Influence on %EE
The polynomial equation obtained for %EE is mentioned below:(y2)=78.26+5.70x1−0.7900x2+1.25x3+3.05x1x2+0.7500x1x3−2.37x2x3−6.40x12−4.54x22−2.31x32
The F-value 77.81 illustrates the model is significant (p < 0.0001). The coefficient of regression (R2) of the design was obtained 0.9901. The “Predicted R2”, “Adjusted R2” (0.8417, 0.9774), and percent CV (1.10%) confirmed good model fit.
The results indicate that as the concentration of PLGA increases there is rise in %EE, however, after a definite point there was reduction in %EE occurred (Fig. 2 ). This might be due to increase in viscosity of the phases with increase in PLGA amount, attributed rapid solidification and slower diffusion of drug from interior to exterior region [24]. On the other hand, as the concentration of surfactant Poloxamer 407 increased the dispersal of drug in both aqueous and non-aqueous phases occurred. Furthermore, surfactant also enhanced drug solubility in exterior phase and improves stability of the w/o emulsion [31]. Thus, decrease in %EE was observed with increase in poloxamer 407 concentration.Fig. 2 3D surface plots of A) showing effects of Favipiravir concentration (X1) and PLGA concentration (X2); B) Favipiravir concentration (X1) and Poloxamer concentration (X3) C) PLGA concentration and Poloxamer concentration response entrapment efficiency (Y2).
Fig. 2
3.3 Evaluation of FAV-nanoparticles
3.3.1 Entrapment efficiency
Entrapment efficiency is an important criterion in preparing the favipiravir loaded polymeric nanoparticles, as well as it has been proved to be an important variable in the optimization of nanoparticles. To obtain precise drug encapsulation, the factors such as PLGA concentration and Poloxamer 407 concentration were found to be important factors. The optimized formula of nanoparticles (Formula 2) has %EE >70%.
3.3.2 ζ-potential, polydispersity index, & particle size
The nanoparticles revealed size distribution in range of 175.6 ± 3 nm with PDI 0.130 suggesting that the particle size was evenly distributed throughout the dispersion and have a uniform size (Fig. 3 ). Zeta potential helps in understanding stability of NPs. Zeta potential values higher than +30 mV and less than −30mV, considered as stable [32]. However, it was observed that PLGA NPs mostly carries negative charge due to presence of carboxyl group [33]. Thus, the zeta potential of FAV-NPs was observed −17.5mv due to the charge on the PLGA molecule. Cui et al. also reported stable PLGA nanoparticles loaded with doxorubicin and paclitaxel which showed zeta potential in the range of −9.0mv to −18.0 mv [34].Fig. 3 Particle size distribution and ζ-potential of FAV NPs.
Fig. 3
3.3.3 Scanning electron microscopy
SEM study demonstrated that FAV-NPs exhibit spherical morphology and possessed a smooth surface which in agreement with dynamic light scattering analysis (Fig. 4 ). The image illustrates absence of aggregation or crysallization of drug molecule and confirmed uniformity of nanoparticles. Moreover, spherical morphology with smooth surface could result into sluggish clearance and good deposition pattern in nasal area. The results were in accordance with prior study by Patil et al. They reported that ropinirole hydrochloride encapsulated PLGA NPs for nasal delivery and revealed that cicular NPs diffuses slowly through nasal mucosa as compared to rod-shaped NPs [35].Fig. 4 SEM image of FAV nanoparticles.
Fig. 4
3.3.4 FTIR analysis
FTIR transforms of FAV, PLGA, Poloxamer, and overlay plot shown in Fig. 5 (A, B, C, D). Favipiravir transform of FAV revealed all typical peaks such as band developed due to vibrations & stretching of carbonyl bond (C <svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="20.666667pt" height="16.000000pt" viewBox="0 0 20.666667 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.019444,-0.019444)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z M0 280 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z"/></g></svg> O) at 1716 cm−1, ether (C–O–C) band at 1180 cm−1 and vibrations stretching of amine N–H at 3347 cm−1 & 3212 cm−1 [36]. The transform of poloxamer 407 showed band of esters (C–O–C) at 1240 cm− 1, vibrations stretching of (O–H at 3647.2 cm− 1 & amine bond (N–H) at 3347 cm− 1). The FTIR of PLGA suggests presence of a band of vibrations stretching of CH2 at 2294.7 cm− 1, band of cyclic alcohols (C–OH) at 1087 cm− 1, and vibrations stretching of carbonyl (CO) at 1716 cm− 1 [37]. Nevertheless, the overlay transforms of FAV with excipients demonstrated all primary peaks of FAV such as band developed due to vibrations & stretching of carbonyl bond (CO) at 1716 cm−1, ether (C–O–C) band at 1180 cm−1 and vibrations stretching of amine N–H at 3347 cm−1 & 3212 cm−1. Thus, FTIR study confirmed absence of interaction between favipiravir and the excipients demonstrating compatibility during formulation development.Fig. 5 FTIR transforms of A) Favipiravir B) PLGA C) Poloxamer D) Favipiravir and excipient E) Over lay plot.
Fig. 5
3.3.5 DSC study
DSC thermometry result of PLGA and FAV represented in Fig. 6 . DSC study demonstrated that FAV showed the presence of endotherms at 189.91 °C, implying crystallinity. The melting endotherms of favipiravir has not been observed in nanoparticles, indicating amorphous form [38]. Additionally, DSC study also suggested formation of nanoparticles with encapsulation of drug inside nanoparticles.Fig. 6 DSC thermogram of a) Favipiravir (b) PLGA (c) Favipiravir-PLGA NPs.
Fig. 6
3.3.6 Atomic force microscopy (AFM)
The structure of FAV NPs was evaluated by AFM (Fig. 7 ). The nanoparticle exhibits spherical morphology which is in corroboration with SEM studies. Additionally, the AFM technique showed that the size of the NPs was lesser than that detected by the DLS system. The possible reason could be due to the procedural differences in measurement of size. Singh et al. mentioned that the DLS method assess the hydrodynamic particle size, which was observed every time bigger than the real size [39]. Kurtosis measures the sharpness of the peak [40]. When the skewness value ranges in 0.5–1 then it is considered skewed. If the value lies in ranges of −0.5 to 0.5 specifies symmetrical distribution. The skewness and kurtosis values were found to be 0.966 and 6.49, respectively which showed that the values are in acceptable range.Fig. 7 AFM studies of favipiravir nanoparticles.
Fig. 7
3.3.7 XRD study
X-ray diffractogram is the fingerprint region which was observed due to peculiar crystalline structures of compounds. Fig. 8 illustrates X-ray diffractogram of FAV and FAV-NPs. The characteristics sharp and intense peak of favipiravir was obsereved by XRD. Whereas, the XRD diffractograms of nanoparticles showed less intensity of FAV suggesting amporphization of drug which is in coroborration with DSC studies [38].Fig. 8 XRD graph of a) Favipiravir and b) Favipiravir loaded nanoparticles.
Fig. 8
3.3.8 Stability studies
The preparations were examined for appearance, particle size, and % entrapment efficiency. At the, end of, six months’ nanoparticles were found to be at stable at 40 °C and 75% RH. Table 5 represents the stability studies of NPs with their % entrapment efficiency and particle size. The stability studies confirmed that NPs were found stable over six months indicating good stability.Table 5 Accelerated stability studies.
Table 5Months % Entrapment efficiency Particle size (nm)
Initial 78.12 ± 0.15% 175±1 nm
1 78.12 ± 0.25% 176±2 nm
2 77.92 ± 0.1% 179±1 nm
3 76.50 ± 0.3% 179±1 nm
4 75.12 ± 0.1% 180±1 nm
5 75.11 ± 0.11% 181±1 nm
6 75.00 ± 0.15% 182±2 nm
3.4 Evaluation of FAV-NPs entrapped gel
3.4.1 Transparency and pH
The nasal gel was found transparent and had no additional particles. Gel showed clarity, good texture, and pH (5.5–6.1) suggesting suitability for delivery by nasal route.
3.4.2 Gelation temperature
The nasal area has temperature ranging 35–37 °C. The sol initiated to convert into gel at 37 °C, suggesting that gel has appropriate gelation temperature required for administration via nasal route. The combination of poloxamer polymers was responsible for formation of transparent gel. It was observed that as the Poloxamer 407 amount rise, the gelation-temperature slowly decreases. However, if the Poloxamer 188 concentration enhanced, the temperature was also increased. Thus, Poloxamer 407 and 188 were incorporated combinely, to control the gelation temperature that provided appropriate mucoadhesion and gelation [18].
3.4.3 Viscosity measurement
The gel viscosity was observed in range of 34592–4568 cps with the torque lies in range of 62–85.5% from 10 to 100 rpm. It was observed that HPMC, poloxamer 407, and poloxamer 188 concentrations contributed in maintaining the gel viscosity. The formation of 3D matrix occurred due to crosslinking of polymers which helps in maintaining the viscosity of gel. Poloxamers and HPMC usually contains hydrogen bonds and electrostatic interactions which aids in viscosity maintenance and mucoadhesion properties [41].
3.4.4 Texture profile
The texture profiling is an imperative property in development of nasal gel. The texture analysis of thermoreversible gel was shown in Fig. 9 and elucidated hardness at 16 g, adhesiveness of 1.01 mJ, gumminess of 14 g, cohesiveness of 2.34, and stringiness of 1.46 mm at 37 °C. Hardness and adhesiveness helps in finding out ease in application and detachment of gel from the surface. The gel strength of the gel was found 1.2 ± 0.3 mJ suggesting good strength. Gel strength is explained as capacity of sol to convert into intact gel [42]. The HPMC and poloxamers combinely contributed in providing strength to gel and adhesive property. Gel strength evidenced that the gel is firm, thus, it could resist the nasal mucociliary clearance.Fig. 9 Texture profiling of thermoreversible gel.
Fig. 9
3.4.5 Mucoadhesion study
Mucoadhesion strength helps to evaluate the strength and the impact of polymers over the biological surface, the mucoadhesion is the amount of adhesion on mucosal surface that the polymer can show (by binding onto the mucosal surface), at a bodily temperature. In this experiment the contact time of the polymer to the nasal membrane with respect to time was measured. If the contact time decreases, then the mucoadhesion strength of the polymer is also less [43]. The mucoadhesion of the thermoreversible gel was found out to be 9373.9 dyne/cm2 indicating good adhesion on mucosa. Therefore, in the nasal cavity, the NPs dispersed gel can hold for an elongated period resulting increase in permeation and FAV absorption.
3.4.6 Dissolution studies
Sample and separate technique was utilized to find out release pattern of FAV from formulations. The study revealed that favipiravir solution showed 90% drug release in 4h. However, nanoparticles showed a sustained release of favipiravir for more than 20h. The NPs dispersed gel showed 79% of FAV release up to 32h (Fig. 10 ). Nanoparticles revealed a prolonged release of FAV because of polymeric matrix. The entrapment of FAV inside PLGA matrices decelerated release rate that could be controlled by diffusion and erosion of polymer membrane. These results were in agreement with prior studies [44,45]. Nevertheless, FAV-NPs dispersed gel showed more prolonged release than polymeric nanoparticles due to the presence of dual barrier of swelled hydrogel phase and polymer [44,46].Fig. 10 Dissolution study of Favipiravir solution, nanoparticles and nanoparticles dispersed gel.
Fig. 10
The release kinetics of preparations were calculated by release models. Table 6 exemplified calculated coefficient of correlation (R2). The FAV solution demonstrated First order release pattern (R2 = 0.9704). Nonetheless, release data of nanoparticles and nanoparticles distributed in gel fitted well in Koresmeyer-Peppa's model (R2 = 0.9771, 0.9578). It suggests that the favipiravir discharge by dual mechanism from NPs i.e. slow diffusion through PLGA mesh and destruction of nanoparticles [47].Table 6 Drug release kinetics.
Table 6 R2 Value
Description Zero Order First Order Higuchi Plot Korsmeyer- Peppa Plot
Favipiravir solution 0.8554 0.9704 0.8979 0.9473
Favipiravir nanoparticles 0.9669 0.7606 0.9647 0.9771
Favipiravir nanoparticles dispersed in Gel 0.9472 0.9089 0.8923 0.9578
3.4.7 Ex vivo permeation study
Permeation study across nasal mucosa of goat suggested higher permeation of FAV nanoparticles distributed in gel as compared to nanoparticles dispersion and drug solution (P < 0.05) (Fig. 11 ). It was indicated by the permeability coefficient (P), flux enhancement ratio, and steady-state flux (fss) (Table 7 ).Fig. 11 Ex vivo permeation studies of favipiravir from solution, nanoparticles, and nanoparticles loaded gel.
Fig. 11
Table 7 Ex vivo permeation study of formulations.
Table 7 FAV solution FAV Nanoparticles FAV Nanoparticles entrapped gel
Flux ss (μg. Cm−2.h−1) 2.011 ± 0.16 5.01 ± 0.31 9.41 ± 0.93
Enhancement ratio – 3.08 ± 0.31 4.6 ± 0.5
Permeability coefficient (cm/h) 0.00287 0.0071 0.01344
The NPs dispersed gel displayed 4.5 times greater permeation than solution and 1.8 times than nanoparticles. The significantly high ex vivo penetration of FAV-NPs distributed gel through the nasal mucosa might be due to the adherence of NPs dispersed gel on the mucosal sheath which resulted high permeation rate. Moreover, mucoadhesive polymer HPMC also contributed in elongated interaction time of formulation and nasal mucosal surface resulting improved permeation of FAV [48]. However, solution and nanoparticles have no mucoadhesion and high fluidity. Thus, contact time with mucosal layer is less resulting small permeability of FAV. Literature confirmed that improvement of contact time statistical enhance drug permeation [49]. These results were in agreement with a previous study in which NLC dispersed gel displayed better permeation (1.5 folds) than drug solution [50].
4 Conclusion
The stable nanoparticles with good entrapment efficiency (>70%) and particle size (175.6 nm) were successfully optimized by using box-behkhen design. Nanoparticles demonstrated spherical morphology and smooth surface. In-vitro release study showed a sustained release property of NPs (20h). Nanoparticles dispersed thermoreversible in-situ gel was developed which showed good texture analysis profile. Additionally, gel showed good mucoadhesive properties supporting nasal permeation. The NPs dispersed gel showed prolonged release for more than 30h. Ex vivo permeation through goat's nasal mucosa confirmed that NPs dispersed gel demonstrated higher permeation of FAV than NPs and solution. Therefore, FAV nanoparticles in-situ gel could be a promising treatment for Covid-19 infection via nasal route. This formulation could reduce high dose and dosing frequency of oral FAV and could be a new patient friendly alternative therapy for Covid-19. Nevertheless, in vivo studies are required to understand the efficacy of formulation.
Consent for publication
I give my consent to publish this research article in this journal.
Funding for publication
Not applicable.
CRedit author statement
Conceptualization: Dr. Shilpa Dawre.
Methodology: Dr. Shilpa Dawre and Vaishnavi Gattani.
Validation: Dr. Shilpa Dawre.
Investigation: Vaishnavi Gattani.
Data Curation: Dr. Shilpa Dawre and Vaishnavi Gattani.
Writing: Dr. Shilpa Dawre and Vaishnavi Gattani.
Supervision: Dr. Shilpa Dawre.
Declaration of competing interest
The authors declare that they have no conflict of interest.
Data availability
Data will be made available on request.
Acknowledgements
Authors want to acknowledge R.C Patel institute Shirpur for the XRD analysis.
==== Refs
References
1 Dawre S. Maru S. Human respiratory viral infections: current status and future prospects of nanotechnology-based approaches for prophylaxis and treatment Life Sci [Internet] 278 2021 Aug 119561 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0024320521005476
2 Worldometer [Internet] Available from: https://www.worldometers.info/coronavirus/
3 Hu B. Guo H. Zhou P. Shi Z.-L. Characteristics of SARS-CoV-2 and COVID-19 Nat. Rev. Microbiol. [Internet] 19 3 2021 Mar 6 141 54. Available from: https://www.nature.com/articles/s41579-020-00459-7 33024307
4 Guy R.K. DiPaola R.S. Romanelli F. Dutch R.E. Rapid repurposing of drugs for COVID-19 May 22 Science (80-) [Internet] 368 6493 2020 829 30. Available from https://www.science.org/doi/10.1126/science.abb9332
5 Yasir M. Sara U.V.S. Chauhan I. Gaur P.K. Singh A.P. Puri D. Solid lipid nanoparticles for nose to brain delivery of donepezil: formulation, optimization by Box–Behnken design, in vitro and in vivo evaluation. Artif Cells, Nanomedicine Biotechnol [Internet] 2017 Nov 3 1 14 Available from: https://www.tandfonline.com/doi/full/10.1080/21691401.2017.1394872
6 Yasir M. Chauhan I. Zafar A. Verma M. Noorulla K.M. Tura A.J. Buspirone loaded solid lipid nanoparticles for amplification of nose to brain efficacy: formulation development, optimization by Box-Behnken design, in-vitro characterization and in-vivo biological evaluation J. Drug Deliv. Sci. Technol. [Internet] 61 2021 Feb 102164 Available from: https://linkinghub.elsevier.com/retrieve/pii/S1773224720314532
7 Yasir M. Zafar A. Noorulla K.M. Tura A.J. Sara U.V.S. Panjwani D. Nose to brain delivery of donepezil through surface modified NLCs: formulation development, optimization, and brain targeting study J. Drug Deliv. Sci. Technol. [Internet] 75 2022 Sep 103631 Available from: https://linkinghub.elsevier.com/retrieve/pii/S1773224722005421
8 Abd Elkodous M. Olojede S.O. Morsi M. El-Sayyad G.S. Nanomaterial-based drug delivery systems as promising carriers for patients with COVID-19 RSC Adv. [Internet 11 43 2021 26463 26480 Available from: http://xlink.rsc.org/?DOI=D1RA04835J 35480012
9 Schoenmaker L. Witzigmann D. Kulkarni J.A. Verbeke R. Kersten G. Jiskoot W. mRNA-lipid nanoparticle COVID-19 vaccines: structure and stability Int. J. Pharm. [Internet] 601 2021 May 120586 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0378517321003914
10 Surnar B. Kamran M.Z. Shah A.S. Dhar S. Clinically approved antiviral drug in an orally administrable nanoparticle for COVID-19 ACS Pharmacol. Transl. Sci. [Internet 3 6 2020 Dec 11 1371 80. Available from: https://pubs.acs.org/doi/10.1021/acsptsci.0c00179 33330844
11 Ucar B. Acar T. Arayici P.P. Derman S. A nanotechnological approach in the current therapy of COVID-19: model drug oseltamivir-phosphate loaded PLGA nanoparticles targeted with spike protein binder peptide of SARS-CoV-2 Nanotechnology [Internet] 32 48 2021 Nov 26 485601 Available from: https://iopscience.iop.org/article/10.1088/1361-6528/ac1c22
12 Wu J. Wang H. Li B. Structure-aided ACEI-capped remdesivir-loaded novel PLGA nanoparticles: toward a computational simulation design for anti-SARS-CoV-2 therapy Phys. Chem. Chem. Phys. [Internet 22 48 2020 28434 9. Available from: http://xlink.rsc.org/?DOI=D0CP04389C
13 Joshi S. Parkar J. Ansari A. Vora A. Talwar D. Tiwaskar M. Role of favipiravir in the treatment of COVID-19 Int. J. Infect. Dis. [Internet] 102 2021 Jan 501 508 Available from: https://linkinghub.elsevier.com/retrieve/pii/S1201971220322736 33130203
14 Agrawal U. Raju R. Udwadia Z.F. Favipiravir: a new and emerging antiviral option in COVID-19 Med. J. Armed Forces India [Internet] 76 4 2020 Oct 370 376 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0377123720301581 32895599
15 Saha T. Quiñones-Mateu M.E. Das S.C. Inhaled therapy for COVID-19: considerations of drugs, formulations and devices Int. J. Pharm. [Internet] 624 2022 Aug 122042 Available from: https://linkinghub.elsevier.com/retrieve/pii/S037851732200597X
16 clinical trials [Internet]. Available from: https://clinicaltrials.gov/ct2/show/NCT04539262
17 Herdiana Y. Wathoni N. Shamsuddin S. Muchtaridi M. Scale-up polymeric-based nanoparticles drug delivery systems: development and challenges OpenNano [Internet] 7 2022 Jul 100048 Available from: https://linkinghub.elsevier.com/retrieve/pii/S2352952022000111
18 [Internet] Wang Y. Jiang S. Wang H. Bie H. A mucoadhesive, thermoreversible in situ nasal gel of geniposide for neurodegenerative diseases 12(12) Xu B. PLoS One 2017 Dec 14 e0189478 Available from: https://dx.plos.org/10.1371/journal.pone.0189478
19 Giannouli M. Karagkiozaki V. Pappa F. Moutsios I. Gravalidis C. Logothetidis S. Fabrication of quercetin-loaded PLGA nanoparticles via electrohydrodynamic atomization for cardiovascular disease 6005 Mater. Today Proc. [Internet 5 8 2018 15998 Available from: https://linkinghub.elsevier.com/retrieve/pii/S2214785318309155
20 Ochi M. Wan B. Bao Q. Burgess D.J. Influence of PLGA molecular weight distribution on leuprolide release from microspheres Int. J. Pharm. [Internet] 599 2021 Apr 120450 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0378517321002556
21 Chorny M. Fishbein I. Danenberg H.D. Golomb G. Lipophilic drug loaded nanospheres prepared by nanoprecipitation: effect of formulation variables on size, drug recovery and release kinetics J. Control Release [Internet] 83 3 2002 Oct 389 400 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0168365902002110 12387947
22 Haider M. Elsherbeny A. Jagal J. Hubatová-Vacková A. Saad Ahmed I. Optimization and evaluation of poly(lactide-co-glycolide) nanoparticles for enhanced cellular uptake and efficacy of paclitaxel in the treatment of head and neck cancer Pharmaceutics [Internet] 12 9 2020 Aug 30 828 Available from: https://www.mdpi.com/1999-4923/12/9/828 32872639
23 Jokicevic K. Kiekens S. Byl E. De Boeck I. Cauwenberghs E. Lebeer S. Probiotic nasal spray development by spray drying Eur. J. Pharm. Biopharm. [Internet] 159 2021 Feb 211 220 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0939641120303416 33238191
24 Huang W. Zhang C. Tuning the size of poly(lactic-co-glycolic acid) (PLGA) nanoparticles fabricated by nanoprecipitation Biotechnol. J. [Internet] 13 1 2018 Jan 1700203 Available from: https://onlinelibrary.wiley.com/doi/10.1002/biot.201700203
25 Cortés H. Hernández-Parra H. Bernal-Chávez S.A. Prado-Audelo ML Del Caballero-Florán I.H. Borbolla-Jiménez F.V. Non-ionic surfactants for stabilization of polymeric nanoparticles for biomedical uses Materials (Basel) [Internet] 14 12 2021 Jun 10 3197 Available from: https://www.mdpi.com/1996-1944/14/12/3197 34200640
26 Sharma D. Maheshwari D. Philip G. Rana R. Bhatia S. Singh M. Formulation and optimization of polymeric nanoparticles for intranasal delivery of lorazepam using box-behnken design: in vitro and in vivo evaluation Biomed. Res. Int. [Internet] 2014 2014 1 14 Available from: http://www.hindawi.com/journals/bmri/2014/156010/
27 Sakhi M. Khan A. Iqbal Z. Khan I. Raza A. Ullah A. Design and characterization of paclitaxel-loaded polymeric nanoparticles decorated with Trastuzumab for the effective treatment of breast cancer Front. Pharmacol. [Internet] 2022 Mar 14;13. Available from: https://www.frontiersin.org/articles/10.3389/fphar.2022.855294/full
28 Pradhan R. Poudel B.K. Ramasamy T. Choi H.-G. Yong C.S. Kim J.O. Docetaxel-loaded polylactic acid-Co-glycolic acid nanoparticles: formulation, physicochemical characterization and cytotoxicity studies J Nanosci Nanotechnol [Internet 13 8 2013 Aug 1 5948 5956 Available from: http://openurl.ingenta.com/content/xref?genre=article&issn=1533-4880&volume=13&issue=8&spage=5948 23882865
29 Draheim C. de Crécy F. Hansen S. Collnot E.-M. Lehr C.-M. A design of experiment study of nanoprecipitation and nano spray drying as processes to prepare PLGA nano- and microparticles with defined sizes and size distributions Pharm. Res. [Internet] 2015 Feb 13 Available from: http://link.springer.com/10.1007/s11095-015-1647-9
30 Snejdrova E. Loskot J. Martiska J. Soukup T. Prokes L. Frolov V. Rifampicin-loaded PLGA nanoparticles for local treatment of musculoskeletal infections: formulation and characterization J. Drug Deliv. Sci. Technol. [Internet] 73 2022 Jul 103435 Available from: https://linkinghub.elsevier.com/retrieve/pii/S1773224722003458
31 Girotra P. Singh S.K. Kumar G. Development of zolmitriptan loaded PLGA/poloxamer nanoparticles for migraine using quality by design approach Int J Biol Macromol [Internet] 85 2016 Apr 92 101 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0141813015302567 26724690
32 Clogston J.D. Patri A.K. Zeta Potential Measurement 2011 63 70 Available from: http://link.springer.com/10.1007/978-1-60327-198-1_6
33 Chiu H.I. Samad N.A. Fang L. Lim V. Cytotoxicity of targeted PLGA nanoparticles: a systematic review RSC Adv. [Internet 11 16 2021 9433 9449 Available from: http://xlink.rsc.org/?DOI=D1RA00074H 35423427
34 Cui Y. Xu Q. Chow P.K.-H. Wang D. Wang C.-H. Transferrin-conjugated magnetic silica PLGA nanoparticles loaded with doxorubicin and paclitaxel for brain glioma treatment Biomaterials [Internet] 34 33 2013 Nov 8511 8520 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0142961213008892 23932498
35 Patil G.B. Surana S.J. Fabrication and statistical optimization of surface engineered PLGA nanoparticles for naso-brain delivery of ropinirole hydrochloride: in-vitro–ex-vivo studies [Internet] J. Biomater. Sci. Polym. 24 15 2013 Oct 24 1740 56. Available from https://www.tandfonline.com/doi/full/10.1080/09205063.2013.798880
36 Tulbah A.S. Lee W.-H. Physicochemical characteristics and in vitro toxicity/anti-SARS-CoV-2 activity of favipiravir solid lipid nanoparticles (SLNs) Pharmaceuticals [Internet] 14 10 2021 Oct 19 1059 Available from: https://www.mdpi.com/1424-8247/14/10/1059 34681283
37 Lu B. Lv X. Le Y. Chitosan-modified PLGA nanoparticles for control-released drug delivery Polymers (Basel) [Internet] 11 2 2019 Feb 12 304 Available from: https://www.mdpi.com/2073-4360/11/2/304 30960288
38 Dawre S. Devarajan P.V. Samad A. Enhanced antibacterial activity of doxycycline and rifampicin combination loaded in nanoparticles against intracellular Brucella abortus Curr. Drug. Deliv. [Internet] 19 1 2022 Jan 104 116 Available from: https://www.eurekaselect.com/193997/article 34151761
39 Singh S.K. Dadhania P. Vuddanda P.R. Jain A. Velaga S. Singh S. Intranasal delivery of asenapine loaded nanostructured lipid carriers: formulation, characterization, pharmacokinetic and behavioural assessment RSC Adv. [Internet 6 3 2016 2032 45. Available from: http://xlink.rsc.org/?DOI=C5RA19793G
40 Bellotti R. Picotto G.B. Ribotta L. AFM measurements and tip characterization of nanoparticles with different shapes Nanomanufacturing Metrol 5 2 2022 127 138
41 Mura P. Mennini N. Nativi C. Richichi B. In situ mucoadhesive-thermosensitive liposomal gel as a novel vehicle for nasal extended delivery of opiorphin Eur. J Pharm. Biopharm. [Internet] 122 2018 Jan 54 61 Available from: https://lininghub.elsevier.com/retrieve/pii/S0939641117307026 29032194
42 Dawre S. Pathak S. Sharma S. Devarajan P.V. Enhanced antimalalarial activity of a prolonged release in situ gel of arteether–lumefantrine in a murine model Eur. J. Pharm. Biopharm. [Internet 123 2018 Feb 95 107 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0939641117306215 29122736
43 Leung S.-H.S. Robinson J.R. Polymer structure features contributing to mucoadhesion. II J. Control Release [Internet] 12 3 1990 May 187 194 Available from: https://linkinghub.elsevier.com/retrieve/pii/016836599090099F
44 Kim D.-H. Nguyen T.N. Han Y.-M. Tran P. Rho J. Lee J.-Y. Local drug delivery using poly(lactic-co-glycolic acid) nanoparticles in thermosensitive gels for inner ear disease treatment 77 Drug Deliv. [Internet]. 28 1 2021 Jan 1 2268 Available from: https://www.tandfonline.com/doi/full/10.1080/10717544.2021.1992041 34668836
45 Kim D.-H. Martin D.C. Sustained release of dexamethasone from hydrophilic matrices using PLGA nanoparticles for neural drug delivery Biomaterials [Internet] 27 15 2006 May 3031 3037 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0142961206000044 16443270
46 Dawre S. Waghela S. Saraogi G. Statistically designed vitamin D3 Encapsulated PLGA microspheres dispersed in thermoresponsive in-situ gel for nasal delivery J. Drug Deliv. Sci. Technol. [Internet] 75 2022 Sep 103688 Available from: https://linkinghub.elsevier.com/retrieve/pii/S1773224722005998
47 Bohrey S. Chourasiya V. Pandey A. Polymeric nanoparticles containing diazepam: preparation, optimization, characterization, in-vitro drug release and release kinetic study Nano. Converg. [Internet] 3 1 2016 Dec 1 3 Available from: http://www.nanoconvergencejournal.com/content/3/1/3 28191413
48 Jain S.A. Chauk D.S. Mahajan H.S. Tekade A.R. Gattani S.G. Formulation and evaluation of nasal mucoadhesive microspheres of Sumatriptan succinate J. Microencapsul. [Internet 26 8 2009 Dec 5 711 721 Available from: http://www.tandfonline.com/doi/full/10.3109/02652040802685241 19888880
49 Pund S. Rasve G. Borade G. Ex vivo permeation characteristics of venlafaxine through sheep nasal mucosa Eur. J. Pharm. Sci [Internet] 48 1–2 2013 Jan 195 201 Available from: https://linkinghub.elsevier.com/retrieve/pii/S0928098712004320 23159662
50 Deshkar S.S. Jadhav M.S. Shirolkar S.V. Development of carbamazepine nanostructured lipid carrier loaded thermosensitive gel for intranasal delivery Adv. Pharm. Bull [Internet] 11 1 2020 Nov 7 150 162 Available from: https://apb.tbzmed.ac.ir/Article/apb-28382 33747862
| 0 | PMC9745979 | NO-CC CODE | 2022-12-15 00:04:04 | no | J Drug Deliv Sci Technol. 2022 Dec 13;:104082 | utf-8 | J Drug Deliv Sci Technol | 2,022 | 10.1016/j.jddst.2022.104082 | oa_other |
==== Front
Inf Sci (N Y)
Inf Sci (N Y)
Information Sciences
0020-0255
1872-6291
Elsevier Inc.
S0020-0255(22)01511-0
10.1016/j.ins.2022.12.017
Article
MIC-Net: A Deep Network for Cross-site Segmentation of COVID-19 Infection in the Fog-assisted IoMT
Ding Weiping ab⁎
Abdel-Basset Mohamed c
Hawash Hossam c
Pedrycz Witold d
a School of Information Science and Technology, Nantong University, Nantong, China
b Faculty of Data Science, City University of Macau, Macau, China
c DEEPOLOGY LAB, Zagazig University, Zagazig, Egypt
d Electrical and Computer Engineering, University of Alberta, Edmonton, AB T6R 2V4, Canada
⁎ Corresponding author.
13 12 2022
13 12 2022
22 3 2022
2 12 2022
7 12 2022
© 2022 Elsevier Inc. All rights reserved.
2022
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The automatic segmentation of COVID-19 pneumonia from a computerized tomography (CT) scan has become a major interest for scholars in developing a powerful diagnostic framework in the Internet of Medical Things (IoMT). Federated deep learning (FDL) is considered a promising approach for efficient and cooperative training from multi-institutional image data. However, the nonindependent and identically distributed (Non-IID) data from health care remain a remarkable challenge, limiting the applicability of FDL in the real world. The variability in features incurred by different scanning protocols, scanners, or acquisition parameters produces the learning drift phenomena during the training, which impairs both the training speed and segmentation performance of the model. This paper proposes a novel FDL approach for reliable and efficient multi-institutional COVID-19 segmentation, called MIC-Net. MIC-Net consists of three main building modules: the down-sampler, context enrichment (CE) module, and up-sampler. The down-sampler was designed to effectively learn both local and global representations from input CT scans by combining the advantages of lightweight convolutional and attention modules. The contextual enrichment (CE) module is introduced to enable the network to capture the contextual representation that can be later exploited to enrich the semantic knowledge of the up-sampler through skip connections. To further tackle the inter-site heterogeneity within the model, the approach uses an adaptive and switchable normalization (ASN) to adaptively choose the best normalization strategy according to the underlying data. A novel federated periodic selection protocol (FED-PCS) is proposed to fairly select the training participants according to their resource state, data quality, and loss of a local model. The results of an experimental evaluation of MIC-Net on three publicly available data sets show its robust performance, with an average dice score of 88.90% and an average surface dice of 87.53%.
Keywords
Internet of Medical Things
Fog Computing
Deep Learning
Data Heterogeneity
COVID-19
==== Body
pmc1 Introduction
The recent COVID-19 pandemic resulting from the spread of the novel coronavirus SARS-CoV-2 has a devastating effect on the health infrastructure of many countries. Thus, there is an urgent need for time-sensitive, precise, fast, and simple technologies for efficient diagnosis in modern maritime transportation systems (MTSs). In response to these requirements, the Internet of Medical Things (IoMT) has been developed as an instantiation of the Internet of Things (IoT) to interconnect a diversity of medical and health-care platforms. The IoMT, therefore, combines a large number of computational units and IoT devices in hospitals and clinical processes to facilitate clinical diagnosis and the monitoring of patient data. The major aim of IoMT services is to computerize diagnosis tasks using intelligent techniques that can process and extract insights from medical records (e.g., medical images). Hence, developing such a service for COVID-19 diagnosis is of great importance and can significantly advance the medical diagnosis process. Based on recent studies, the diagnosis of COVID-19 should be established using the reverse transcription–polymerase chain reaction (RT-PCR) [7]. However, owing to the practical problems in sample aggregations and transportation, and the variations in testing kits, specifically at the time of the outbreak, RT-PCR testing has a high false-negative ratio. As an alternative, computerized tomography (CT) scans are a reliable tool for the diagnosis of COVID-19 infection in the medical domain.
Infection segmentation is, therefore, a primary phase of diagnosis, infection quantification, infection progression monitoring, and severity assessment. A wide variety of artificial intelligence (AI) solutions have been investigated for infection segmentation from medical images, such as fuzzy intelligence, machine learning, and mathematical modeling [27], [39]. However, given the inherent complexity of infection representations and the high volume and high-dimensionality of medical data, these approaches cannot achieve good performance. Deep learning (DL), as a subfield of machine learning, has overcome the limitations of the above methods and achieved excellent segmentation results from a different form of data, irrespective of the size and dimensionality of data. Motivated by these results, this work emphasizes studying DL to address COVID-19 segmentation in the real IoMT.
The grouping of communicating medical apparatus, devices, and applications is termed the IoMT. It includes instantaneous data, the remote nursing of patients, the situational data of patients, and diagnostic decision-making according to aggregated information. The key advantage of the online interconnections between medical appliances in the IoMT is the reduction of the time needed by specialists through offering immediate analysis of a patient’s circumstances; moreover, patient assessment can be offered at home, in a clinic, or in a hospital. However, the aggregation of a high volume of medical data is laborious and time-consuming, making it a huge problem in the IoMT environment. Moreover, relying on a single source of data limits the development of an efficient DL model [8]. Thus, it is essential to share the high volume of training data dispersed among several medical organizations. With the advancements in cloud computing of ample storage space and unrestrained computing resources, the DL approaches applied to medical data have been broadly investigated by developing IoMT applications based on cloud computing. Nevertheless, the continuous increase in the number of connections between the cloud platform and mobile individuals causes unwanted and untimely responses to user requests, along with broadcast latency. A late response to diagnostic requests can impact the health and lives of patients, particularly patients suffering from infectious or serious diseases. As a remedy, the paradigms of fog/edge computing have been introduced to provide resources positioned closer to clients and to reduce latency by making computation adjacent to clients. These computing paradigms have received increasing attention across a variety of IoT-enabled applications and have shown great potential in delivering the required response. These evolving paradigms afford edge devices, including a variety of storage space, processing, and communication technologies, efficient network bandwidth, scalability, privacy preservation, mobility, security, and short latency, thus providing a better option for time-sensitive applications.
The introduction of edge and fog computing has shifted the attention of scholars to the distributed training of DL solutions on geographically distributed devices. In this regard, federated learning (FL) was proposed to enable distributed entities to cooperatively train complex deep networks using their locally stored data by communicating and aggregating the local training parameters in a regular fashion. Contrasted with the traditional cloud-centralized learning strategies, FL solution has a design that can bring many benefits to the IoMT environment, including efficient network bandwidth, privacy preservation, and communication efficiency (in terms of low overhead and low latency). Recently, the field of medical image analysis has witnessed an active shift toward FL in a broad range of tasks, such as federated segmentation, disease prediction, image reconstruction, classification, and computational pathology. However, the research has overlooked many important design aspects, such as complexity, communication efficiency, and heterogeneity, which limit their use in real-world health-care applications.
1.1 Research Gaps
Although much effort has been devoted to developing precise and reliable segmentation of COVID-19 lesions, multiple research gaps have still been uncovered in the literature. These gaps can be concisely described as follows:(1) Complex manifestations: Owing to the low contrast problem, the border between COVID-19 infection lesions and the adjacent normal tissues is frequently hazy. As a result, precise lung infection segmentation is severely hampered. Moreover, COVID-19 lung infection displays a wide range of morphological characteristics, such as size and shape, which adds to the challenge of precise segmentation.
(2) Lung CT scans from various domains introduce superficial domain heterogeneity owing to changes in imaging techniques, scanning devices, and demography. Because better DL network performance often requires a well-standardized distribution of data, the discrepancies inherent in heterogeneous domain data can cause problems when compounding examples from various sources during the training.
(3) Robust segmentation performance always requires complex deep networks, which makes the training and inferencing process computationally intensive, thus limiting the responsiveness of medical image analysis and diagnosis. This phenomenon represents a great barrier to applying and deploying segmentation models in real-world distributed IoMT systems.
(4) The existing research has proved that the data heterogeneity of participants leads to significant drift in local and global training optimizations, slowing the convergence and reducing the stability of training, as the local training is optimized according to the local objective rather than the global one.
1.2 Main Contributions
This work proposes a new multi-institutional COVID-19 segmentation framework called MIC-Net to fill in the abovementioned gaps. The proposed MIC-Net contributes to the body of knowledge as follows:(1) A novel lightweight DL technique called MIC-Net is presented for fine-tuning lung infection segmentation by exploiting heterogeneous-source CT scans aggregated in an IoMT environment. MIC-Net is integrated into a domain-adaptive FL scheme to further support the applicability of real-world distributed IoMT solutions.
(2) The down-sampling module of MIC-Net is innovatively designed to effectively capture lesion features from cross-domain CT images, whereby inter-source data heterogeneity is further addressed via a new switchable domain-adaptive normalization layer.
(3) In MIC-Net, a novel contextual enrichment (CE) module is introduced to empower the network to capture the combinations of static and dynamic contextual representations using a lightweight attention mechanism. The mined information CE module is later used to enrich the up-sampling path so as to accurately reconstruct the segmentation maps.
(4) A novel federated periodic client selection protocol (FED-PCS) is proposed to select the training participants fairly and periodically according to multiple criteria including the resource state (time, CPU, GPU, Memory, bandwidth), data quality, and loss of a local model.
(5) The results of experimental simulations on real-world cross-domain COVID-19 data validate the increased efficiency and effectiveness of the proposed MIC-Net over state-of-the-art approaches in segmenting infection lesions, making the method better suited for deployment in real-world health-care systems.
1.3 Paper Organization
The rest of the article is organized as follows. Section II reviews the recent studies relevant to this work. Section III details the system design. Then, a detailed explanation of the methodology of MIC-Net is presented in Section IV. Next, Section V describes the experimental specifications of this work. Section VI discusses and analyzes the numerical results of our experiments. Finally, Section 6 concludes this work and lays out promising future directions.
2 Related Work
This section starts by reviewing the latest research relevant to the subject matter of this work to better position the current work. The discussion of the research is decomposed into three main subsections, as described below:
2.1 COVID-19 Segmentation
There have many been research efforts to use deep learning to improve the segmentation of COVID-19 by locating and segmenting the infection region in either two-dimensional (2D) or volumetric CT scans. For example, Fan et al. [6] developed a semi-supervised segmentation approach that utilizes a parallel partial decoder to fuse the complex features and produce a global map that is subsequently fed into attention modules to detect infection boundaries and improve representations. However, the reliance on generated pseudo labels limited the performance. In [9], the authors presented a dual-branch deep learning framework to concurrently accomplish two tasks of COVID-19 diagnosis: patient-based classification and lesion segmentation. In this work, a lesion attention mechanism was proposed to integrate the segmentation outcomes so as to guide the classification of the infection lesions. Moreover, one study [44] presented a convolutional network for COVID-19 segmentation that integrates a self-attention layer to extend the receptive field so as to improve the representational power by distilling valuable semantic patterns from deeper layers, with no additional training time. It also presents semi-supervised shots to ease the deficiency of annotated multi-class data and the unbalanced training data by adopting a re-weighting of the loss to precisely categorize different types of labels. Furthermore, a study [26] developed a deep segmentation network named CovTANet that integrates tri-level attention techniques to combine channel, spatial, and pixel information for precise segmentation of COVID lesions from CT scans using multistage optimization methods. In [41], the authors developed a noise-robust dice loss as a generalized variant of dice loss and mean absolute error (MAE) loss to make the segmentation network more robust against noisy COVID-19 originating from different scales and presences. They also designed a segmentation network called COPLE-Net. It is integrated into an adaptive self-ensembling architecture in which the exponential moving average (EMA) of a student network is leveraged as a teacher network that performs an adjustive update by overpowering the involvement of the student to the EMA at the time the student has a large, noise-robust dice loss. Moreover, the authors of [4] proposed a segmentation framework that first integrates the patch mechanism to extract the region of interest so as to eliminate unrelated background regions and then applies a three-dimensional (3D) network to capture spatial features from important target areas. They applied a combinatorial loss to accelerate the training convergence. Speaking of 3D networks, the work [43] presented a joint learning framework that combines classification and volumetric segmentation of COVID-19 from a large-scale CT database, whereby a 3D lesion subnetwork is applied for lesion segmentation and another subnetwork, while task-aware loss was developed to regularize the interaction between two subnetworks during the training. The most common aspect of the studies described above for COVID-19 diagnosis is that all models were trained and evaluated on homogeneous (single-source) data. Consequently, their performance degraded on unseen data from another source. Such data heterogeneity is prevalent in the IoMT environment. Additionally, all approaches exhibited high computational complexity, making them unsuitable for deployment in an IoMT environment.
2.2 Cross-Domain Medical Image Analysis
The heterogeneity of domains in medical images has gained increasing interest in recent years owing to its considerable consequences for the efficiency of DL models. For example, in [17], the authors proposed a two-stage transfer learning approach called nCoVSegNet for efficient cross-domain segmentation of COVID-19 lesions from CT scans. nCoVSegNet was developed to extract multilevel features from the received CT scans using global context-aware modules. To improve the segmentation, nCoVSegNet adopts two-stage transfer learning methods: One uses a network trained on ImageNet, while the other conducts transfer learning based on a large public CT data set. Besides, in [19], the authors proposed a DL framework called MS-Net for enhancing the accuracy of prostate segmentation from multidomain magnetic resonance images (MRIs) by compensating for the intersite data heterogeneity via domain-related batch normalization (BN) operations. The MS-Net presents knowledge transfer to direct joint convolutional kernels to learn powerful representations from cross-domain MRIs. Similarly, the work [45] proposed explicitly addressing the problem of domain shift in cross-site COVID-19 classification by performing distinct feature normalization and applying contrastive training loss to improve the domain invariance of semantic embeddings. The aim is to improve the classification performance across different domains. Moreover, the work [42] presented a cycle-consistent framework for efficient semantic segmentation from cross domains in the medical field, which integrates consistency regularization and online diverse image translation into a single framework to promote modeling complicated relationships between domains such as many-to-many mappings. In [48], the authors presented a wholly automated machine-agnostic approach for segmenting and quantifying COVID-19 lesions from multisource CT scans. They first designed a CT scan simulator by fitting the dynamic alteration of actual patients' information captured at various time intervals, which significantly eliminates the data shortage problem; then, a deep network was proposed to decompose the volumetric segmentation into three 2D ones, thus improving the model complexity and segmentation efficiency.
2.3 IoMT
Owing to the robust performance of DL in processing and learning from large data incidents in the IoT-based system, several studies have investigated applying DL techniques in the IoMT environment for disease diagnosis. For example, in [33], the authors presented an inclusive survey of potential IoT-based solutions that can help tackle COVID-19-like pandemics by highlighting the social effects of such pandemics and recognizing the particular gaps in existing IoT systems. The authors of [31] presented an overview of the latest innovations in FL in making smart health care more intelligent and efficient by studying different federated intelligence paradigms such as resource-aware paradigms, privacy-preserved federated intelligence, personalized federated intelligence, and incentive federated intelligence. The authors of [24] proposed to detect kidney disease diagnosis using a novel heterogeneous DL segmentation approach. Furthermore, the authors of [21] discussed multiple potential deep reinforcement learning techniques for identifying and diagnosing lung cancer from CT images, and for precisely detecting tumors. Additionally, the authors of [46] used a convolutional model for the diagnosis of gallbladder stones utilizing aggregated image data in the IoMT environment. Moreover, the authors of [3] developed an integrated IoT framework to support instantaneous communication and identification of emotions from physiological signals using long-short-term memory (LSTM), where real-time health observation and distance learning provide support during epidemics. Moreover, the authors of [2] presented an IoT-empowered approach for the early evaluation of COVID-19 by applying a faster region-based CNN (Faster-RCNN) to detect the disease from chest x-rays. In [12], the authors developed a blockchain-managed FL framework in which a data normalization method is applied to lessen the impact heterogeneity of CT data originating from various hospitals. Then, a capsule network is adopted to perform the classification and segmentation of COVID-19. Blockchain technology enables the collaborative and privacy-preserved training of the model in real-life health care. Similarly, FL was used for semi-supervised training of COVID-19 from multinational CT scans; however, the cloud server takes the responsibility of coordinating the training instead of a blockchain. In [47], the authors designed a dynamic fusion-based FL framework for detecting COVID-19 from medical images through the dynamic selection of the contributing clients based on the performance of their local model while scheduling the model fusion according to the training time of participants.
In sum, although there have been many studies addressing COVID-19 segmentation from CT scans using various DL models, only a few of them have seriously considered data heterogeneity during federated training with multi-institutional data. They have not considered simulating their models physically or virtually in an IoMT context, which limits the applicability in the real-world settings with few resources.
3 System Design
Real-time semantic segmentation is crucial in different IoMT tasks, such as diagnosis, patient follow-up, and detection. Generally, the application of DL requires large storage resources and powerful computing resources. Regrettably, implementing and training the segmentation model on the cloud is frequently subject to high bandwidth utilization, unanticipated latency, poor dependability, and privacy concerns. The introduction of fog and edge computing paradigms has offered a great opportunity to address the above problems by bringing the segmentation models close to the data source in the IoMT (end devices), as a supplement to the cloud.
3.1 System Model
This section provides a detailed description of the system model of the proposed framework, which is composed of three main layers: the edge, fog, and cloud.
3.1.1 Cloud computing layer
In the cloud layer, the cloud server is responsible for coordinating the collaboration between the edge and fog nodes during the distributed training of segmentation models. As the collaborative training between the edge and fog layer may be considerably compromised because of local experience, the cloud layer combines different well-trained local segmentation models to achieve global learning. If the edge cannot reliably offer the service, the cloud can leverage its extensive computing resources and global expertise to help edge nodes update their DL models. The proposed solution exploits the resources of the cloud layer to avoid the problems of an overloaded fog infrastructure, latency, a larger input size, network congestion, and limited scalability. This offers extra robustness and reliability amidst heavy load requests and autonomous computation.
3.1.2 Edge/Fog computing layer
This layer contains different devices that lie in the vicinity of data owners (e.g., medical institutes, hospitals, research labs), which directly store the captured CT scans from COVID-19 patients. The data stored in this layer are usually heterogeneous because of the discrepancy in CT scanners, scanning protocols, and other conditions. Compared to the cloud layer, the edge and fog devices have far fewer resources and less computing power and storage. However, they bring the training of the DL model (i.e., MIC-Net) closer to the source of data by performing collaborative model training using the local CT data.
3.2 Resource allocation
The participation of heterogeneous devices is a common case for applying FL in the real-world IoMT. These devices can vary in the quality of their data sets, the capability of their computations, their energy states, and their opportunity to engage. Because different devices have different energy states and communication bandwidths, resource allocation needs to be optimized so that the learning process can be as effective as possible. This is necessary to make the most of the available resources. Participant selection, joint radio, computation resource management, and incentive mechanisms are common directions for addressing resource allocation in federated training scenarios. To alleviate the training traffic jam, this paper introduces a multi-criteria periodic selection protocol in federated settings. The systematic diagram of this protocol is shown in Fig. 1 , in which the cloud server acts as a coordinator of the collaborative training of the segmentation model on a heterogenous network of IoMT devices. The proposed selection protocol is operated in a periodic fashion rather than a selection per-round manner, whereas the length of the period at which the selection is performed is heuristically determined via the cloud coordinator. As shown, the proposed selection is executed as follows:Fig. 1 Resource allocation in the proposed system model.
Step 1: The cloud server initiates a call for a resource to acquire information about the resources of local participants that can help judge the willingness of heterogeneous health-care devices to participate in federated training.
Step 2: The Profiler of local devices is in charge of monitoring the resource consumption during the local training and generating the resource profile of its devices. The protocol uses the following metadata: time, CPU, GPU, RAM, energy, wireless channel states, bandwidth, and quality of data (e.g., number of samples, number of sources). The collected resource information is exploited to determine the highest probable number of local clients (e.g., hospitals, institutions) that can complete the training process in the preset time intervals for a later global aggregation period. Besides, the inclusion of data as selection criteria helps avoid bias toward devices with better computational resources, and enables the selection of clients hosting representative data from a population distribution.
Step 3: By choosing the highest viable number of participating clients per communication round, the segmentation performance can be easily retained and improved during the training. The cloud can regard this as a maximization problem, where greedy algorithms are widely used to select the clients with the smallest time for model uploads. Generally, the IoMT network is known to be dynamic and ambiguous, with varying conditions such as power conditions and network conditions. Thus, the proposed protocol adopts double deep q-learning (DDQL) [38] to optimize the abovementioned optimization problem, in which the server acts as the agent, the profile of client devices acts as state space, and the number of data and power items act as action space. The collected data, power usage, and training time are adopted as a reward function. Typically, the cloud coordinator seeks to reach the optimal resource allocation that maximizes training speed and simultaneously minimizes the usage of client resources.
Step 4: Given the selected clients, the cloud server starts broadcasting the parameters of the global model to those clients to perform local training and share their local updates to the cloud server so as to aggregate their acquired knowledge from the distributed cross-domain devices.
Unlike previous protocols designed and evaluated for simple deep models [32], the proposed selection protocol is suitable for segmentation models that are relatively more complex than previous ones. Another difference is that our method factors in GPU acceleration resources. Another problem to encounter is fraudulent resource allocation, which means that the FL is overrepresented by the distribution of data held by clients possessing higher computing resources. To address this problem, inspired by previous research [13], fairness of selection is regarded as an extra objective in DDQL, which is actually defined as the difference in the performance of the segmentation model (MIC-Net) across different clients. If the variation in the segmentation performance is high, it implies the existence of high bias or less equality, as the learned model may be extremely precise for particular applicants and less so for other underrepresented participants.
3.3 System Security
Although FL avoids the need to communicate training data with a remote cloud by communicating only the learned parameters, it is still subject to a broad range of attacks, such as model and data poisoning, wherein a malevolent client can deliver inaccurate parameters or degraded models to forge the learning procedure throughout global aggregation. Therefore, the vulnerability of FL to such attacks can corrupt the entire learning system. Moreover, FL still suffers from critical privacy challenges, which can be caused by a malicious client trying to infer sensitive information about clients. Security and privacy are major concerns in the integration of FL in health care; however, these topics are out of the scope of this work. Hence, the design of our system assumes that honest clients participate in training and an honest cloud is involved in the coordination for training. To imitate a real scenario, the security administrator module is settled on the cloud server to manage the security mode of the federated training process and decide the best security and/or privacy-preservation mechanisms in dishonest and semi-honest scenarios.
4 MIC-Net
In this section, we provide a detailed explanation of MIC-Net. For convenience, the architecture is illustrated in Fig. 2 . As a pre-training step for the MIC-Net, an intelligent harmonization mechanism is applied to lessen the impact of distribution shifts in the training data. As shown, the MIC-Net has a U-shaped structure, which is a common design for segmentation networks. In particular, the MIC-Net consists of three main building blocks: the encoding path, decoding path, and contextual enrichment (CE) module. Unlike previous approaches, our method has an encoding path consisting of new down-sampler modules instead of conventional commotional encoders. To avoid losing the contextual representations between the encoding and decoding path, the CE module is integrated to enable the network to capture both static and dynamic representations from multi-distribution data. Similarly, instead of a de-convolutional decoder, the decoding path was designed with new up-sampler modules to precisely reconstruct the segmentation maps from the learning solution. In the following subsections, we detail each building module.Fig. 2 System model of the proposed federated MIC-Net in the IoMT environment.
4.1 Feature Encoding Path (Down-Samplers)
The structure of the down-sampler module is composed of two parallel modules, the lightweight attention encoding (LAE) module and lightweight convolutional encoding (LCE) module. These modules bring two advantages to the encoding: the robust feature extraction capabilities of convolution and the ability to capture global context with attention.
The design of the LEA module is largely inspired by vision transformers [20], albeit with reduced complexity (See Fig. 3 ). Assume that we have X∈RC×H×W as the input map of the LCE block, with C,H,andW denoting the number of channels, height, and width, respectively. The LAE passes input X to a point-wise convolution (C1×1), as follows:(1) x=F0=C1x1X
Fig. 3 Architecture of the down-sampler module
Then, the generated map is passed to the sparse self-attention layer (SSA), whose input is decomposed into three paths: input I, key (K), and value (V). A linear layer is used to map the input (I) to the d-dimensional token into a scalar value. This layer comes with a weight matrix (WI∈Rd) that acts as the latent node (L). The output of this linear projection is a k-dimensional vector representing the calculated distance between input k and L. These k-dimensional vectors are passed to a SoftMax function to generate attention scores (As∈Rk). Unlike standard vision transformers, the SSA calculates the attention score for every token in regard to L rather than all tokens (k). This, in turn, implies linear time computation, as demonstrated by [29]. The context scores (As) are then adopted to calculate a context vector (cv∈Rd). In particular, input x is linearly encoded into a d-dimensional representation through K and weighted by the parameters Wk∈Rdxd to generate an outcome (xk∈Rkxd). Follow, the context vector (cv) is calculated as a weighted combination of xk, as follows:(2) cv=∑i=1kAs(i)xK(i).
Similar to the attention matrix, cv encapsulates the knowledge from all tokens, but in a computationally cheap manner. The encoded knowledge in cv is later combined with all tokens in input x by linearly projecting input x to a d-dimensional representation utilizing V weighted by the parameters WV∈Rdxd, then passed to the ReLU function to generate xv∈Rkxd. The context representations in cv are later transmitted to xv through transmitted Hadamard products. The obtained outcome is later passed to one more linear projection operation weighted by the parameters Wo∈Rdxd to generate the ultimate outcome (y∈Rkxd). In mathematical terms, the SSA can be designated separately, as follows:(3) F1=y=∑σxWI⏞ck∈Rk∗xWK⏟Cv∈Rd∗ReLUxWvWo
where * and ∑ denote the are broadcast-capable Hadamard product and summation functions, respectively. At the end of the LAE, a point-wise convolution is applied, followed by an adaptive and switchable normalization (ASN) layer, as follows:(4) F2=ASNC1×1F1
The ASN is described in more detail later in this section.
Although the factorization of a convolution operation into a point-wise convolution and a depthwise separable convolution (DwS-Conv) can considerably decrease the number of learning parameters, and thereby the computing complexity, it generally results in a performance drop. Motivated by the reality that successful multi-scale learning has a significant role in enhancing the segmentation performance [1], [10], we introduce a lightweight convolutional encoding (LCE) module for efficient and effective multiscale extraction of lesion features using dilated versions of DwS-Conv layers. Each DwS-Conv layer is denoted as CrK×K, where K represents the kernel size and r denotes the dilation rate. For convenience, when r=1, the layer can be denoted as CK×K. Given the aforementioned definitions, the structural design of the LCE block can be illustrated as shown in Fig. 2. Given X∈RC×H×W as the input map of the LCE block, with C,H,andW denoting the number of channels, height, and width, respectively, the LCE block generates the output map FX∈RC′×H′×W′, whereas F designates the function of the transformation of the input. C′,H′,andW′ follow the above definition, but for an output map.
Let us dive into the working F. At an early stage, the number of channels is reduced to C′/P by passing input X to a point-wise convolution (C1×1), in which P is the number of concurrent paths (see Fig. 1). This layer can be mathematically expressed as follows:(5) F′=C1x1X
The output maps (F′) from the above operation are passed to P parallel dilated DwS-Conv, whereby the output of each path is added to the input of the next path such that(6) Fp″=max-poolF′if(p=0)C2p-13×3F′if(p=1)C2p-13×3F′+Fp-1′if(p=2,⋯,P)
whereby the dilation factor grows in an exponential manner to expand the receptive field. These parallel convolutions shape the core of multi-scale representational, as high dilation factors enable extracting large-scale representations, while small-scale representations can be extracted by low dilation factors. In the above operation, the generated output maps are down-sampled by setting a stride of 2 for each layer. In the LCE block, the breaks are handled in a multi-scale manner, which enables the network to capture both the local and global representations from multi-distribution inputs. The output of each DwS-Conv layer in each path is passed to the point-wise convolution preceded by adaptive normalization, as follows:(7) Fp″′=F″if(p=0)GELUC1×1ASNFp″if(p=1,⋯,P)
The Gaussian error linear unit (GELU) activation is applied after the above convolution to obtain smooth activation. Next, a spatial squeeze and excite (sSE) [36] is applied to enable the LCE block to automatedly learn to emphasize target patterns of different scales. Moreover, the sSE acts as an attention method that can also learn to overturn unrelated representations at certain scales and focus on important representations on other scales. The sSE enables each scale to express itself to determine which one it contributes to in the multi-scale learning procedure. The transformation incurred by the sSE can be expressed as(8) Fp″″=F″′if(p=0)Fp″′+Fp″′⨂σC1×1Fp″′if(p=1,⋯,P)
where σ is a sigmoid function and ⨂ denotes the Hadamard product.
By the end, information from various scales is fused by concatenating the output of all paths and processing them with point-wise convolution, as follows:(9) F∗=ConcatF0″″,⋯,FP″″
(10) FX=C1×1F∗
The convolution in the above formula is a P-grouped convolution generating a feature map with C′ channels, which acts as a fuser of features from parallel paths with DwS-Conv layers. In the next LCE blocks, the fusion can be performed via point-wise convolution at the beginning. The design of the LCE block can considerably decrease the number of learning parameters by P times when contrasted with the approach of applying a standard convolution. Notably, increasing the number of paths P is beneficial for reducing the number of parameters of LCE blocks. To maintain a better balance of computing cost and segmentation precision, we set the value of P to 4. Thus, LCE block brings advantages. One is a significant reduction in the number of learning parameters, and the other is the ability to efficiently learn effective multi-scale features. Thus, it can precisely address the variable size COVID-19 lesions.
The ASN layer can be described as follows. Generally, an input instance (h) to a specific normalization layer is shaped into a four-dimensional tensor (N;C;H;W), representing the number of examples per batch, number of input channels, height, and width, respectively. In this context, hncij represents the pixel value before normalization, while h∼ncij represents the corresponding normalized pixel value, which is computed as follows:(11) h∼ncij=γhncij-μσ2+∊+β,
where n∈1,N, c∈1,C, i∈1,H, and j∈[1,W]. γ and β denote a scale and a shifting parameter, respectively. ∊ denotes the tiny constant that maintains numerical constancy, while μ and σ represent the mean and standard deviation statistics, respectively. The above formula is the same for instance normalization (IN), layer normalization (LN), and BN; however, they adopt diverse groups of pixels to approximate μ and σ. This can be generally represented as(12) μk=1|Ik|∑ncij∈Ikhncij
(13) σk2=1|Ik|∑ncij∈Ikhncij-μk2
where k∈{BN,LN,IN} is adopted to differentiate diverse normalization layers methods.
In this formulation, |Ik| signifies the number of pixels. In particular, IBN, ILN, and IIN denote the group of pixels applied to calculate statistics in various layers.
Inspired by switchable normalization (SN) [23], the ASN layer was designed based on adaptive IN (AdaIN) [11], [16], in which the mean and variance of the input features can be aligned with those of the style features. The normalization of the pixel value can be formulated as follows:(14) h∼ncij=γhncij-∑k∈Ωwkμk∑k∈Ωwk′σk2+∈+β,
where Ω∈{BN,LN,AdaIN} represents the statistics calculated from integrated normalization methods. However, this formulation exhibits highly repetitive calculations. To avoid this problem, the dependency between statistics can be exploited to reuse computations, as follows:(15) μAdaIN=1HW∑ijHWhncij,σAdaIN2=1HW∑ijHWhncij-μAdaIN2,
(16) μLN=1C∑c=1CμAdaIN,σLN2=1C∑c=1CσAdaIN2+μAdaIN2-μLN2,
(17) μBN=1N∑n=1NμAdaIN,σk2=1N∑n=1NσAdaIN2+μAdaIN2-μBN2.
This design leads to a computing cost of ONCHW. Moreover, wk and wk′ denote the significance fractions applied to calculate the weighted average of μk and σk2, respectively. Both wk and wk′ are designated as a scalar instance that is commonly used in each channel. Similar to SN, the ASN layer contains a total of six significance weights, while wk is calculated as follows:(18) wk=eλk∑z∈BN,LN,INeλz
where each wk is calculated with the SoftMax function. λAdaIN, λLN , and λBN are the monitor parameters to be updated during the training. The definition of wk′ is the same as that of wk, with three control parameters: λAdaIN′, λLN′ , and λBN′. In ASN, weight standardization (WS) [34] is applied to standardize the weights of the encoding layers with the aim of smoothing the landscape. Specifically, weight standardization alleviates the concern about transmitting smoothing outcomes from activations to weights. The traditional convolutional layer is
y=W^∗x, (19)
where W^∈RO×I represents the weight matrix, and ∗ designates the convolution layer. O and I denote the number of the output channels and input channels, respectively. In this setting, rather than explicitly optimizing the loss with the original weight’s matrix (W^), the network reparametrizes the weights as follows:(20) W^=WS(W)
(21) W^=W^i,j|=W^i,j=Wi,j-μwiσwi,
whereas,(22) μwi=1I∑j=1IWi,j,σwi=1I∑j=1IWi,j-μwi2+∊
In the way that BN manipulates the first and second moments of the weights in convolutional layers for each individual output channel, WS does the same for each output channel. Keep in mind that the weights are of ten initialized in a consistent way across multiple initialization strategies. When compared to these other methods, WS attempts to normalize gradients through back-propagation by standardizing the weights in a differentiable way. The absence of affine transformations in W^ is important to keep in mind. It is expected that subsequent normalization layers, such as a BN or group normalization (GN), can re-normalize the output of this convolutional layer. Thus, including an affine transformation generates confusion and slows down the training process.
4.2 CE Module
A popular shortcoming of encoder–decoder networks is that successive convolution striding or pooling layers lead to a considerable reduction in feature resolution during the learning of encoded feature representation, resulting in loss of contextual information. The CE module is introduced to encode the contextual representations from multi-scale features fused from the down-samplers with no more learning weights. In particular, the CE module was designed to extract semantic representations with an intelligent multi-headed window-based attention layer, aiming to enrich the up-samplers with contextual information necessary for improving the accuracy of segmentation results. The CE equips MIC-Net with an elegant method for improving the segmentation performance through better exploitation of the abundance of contexts amongst input keys over multi-scale 2D encoding maps. The design of the CE module combines both contexts digging between keys and SSA over feature maps from down-samplers, thus preventing the creation of another path for context fusion (See Fig. 4 ). Technologically, the CE module first contextualizes the representation of the key via applying C3×3 to adjacent keys inside the 3×3 grid. In mathematical terms, we have feature map X∈RC×H×W with query X=Q, key (X=K), and value (X=V). Rather than encoding each key with point-wise convolution in traditional SA, the CE module applies DwS-Conv to generate a contextualized key representation, as follows:(23) K′=C3×3K
Fig. 4 Architecture of CE module
The obtained contextualized keys (K′∈RC×H×W) inherently indicate the static contextual information between local neighboring keys, so K′ can be considered as the static contextual mapping of multi-scale input X.
Next, these contextualized key maps and input query are concatenated and passed to two successive point-wise convolutions. The aim is to calculate the attention scores by exploiting the joint relationships between query and latent keys (as described in SSA) under the supervision of the static context.(24) A=GELUC1×1C1×1concatK′,Q
For each attention head, the local attention score for every position of A is learned according to the query information and the context key information instead of the solitary query-key couples. This improves the SSA with the extra supervision from the fused static context K′. Later, matrix A is adopted to compute attention feature map K″ by the fusion of V, as in SSA:(25) K″=V⊗A
This way, the calculated attention scores can be exploited to fuse all input elements and learn a contextual dynamic representation of encoding maps to represent the dynamic context. Thus, the output of the CE module is calculated as a combination of the dynamic contextual representation (K″) and the static contextual representation. As shown in Fig. 5 , the output of CE is scaled to various resolutions such that CEiX,i=1,2,3,4. Each scaled output is later exploited to enrich the semantic knowledge transferred from the down-sampler to the up-sampler through the skip connections.(26) ENii=CEiX+FiX
Fig. 5 Systematic diagram of the proposed MIC-Net with three main constituents: the down-sampling path, CE module, and up-sampling path.
4.3 Feature Decoding Path (Up-Sampler)
Given that the last feature map in the encoding path comes with 1/16 the scale of the network input, it is not ideal to model COVID-19 infection lesions explicitly because of the loss of fine-grained information. As an alternative, the decoding path is introduced to re-establish the complex semantic multi-scale features captured by the feature down-sampling modules. Skip linking is employed to transmit detailed information from the encoder modules to the corresponding decoder modules to avoid losing semantic information. The decoding path structure consists of up-sampler modules that progressively up-sample and aggregate the learned representational map at every down-sampling stage. An up-sampler module is in charge of aggregating learned features. We denote the transformation function of the up-sampler as(27) Si′=C1×1X
(28) DiX=ASNC3×3Si′+C3×3Si′
where DiX, i=1,2,3 applies the point-wise convolution to adapt the number of channels. The up-sampling feature map of the decoder is Di∈RCi×H2i×W2i. In the first stem of the decoding path, we have(29) D4=ASNC1×1EN44
Thereby, Dii={1,2,3} is calculated as follows:(30) Si″=DiUp-sampleDi+1,2,
(31) Di=GELUSi″+ASNC1×1ENii
where Up-sample,t denotes the up-sampling of the feature map with scale t applying bilinear interpolation. Thus, the decoding path can fuse the fine-grained semantics to enable the network to perform precise segmentation of COVID-19 infection lesions. Given Di (i=1,2,3,4), the dense estimation map is computed with a point-wise convolution, as follows:(32) Oi=SoftMaxUpSampleC1×1Di,2i
The predicted output map (Oi∈RH×W) is calculated with two channels indicative of two classes: background (black) and COVID-19 infections lesion (white). Hence, the final segmentation map can be designated with Oi.
5 Experimental Design
In this section, a detailed explanation of the experimental design is given to enable reproducing the experiments in terms of the implementation setup, data set description, preprocessing, and performance metrics.
5.1 Implementation Setup
The overall implementation of MIC-Net is conducted with the TensorFlow library using three NVIDIA Quadro GPUs, one for every data source. Table 1 summarizes the best hyper-parameters of our simulation model. Notably, the optimal values for each parameter are selected after an exhaustive experiment with different possible sets of parameters.Table 1 Hyper-parameters of the proposed MIC-NET.
Hyper-parameter Optimal values
β1 0.9
β2 0.999
σ 0.6
η 0.0001
Number of iterations 25000
Dropout 0.1
Optimizer Adam
Learning rate 0.0001
Total clients 100
Aggregation q-FedAVG [13]
Local Epochs 50
Batch-size 4
Communication rounds 200
5.2 Data Sets
In order to assess the performance of the proposed MIC-Net on heterogeneous multi-source data, we selected three publicly available COVID-19 CT data sets. The first is the COVID-19-CT-Seg data set [25], which comprises 20 public COVID-19 CT volumes from the Coronacases Initiative and Radiopaedia, with more than 1,800 annotated slices. In our experiments, we refer to this data set as D1. Second, we use the 50 CT volumes published in the MosMedData data set [30], which was aggregated from municipal hospitals in Moscow, Russia. In our experiments, we refer to this data set as D2. The third is a larger data set, MedSeg [28], which consists of nine CT volumes that comprise a total of 829 slices, with 373 slices confirmed as positive. We refer to this data set as D3. The specification of each data set is reported in corresponding studies.
5.3 Data Preparation
Following [19], all three data sets were prepared by slicing whole volumes into 2D images and applying some standard augmentation (i.e., cropping, rotating, and scaling) to alleviate the observed data imbalances. Then, we resized all the generated CT slices to 384×384 in order to minimize the discrepancy of intensity among CT scans from diverse sources. We allocated 80% and 20% of the data sets into training and testing sets, respectively. Once again motivated by [45], three types of intensity normalization were adopted: bias field correction, noise filtering, and whitening. Moreover, harmonization mechanisms (HMs) [18] have been demonstrated as an effective way of lessening inter-site heterogeneity through continuous frequency space interpolation. Thus, an HM was applied to prepare the data before training. All experiments of the proposed framework and competing 2D segmentation models are performed on an axial view of a CT scan. For every CT image, the intensity score is stabilized to exhibit unit variance and a zero average before being fed into MIC-Net.
5.4 Evaluation Metrics
Inspired by the extensively used approaches for evaluating segmentation techniques, we assess the segmentation performance of MIC-Net using two complementary metrics: the normalized surface dice (NSD) and the dice similarity coefficient (DSC). Let GandS denote the GT and model outcomes, respectively. The two measures are given as follows:(33) DSC=2S∩GS+G,
(34) NSD=2∂S∩B∂Sτ+2∂S∩B∂Gτ∂S+∂G,
where B∂SτandB∂Gτ designate the border area of the GT and segmentation surface at τ tolerance, respectively, and are expressed as B∂Sτ={x∈R3|∃x∈∂S,‖x-x^‖≤τ^} and B∂Gτ={x∈R3|∃x∈∂G,‖x-x^‖≤τ^}, respectively. Herein, τ is set to 1 mm and 3 mm for segmenting the lung and lesions, respectively. The acceptance is calculated by estimating the difference between two diverse radiologists.
6 Results and Analysis
This section comprehensively discusses and analyzes the findings obtained from the experimental evaluation of the proposed framework. The details of each experiment and the corresponding results can be found in the following subsections.
6.1 Domain shift analysis
To begin, MIC-Net experiments are done under two scenarios. One trains MIC-Net with merged multi-source CT scans (mixed), and the other scenario emphasizes training MIC-Net on single-source data (independent). To experiment with data from different domains in the IoMT, it is important to numerically investigate the data heterogeneity from different domains. Motivated by a recent paradigm for analyzing domain shift [23], we perform cross-source validation among D1, D2, and D3 by training independent networks on each separately, and then evaluating the networks using different samples from the three data sets. It is obvious from Table 2 that the independent models achieve better performance when tested on samples from the same data set, and their performance degrades when tested using samples from another domain. Conversely, the mixed model realizes a slight performance improvement compared to the independent model on D1 and has comparable performance to the independent model on D2. In some cases, the mixed training may lead to lower segmentation performance than that achieved under independent settings. Table 2 Performance comparison between independent and mixed approach, trained and evaluated on different data sets (mean ± standard deviation).
Methods D1 D2 D3
DSC
Independent (D1) 79.3±13.1 69.5±10.04 64.8±12.3
Independent (D2) 72.9±11.40 80.7±08.91 72.2±11.18
Independent (D3) 76.67±9.17 70.8±12.50 81.3±08.19
Mixed 81.8±10.60* 79.6±13.10 81.1±09.13
NSD
Independent (D1) 78.6±8.66 67.5±5.33 61.8±9.46
Independent (D2) 72.5±9.38 76.8±8.01 72.4±7.15
Independent (D3) 74.6±6.89 70.3±8.78 78.3±6.68
Mixed 79.8±7.14* 79.6±10.29* 77.8±7.71*
6.2 Comparative Analysis
As a common research practice, the MIC-Net is compared with state-of-the-art segmentation methods to evaluate its competitive capabilities under the same experimental settings. The numerical results obtained from these experiments are displayed in Table 3 . Specifically, the MIC-Net is compared with the common 2D segmentation models U-Net [35], FCN-8 [22], and Inf-net [6]. They are also in compared with —the 3D segmentation models 3D U-Net [5] and 3D V-Net [33], with patch size dimensions of 64×64×64, and 2.5-dimensional-based approaches H-DUnet [14] and MultiPlanar UNet (MPUnet) [37], which perform view accumulation from 2D patch architectures. The inclusion of models with different dimensionalities in comparative experiments is beneficial to evaluate the competitiveness of the proposed solution. For multi-site approaches, the proposed framework is compared with MS-Net [19], MS-Fed [15], and FED-DG [18].Table 3 Quantitative results (mean ± standard deviation) for COVID-19 lesion segmentation on test sets.
DSC↑ NSD↑ # Parameters
Methods D1 D2 D3 Average D1 D2 D3 Average
2D-U-Net [21] 82.1±8.9 86.1±9.9 87.1±8.9 85.10±9.23 80.9±10.3 84.9±11.1 87.1±11.5 84.3±10.97 7.8530 M
2D-FCN-8[11] 82.2±9.1 86.2±9.1 86.9±10.1 85.10±9.43 80.3±11.2 84.4±10.3 86.9±10.3 83.9±10.60 41.530 M
3D-U-Net [16] 80.8±11.4 84.0±8.4 85.9±14.3 83.57±11.37 79.7±14.5 83.5±9.8 85.2±13.4 82.8±12.57 22.577 M
3D-V-Net [33] 80.1±11.2 84.7±9.2 85.3±15.1 83.37±11.83 79.9±13.1 84.6±10.7 85.3±13.3 83.2±12.37 46.048 M
H-DUnet [34] 81.2±9.2 81.9±7.2 86.1±9.7 83.07±8.70 80.3±8.7 80.1±9.3 86.1±10.5 82.2±9.50 45.082 M
MPUnet [40] 80.4±10.2 81.1±9.7 84.7±13.3 82.07±11.07 79.9±9.3 79.1±12.2 84.7±9.7 81.2±10.40 62.001 M
Inf-Net [9] 81.4±7.2 85.6±10.3 86.0±11.1 84.99±9.77 81.9±7.9 85.8±6.3 87.9±7.16 85.1±8.44 33.122 M
MS-Net [9] 83.8±6.5 87.8±6.4 88.5±5.9 86.70±6.27 82.9±9.4 86.6±8.7 87.9±6.5 85.9±8.20 18.841 M
MS-Fed [15] 81.6±4.7 85.1±7.1 84.9±3.8 83.87±5.20 81.8±3.7 87.6±4.3 86.2±7.4 85.20±5.13 7.8530 M
FEDDG [18] 84.7±9.1 87.3±8.8 89.2±6.6 87.06±8.17 83.3±5.6 85.6±9.2 88.3±8.8 85.73±7.87 18.841 M
* MIC-Net 85.8±7.8 90.1±7.3 90.8±5.6 88.90±6.90 85.1±6.7 88.8±7.8 88.7±9.2 87.53±7.90 980.131K
Table 3 presents the quantitative results of the proposed MIC-Net against existing approaches for COVID-19 lesion segmentation. We observe that MPUnet achieves the lowest performance, with DSCs of 82.07 and 81.23. 3D-U-Net and 3D-VNet attain around 2% and 1% reductions in DSC and NSD, respectively, in comparison with the 2D models and realize comparable NSD. It is also noted that the recently proposed DSBN and MS-Net attain better performance, with 86.93 and 86.70 DSCs, respectively; furthermore, they yield the highest NSDs, of 85.87 and 86.87, respectively. Compared to the MS-Net, MIC-Net achieves a 2.3% and 1.6% improvement on DSC and NSD, respectively. Compared to the recent FL approaches, the proposed MIC-Net achieves a 1%–2% improvement across different metrics. The results demonstrate the superiority of MIC-Net in the supervised segmentation of pneumonia lesions from heterogeneous CT scans; its better generalization performance qualifies it for integration in the IoMT environment.
6.3 Statistical & Visual Analysis
To further validate the significance of results obtained in the comparative analysis, a statistical significance test is performed to assess how the obtained results significantly differ from those of the previous one. The p-value obtained from this test is shown in Table 4 . In this setting, the significance threshold is set to be 0.05. As shown, the majority of statistical results are beyond the significance threshold. This, in turn, demonstrates the competing advantages of MIC-Net.Table 4 Results of the statistical significance test (p-value) for comparing MIC-Net with competing methods.
DSC↑ NSD↑
Method D1 D2 D3 D1 D2 D3
MIC-Net vs 2D-U-Net 8.112E-05 6.930E-04 3.490E-04 3.114E-03 8.975E-06 8.820E-04
MIC-Net vs Inf-Net 8.846E-02 6.263E-05 1.912E-03 2.323E-03 2.342E-03 2.586E-03
MIC-Net vs MS-Net 2.155E-03 2.827E-03 8.014E-02 1.509E-03 2.832E-03 3.285E-04
MIC-Net vs MS-Fed 1.762E-02 2.464E-03 6.680E-04 9.858E-02 2.454E-03 1.745E-03
MIC-Net vs FEDDG 2.549E-03 2.459E-02 3.425E-05 1.709E-02 2.405E-03 2.111E-03
Moreover, Fig. 6 provides a visual comparison of the segmentation outcomes from different competing methods on samples belonging to different domains. The illustration provides an obvious indication of the precision of the segmentation maps compared to that of competing methods. Notably, MIC-Net achieves accurate segmentation for various lesions of different shapes and sizes.Fig. 6 Comparison of the segmentation outcomes from different methods against the ground-truth label.
6.4 Ablation Analysis
This section analyzes and discusses the ablation experiments to understand the contribution of different building blocks to the final segmentation performance. In this scenario, common U-Net is chosen as the baseline method. Table 5 shows the results of ablation experiments across different performance metrics. For convenience, a nickname is given to each ablation experiment in each row of Table 5. Obviously, the harmonization of training CT images shows considerable performance improvements (DSC: NSD:). Speaking of building blocks, note that redesigning the encoder path with the proposed down-sampler module improves the ability of the network to learn from cross-domain CT slices. This can be attributed to the ability of the ASN to adaptively lessen the impact of domain shift during the training process. To validate this justification, we implement the down-sampler without ASN (See V4), which leads to a significant drop in the performance of MIC-Net. Moreover, eliminating the SSA from the down-sampler (See V5) also results in a significant drop in segmentation performance, which indicates that the attention enables the network to effectively encode lesion features under distribution shifts. In addition, the inclusion of the CE module improves the performance over the baseline (See V6), and similarly, when combined with a down-sampler or up-sampler, it shows notable improvements in all data sets. This can be attributed to the fusion of static and dynamic contextual features by the CE module, enabling the network captures the features of lesions from different size and shapes in CT images from varied distribution. Finally, the up-sampler module is elegantly designed to reconstruct the segmentation maps given the down-sampling maps as well as contextual representations from the CE module, leading to notable improvements compared with the standard deconvolution in the baseline.Table 5 Results of ablation experiments for the proposed MIC-NET.
Variants Baseline +HM Down-sampler CE Up-sampler DSC↑ NSD↑
All w/o ASN w/o SSA D1 D2 D3 D1 D2 D3
V1 × 82.1±8.9 86.1±9.9 87.1±8.9 80.9±10.3 84.9±11.1 87.1±11.5
V2 × 83.3±8.2 87.3±10.3 87.3±9.3 82.1±7.2 86.3±10.9 87.7±8.4
V3 × × 84.6±11.3 88.3±9.1 88.5±8.2 83.4±5.9 86.9±6.2 87.3±3.7
V4 × × 83.8±9.2 87.8±8.8 87.31±7.7 83.1±6.8 86.6±4.7 87.4±6.3
V5 × × 83.9±8.5 87.8±10.2 87.9±9.4 82.9±9.1 86.7±5.4 87.1±5.7
V6 × × 84.4±10.1 88.1±9.9 88.2±8.1 83.1±8.4 86.6±8.2 87.2±7.2
V7 × × × 85.4±9.7 89.1±6.8 89.3±4.9 84.2±7.5 87.6±6.6 87.5±4.6
V8 × × × 84.9±8.8 88.5±8.1 88.4±6.3 83.8±4.9 87.2±7.4 87.4±8.1
V9 × × × × 85.8±7.8 90.1±7.3 90.8±5.6 85.1±6.71 88.8±7.81 88.7+9.2
Given the notable performance gain achieved by the ASN, an additional experimental analysis is performed to analyze and compare the performance of MIC-Net when implemented using different normalization layers such as BN, IN, LN, GN, SN, and sparse switchable normalization (SSN). The numerical results obtained from these experiments are shown in Fig. 7 . Notably, the proposed ASN is conducive to improving the segmentation performance under distribution shifts owing to its ability to perform adjustive normalization according to the underlying distribution.Fig. 7 Comparison of the performance of MIC-Net under different normalization layers.
6.5 Computational Analysis
The framework setup for the proposed IoMT framework assessment and the employed computer hardware configurations are summarized in Table 6 .Table 6 Hardware setup for the proposed COVID-19-Fog.
Fog Component Configuration Setup
Gateway Device Samsung J4, 4GB RAM, 16G storage, android 8
Broker Laptop: Dell inspiron 3500, with Intel(R)Core (TM) i5-7200 CPU @ 1.7 GHZ, 8.00 GB DDR4 RAM, 64-bit system bus, and Windows 10. Apache HTTP Server 2.4.34 utilized for deployment
Worker Node Laptop: Toshiba PS582E-002002AR, with Intel(R)Core (TM) i7-7200 CPU @ 2.5 GHZ, 8.00 GB, 32-bit system bus, and Windows 8.1. Apache HTTP Server 2.4.34.
CDC Microsoft Azure Engine, 2 GB SSD, 1vCPU, 2 GB RAM, Windows Server 2016.
Latency analysis. Latency is a major concern for the health-care community, especially when it comes to IoMT-aided diagnosis. Thus, the latency of the proposed framework is compared under different FL schemes, namely vanilla FL, FED-CS [32], FED-MCCS, q-FFL, and FED-PCS, computed by combining the queueing delay and communication delay. Note that the FED-PCS can achieve the lowest latency compared to the other approaches, as shown in Fig. 8 . This may be because the FED-PCS can adaptively and fairly select the clients according to their resources and data statistics. In general, the lightweight nature of the MIC-Net enables attaining low latency during the training under different FL schemes, thereby leading to significant differences in latency performance.Fig. 8 Latency of the MIC-Net under different resource allocation protocols.
Execution time analysis. When it comes to model training in the real-world IoMT, the Execution time is essential to judging the ability of the model to generate real-time segmentation of the uploaded COVID-19 CT scan. Thus, the Execution time of MIC-Net is calculated and compared under different resource allocation protocols, as shown in Fig. 9 . As anticipated, the FED-PCS protocol configuration exhibits the shortest Execution time (188.3 ms) owing to its powerful resource accessibility. However, the average inference times at edge devices are relatively larger, but it is still acceptable to be considered a real-time response.Fig. 9 Execution time of the MIC-Net under different resource allocation protocols.
Network Bandwidth Analysis. As shown in Fig. 10 , the average network bandwidth consumption is compared under different FL allocation protocols. The bandwidth consumption reaches the highest degree when MIC-Net is trained with vanilla FL (46.3 kbps). However, the FED-PCS protocol achieves relatively lower bandwidth consumption, as the parameters are communicated intelligently based on the profile state considered during the client selection. The intelligent FED-PCS achieves the lowest consumption, as it adaptively selects participants according to the resource profile of the clients.Fig. 10 The bandwidth consumption of the MIC-Net under different resource allocation protocols.
Power Consumption Analysis: Investigating power/energy utilization is important for any IoT framework. Our analysis for the power proposed IoMT framework is shown in Fig. 11 . It can be seen that training MIC-Net under the FED-PCS protocol results in the lowest power consumption (14.3 W) compared to the other competing methods. Furthermore, the lightweight nature of MIC-Net helps it maintain low power consumption under different FL allocation protocols.Fig. 11 The power consumption of the MIC-Net under different resource allocation protocols.
7 Conclusions and Future Work
This work presented a framework, MIC-Net, for federated segmentation from multi-domain CT data. The concept of drift phenomena is tackled through two stages: one is the attentive learning of cross-domain images, and the other is the adaptive normalization layer. A multi-criteria selection FL protocol is presented to train the MIC-Net in such a way that we can preserve the resources and maximize the training efficiency in the IoMT environment. The experimental findings demonstrate the efficiency and effectiveness of our solution over recent centralized and federated segmentation approaches. The lightweight nature of MIC-Net enables low resource consumption (in terms of power, CPU, memory, and communication bandwidth), making it a robust candidate to be deployed as a distributed pneumonia diagnosis tool, especially in resource-constrained IoMT environments.
In our plans for future works, the proposed framework will be extended to support a multitask diagnosis of COVID-19 patients by including extra tasks such as automatic classification, multi-class segmentation, severity assessment, and other follow-up functionalities. Responsible AI is currently regarded as the most interesting topic for both academia and industry; hence, this work will be updated to satisfy the principles of responsibility in terms of security against adversaries, precision, interpretability, fairness, and privacy preservation. Moreover, the efficiency-cost tradeoff is still a major concern for IoMT application; thus, extending the proposed framework to be more lightweight and efficient will be an important step in our future work. Furthermore, estimating uncertainty in the decisions generated by DL has become a major concern in the majority of IoMT applications; thus, type-2 and type-3 fuzzy sets will be investigated to expand the proposed model to include this functionality.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Data availability
No data was used for the research described in the article.
Acknowledgments
The authors would like to express sincere appreciation to the editor and anonymous reviewers for their insightful comments, which greatly improved the quality of this paper. This work is in part by the National Natural Science Foundation of China under Grants 61300167 and 61976120, the Natural Science Foundation of Jiangsu Province under Grant BK20191445, and the Natural Science Key Foundation of Jiangsu Education Department under Grant 21KJA510004.
==== Refs
References
1 Abdel-Basset M. Chang V. Hawash H. Chakrabortty R.K. Ryan M. FSS-2019-nCov: A deep learning architecture for semi-supervised few-shot segmentation of COVID-19 infection Knowledge-Based Systems 212 2021 106647
2 Ahmed I. Ahmad A. Jeon G. An IoT-Based Deep Learning Framework for Early Assessment of COVID-19 IEEE Internet Things Journal 8 21 2020 15855 15862
3 Awais M. Raza M. Singh N. Manzoor U. Islam S.U. Rodrigues J.J.P.C. LSTM-based emotion detection using physiological signals: IoT framework for healthcare and distance learning in COVID-19 IEEE Internet of Things Journal 8 23 2020 16863 16871 35582634
4 Chen C. Zhou K.N. Zha M.X. Qu X.Y. Guo X.Y. Chen H.Y. Wang Z. Xiao R. An effective deep neural network for lung lesions segmentation from COVID-19 CT images IEEE Transactions on Industrial Informatics 17 9 2021 6528 6538
5 Çiçek Ö. Abdulkadir A. Lienkamp S.S. Brox T. Ronneberger O. 3D U-Net: learning dense volumetric segmentation from sparse annotation 2016 Springer Cham 424 432
6 Fan D.P. Zhou T. Ji G.P. Zhou Y. Chen G. Fu H.Z. Shen J.B. Shao L. Inf-net: Automatic covid-19 lung infection segmentation from ct images IEEE Transactions on Medical Imaging 39 8 2020 2626 2637 32730213
7 Y.C. Fang, H.Q. Zhang, J.C. Xie, M.J. Lin, L.J. Ying, P.P. Pang, W.J. Ji, Sensitivity of chest CT for COVID-19: comparison to RT-PCR, Radiology, 296 (2020) E115–E117.
8 Ganesh P. Chen Y. Lou X. Khan M.A. Yang Y. Sajjad H. Nakov P. Chen D.M. Winslett M. Compressing large-scale transformer-based models: A case study on bert, Transactions of the Association for Computational Linguistics 9 2021 1061 1080
9 Gao K. Su J.P. Jiang Z.B. Zeng L.L. Feng Z.C. Shen H. Rong P.F. Xu X. Qin J. Yang Y.X. Wang W. Hu D.W. Dual-branch combination network (DCN): Towards accurate diagnosis and lesion segmentation of COVID-19 using CT images Medical Image Analysis 67 2021 101836
10 Gao S.H. Cheng M.M. Zhao K. Zhang X.Y. Yang M.H. Torr P. Res2Net: A New Multi-Scale Backbone Architecture IEEE Transactions on Pattern Analysis and Machine Intelligence 43 2 2019 652 662
11 Huang X. Belongie S. Arbitrary style transfer in real-time with adaptive instance normalization Proceedings of the IEEE International Conference on Computer Vision 2017 1501 1510
12 Kumar R. Khan A.A. Kumar J. Zakria N.A. Golilarz S.M. Zhang Y. Ting C.Y. Zheng W.Y.W. Blockchain-federated-learning and deep learning models for covid-19 detection using ct imaging IEEE Sensors Journal 21 14 2021 16301 16314 35789224
13 Li T. Sanjabi M. Beirami A. Smith V. Fair Resource Allocation in Federated Learning, International Conference on Learning Representations 2020 1 13
14 Li X.M. Chen H. Qi X.J. Dou Q. Fu C.W. Heng P.A. H-DenseUNet: Hybrid Densely Connected UNet for Liver and Tumor Segmentation from CT Volumes IEEE Transactions on Medical Imaging 37 12 2018 2663 2674 29994201
15 Li X.X. Gu Y.F. Dvornek N. Staib L.H. Ventola P. Duncan J.S. Multi-site fMRI analysis using privacy-preserving federated learning and domain adaptation: ABIDE results Medical Image Analysis 65 2020 101765
16 Ling J. Xue H. Song L. Xie R. Gu X. Region-aware adaptive instance normalization for image harmonization Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2021 9361 9370
17 Liu J.N. Dong B. Wang S. Cui H. Fan D.P. Ma J.Q. Chen G. COVID-19 lung infection segmentation with a novel two-stage cross-domain transfer learning framework Medical Image Analysis 74 2021 102205
18 Liu Q.D. Chen C. Qin J. Dou Q. Heng P.A. Feddg: Federated domain generalization on medical image segmentation via episodic learning in continuous frequency space Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition 2021 1013 1023
19 Liu Q.D. Dou Q. Yu L.Q. Heng P.A. MS-Net: Multi-Site Network for Improving Prostate Segmentation with Heterogeneous MRI Data IEEE Transactions on Medical Imaging 39 9 2020 2713 2724 32078543
20 Liu Z. Lin Y.T. Cao Y. Hu H. Wei Y.X. Zhang Z. Lin S. Guo B.N. Swin transformer: Hierarchical vision transformer using shifted windows Proceedings of the IEEE/CVF International Conference on Computer Vision 2021 10012 10022
21 Liu Z. Yao C.H. Yu H. Wu T.H. Deep reinforcement learning with its application for lung cancer detection in medical Internet of Things Future Generation Computer Systems 97 2019 1 9
22 Long J. Shelhamer E. Darrell T. Fully convolutional networks for semantic segmentation Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition 2015 3431 3440
23 Luo P. Zhang R.M. Ren J.M. Peng Z.L. Li J.Y. Switchable Normalization for Learning-to-Normalize Deep Representation IEEE Transactions on Pattern Analysis and Machine Intelligence 43 2 2019 712 728
24 Ma F.Z. Sun T. Liu L.Y. Jing H.Y. Detection and diagnosis of chronic kidney disease using deep learning-based heterogeneous modified artificial neural network Future Generation Computer Systems 111 2020 17 26
25 J. Ma, Y.X. Wang, X. An, C. Ge, Z. Yu, J. Chen, Q. Zhu, G. Dong, J. He, Z. He, Towards Efficient COVID-19 CT Annotation: A Benchmark for Lung and Infection Segmentation, arXiv preprint, 2020, Available: http://arxiv.org/abs/2004.12537.
26 Mahmud T. Alam M.J. Chowdhury S. Ali S.N. Rahman M.M. Fattah S.A. Saquib M. CovTANet: a hybrid tri-level attention-based network for lesion segmentation, diagnosis, and severity prediction of COVID-19 chest CT scans IEEE Transactions on Industrial Informatics 17 9 2020 6489 6498
27 Mansour R.F. Escorcia-Gutierrez J. Gamarra M. Gupta D. Castillo O. Kumar S. Unsupervised deep learning based variational autoencoder model for COVID-19 diagnosis and classification Pattern Recognition Letters 151 2021 267 274 34566223
28 “MedSeg” [Online]. Available: https://medicalsegmentation.com/covid19/.
29 S. Mehta, M. Rastegari, Separable Self-attention for Mobile Vision Transformers, arXiv preprint, 2022, Available: http://arxiv.org/abs/2206.02680.
30 Morozov S.P. Andreychenko A.E. Blokhin I.A. Gelezhe P.B. Gonchar A.P. Nikolaev A.E. Pavlov N.A. Chernina V.Y. Gombolevskiy V.A. MosMedData: data set of 1110 chest CT scans performed during the COVID-19 epidemic Digital Diagnostics 1 1 2020 49 59
31 Nguyen D.C. Pham Q.V. Pathirana P.N. Ding M. Seneviratne A. Lin Z.H. Dobre Q. Hwang W. Federated learning for smart healthcare: A survey ACM Computing Surveys (CSUR) 55 3 2022 1 37
32 T. Nishio, R. Yonetani, Client selection for federated learning with heterogeneous resources in mobile edge, ICC 2019-2019 IEEE International Conference on Communications (ICC), 2019, pp. 1-7.
33 Pathak N. Deb P.K. Mukherjee A. Misra S. IoT-to-the-rescue: A survey of IoT solutions for COVID-19-like pandemics IEEE Internet of Things Journal 8 17 2021 13145 13164
34 S.Y. Qiao, H.Y. Wang, C.X. Liu, W. Shen, A. Yuille, Micro-Batch Training with Batch-Channel Normalization and Weight Standardization, arXiv preprint, 2019, Available: http://arxiv.org/abs/1903.10520.
35 Ronneberger O. Fischer P. Brox T. U-net: Convolutional networks for biomedical image segmentation International Conference on Medical Image Computing and Computer-assisted Intervention 2015 Springer Cham 234 241
36 Roy A.G. Navab N. Wachinger C. Recalibrating Fully Convolutional Networks With Spatial and Channel ‘Squeeze and Excitation’ Blocks IEEE Transactions on Medical Imaging 38 2 2018 540 549
37 Valanarasu J.M.J. Yasarla R. Wang P.Y. Hacihaliloglu I. Patel V.M. Learning to Segment Brain Anatomy from 2D Ultrasound with Less Data IEEE J. Sel. Top. Signal Process. 14 6 2020 1221 1234
38 H.V. Hasselt, A. Guez, D. Silver, Deep reinforcement learning with double q-learning, Proceedings of the AAAI Conference on Artificial Intelligence, 2016, pp. 2094-2100.
39 Varela-Santos S. Melin P. A new approach for classifying coronavirus COVID-19 based on its manifestation on chest X-rays using texture features and neural networks Information Sciences 545 2021 403 414 32999505
40 Vaswani A. Shazeer N. Parmar N. Uszkoreit J. Jones L. Gomez A.N. Kaiser L. Polosukhin I. Attention is all you need Advances in Neural Information Processing Systems 2017 5998 6008
41 Wang G.T. Liu X.L. Li C.P. Xu Z.Y. Ruan J.G. Zhu H.F. Meng T. Li K. Huang N. Zhang S.T. A noise-robust framework for automatic segmentation of COVID-19 pneumonia lesions from CT images IEEE Transactions on Medical Imaging 39 8 2020 2653 2663 32730215
42 Wang R.Z. Zheng G.Y. CyCMIS: Cycle-consistent Cross-domain Medical Image Segmentation via diverse image augmentation Medical Image Analysis 76 2022 102328
43 Wang X.F. Jiang L. Li L. Xu M. Deng X. Dai L.S. Xu X.Y. Li T.Y. Guo Y.C. Wang Z.L. Dragotti P.L. Joint learning of 3D lesion segmentation and classification for explainable COVID-19 diagnosis IEEE Transactions on Medical Imaging 40 9 2021 2463 2476 33983881
44 Wang X.Y. Yuan Y.W. Guo D.Y. Huang X.J. Cui Y. Xia M. Wang Z.H. Bai C. Chen S.Y. SSA-Net: Spatial self-attention network for COVID-19 pneumonia infection segmentation with semi-supervised few-shot learning Medical Image Analysis 79 2022 102459
45 Wang Z. Liu Q.D. Dou Q. Contrastive Cross-Site Learning with Redesigned Net for COVID-19 CT Classification IEEE Journal of Biomedical and Health Informatics 24 10 2020 2806 2813 32915751
46 Yao C.H. Wu S.D. Liu Z. Li P. A deep learning model for predicting chemical composition of gallstones with big data in medical Internet of Things Future Generation Computer Systems 94 2019 140 147
47 Zhang W.S. Zhou T. Lu Q.H. Wang X. Zhu C.S. Sun H.Y. Wang Z.P. Lo S.K. Wang F.Y. Dynamic-fusion-based federated learning for COVID-19 detection IEEE Internet of Things Journal 8 21 2021 15884 15891 35663640
48 Zhou L.X. Li Z.X. Zhou J.X. Li H.Y. Chen Y.P. Huang Y.X. Xie D.X. Zhao L.T. Fan M. Hashmi S. Abdelkareem F. Eiada R. Xiao X.G. Li L.H. Qiu Z.W. Gao X. A rapid, accurate and machine-agnostic segmentation and quantification method for CT-based COVID-19 diagnosis IEEE Transactions on Medical Imaging 39 8 2020 2638 2652 32730214
| 0 | PMC9745980 | NO-CC CODE | 2022-12-15 00:03:20 | no | Inf Sci (N Y). 2023 Apr 13; 623:20-39 | utf-8 | Inf Sci (N Y) | 2,022 | 10.1016/j.ins.2022.12.017 | oa_other |
==== Front
Bull Acad Natl Med
Bull Acad Natl Med
Bulletin De L'Academie Nationale De Medecine
0001-4079
2271-4820
Published by Elsevier Masson SAS on behalf of l'Académie nationale de médecine.
S0001-4079(22)00379-X
10.1016/j.banm.2022.12.003
Correspondance
Discussion à propos de la communication: «Impact de la première année de la pandémie de COVID-19 sur l'épidémiologie des infections invasives (bactériémies) dans les hôpitaux de l’Assistance Publique – Hôpitaux de Paris»**
Discussion about the presentation: “Impact of the first year of the COVID-19 pandemic on the epidemiology of invasive infections (bacteremia) in the hospitals of the Assistance Publique - Hôpitaux de Paris”Jarlier Vincent 12
1 Membre correspondant de l’Académie nationale de médecine. Académie nationale de médecine, 16 rue Bonaparte, 75006 Paris, France
2 Service de Bactériologie-Hygiène et CIMI-Paris, Inserm U1135, Sorbonne Université, Paris, France
13 12 2022
13 12 2022
© 2022 Published by Elsevier Masson SAS on behalf of l'Académie nationale de médecine.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc* Séance du 15/11/2022
| 0 | PMC9746018 | NO-CC CODE | 2022-12-15 00:03:22 | no | Bull Acad Natl Med. 2022 Dec 13; doi: 10.1016/j.banm.2022.12.003 | utf-8 | Bull Acad Natl Med | 2,022 | 10.1016/j.banm.2022.12.003 | oa_other |
==== Front
Am J Obstet Gynecol
Am J Obstet Gynecol
American Journal of Obstetrics and Gynecology
0002-9378
1097-6868
Elsevier Inc.
S0002-9378(21)00020-X
10.1016/j.ajog.2020.12.1218
Letter to the Editor
Hydroxychloroquine early in pregnancy and risk of birth defects: don’t throw out the baby with the bathwater
Bermas Bonnie L. MD
Division of Rheumatic Diseases, Department of Internal Medicine, The University of Texas Southwestern Medical Center, Dallas, TX 75390
Chambers Christina PhD, MPH
Department of Pediatrics, University of California, San Diego, San Diego, CA
9 1 2021
5 2021
9 1 2021
224 5 548549
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcTo the Editors:
We read with interest the article by Huybrechts et al1 that reported increased major congenital anomalies among first-trimester hydroxychloroquine-exposed pregnancies. The methods used by this excellent research group are sound. Using claims data, they compared hydroxychloroquine-exposed pregnancies with matched unexposed pregnancies and found a 1.26 adjusted relative risk of all major malformations among exposed pregnancies. No particular pattern of anomalies was found. The authors concluded that first-trimester hydroxychloroquine exposure leads to a small increased risk of congenital anomalies. We worry that the findings presented in this paper will be interpreted by practicing clinicians as causally related. This information could tip the balance in clinical decision-making toward anticipated risk rather than the established benefit of hydroxychloroquine during pregnancy.
The authors’ conclusions imply that hydroxychloroquine is potentially teratogenic. Establishment of teratogenicity requires several criteria.2 This study fulfills the tenet of identifying exposure to an agent at a critical time in gestation. However, the study does not satisfy other criteria, such as careful delineation of the clinical cases, exposure associated with a specific pattern of defects, and the association making biologic sense. The study’s large sample size provided sufficient statistical power to demonstrate a small but significant increased risk of congenital anomalies with hydroxychloroquine exposure. Although the authors used state-of-the art methods to adjust for confounders, small effect sizes can sometimes be explained by unmeasured confounding. As in all claims data, there is insufficient information to control for exposures to tobacco, alcohol, other drugs, folic acid supplements, and over-the-counter medications that may have impacted the risk of malformations. Most importantly, given the lack of any specific pattern of congenital anomalies, these results could be spurious. Evaluation of increased risks for specific congenital anomalies with hydroxychloroquine exposure in case-control studies could help clarify this issue.
In contrast, several studies have now clearly established the critical role of hydroxychloroquine in controlling systemic lupus erythematosus disease activity in pregnancy and improving outcomes. Currently, several professional societies recommend continuing hydroxychloroquine during lupus pregnancies.3 Moreover, this medication reduces the risk of congenital heart block in offspring of women who are anti-Ro and anti-La positive.4
We would caution readers in their interpretation of the results presented. This study has not proven that hydroxychloroquine is teratogenic, whereas data supporting the benefits of hydroxychloroquine during pregnancy for malarial prophylaxis, lupus pregnancy outcome, and prevention of congenital complete heart block are sound.
B.L.B. reports no conflict of interest.
C.C. receives research funding from the following industry sponsors and a foundation: Amgen Inc; AstraZeneca; Celgene; GlaxoSmithKline; Janssen Pharmaceuticals, Inc; Pfizer Inc; Regeneron Pharmaceuticals Inc; Hoffmann-La Roche-Genentech; Genzyme Sanofi-Aventis; Takeda Pharmaceutical Company Limited; Sanofi; UCB Pharma, Smyrna, GA; Sun Pharma Global FZE; and the Gerber Foundation.
==== Refs
References
1 Huybrechts K.F. Bateman B.T. Zhu Y. Hydroxychloroquine early in pregnancy and risk of birth defects Am J Obstet Gynecol 224 2021 290.e1 290.e22
2 Shepard T.H. “Proof” of human teratogenicity Teratology 50 1994 97 98 7801306
3 Sammaritano L.R. Bermas B.L. Chakravarty E.E. 2020 American College of Rheumatology Guideline for the management of reproductive health in rheumatic and musculoskeletal diseases Arthritis Care Res (Hoboken) 72 2020 461 488 32090466
4 Izmirly P.M. Costedoat-Chalumeau N. Pisoni C.N. Maternal use of hydroxychloroquine is associated with a reduced risk of recurrent anti-SSA/Ro-antibody-associated cardiac manifestations of neonatal lupus Circulation 126 2012 76 82 22626746
| 33434555 | PMC9746025 | NO-CC CODE | 2022-12-15 00:03:23 | no | Am J Obstet Gynecol. 2021 May 9; 224(5):548-549 | utf-8 | Am J Obstet Gynecol | 2,021 | 10.1016/j.ajog.2020.12.1218 | oa_other |
==== Front
Powder Technol
Powder Technol
Powder Technology
0032-5910
1873-328X
Published by Elsevier B.V.
S0032-5910(22)01049-X
10.1016/j.powtec.2022.118168
118168
Article
Spray freeze dried niclosamide nanocrystals embedded dry powder for high dose pulmonary delivery
Zhang Shengyu 1
Yan Shen 1
Lu Kangwei
Qiu Shixuan
Chen Xiao Dong
Wu Winston Duo ⁎
Engineering Research Centre of Advanced Powder Technology (ERCAPT), School of Chemical and Environmental Engineering, College of Chemistry, Chemical Engineering and Materials Science, Soochow University, Suzhou, Jiangsu Province 215123, PR China
⁎ Corresponding author.
1 These authors contributed equally to this work
13 12 2022
13 12 2022
11816831 10 2022
9 12 2022
12 12 2022
© 2022 Published by Elsevier B.V.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Based on the drug repositioning strategy, niclosamide (NCL) has shown potential applications for treating COVID-19. However, the development of new formulations for effective NCL delivery is still challenging. Herein, NCL-embedded dry powder for inhalation (NeDPI) was fabricated by a novel spray freeze drying technology. The addition of Tween-80 together with 1,2-Distearoyl-sn-glycero-3-phosphocholine showed the synergistic effects on improving both the dispersibility of primary NCL nanocrystals suspended in the feed liquid and the spherical structure integrity of the spray freeze dried (SFD) microparticle. The SFD microparticle size, morphology, crystal properties, flowability and aerosol performance were systematically investigated by regulating the feed liquid composition and freezing temperature. The addition of leucine as the aerosol enhancer promoted the microparticle sphericity with greatly improved flowability. The optimal sample (SF−80D-N20L2D2T1) showed the highest fine particle fraction of ~47.83%, equivalently over 3.8 mg NCL that could reach the deep lung when inhaling 10 mg dry powders.
Graphical abstract
Unlabelled Image
Keywords
NCL-embedded dry powder for inhalation (NeDPI)
Spray freeze drying
Fine particle fraction
Particle formation mechanism
Structure-performance relationship
==== Body
pmc1 Introduction
Coronavirus Disease 2019 (COVID-19) is a global pandemic caused by the Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2), which has caused >624 million infections and 6.5 million deaths worldwide [1] until October of 2022. The potential to inhibit SARS-CoV-2 has been confirmed for chloroquine [2,3], favipiravir [4], abidol [5], remdesivir [6,7], interferon [8,9], etc. Niclosamide (NCL) was proven to be the best candidate for effectively preventing the replication and destruction of SARS-CoV-2 after screening >3000 kinds of FDA (U.S. Food & Drug Administration) approved drugs [10]. As a kind of salicylamide derivative, NCL has been developed since the 1950s as a molluskicide against oncomelania snails, exhibiting well-known lethality against a large number of viruses [11,12]. However, a clinical study has shown that oral administration of 500 mg NCL three-times-daily only reached Cmax values of 182 ng/mL [13]. Therefore, the exploration of more effective delivery routes for improving the NCL bioavailability, especially for treating pulmonary viral infection is still urgently needed.
Recently, animal studies have shown that pulmonary administration of NCL was generally of superior bioavailability to oral administration [14]. Furthermore, a randomized, double-blind, parallel and placebo-controlled phase I clinical study (NCT04576312) confirmed the safety of NCL in liquid (solution) form via inhalation, suggesting the promising perspective of inhalable formulation for delivering NCL [15]. Costabile et al. developed NCL nanosuspension in the presence of stabilizers of Tween-20 and Tween-80 for pulmonary administration by commercial nebulizers [16]. However, common issues of inconvenience and time consumption of using nebulizers, as well as the risk of cross infection largely jeopardize patient compliance [17]. To avoid those problems, dry powder inhaler (DPI) provides an alternative and promising means of pulmonary NCL delivery, due to its advantages of drug stability, dosing precision and portability [18,19]. Most of current DPI products are carrier-based, comprising fine micronized active pharmaceutical ingredients (APIs) and coarse carrier powders, typically lactose monohydrates. Nevertheless, the nominal dose of the carrier-based DPI is extremely low (from few μg to mg per dose), which could not satisfy the requirement of pulmonary administration of high dose APIs such as antibiotics，antibacterial and antiviral drugs [19,20]. To pursue high drug doses, carrier-free DPI, especially the API-embedded dry powder for inhalation (AeDPI) has advanced remarkably in recent decades [21], which can be fabricated through bottom-up particle engineering techniques, such as spray drying [22,23], spray freeze drying [24], thin-film freezing drying [25] and routine freeze drying [26]. Tse et al. fabricated levofloxacin embedded dry powder for inhalation (LeDPI) using phytoglycogen (PyG) and leucine as excipient matrix via spray drying, and apparently improved drug mobility and distribution [27]. Yu et al. developed the spray freeze dried (SFD) curcumin-chitosan nanocomplex embedded dry powder in mannitol/leucine mixed matrix. The obtained SFD microparticles maintained highly spherical morphology, and thus possessed outstanding flowability [28].
Regarding to NCL, Jara et al. developed NCL-embedded DPI (NeDPI) by TFFD tert-butanol (TBA)-based NCL solution. The maximum fine particle fraction (FPF) reached 86.0%, however, effective drug deposition to the respiratory tract was still quite low as the drug loading was below 22%, limited by poor solubility of NCL in TBA-based solvent, which may impede its actual efficacy [14]. Apart from using toxic solvents, TFFD process is high energy-consuming due to the low heat transfer efficiency, detrimental to large-scale production of AeDPI. Comparably, spray freeze drying enables to enhance drying efficiency and amplify the production yield and capacity, and meanwhile the feed solution for spray freeze drying could be non-toxic aqueous-based suspension containing dispersed nano- or micro-sized hydrophobic API. Up to now, liquid nitrogen (LN2) is the most widely used refrigerant for spray freeze drying [29,30], whereas some issues derived from using LN2 must be considered for fabricating AeDPI. Using LN2 would cause increased operation cost and risk, and also restrict the production scale. Additionally, the ultralow and unchangeable temperature (−196 °C) of LN2 could be not favorable or necessary for obtaining desirable structural features of the final particles. Also importantly, Leidenfrost phenomenon would occur when the atomized droplets directly contacting LN2, causing droplet random fast motion, collisions among droplets and varied heat transfer behaviors of droplets [31], consequently resulting in uncontrollable particle structure. To avoid such drawbacks of using LN2, a self-designed micro-fluidic jet spray freeze tower (MFJSFT) with cooled air as the freezing medium has been recently developed [32,33], which can realize controllable freezing temperature as low as −90 °C via tuning the temperature and flow rate of the refrigerant cycling in the interlayer of the tower, enabling fabrication of AeDPI with tailorable characteristics, e.g. size, morphology, porosity, surface composition, etc.
This study aims to fabricate high dose NeDPI with good aerosol performance via MFJSFT for the first time. To the best of our knowledge, there have not been reported studying SFD NeDPI. Tween-80 and 1,2-distearoyl-sn-glycero-3-phosphocholine (DSPC) as FDA approved excipients for inhalable formulations [34] were used to achieve a stable NCL aqueous nanosuspension for spray freeze drying. The suspension stability was evaluated according to the size and morphology of primary NCL nanoparticles characterized using laser diffraction and inverted phase contrast microscopy. Leucine was also added as the excipient to potentially improve the physical stability and aerosolization performance of NeDPI. The effects of the formulation and freezing temperature were investigated on the physicochemical properties, including size, morphology, crystal property and flowability of the obtained samples characterized by scanning electron microscopy (SEM), X-ray powder diffraction (XRPD), attenuated total reflection Flourier transformed infrared spectroscopy (ATR-FTIR) and FT-4 powder rheometry. The aerosol performance was evaluated by next-generation impactor (NGI), and the drug concentration was determined by high-performance liquid chromatography (HPLC).
2 Material and methods
2.1 Materials
Niclosamide (NCL) was purchased from Shanghai Yuanye Bio-Technology Co., Ltd. (Shanghai, China). 1,2-Dioctadecanoyl-sn-glycero-3-phophocholine (DSPC), polysorbate 20 (Tween-20) and polysorbate 80 (Tween-80) were purchased from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). L-Leucine (Leu) and methanol were obtained from Sigma–Aldrich (Saint Louis, USA). Acetonitrile (ACN) and formic acid (FA) for HPLC analysis were purchased from Fisher Scientific Co., Ltd. (Pittsburgh, PA, USA) and Aladdin Co., Ltd. (Shanghai, China), respectively. Distilled water (18.2–18.5 MΩ·cm) from a Milli-Q apparatus (Merck, Darmstadt, Germany) was used throughout the experiment.
2.2 Preparation of the NCL nanosuspension
Typically, Tween-80 (0.16 g) was placed into a beaker containing 96 g of ultrapure water. After processing by an ultrasonic processor (NE-500Z, Fangxu Technology Co., Ltd., Shanghai, China) at a frequency of 26 kHz for 5 mins, other substances of DSPC (0.32 g), L-leucine (0.32 g) and NCL (3.20 g) were added and processed by ultrasonication again. To enhance the dispersity of primary NCL nanocrystals (pNNc), the ultrasonically processed suspensions were further homogenized by high-pressure homogenization (AH-Basic Plus, ATS Engineering Co., Ltd., Suzhou, China) under 600 bar for 30 mins. The obtained NCL nanosuspension, namely, NxLyDzTu, where N, L, D and T refer to NCL, Leu, DSPC and Tween-80, respectively, and x:y:z:u is the mass ratio of the corresponding ingredient. Other NCL nanosuspensions, as listed in Table 1 , were prepared using the same procedure. (See Table 2 .)Table 1 Summary of NCL suspensions for spray freeze drying.
Table 1Sample Freezing temp. (°C) NCL (w/w %) Leucine (w/w %) DSPC (w/w %) Tween-80 (w/w %) Water (w/w %) Solid content (w/w %) Drug loading (%)
SF−40D-N20T1 −40 3.81 – – 0.19 96 4 95.24
SF−40D-N20D1T1 3.64 0.18 0.18 90.91
SF−40D-N20D2T1 3.48 0.35 0.17 86.96
SF−40D-N20D4T1 3.20 0.64 0.16 80.00
SF−40D-N20L2D2T1 3.20 0.32 0.32 0.16 80.00
SF−40D-N20L8D2T1 2.58 1.03 0.26 0.13 64.52
SF−80D-N20D2T1 −80 3.48 – 0.35 0.17 86.96
SF−80D-N20L2D2T1 3.20 0.32 0.32 0.16 80.00
SF−80D-N20L8D2T1 2.58 1.03 0.26 0.13 64.52
Table 2 Summary of flowability parameters of SFD NCL-embedded dry powder formulations (n = 3; mean ± SD).
Table 2Sample BFE (mJ) SI FRI SE (mJ/g)
SF−40D-N20T1 48.35 ± 2.66 0.90 ± 0.05 2.29 ± 0.54 13.06 ± 4.71
SF−40D-N20D2T1 10.54 ± 3.68 0.85 ± 0.10 2.56 ± 0.30 5.40 ± 0.79
SF−40D-N20L2D2T1 3.22 ± 0.74 1.35 ± 0.06 2.60 ± 1.13 8.40 ± 2.80
SF−40D-N20L8D2T1 3.43 ± 0.86 1.31 ± 0.17 2.33 ± 0.83 10.36 ± 1.95
SF−80D-N20D2T1 9.32 ± 3.01 0.94 ± 0.13 2.54 ± 0.45 8.67 ± 2.80
SF−80D-N20L2D2T1 4.65 ± 1.38 0.84 ± 0.07 2.32 ± 0.22 5.63 ± 1.67
SF−80D-N20L8D2T1 5.51 ± 0.94 0.88 ± 0.27 2.23 ± 0.56 5.56 ± 1.17
2.3 Characterization of the NCL nanosuspension
The particle size and morphology of pNNc dispersed in the nanosuspension were analyzed by the laser diffraction method (LA960, HORIBA Ltd., Kyoto, Japan) and inverted phase contrast microscopy (MI52-N, MSHOT photoelectric technology Co. Ltd., Guangzhou, China) equipped with a high-speed camera (MS23, MSHOT photoelectric technology Co. Ltd., Guangzhou, China), respectively. Regarding to the laser diffraction method, distilled water was used as the dispersion media, and the nanosuspension was dropped into the media under rotation of a paddle for analyzing.
2.4 Fabrication of NCL-embedded dry powder for inhalation (NeDPI)
A syringe containing 50 mL of NCL nanosuspension was fixed on a syringe pump (SPLab01, Shenchen Precision Pump Co., Ltd., Baoding, China) with a flow rate of 5 mL∙min−1. Then, the precursor liquid was atomized into the self-designed MFJSFT (Fig. 1 ) via a self-designed ultrasonic atomizing nozzle. The freezing temperature of the MFJSFT was set as −40 or − 80 °C. Frozen ice spheres were collected at the bottom of the MFJSFT by a special collection cup, which was directly transferred into a freeze dryer (DGJ-30H, Shanghai Bodeng Biological Science and Technology Co., Ltd., Shanghai, China). Primary drying was carried out at −40, −20 and − 10 °C for 12 h at each temperature, and then secondary drying was carried out at 0, 10 and 20 °C for 8 h at each temperature. The pressure was maintained below 10 Pa during the lyophilization process. Finally, the collected powders were stored in a desiccator at room temperature for further characterization.Fig. 1 The photographs (a) and schematic graph (b) of the micro-fluidic jet spray freeze tower (US10436493B2, US1033776B2).
Fig. 1
2.5 Scanning electron microscopy (SEM)
The microparticle morphology of the SFD NeDPI was characterized by SEM (Hitachi Ltd., Japan) with an accelerating voltage of 15 kV and working distance of ~12 mm. Before observation, powder was fixed on an aluminum stub by conducting carbon tape and sputter-coated with platinum at 20 mA for 1 min to produce a conductive surface via a sputter coater (MC1000, Hitachi High Technologies Corporation, Japan).
2.6 Powder rheometer
The powder flowability was assayed by powder rheometer (FT-4, Freeman Technology, Gloucestershire, UK) equipped with a 25 mm vessel and a 23.5 mm blade. Briefly, the powders were freely poured into the vessels and pretreated under the specific programs by the blade. The test was simply a combination of seven conditioning conditions (blade tip speed of 100 mm/s) for the stability test and four conditioning conditions (blade tip speed of 100, 70, 40 and 10 mm/s) for the variable flow rate test. The basic flowability (BFE), stability index (SI), specific energy (SE) and flow rate index (FRI) were calculated according to Eq. (1), (2), (3), (4).(1) BFE=ETest7mJ
(2) SI=ETest7ETest1
(3) SE=ETest6+ETest72∙msamplemJ/g
(4) FRI=ETest11ETest8
where, E Testn and m sample are the energy of test number (n, 1– 11) and mass of sample, respectively.
2.7 X-ray powder diffraction (XRPD)
The crystallographic analysis of SFD NCL powders was conducted using a desktop X-ray diffractometer (D2 PHASER, Bruker, Germany) equipped with a Cu Kα radiation source (λ = 1.54056 Å, 30 kV, 10 mA). The 2-theta angle was set at 5–90°, with 0.05° per step and 0.3 s per step.
2.8 Attenuated total reflection Flourier transformed infrared spectroscopy (ATR-FTIR)
ATR-FTIR measurements were carried out using an infrared spectrometer (Nicolet™ iS50R, Thermo Fisher Scientific, Pittsburgh, PA, USA). For each sample, 32-scan interferograms were collected at room temperature in the range of 4000 to 400 cm−1 with a spectral resolution of 4 cm−1.
2.9 High-performance liquid chromatography (HPLC) analysis
The NCL content was quantified at 331 nm via an HPLC system (Prominence-i LC-2030C Plus, Shimadzu, Kyoto, Japan) with a Welch Ultimate® LP-C18 column (4.6 × 250 mm, 5 μm, Welch Materials Inc., Jinhua, China) at a 1.2 mL∙min−1 flow rate. The column temperature was set as 35 °C. An isocratic method was used with a mobile phase consisting of FA aqueous solution (0.3%, v/v) and ACN mixed in a volume ratio of 20:80. The calibration curve of NCL was linear over the concentration range of 0.025–100 μg∙mL−1 (R2 > 0.9999).
2.10 Laser diffraction (LD)
The geometric diameter of aerosolized SFD-derived NCL powders was determined by laser diffraction (HELOS-INHALER, SYMPATEC GmbH, Germany) with a 5 mW He-Ne laser and wavelength of 632.8 nm. Size #3 hydroxypropyl methylcellulose (HPMC) capsules (Capsugel, Suzhou, China) containing a certain amount of powder were loaded into Breezhaler®. Empty capsule was punctured and used as control and the flow rate was set as 60 L/min.
2.11 In vitro aerosol performance
The in vitro aerosol performance of inhalable NCL dry powder was evaluated by next generation impactor (NGI, Copley Scientific, Nottingham, UK) connected with two high-capacity pumps (KRX5-P-V-01, ORION Machinery Co., Ltd., Nagano-ken, Japan) in parallel, and a critical flow controller (TPK 2000, Copley Scientific, Nottingham, UK). Briefly, 2–3 mg of powder was loaded into a size #3 HPMC capsule and fixed into the chamber of Breezhaler®. The flow rate was set as 60 L/min for 4 s per actuation, corresponding to a total flow volume of 4 L. Meanwhile, a solution of Tween-20 in methanol (1.5%, w/v) was applied and air dried onto the NGI collection cups to prevent the bouncing of particles. After actuation, the powder on the collection cups was washed with a mixture of water/acetonitrile (20:80, v/v). Then, the content of NCL was quantitated by HPLC. The fine particle fraction (FPF, the fraction of particles with aerodynamic size < 5 μm), extra fine particle fraction (eFPF, the fraction of particles with aerodynamic size < 2 μm), mass median aerodynamic diameter (MMAD) and geometric size distribution (GSD) were analyzed by Copley Inhaler Testing Data Analysis Software (CITDAS, version 3.10, Copley Scientific, Nottingham, UK).
3 Results and discussion
3.1 Preparation of NCL nanosuspensions
The development of formulations was carried out to optimize the dispersity of nanosuspensions suitable for spray freeze drying. Tween-80 and DSPC were chosen as the dispersants and used at different ratios by weight to NCL (Table 1). Fig. 2A and B show the microscope images and size distribution of the pNNc dispersed in the aqueous suspension, respectively. In absence of any dispersant, pNNc showed D50 of 25.87 ± 1.28 μm (Fig. 2B-control, Table S1), and rapidly re-aggregated and settled after ultrasonication (Fig. S1). By solely adding Tween-80 (N20T1), after ultrasonication the suspended pNNc exhibited the reduced D50 of 11.40 ± 0.28 μm, but were still obviously aggregated (Fig. 2A-a1). After further homogenization, most of the pNNc aggregates disassembled (Fig. 2A-a2), showing D50 of 0.81 ± 0.04 μm with a bimodal size distribution (Fig. 2B-a2, Table S1). It indicated the positive effect of the two-step processing on improving the dispersity of pNNc, and yet Tween-80 could not sufficiently suppress the pNNc aggregation. By introducing DSPC into the Tween-80 stabilized suspension, NCL nanosuspension showed a better dispersity, and increasing the DSPC amount could further facilitate dispersing the pNNc (Fig. 2A-b-d). It was evidenced by the gradually decreased D50 and extremely narrow size distribution of the pNNc after ultrasonication and homogenization, i.e. 0.29 ± 0.03 μm for N20D1T1, 0.21 ± 0.02 μm for N20D2T1 and 0.18 ± 0.00 μm for N20D4T1 (Fig. 2B-b-d, Table S1). Besides, the addition of leucine did not significantly affect the state of the suspensions (Fig. 2A-e, f), while D50 of the pNNc maintained 0.16 ± 0.00 μm for both N20L2D2T1 and N20L8D2T1 (Fig. 2B-e, f, Table S1).Fig. 2 Microscope images (scale bar = 100 μm) (A) and particle size distribution (B) of NCL suspension after ultrasonic treatment (1 & broken line) followed by further homogenization (2 & solid line): N20T1 (a), N20D1T1 (b), N20D2T1 (c), N20D4T1 (d), N20L2D2T1 (e) and N20L8D2T1 (f).
Fig. 2
Scheme 1 illustrates the possible underlying mechanism for the dispersity of the NCL nanosuspension. From the point of thermodynamics, the Gibbs free energy change (ΔG) of a suspension system can be given by Eq. (5):(5) ∆G=γ∆A−T∆S
where γ, ΔA, T and ΔS are the surface tension, change in surface area, absolute temperature and change in entropy, respectively [35]. Without dispersant, an increasing number of small nanocrystals with enlarged ΔA appeared when processing the suspension via ultrasonication and homogenization. The ΔG increased due to the enlarged ΔA, probably resulting in the re-aggregation of pNNc and disruption of suspension stability, as indicated by Fig. 2B-control. The addition of appropriate dispersants to decrease ΔG and meanwhile avoiding the particle re-aggregation can be a feasible means to improve suspension stability. When using the surfactant Tween-80, its hydrophobic segment could partially occupy the surface of suspended NCL particles, while the hydrophilic segment extended in the aqueous medium to reduce the interfacial tension of the suspended nanocrystals [36,37]. As a result, the nanosuspension dispersity was improved. After adding DSPC into the Tween-80 stabilized suspension, DSPC could penetrate the voids of pNNc aggregates and adsorb on the nanocrystals surface, induced by the hydrogen bonds between the -NH group of NCL and the C <svg xmlns="http://www.w3.org/2000/svg" version="1.0" width="20.666667pt" height="16.000000pt" viewBox="0 0 20.666667 16.000000" preserveAspectRatio="xMidYMid meet"><metadata> Created by potrace 1.16, written by Peter Selinger 2001-2019 </metadata><g transform="translate(1.000000,15.000000) scale(0.019444,-0.019444)" fill="currentColor" stroke="none"><path d="M0 440 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z M0 280 l0 -40 480 0 480 0 0 40 0 40 -480 0 -480 0 0 -40z"/></g></svg> O and PO2 − groups of DSPC, which could facilitate dispersing the pNNc [38].Scheme 1 Schematic illustration of the dispersed NCL suspension in the presence of Tween-80 or/and DSPC.
Scheme 1
3.2 Microparticle morphology of SFD NeDPI
Fig. 3 displays the morphology of the commercially available raw NCL and SFD NeDPI. The raw NCL was presented as aggregates of large crystals with irregular shapes (Fig. 3a). Generally, in order to obtain the intact and disperse microparticles embedding the pNNc, the binding force among the pNNc within the individual microparticles needs to be sufficiently strong, and agglomeration of the microparticles needs to be avoided [28]. Especially for the SFD microparticles, it is even more difficult due to their high porosity induced poor structure stability and ease of moisture absorption. Tween-80 was thus selected, as it could act like “glue” to strengthen the packing of pNNc within the SFD microparticle. However, by solely adding Tween-80 (SF−40D-N20T1), the SFD sample appeared loose and lump morphology with agglomerated particulate and needle-like crystals (Fig. 3b). It could be ascribed to the poor stability of the feed suspension (Fig. 2B-a2) as Tween−80 was unable to prevent the aggregation of pNNc, thus probably resulting in the phase separation between NCL particles and ice crystals during freezing process. In addition, the strong hygroscopicity of both NCL and Tween-80 could induce agglomeration of the SFD sample due to the moisture absorption.Fig. 3 SEM images of raw NCL powder (a) and SFD NCL-embedded dry powders of different formulations and freezing temperatures: SF−40D-N20T1 (b), N20D1T1 (c), N20D2T1 (d), N20D4T1 (e), N20L2D2T1 (f), N20L8D2T1 (g) and SF−80D-N20D2T1 (h), N20L2D2T1 (i), N20L8D2T1 (j).
Fig. 3
Therefore, DSPC as an inhalable surfactant ingredient [39] was further added, but small additive amount of DSPC (SF−40D-N20D1T1) seemed still unable to either fabricate intact individual microparticles or avoid the microparticle agglomeration (Fig. 3c). By adding more DSPC (SF−40D-N20D2T1), spherical and porous microparticles with intact structure and good dispersity were obtained (Fig. 3d), revealing the apparently positive effect of DSPC. Excessive DSPC (SF−40D-N20D4T1) however, led to loose and lump morphology (Fig. 3e). It was speculated that excessive DSPC could assemble with Tween-80 into vesicles [40], which may weaken the ‘glue’ effect of Tween-80, thus jeopardizing the packing strength among the NCL particles. Hence, during sublimation of ice crystals, the pNNc could not support the integrity of spherical microparticle that consequently collapsed and piled up to form a lump.
In addition, the formulation containing leucine generated intact and porous microparticles with spherical shapes and some fragments, i.e. SF−40D-N20L2D2T1 (Fig. 3f). The microparticle integrity was further improved as the leucine content increased, i.e. SF−40D-N20L8D2T1 (Fig. 3g). It could be ascribed that hydrophobic leucine was inclined to occupy the surface of atomized droplets, with their hydrophobic isobutyl group pointing toward the air and their hydrophilic amino group and carboxyl group oriented toward the water [41,42]. And the leucine would quickly precipitate and remain on the microparticle surface during freezing process, thus inhibiting the growth and diffusion of the ice crystals [43,44], ultimately obtaining intact SFD microparticles.
Decreasing the freezing temperature to −80 °C led to porous microparticles with better integrity and sphericity, and more open pores on the microparticles surface (Fig. 3h-j). Because the lower freezing temperature resulted in higher supercooling degree, hydrophobic leucine and/or DSPC had less time to migrate to the gas-liquid interface during the freezing process, and consequently formed a homogeneous phase containing the pNNc together with the ice crystals. After freeze drying, more pores on the microparticle surface were left by ice sublimation, and the well-dispersed leucine could also protect the microparticles from absorbing water, facilitating their sphericity and dispersity. Moreover, compared to dendritic and lamellar pore structure attained at − 40 °C (Fig. 3 d3), cellular pore structure was found on the surface of SF−80D-N20D2T1 microparticles (Fig. 3 h3), possible ascribing to the rapid growth regime with high nucleation site density [45]. Surprisingly, such phenomenon was not observed in those samples containing Leu (Fig. 3 i3 & j3), which could be explained by the rapid precipitation of Leu. The surface enriched Leu could occupy the nucleation site of ice crystals, thus inducing slower crystal growth and consequently lamellar pores.
3.3 Crystal properties of SFD NeDPI
Crystallized NCL is commonly identified as anhydrate, monohydrate forms of HA and HB. Fig. 4A shows the XRPD patterns of the raw NCL and SFD NeDPI. Raw NCL showed intense diffraction peaks located at approximately 13.5° for anhydrate and 10.5° for HB [46], indicating a mixture of anhydrate and HB (Fig. 4A-a). Comparably, new diffraction peaks at 9.4° and 11.4° corresponding to HA [46] were observed for the microparticles containing either Tween-80 (Fig. 4A-b) or Tween-80/DSPC (Fig. 4A-c-g), indicating the formation of HA after granulation. Compared to SF−40D-N20D2T1 (Fig. 4A-d), the relative intensity between the feature peaks of 11.4° and 13.5° significantly increased for the samples of SF−40D-N20L2D2T1 (Fig. 4A-f) and -N20L8D2T1 (Fig. 4A-g), suggesting that more HA formation could be induced by the addition of leucine. Moreover, at a freezing temperature of −80 °C, the lack of double peaks at approximately 13.5° for anhydrate NCL but remarkable signals at 9.4 and 11.4° in the XRPD patterns (Fig. 4A-h-j) revealed that the obtained NeDPI were predominated by HA.Fig. 4 XRPD pattern (A) and IR spectra (B) of raw NCL (a) and SFD NCL-embedded dry powder formulations: SF−40D-N20T1 (b), N20D1T1 (c), N20D2T1 (d), N20D4T1 (e), N20L2D2T1 (f), N20L8D2T1 (g) and SF−80D-N20D2T1 (h), N20L2D2T1 (i), N20L8D2T1 (j). The ♣, ● and ▼ represent the anhydrate, monohydrate HA and monohydrate HB crystal form of NCL, respectively.
Fig. 4
To investigate the NCL crystal transformation under different conditions, ATR-FTIR was utilized for further characterization. As shown in Fig. 4B, the feature peak of 1651 cm−1 attributed to the CO stretching vibration of anhydrate NCL [47] was observed for the raw NCL (Fig. 4B-a). After adding Tween-80, DSPC and leucine, it shifted to 1680 cm−1, mainly due to the formation of HA (Fig. 4B-b-g) [46,47]. Moreover, the feature peaks at 1651 cm−1 completely disappeared when lowering the freezing temperature to −80 °C, suggesting the absence of the anhydrate NCL (Fig. 4B-h-j). Those FTIR results were all in good agreement with XRPD ones. Generally, NCL molecules tended to interact with water molecules spontaneously via hydrogen bonds during the preparation of NCL nanosuspensions, which could exert considerable effect on the growth of NCL crystals. During the following SFD process, the hydrophobic segments of surfactants of Tween-80 and DSPC, as well as leucine molecules could be enriched on the SFD microparticle surface and inhibited the desorption of water molecules, thereby reducing the dehydration of HA. With regards to the effect of freezing temperature, more ice crystal domains were created at extremely low freezing temperatures, meanwhile resulting in ultranarrow channels that could in turn inhibit the ice sublimation, so that HA was preferably formed with more water molecules bonding to the NCL crystals.
3.4 In vitro aerosol performance of SFD NeDPI
The aerosol performance of NeDPI was evaluated by NGI. As shown in Table 3 , all SFD samples showed a high EF above 90%, indicating that most of the microparticles were emitted efficiently from the device, probably resulted from the low electrostatic force between the NeDPI and the device. The FPF of SF−40D-N20T1 and SF−40D-N20D1T1 was about 22.54% and 17.93%, respectively, much lower than SF−40D-N20D2T1, i.e. ~ 37.32%, implying that the spherical morphology of microparticle could play an important role in pursuing better aerosol performance. It could be ascribed to low cohesive force among spherical microparticles and their good flowability, indicated by the apparently lower BFE value of SF−40D-N20D2T1 (~ 10.54 mJ) than that of SF−40D-N20T1 (~ 48.35 mJ). Furthermore, SF−40D-N20D2T1 was of much smaller D50 (~ 13.19 μm) than SF−40D-N20D1T1 (~ 21.55 μm) and SF−40D-N20T1 (~ 40.84 μm) characterized under the inhalation condition, also resulted from better dispersity of the spherical microparticles than that of the agglomerates. Not surprisingly, SF−40D-N20D4T1 was of fairly low FPF (~ 19.15%) due to its lump morphology (Fig. 3e).Table 3 Summary of the aerodynamic properties of SFD NCL-embedded dry powder formulations (n = 3; mean ± SD).
Table 3Sample D50 (μm) MMAD (μm) Emitted fraction (%, of recovered dose) Fine particle fraction (%, of recovered dose) Extra fine particle fraction (%, of recovered dose) Fine particle dose⁎ (mg, of 10 mg powder) GSD (μm)
SF−40D-N20T1 40.84 ± 10.90 NA 90.71 ± 5.95 22.54 ± 6.60 4.13 ± 1.19 2.15 ± 0.63 NA
SF−40D-N20D1T1 21.55 ± 0.23 7.67 ± 0.48 95.19 ± 0.13 17.93 ± 0.19 3.60 ± 0.16 1.63 ± 0.02 2.58 ± 0.08
SF−40D-N20D2T1 13.19 ± 0.36 4.74 ± 0.57 91.46 ± 4.21 37.32 ± 2.19 7.69 ± 0.41 3.25 ± 0.19 2.08 ± 0.23
SF−40D-N20D4T1 32.36 ± 3.63 NA 93.54 ± 1.83 19.15 ± 3.05 4.52 ± 0.74 1.53 ± 0.24 NA
SF−40D-N20L2D2T1 20.16 ± 1.83 4.92 ± 0.24 91.13 ± 2.48 42.01 ± 1.00 9.29 ± 0.38 3.36 ± 0.08 2.13 ± 0.05
SF−40D-N20L8D2T1 16.14 ± 1.47 4.89 ± 0.07 90.23 ± 0.95 42.53 ± 0.59 9.21 ± 0.51 2.74 ± 0.04 2.10 ± 0.03
SF−80D-N20D2T1 18.67 ± 1.95 4.80 ± 0.33 90.59 ± 2.14 38.39 ± 4.41 9.31 ± 1.35 3.34 ± 0.38 2.16 ± 0.06
SF−80D-N20L2D2T1 20.53 ± 0.59 4.29 ± 0.18 90.32 ± 0.61 47.83 ± 3.18 11.80 ± 1.12 3.82 ± 0.25 2.05 ± 0.07
SF−80D-N20L8D2T1 24.85 ± 0.95 4.65 ± 0.31 93.01 ± 0.44 47.29 ± 1.77 13.01 ± 1.80 3.05 ± 0.11 2.24 ± 0.05
⁎ Fine particle dose (mg, of 10 mg powder) = Fine particle fraction (%, of recovered dose) × Drug loading (%) × 10 (mg).
The addition of leucine (SF−40D-N20L2D2T1) led to enhanced FPF (~ 42.01%) with more extra fine particles deposited in Stage-5 and below (eFPF ~ 9.29%) (Fig. 5 ). It could be ascribed that the leucine enriched on the surface could form a hydrophobic film that could inhibit the microparticle hygroscopicity [41], which would further improve the dispersity of the microparticles (BFE ~ 3.22 mJ) superior to SF−40D-N20D2T1, consequently leading to better aerosol performance. However, SF−40D-N20L8D2T1 was of quite similar FPF (~ 42.53%) and eFPF (~9.21%) with almost unchanged BFE (~ 3.43 mJ), possibly representing the threshold of the positive effect of leucine on the microparticle aerosol performance.Fig. 5 In vitro deposition profile of SFD NCL-embedded dry powder formulations (n = 3; mean ± SD).
Fig. 5
Apart from the formulation optimization, the spray freezing temperature was adjusted to pursue the better aerosol performance. By using the freezing temperature of −80 °C, the sample FPF was elevated with more extra fine particles deposited in Stage-5 compared to the counterpart fabricated at −40 °C. The improvement of aerosol performance influenced by the freezing temperature needs to be further investigated in the future work. Increasing leucine content (N20L8D2T1) seemed also not to improve the aerosol performance of the NeDPI fabricated under −80 °C, implying that the threshold of the positive effect of leucine may not depend on the freezing temperature. Considering the effective API content delivered to the deep lung per dose, SF−80D-N20L2D2T1 was the optimal sample of the highest FPF (~ 47.83%) with the drug loading up to 80%, equivalently 38.26% per dose of API, i.e. >3.8 mg NCL that could reach the deep lung when inhaling 10 mg dry powders. Meanwhile, the eFPF of SF−80D-N20L2D2T1 could reach ~ 11.80%, meaning that 11.80% of NCL may be delivered to the alveoli, possibly achieving maximum efficacy.
4 Conclusion
In this work, highly respirable NCL-embedded microparticles were successfully fabricated by a novel spray freeze drying technology. The dispersity of suspended pNNc in the precursor liquid was apparently improved by adding Tween-80 together with DSPC, possibly due to the hydrogen bonds formed between the -NH group of NCL and the CO and PO2 − groups of DSPC. When the mass ratio of DSPC to Tween-80 was 2:1 and freezing temperature of −40 °C was used, the SFD microparticles (SF−40D-N20D2T1) appeared spherical, whereas less or excessive amount of DSPC led to agglomerates of loose and lump morphology. The addition of leucine (SF−40D-N20L2D2T1 and SF−40D-N20L8D2T1) could facilitate the integrity and sphericity of SFD microparticles, inducing the excellent flowability as revealed by the extremely low BFE. By using the lower freezing temperature of −80 °C, the SFD microparticles (SF−80D-N20L2D2T1 and SF−80D-N20L8D2T1) were also spherical but with more open pores on the surface, and the NCL crystal form of HA was obtained in comparison with the mixture of anhydrate and HA form of SF−40D samples. It indicated that both the formulation and freezing temperature played important roles in the SFD microparticle morphology and crystal forms. The FPF of different samples varied from about 17.93% to 47.83%, where not surprisingly the spherical and dispersed SFD microparticles showed better aerosol performance. To pursue high API dose effectively delivered, SF−80D-N20L2D2T1 was optimal, i.e., >3.8 mg NCL that could reach the deep lung when inhaling 10 mg dry powders. This study demonstrated a promising strategy to develop DPI products for high dose pulmonary drug delivery via a novel spray freeze drying technology.
CRediT authorship contribution statement
Shengyu Zhang: Conceptualization, Methodology, Validation, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Shen Yan: Conceptualization, Methodology, Validation, Investigation, Data curation, Writing – original draft, Writing – review & editing, Visualization. Kangwei Lu: Data curation, Validation, Investigation. Shixuan Qiu: Validation, Visualization. Xiao Dong Chen: Methodology, Supervision. Winston Duo Wu: Conceptualization, Methodology, Writing – review & editing, Supervision, Project administration, Funding acquisition.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supplementary data
Supplementary material
Image 1
Data availability
No data was used for the research described in the article.
Acknowledgements
This research was funded by the 10.13039/501100001809 National Natural Science Foundation of China (No. 21878197). The support from the Priority Academic Program Development (PAPD) of Jiangsu Higher Education Institutions is appreciated. We also thank Capsugel (Suzhou, China) for kindly supplying the HPMC capsules.
Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.powtec.2022.118168.
==== Refs
References
1 Johns Hopkins Coronavirus Resource Center (CRC) COVID-19 Data Dashboard https://coronavirus.jhu.edu/ 2022 (accessed 16 October 2022)
2 Gautret P. Lagier J.C. Parola P. Hoang V. Meddeb L. Mailhe M. Doudier B. Courjon J. Giordanengo V. Vieira V.E. Dupont H.T. Honore S. Colson P. Chabriere E. La Scola B. Rolain J.M. Brouqui P. Raoult D. Hydroxychloroquine and azithromycin as a treatment of COVID-19: results of an open-label non-randomized clinical trial Int. J. Antimicrob. Agents 56 2020 105949 10.1016/j.ijantimicag.2020.105949
3 Wang M.L. Cao R.Y. Zhang L.K. Yang X.L. Liu J. Xu M.Y. Shi Z.L. Hu Z.H. Zhong W. Xiao G.F. Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro Cell Res. 30 2020 269 271 10.1038/s41422-020-0282-0 32020029
4 Cai Q.X. Yang M.H. Liu D.J. Chen J. Shu D. Xia J.X. Liao X.J. Gu Y.B. Cai Q. Yang Y. Shen C.G. Li X.H. Peng L. Huang D.L. Zhang J. Zhang S.R. Wang F.X. Liu J.Y. Chen L. Chen S.Y. Wang Z.Q. Zhang Z. Cao R.Y. Zhong W. Liu Y.X. Liu L. Experimental treatment with Favipiravir for COVID-19: an open-label control study Engineering 6 2020 1192 1198 10.1016/j.eng.2020.03.007 32346491
5 Zeng H. He X. Liu W. Kan J. He L. Zhao J. Chen C. Zhang J. Chen S. Antiviral Abidol is associated with the reduction of in-hospital mortality in COVID-19 patients Cardiol. Discov. 1 2021 37 43 10.1097/cd9.0000000000000014 34977907
6 Beigel J.H. Tomashek K.M. Dodd L.E. Mehta A.K. Zingman B.S. Kalil A.C. Hohmann E. Chu H.Y. Luetkemeyer A. Kline S. de Castilla D.L. Finberg R.W. Dierberg K. Tapson V. Hsieh L. Patterson T.F. Paredes R. Sweeney D.A. Short W.R. Touloumi G. Lye D.C. Ohmagari N. Oh M.D. Ruiz-Palacios G.M. Benfield T. Fatkenheuer G. Kortepeter M.G. Atmar R.L. Creech C.B. Lundgren J. Babiker A.G. Pett S. Neaton J.D. Burgess T.H. Bonnett T. Green M. Makowski M. Osinusi A. Nayak S. Lane H.C. Grp A.-S. Remdesivir for the treatment of COVID-19-final report N. Engl. J. Med. 383 2020 1813 1826 10.1056/NEJMoa2007764 32445440
7 Wang Y.M. Zhang D.Y. Du G.H. Du R.H. Zhao J.P. Jin Y. Fu S.Z. Gao L. Cheng Z.S. Lu Q.F. Hu Y. Luo G.W. Wang K. Lu Y. Li H.D. Wang S.Z. Ruan S.N. Yang C.Q. Mei C.L. Wang Y. Ding D. Wu F. Tang X. Ye X.Z. Ye Y.C. Liu B. Yang J. Yin W. Wang A.L. Fan G.H. Zhou F. Liu Z.B. Gu X.Y. Xu J.Y. Shang L.H. Zhang Y. Cao L.J. Guo T.T. Wan Y. Qin H. Jiang Y.S. Jaki T. Hayden F.G. Horby P.W. Cao B. Wang C. Remdesivir in adults with severe COVID-19: a randomised, double-blind, placebo-controlled, multicentre trial Lancet 395 2020 1569 1578 10.1016/s0140-6736(20)31022-9 32423584
8 Monk P.D. Marsden R.J. Tear V.J. Brookes J. Batten T.N. Mankowski M. Gabbay F.J. Davies D.E. Holgate S.T. Ho L.P. Clark T. Djukanovic R. Wilkinson T.M.A. Inhaled interferon Beta, safety and efficacy of inhaled nebulised interferon beta-1a (SNG001) for treatment of SARS-CoV-2 infection: a randomised, double-blind, placebo-controlled, phase 2 trial Lancet Respir. Med. 9 2021 196 206 10.1016/s2213-2600(20)30511-7 33189161
9 Mantlo E. Bukreyeva N. Maruyama J. Paessler S. Huang C. Antiviral activities of type I interferons to SARS-CoV-2 infection Antivir. Res. 179 2020 4. 104811 10.1016/j.antiviral.2020.104811
10 Braga L. Ali H. Secco I. Chiavacci E. Neves G. Goldhill D. Penn R. Jimenez-Guardeno J.M. Ortega-Prieto A.M. Bussani R. Cannata A. Rizzari G. Collesi C. Schneider E. Arosio D. Shah A.M. Barclay W.S. Malim M.H. Burrone J. Giacca M. Drugs that inhibit TMEM16 proteins block SARS-CoV-2 spike-induced syncytia Nature 594 2021 88 93 10.1038/s41586-021-03491-6 33827113
11 Xu J.M. Shi P.Y. Li H.M. Zhou J. Broad Spectrum antiviral agent Niclosamide and its therapeutic potential ACS Infect. Dis. 6 2020 909 915 10.1021/acsinfecdis.0c00052 32125140
12 Li Y.H. Li P.K. Roberts M.J. Arend R.C. Samant R.S. Buchsbaum D.J. Multi-targeted therapy of cancer by niclosamide: a new application for an old drug Cancer Lett. 349 2014 8 14 10.1016/j.canlet.2014.04.003 24732808
13 Schweizer M.T. Haugk K. McKiernan J.S. Gulati R. Cheng H.H. Maes J.L. Dumpit R.F. Nelson P.S. Montgomery B. McCune J.S. Plymate S.R. Yu E.Y. A phase I study of niclosamide in combination with enzalutamide in men with castration-resistant prostate cancer PLoS One 13 2018 12 e0198389 10.1371/journal.pone.0198389
14 Jara M.O. Warnken Z.N. Sahakijpijarn S. Moon C. Maier E.Y. Christensen D.J. Koleng J.J. Peters J.I. Maier S.D.H. Williams R.O. Niclosamide inhalation powder made by thin-film freezing: multi-dose tolerability and exposure in rats and pharmacokinetics in hamsters Int. J. Pharm. 603 2021 120701 10.1016/j.ijpharm.2021.120701
15 Backer V. Sjobring U. Sonne J. Weiss A. Hostrup M. Johansen H.K. Becker V. Sonne D.P. Balchen T. Jellingso M. Sommer M.O.A. A randomized, double-blind, placebo-controlled phase 1 trial of inhaled and intranasal niclosamide: a broad spectrum antiviral candidate for treatment of COVID-19 Lancet Region. Health Europe 4 2021 100084 10.1016/j.lanepe.2021.100084
16 Costabile G. d'Angelo I. Rampioni G. Bondi R. Pompili B. Ascenzioni F. Mitidieri E. Bianca R. Sorrentino R. Miro A. Quaglia F. Imperi F. Leoni L. Ungaro F. Toward repositioning Niclosamide for Antivirulence therapy of Pseudomonas aeruginosa lung infections: development of inhalable formulations through Nanosuspension technology Mol. Pharm. 12 2015 2604 2617 10.1021/acs.molpharmaceut.5b00098 25974285
17 Ari A. Fink J.B. Recent advances in aerosol devices for the delivery of inhaled medications Expert Opin. Drug Deliv. 17 2020 133 144 10.1080/17425247.2020.1712356 31959028
18 Brunaugh A.D. Smyth H.D.C. Formulation techniques for high dose dry powders Int. J. Pharm. 547 2018 489 498 10.1016/j.ijpharm.2018.05.036 29778822
19 Scherliess R. Etschmann C. DPI formulations for high dose applications - challenges and opportunities Int. J. Pharm. 548 2018 49 53 10.1016/j.ijpharm.2018.06.038 29940300
20 Ke W.-R. Chang R. Yoon Kyung Chan H.-K. Engineering the right formulation for enhanced drug delivery Adv. Drug Deliv. Rev. 2022 114561 10.1016/j.addr.2022.114561 36191861
21 Ke W.R. Chang R.Y.K. Chan H.K. Engineering the right formulation for enhanced drug delivery Adv. Drug Deliv. Rev. 191 2022 114561 10.1016/j.addr.2022.114561
22 Alhajj N. O'Reilly N.J. Cathcart H. Designing enhanced spray dried particles for inhalation: a review of the impact of excipients and processing parameters on particle properties Powder Technol. 384 2021 313 331 10.1016/j.powtec.2021.02.031
23 Son Y.J. Miller D.P. Weers J.G. Optimizing spray-dried porous particles for high dose delivery with a portable dry powder inhaler Pharmaceutics 13 2021 32 1528 10.3390/pharmaceutics13091528
24 Yu S.H. Pu X.H. Ahmed M.U. Yu H.D.H. Mutukuri T.T. Li J. Zhou Q.T. Spray-freeze-dried inhalable composite microparticles containing nanoparticles of combinational drugs for potential treatment of lung infections caused by Pseudomonas aeruginosa Int. J. Pharm. 610 2021 9. 121160 10.1016/j.ijpharm.2021.121160
25 Watts A.B. Wang Y.B. Johnston K.P. Williams R.O. Respirable low-density microparticles formed in situ from aerosolized brittle matrices Pharm. Res. 30 2013 813 825 10.1007/s11095-012-0922-2 23229856
26 Mukhtar M. Szakonyi Z. Farkas A. Burian K. Kokai D. Ambrus R. Freeze-dried vs spray-dried nanoplex DPIs based on chitosan and its derivatives conjugated with hyaluronic acid for tuberculosis: in vitro aerodynamic and in silico deposition profiles Eur. Polym. J. 160 2021 110775 10.1016/j.eurpolymj.2021.110775
27 Tse J.Y. Kadota K. Imakubo T. Uchiyama H. Tozuka Y. Enhancement of the extra-fine particle fraction of levofloxacin embedded in excipient matrix formulations for dry powder inhaler using response surface methodology Eur. J. Pharm. Sci. 156 2021 105600 10.1016/j.ejps.2020.105600
28 Yu H. Tran T.T. Teo J. Hadinoto K. Dry powder aerosols of curcumin-chitosan nanoparticle complex prepared by spray freeze drying and their antimicrobial efficacy against common respiratory bacterial pathogens Coll. Surf. A Physicochem. Eng. Asp. 504 2016 34 42 10.1016/j.colsurfa.2016.05.053
29 Okuda T. Morishita M. Mizutani K. Shibayama A. Okazaki M. Okamoto H. Development of spray-freeze-dried siRNA/PEI powder for inhalation with high aerosol performance and strong pulmonary gene silencing activity J. Control. Release 279 2018 99 113 10.1016/j.jconrel.2018.04.003 29627404
30 Ito T. Fukuhara M. Okuda T. Okamoto H. Naked pDNA/hyaluronic acid powder shows excellent long-term storage stability and gene expression in murine lungs Int. J. Pharm. 574 2020 118880 10.1016/j.ijpharm.2019.118880
31 Song Y.S. Adler D. Xu F. Kayaalp E. Nureddin A. Anchan R.M. Maas R.L. Demirci U. Vitrification and levitation of a liquid droplet on liquid nitrogen Proc. Natl. Acad. Sci. U. S. A. 107 2010 4596 4600 10.1073/pnas.0914059107 20176969
32 Wu D. Zhang S. Liao Z. Wu Z. Xiao J. Chen X. Freezing Temperature Controllable Spray Freezing Tower for Preparing Micron-Sized Spherical Ice Particles 2019 US 10,436,493 B2
33 Wu D. Liao Z. Zhang S. Wu Z. Xiao J. Chen X. Double-Sealing Type Apparatus for Collecting Spray Freeze Ice Ball Particles and Collecting Method Thereof 2019 US 10,337,796 B2
34 U.S. Food and Drug Administration Inactive Ingredients Database Download https://www.fda.gov/drugs/drug-approvals-and-databases/inactive-ingredients-database-download 2022 (accessed 16 October 2022)
35 Muhammad R.S. Muhammad I. Shafi U. Lipid-Based Nanocarriers for Drug Delivery and Diagnosis First ed. 2017 William Andrew Oxford
36 Pandey K.U. Poornachary S.K. Dalvi S.V. Insights to the action of additives for stabilization of ultrafine particles of Fenofibrate in aqueous suspensions produced by Sonoprecipitation Powder Technol. 363 2020 310 325 10.1016/j.powtec.2019.12.059
37 Yu Q. Wu X.Y. Zhu Q.G. Wu W. Chen Z.J. Li Y. Lu Y. Enhanced transdermal delivery of meloxicam by nanocrystals: preparation, in vitro and in vivo evaluation, Asian J. Pharm. Sci. 13 2018 518 526 10.1016/j.ajps.2017.10.004
38 Miller D.P. Tan T. Tarara T.E. Nakamura J. Malcolmson R.J. Weers J.G. Physical characterization of tobramycin inhalation powder: I. Rational design of a stable engineered-particle formulation for delivery to the lungs Mol. Pharm. 12 2015 2582 2593 10.1021/acs.molpharmaceut.5b00147 26052676
39 Weers J. Tarara T. The PulmoSphere platform for pulmonary drug delivery Ther. Deliv. 5 2014 277 295 10.4155/tde.14.3 24592954
40 Lim W.H. Lawrence M.J. Influence of surfactant and lipid chain length on the solubilisation of phosphatidylcholine vesicles by micelles comprised of polyoxyethylene sorbitan monoesters Coll. Surf. A Physicochem. Eng. Asp. 250 2004 449 457 10.1016/j.colsurfa.2004.06.042
41 Chen Y.J. Yan S. Zhang S.Y. Yin Q.Y. Chen X.D. Wu W.D. Micro-fluidic spray freeze dried ciprofloxacin hydrochloride-embedded dry powder for inhalation AAPS PharmSciTech 23 2022 211 10.1208/s12249-022-02371-0 35915199
42 Alhajj N. O'Reilly N.J. Cathcart H. Leucine as an excipient in spray dried powder for inhalation Drug Discov. Today 26 2021 2384 2396 10.1016/j.drudis.2021.04.009 33872799
43 Capicciotti C.J. Leclere M. Perras F.A. Bryce D.L. Paulin H. Harden J. Liu Y. Ben R.N. Potent inhibition of ice recrystallization by low molecular weight carbohydrate-based surfactants and hydrogelators Chem. Sci. 3 2012 1408 1416 10.1039/c2sc00885h
44 Weng L.D. Stott S.L. Toner M. Molecular dynamics at the Interface between ice and poly(vinyl alcohol) and ice recrystallization inhibition Langmuir 34 2018 5116 5123 10.1021/acs.langmuir.7b03243 29199836
45 Kamarainen T. Kadota K. Tse J.Y. Uchiyama H. Yamanaka S. Tozuka Y. Modulating the pore architecture of ice-templated dextran microparticles using molecular weight and concentration Langmuir 38 2022 6741 6751 10.1021/acs.langmuir.2c00721 35579967
46 van Tonder E.C. Maleka T.S.P. Liebenberg W. Song M.N. Wurster D.E. de Villiers M.M. Preparation and physicochemical properties of niclosamide anhydrate and two monohydrates Int. J. Pharm. 269 2004 417 432 10.1016/j.ijpharm.2003.09.035 14706253
47 Sanphui P. Kumar S.S. Nangia A. Pharmaceutical cocrystals of niclosamide Cryst. Growth Des. 12 2012 4588 4599 10.1021/cg300784v
| 0 | PMC9746026 | NO-CC CODE | 2022-12-15 00:03:23 | no | Powder Technol. 2022 Dec 13;:118168 | utf-8 | Powder Technol | 2,022 | 10.1016/j.powtec.2022.118168 | oa_other |
==== Front
Appl Soft Comput
Appl Soft Comput
Applied Soft Computing
1568-4946
1872-9681
Elsevier B.V.
S1568-4946(22)00975-9
10.1016/j.asoc.2022.109926
109926
Article
Multi-objective automatic analysis of lung ultrasound data from COVID-19 patients by means of deep learning and decision trees
Custode Leonardo Lucio a
Mento Federico a
Tursi Francesco b
Smargiassi Andrea c
Inchingolo Riccardo c
Perrone Tiziano de
Demi Libertario a
Iacca Giovanni a⁎
a Dept. of Information Engineering and Computer Science, University of Trento, Italy
b UOS Pneumologia di Codogno, ASST Lodi, Lodi, Italy
c Dept. of Medical and Surgical Sciences, Fondazione Policlinico Universitario A. Gemelli IRCCS, Rome, Italy
d Dept. of Internal Medicine, IRCCS San Matteo, Pavia, Italy
e Emergency Dept., Humanitas Gavazzeni, Bergamo, Italy
⁎ Corresponding author.
13 12 2022
1 2023
13 12 2022
133 109926109926
10 12 2021
26 10 2022
8 12 2022
© 2022 Elsevier B.V. All rights reserved.
2022
Elsevier B.V.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
COVID-19 raised the need for automatic medical diagnosis, to increase the physicians’ efficiency in managing the pandemic. Among all the techniques for evaluating the status of the lungs of a patient with COVID-19, lung ultrasound (LUS) offers several advantages: portability, cost-effectiveness, safety. Several works approached the automatic detection of LUS imaging patterns related COVID-19 by using deep neural networks (DNNs). However, the decision processes based on DNNs are not fully explainable, which generally results in a lack of trust from physicians. This, in turn, slows down the adoption of such systems. In this work, we use two previously built DNNs as feature extractors at the frame level, and automatically synthesize, by means of an evolutionary algorithm, a decision tree (DT) that aggregates in an interpretable way the predictions made by the DNNs, returning the severity of the patients’ conditions according to a LUS score of prognostic value. Our results show that our approach performs comparably or better than previously reported aggregation techniques based on an empiric combination of frame-level predictions made by DNNs. Furthermore, when we analyze the evolved DTs, we discover properties about the DNNs used as feature extractors. We make our data publicly available for further development and reproducibility.
Keywords
COVID-19
Lung ultrasound
Decision trees
Grammatical evolution
Evolutionary algorithms
Neuro-symbolic artificial intelligence
==== Body
pmc1 Introduction
Since the outbreak of the coronavirus disease 2019 (COVID-19) pandemic, the use of lung ultrasound (LUS) has been globally and fastly spreading. Indeed, the main advantages of LUS (portability, cost-effectiveness, real-time imaging, and safety) compared to other imaging technologies such as, e.g., Computed Tomography (CT), allowed LUS to be widely adopted to evaluate the state of lungs in patients affected by COVID-19 [1], [2], [3], [4], [5], [6], [7], [8], [9]. Moreover, LUS can be nowadays used for patients’ monitoring and for the triage of symptomatic patients [1]. In particular, LUS is often exploited to detect COVID-19 associated interstitial pneumonia and follow its evolution [2], [10]. To perform this task, different imaging protocols have been proposed together with semi-quantitative scoring systems [11]. Indeed, even though quantitative approaches aiming at assessing the condition of lung parenchyma with ultrasound are emerging [10], [12], [13], [14], [15], [16], these strategies are not available for emergency contexts, due to their current preliminary state. Therefore, semi-quantitative scoring systems based on specific LUS imaging patterns (e.g., vertical and horizontal artifacts, or consolidations) have been extensively exploited during the pandemic [2].
Even though a LUS quantitative analysis cannot be performed with the currently available technologies, the use of artificial intelligence (AI) for the classification of LUS frames according to a semi-quantitative scoring system can be exploited to reduce subjectivity in the evaluation and to reduce the time required to perform the analysis [17], [18], [19], [20].
In the hereby study we exploit a standardized imaging protocol based on 14 scanning areas and on a four-level scoring system, which allows the grading of the state of lungs [2]. A recent study demonstrated how this standardized protocol and scoring system have a prognostic value when evaluating the cumulative score (sum of scores obtained in the 14 scanning areas) at exam-level [21]. We acquire 1808 LUS videos from 100 COVID-19 positive patients, which consist of 366,301 frames in total. These frames are then fed to two DNNs [17] that were previously trained to perform automatic scoring and segmentation of LUS frames according to the above-mentioned four-level scoring system [2]. We successively use the scores given as output by two DNNs (respectively for segmentation and labeling) [17] to train and test a novel automatic approach, based on decision trees (DTs) automatically synthesized by evolutionary computation, aiming at passing from frame-based labeling to video-based labeling. Specifically, we compare the video-level scores given by our automatic approach with scores given by expert clinicians. Indeed, to perform their evaluation, clinicians associate a score to each video rather than to each frame. We then assess the performance of our aggregation approach (both at video-level and exam-level) by comparing the results obtained by the proposed method with the empirical aggregation technique previously reported in [22], which represents the current state-of-the-art. We hypothesize that, even though this existing technique achieves good performance, the fact that its decisions are obtained by aggregating the outputs of the DNNs by means of a simple threshold-based approach may be sub-optimal. To overcome this limitation, we instead use a fully data-driven DT-based approach, that is in principle more flexible and does not require empirical choices of thresholds. To summarize, the main contributions of this work are the following:
1. we propose a neuro-symbolic approach to the automatic scoring of COVID-19 patients by combining DNNs and interpretable DTs;
2. we compare single-objective and multi-objective evolutionary approaches to synthesize DTs optimized w.r.t. three different metrics of interest;
3. we interpret the evolved DTs to understand their decision policies;
4. we obtain decision support systems that have both higher prognostic agreement and less variance w.r.t. the approach previously proposed in the state-of-the-art work in the field [22].
The paper is organized as follows. Firstly, we present the dataset and the design of our study, as well as the proposed method aiming at aggregating LUS frame-based predictions to obtain video-level predictions (Section 2). Successively, the results are presented (Section 3), followed by a detailed analysis of the evolved DTs (Section 4). Finally, the conclusions are derived and discussed (Section 5).
2 Materials and methods
We use the two models from [17] as feature extractors, whose outputs are aggregated and given in input to an evolved DT, which will then make a prediction of the score related to the video. A block diagram of the process is shown in Fig. 1.
Fig. 1 Block diagram of the prediction process.
2.1 Data
The investigated population consists of 100 patients diagnosed as COVID-19 positive by a reverse transcription polymerase chain reaction (RT-PCR) swab test. Of the 100 patients, 63 (35 male, 28 female; ages ranging from 26 to 92 years, and average age equal to 63.72 years) were examined within the Fondazione Policlinico San Matteo (Pavia, Italy), 19 (16 male, 3 female; ages ranging from 34 to 84 years, and average age equal to 63.95 years) within the Lodi General Hospital (Lodi, Italy), and 18 (8 male, 10 female; ages ranging from 23 to 95 years, and average age equal to 52.11 years) within the Fondazione Policlinico Universitario Agostino Gemelli (Rome, Italy). As a subgroup of patients was examined multiple times, on different dates, a total of 133 LUS exams were performed (94 at Pavia, 20 at Lodi, and 19 at Rome). A total of 1808 LUS videos were thus acquired (1290 at Pavia, 276 at Lodi, 242 at Rome), which consist of 366,301 frames (292,943 at Pavia, 44,288 at Lodi, 29,070 at Rome).
The data from Pavia have been acquired using a convex probe with an Esaote MyLab Twice scanner, and an Esaote MyLab 50, setting an imaging depth from 8 to 12 cm (depending on the patient) and an imaging frequency from 5.0 to 6.6 MHz (depending on the scanner). The data from Lodi have been acquired using a convex probe with an Esaote Mylab Sigma scanner, and a MindRay TE7, setting an imaging depth from 8 to 12 cm (depending on the patient) and an imaging frequency from 3.5 to 5.5 MHz. The data from Rome have been acquired using a convex probe with an Esaote MyLab 50, an Esaote MyLab Alpha, and a Philips IU22, setting an imaging depth from 8 to 12 cm (depending on the patient), and an imaging frequency from 3.5 to 6.6 MHz (depending on the scanner).
This study was part of a protocol that has been registered (NCT04322487) and received approval from the Ethical Committee of the Fondazione Policlinico Universitario San Matteo (protocol 20200063198), of Milano area 1, the Azienda Socio-Sanitaria Territoriale Fatebenefratelli-Sacco (protocol N0031981), of the Fondazione Policlinico Universitario Agostino Gemelli, Istituto di Ricovero e Cura a Carattere Scientifico (protocol 0015884/20 ID 3117). All patients gave informed consent.
The patients were examined by applying a standardized acquisition protocol based on 14 scanning areas [2]. This protocol is based on a four-level scoring system consisting in assigning a score that ranges from 0 to 3, depending on the observed LUS patterns, with score 0 indicating a healthy lung surface, and 1, 2, 3 an increasingly altered lung surface [2]. All the 1808 LUS videos were thus scored by LUS medical experts (in this case, the authors T.P., F.T., and A.S.). Each expert labeled the videos acquired by himself, i.e., T.P. labeled videos from Pavia, F.T. from Lodi, and A.S. from Rome. The distribution of scores assigned at video-level by the experts is shown in Fig. 2.
Fig. 2 The distribution of scores assigned at video-level by the three clinical experts is shown. The percentage of scores 0, 1, 2, and 3 is shown for each hospital (Pavia, Lodi, and Rome) and for the entire dataset (overall). The total number of videos for each group is provided on top.
Fig. 3 Examples of frames labeled as scores 0, 1, 2, and 3 are shown.
Fig. 4 The distribution of scores assigned at frame-level by the labeling and segmentation DNNs presented in [17], exploited in this work, is shown. The percentage of scores 0, 1, 2, and 3 is shown for each hospital (Pavia, Lodi, and Rome) and for the entire dataset (overall). The frame-level scores given by each architecture are shown separately. As the segmentation DNN can provide multiple scores for the same frame, we scored each frame with the worst-case (i.e., maximum) score predicted by the segmentation DNN. As the segmentation DNN could provide no output (if it does not find any relevant LUS pattern), an extra distribution represented as score −1 is observable in the left bars. The total number of frames for each group is provided on top.
2.1.1 Inputs
All the 1808 videos are fed to the two DNNs presented by Roy et al. [17], i.e., a labeling DNN derived from Spatial Transformer Networks and a segmentation DNN derived from U-Nets and DeepLab v3+. The former provides as output a score for each input frame, whereas the latter provided semantic segmentation and assigned one or multiple scores to each frame [17]. As the segmentation DNN can provide multiple scores for the same frame, we assign only the highest score predicted by this DNN to each considered frame (i.e., the worst-case score). Moreover, it is important to highlight how the segmentation DNN could provide no scores in output (if it does not find any relevant LUS pattern). Therefore, an extra score indicating the absence of LUS patterns (characteristic of the four scores) is considered when evaluating the output provided by segmentation DNN (we called it score -1). These two DNNs have been previously trained with the dataset presented by Roy et al. [17], which does not depend on the dataset exploited in the hereby work. Fig. 3 shows examples of frames labeled as scores 0, 1, 2, and 3. Fig. 4 shows the distribution of scores assigned at frame-level by the exploited two DNNs. Considering the entire dataset (see the overall distribution, Fig. 4, right), there is a high percentage of score 0 and 2 for both labeling and segmentation DNNs, whereas score 1 is less frequently given as output. It is also observable how the percentage of score 3 is significantly higher when looking at the segmentation DNN.
2.1.2 Targets
Given the frame-level labeling provided by the two DNNs, our target consists in finding an aggregation technique that allows us to pass from a frame-level score to a video-level score, which is the output needed by physicians to perform their clinical evaluation. Therefore, the goal of the proposed technique is to optimize three metrics of interest that compare the video-level scores obtained by our algorithm and the ones assigned by clinical experts (see Fig. 2). These three metrics are: the video-level agreement, the exam-level agreement, and the prognostic-level agreement [22].
The video-level agreement consists of the percentage of videos that are correctly classified by the algorithm (i.e., the score assigned by the expert coincides with the score assigned by algorithm) [22]. We also evaluate the video-level agreement when allowing a disagreement up to 1 point (e.g., if the algorithm classified a video as score 2 and the expert as score 1 or 3, the evaluation is correct) [22]. To distinguish these two video-level agreements, we denote the agreement characterized by exact match between video-level scores as video-level agreement with threshold (Th) equal to 0, whereas we refer to the video-level agreement allowing a disagreement up to 1 point as video-level agreement with Th equal to 1.
The exam-level agreement is instead computed by considering the cumulative score obtained by summing the video-level scores assigned to each of the 14 scanning areas [2], [22]. Specifically, we compute the exam-level agreement as the percentage of LUS exams (133 in total) having a cumulative score (ranging from 0 to 3×14=42) allowing a disagreement between algorithm and clinical experts of up to 2, 5, and 10 points (i.e., Th equal to 2, 5, and 10, respectively) [22]. To support the stratification between patients at high risk of clinical worsening and patients at low risk, we need to consider the prognostic value of the aforementioned protocol [2], which has been recently proven in a single-center study on 52 patients [21]. In particular, the patient is at low risk of clinical worsening when the exam-based cumulative score is less than or equal to 24, whereas the patient is at high risk of worsening when the exam-based cumulative score is greater than 24 [21].
We thus evaluate the algorithm capability of automatically stratifying these two categories of patients by measuring the prognostic-level agreement [22]. Specifically, clinical experts and algorithm are considered in prognostic agreement when both cumulative scores are less than or equal to 24 (low risk) or greater than 24 (high risk) (Th equal to 24) [22].
As we will show in detail in Section 2.3.1, it should be noted that we use a proxy for the first two metrics, i.e., the Mean Square Error (MSE), which gives us an advantage over optimizing directly the agreement. In fact, in [22] the authors use several tolerances for the video-level and exam-level agreement: this implies that in practical scenarios one should either use n objectives for each metric, where n is the number of tolerances (e.g., maximize the exam-level agreement with tolerances 2, 5, and 10); or, one should choose one tolerance among all the tolerances, which may lead to DTs that perform well only for that particular tolerance. Instead, using the MSE as we do here allows us to maximize simultaneously the agreement for all the tolerances.
2.1.3 Splitting of the data
We split the data randomly in 5 folds. To prevent data leakage, we make sure that all the data belonging to a patient are assigned to the same fold. Moreover, we use 4 folds for the training phase (i.e., to compute the fitness of the individuals) and the remaining one to assess the generalization capabilities of the best evolved DTs (i.e., as test set).
2.2 Feature extraction and aggregation of the outputs
The DT (as shown in Fig. 1) expects as input video-level features. Instead, the two DNNs’ input consists in single frames of each video. To convert the features from frame-level to video-level, we aggregate the outputs of each DNN. For the labeling DNN, we simply use as features the relative frequency of the prediction of each class (i.e., the argmax of the output vector of each frame). For the segmentation DNN, instead, we aggregate the features by computing the relative frequency of the worst-case (i.e., maximum) predicted classes inside each frame. This distinction between the two DNNs is needed since the segmentation DNN does not produce a single prediction but, instead, it produces a mask for the frame taken in input. Moreover, we add to the feature vector also the minimum and the maximum prediction made by the two DNNs.
The resulting feature vector is thus composed by 12 features: l0, l1, l2, l3, lmin, lmax, s0, s1, s2, s3, smin, smax where: li, i=0,1,2,3, represents the relative frequency of the prediction of the class i made by the labeling DNN; lmin and lmax represent the minimum and the maximum gravity level predicted by the labeling DNN; si, i=0,1,2,3, represents the number of cases in which the class i corresponds to the worst-case in the predictions made by the segmentation DNN; smin and smax represent the minimum and the maximum gravity level predicted by the segmentation DNN. Note that, while lmin and lmax range in [0,3], the segmentation DNN can also detect the absence of LUS patterns (i.e. the pixel is assigned a score of −1). For this reason, smin and smax range in [−1,3].
Table 1 Grammar used to evolve the DTs. The symbol “∣” denotes the possibility to choose between different symbols. When using the grammar to translate a genotype into a phenotype, the rules are expanded in one of the possible choices listed in their production, depending on the value of the genotype.
Rule Production
dt 〈if〉
if if〈condition〉then〈output〉else〈output〉
condition 〈var〉〈op〉〈const〉∣〈var〉〈op〉〈var〉
var {inputi}; i∈[0,12[
op <∣>∣==
output 0∣1∣2∣3∣〈if〉
const [0,1] with step 10−2
Table 2 Parameters used for the grammatical evolution algorithm.
Parameter Value
Pop size 1000
Generations 1000
Genotype length 50
Crossover probability 0.8
Mutation probability 1
Crossover One-point
Mutation Uniform with pcodon=0.05
Selection Best
2.3 Evolutionary settings
We use Grammatical Evolution (GE) [23] to evolve programs that resemble DTs (i.e., they are based on an if-then-else structure). GE is an evolutionary algorithm that allows the evolution of grammars, encoded in the Backus–Naur form. It makes use of a genotype, which consists in a list of integers (called codons). When the genotype has to be evaluated, it makes use of a translator, which allows to convert the grammar to the corresponding phenotype. The grammar we employ is shown in Table 1.
We consider two GE settings, namely: (1) a single-objective one, in which we optimize either the video-level MSE, the exam-level MSE, or the prognostic-level agreement (see Section 2.3.1 for details on the three metrics); and (2) a multi-objective one, in which we optimize simultaneously all the three objectives stated previously.
The pseudo-code of the algorithm is shown in Algorithm 1. The algorithm consists in an initialization step (Line 1) followed by an evolutionary loop (Lines 4–10). The evolutionary loop starts with the evaluation of the population (Line 4), followed by the replacement of the individuals in the population (Line 5). Then, we performs the usual evolutionary steps, i.e., selection (Line 6), crossover (Line 7), and mutation (Line 8).
We should note that the GE algorithm we use here has some differences with respect to the original one described in [23]. First of all, we do not make use of a variable-length genotype but, instead, we fix its length (as shown in Table 2). Fixing the length of the genotype to small values constrain the resulting DTs to be small and, thus, more interpretable than the ones we can obtain by having longer genotypes. Moreover, instead of using the genetic operators described in [23], we use traditional operators for genetic algorithms. The reason underlying this choice is due to the fact that, from preliminary experiments, the original operators seem to achieve worse performance than traditional genetic operators. For this reason, we employ standard operators, described below.
We make use of the replacement operator (Line 5) described in [24], which replaces a parent from the population only if there is an offspring that outperforms it. In case there are two offspring whose performance are better than only one of the two parents, then the best offspring replaces the worst parent.
Moreover, we use of two different parent-selection operators (for Line 6), depending on whether we are working in the single-objective or multi-objective setting. In the single-objective setting, we use the “best-wise” selection operator, i.e., a selection operator that reorders the population by descending fitness such that, when performing crossover, the (2i)th best mates with the (2i+1)th best. Conversely, in the multi-objective setting, the selection operator we use is the NSGA-II [25] operator, which proved to work very well for multi-objective problems.1
The crossover operator (Line 7), instead, is the one-point crossover, which produces two offspring from two parents by splitting their genotypes in a randomly chosen point and mixing the corresponding sub-strings obtained from the two parents.
While mutating a solution (Line 8), we employ a uniform mutation, which mutates each codon of the genotype according to a given probability pcodon. Its new value is sampled randomly from the possible values.
The parameters we use are presented in Table 2. The parameters shown in the table were obtained by manual tuning.
2.3.1 Fitness evaluation
The fitness evaluation phase works as follows. For each training fold, we feed the features of each video to the DT and record its predictions. Then, we compute the following metrics of interest:
1. Video-level MSE (to be minimized): (1) 1Nvideos∑i=1Nvideos(yiv−yˆiv)2;
2. Exam-level MSE (to be minimized): (2) 1Nexams∑j=1Nexams(yje−yˆje)2;
3. Prognostic-level agreement (to be maximized): (3) 1Nexams∑j=1NexamsI(yjp=yˆjp)2;
where: yiv is the ground truth for the video i; yje=∑i=114yj,iv is the ground truth for exam j, i.e., the sum of the scores of each video of the exam; yjp=I(yje>24) is the ground truth for the prognosis j; I is the indicator function, i.e., it outputs 1 if the argument is true, otherwise 0. The notation yˆba refers to the output of the DT given the output b in setting a, i.e., it is the approximation made by the DT of the variable yba.
The pseudo-code for the fitness evaluation function (in the most general case, i.e., multi-objective) is shown in Algorithm 2. In the pseudo-code, a lowercase bold variable represents a vector, while an uppercase bold variable represents a matrix. Otherwise the variable is assumed to be scalar.
The reason underlying the optimization of different metrics is the following. Our overall goal is to maximize the agreement for all the three metrics, as done in [22]. However, as we discussed earlier optimizing the video- and exam-level agreement requires also a specification of the tolerances to use for the computation of the agreement (e.g., in [22], the authors compute the exam-level agreement with a tolerance of 2, 5, and 10 points). Instead, by optimizing the MSE for these two metrics (Lines 6–7 of Algorithm 2) allows us to evolve DTs that minimize the distance of the predictions from the ground truth, no matter the threshold. Finally, for the prognostic-level agreement, we cannot use the MSE because this variable is not a score but, instead, it is a binary variable. So, in this case, using the MSE does not give any advantage over directly optimizing the agreement (Line 8).
Then, for each metric, we use as fitness the worst value obtained on the 4 folds used for training (Lines 11–13).
While the single-objective fitness corresponds to a scalar value that consists in the value of a single metric, in the multi-objective setting it is composed of a list of three values, i.e., the video-level MSE, the exam-level MSE, and the prognostic-level agreement.
Table 3 Descriptive statistics of the video-level agreement on all the folds. “Th” stands for the video-level threshold (i.e., the tolerance) used for the evaluation of the results.
Method Th Min Mean Std Med Max
JASA L 0 44.70 50.40 4.28 50.68 57.38
1 82.74 85.87 2.49 85.91 89.76
JASA L+S 0 42.35 50.10 5.48 51.43 56.67
1 82.74 85.87 2.43 85.36 89.29
Video 0 42.75 46.08 2.41 46.07 49.88
1 88.63 92.40 2.14 93.41 94.52
Exam 0 42.35 47.82 4.74 46.14 54.29
1 80.78 84.42 3.00 85.91 87.86
Prognostic 0 41.18 48.21 4.46 50.68 52.62
1 79.61 83.71 2.79 85.48 86.43
3 objectives 0 45.88 49.48 2.24 49.77 52.86
1 85.47 88.31 2.29 87.84 92.14
Fig. 5 Best DT evolved on the video-level MSE.
Table 4 Statistical comparison of the different approaches on the video-level agreement. “+” means “statistically better”, “=” means “statistically equivalent”, and “–” means “statistically worse” (Welch T-test, α=0.05).
Method Th JASA L JASA L+S Video Exam Prognostic 3 objectives
JASA L 0 = = = = = =
1 = = – = = =
JASA L+S 0 = = = = = =
1 = = – = = =
Video 0 = = = = = –
1 + + = + + +
Exam 0 = = = = = =
1 = = – = = =
Prognostic 0 = = = = = =
1 = = – = = –
3 objectives 0 = = + = = =
1 = = – = + =
Table 5 Descriptive statistics of the exam-level agreement on all the folds. “Th” stands for the exam-level threshold (i.e., the tolerance) used for the evaluation of the results.
Method Th Min Mean Std Med Max
JASA L 2 21.05 32.30 6.75 32.26 41.94
5 45.00 56.95 9.43 57.90 70.97
10 80.00 89.44 6.44 89.47 100.00
JASA L+S 2 21.05 30.97 7.58 35.00 38.71
5 45.16 59.44 13.23 52.63 80.64
10 80.00 84.58 6.27 81.25 96.77
Video 2 21.05 29.65 8.41 25.81 45.16
5 40.00 57.75 12.14 63.16 70.97
10 80.00 85.85 4.47 84.38 93.55
Exam 2 25.00 31.90 7.89 29.03 46.88
5 61.29 64.27 3.28 63.16 70.00
10 73.68 88.01 7.62 90.32 95.00
Prognostic 2 26.32 30.38 3.77 29.03 37.50
5 45.00 58.50 8.81 59.38 67.74
10 78.95 87.06 4.96 87.50 93.55
3 objectives 2 21.05 36.36 9.68 38.71 46.88
5 45.00 63.98 11.35 62.50 77.42
10 85.00 91.51 5.34 90.32 100.00
Table 6 Statistical comparison of the different approaches on the exam-level agreement. “+” means “statistically better”, “=” means “statistically equivalent”, and “–” means “statistically worse” (Welch T-test, α=0.05).
Method Th JASA L JASA L+S Video Exam Prognostic 3 objectives
JASA L 2 = = = = = =
5 = = = = = =
10 = = = = = =
JASA L+S 2 = = = = = =
5 = = = = = =
10 = = = = = –
Video 2 = = = = = =
5 = = = = = =
10 = = = = = =
Exam 2 = = = = = =
5 = = = = = =
10 = = = = = =
Prognostic 2 = = = = = =
5 = = = = = =
10 = = = = = =
3 objectives 2 = = = = = =
5 = = = = = =
10 = + = = = =
3 Results
We perform 10 independent runs for the proposed method in each of the four settings: single-objective, video-level MSE; single-objective, exam-level MSE; single-objective, prognostic-level agreement; multi-objective. For each run, we test (on the test fold) only the best evolved DT, i.e., for the single-objective runs it is the individual with the best fitness, while for the multi-objective runs it is the one with the maximum prognostic-level agreement (i.e., the most important metric).
Tables 3, 5, 7 show the descriptive statistics computed on the agreement (%) (with the physicians opinion) computed across all the 5 folds. Bold values represent the best score across all the methods. We compute the statistics on all the 5 folds for the following reason. Since this work is meant to work in a medical scenario, we are not really interested in knowing the training and the test agreements as such. Instead, we are interested in knowing the worst-case scenario and, to compute it, we need to compute the statistic on all the 5 folds. Note that, for each fold, we compute the three agreement scores on that fold and then we use these scores to compute the statistics across folds. In the tables, “Video”, “Exam”, “Prognostic” and “3 objectives” refer to, respectively, our method evolved on video, exam, prognostic and three objectives. In the same tables, we also report the state-of-the-art results presented in [22] listed as “JASA L” and “JASA L+S”, which refer to, respectively, the approach using only the labeling DNN, and the one using both the labeling and the segmentation DNNs. Note that we do not use the results shown in [22] but rather we evaluate them on the same folds used for the DTs, to guarantee a uniform evaluation of all the methods.
Finally, it should be considered that even if we use MSE as the metric for the first two objectives, we are still interested in evaluating the agreement with the physicians. For this reason, we do not show the MSE in the tables, but the agreement at video- and exam-level. This allows us also to use the same thresholds as in [22], keeping a consistency on the method used for evaluating such models.
Analyzing the results obtained in the single-objective setting, we observe that, for each metric, the model that achieves the maximum performance (especially for higher thresholds) is the one that has been specifically evolved for that metric. On the other hand, we observe that the model evolved on three objectives in some cases has a smaller minimum/mean agreement than that of the model specifically evolved on each metric. However, it is never smaller than the minimum/mean agreement of the other two models evolved on the other metrics. This suggests that the model evolved by means of multi-objective optimization has a good trade-off between the three objectives.
Surprisingly, we observe that in some cases the performance of the DTs evolved in the multi-objective setting (i.e., the ones evolved on the three objectives simultaneously) exceeds even the performance of the best DTs found in the single-objective setting. This suggests that optimizing for all the metrics simultaneously can keep a “consistency” between the different metrics and allows the DT to learn better strategies for classifying the samples. In fact, we find that the DTs evolved on single objectives do not generalize well to the other objectives. Instead, the DT evolved in the multi-objective setting is able to keep a good trade-off between the objectives.
In Table 4, Table 6, Table 8 show the results of a statistical comparison performed with a Welch T-test with confidence level α=0.05. The “+”, “=”, and “−” in the tables indicate respectively statistically better, equal or worse performance of the method on the row w.r.t. the method on the column. We observe that we can reject the null hypotheses (i.e., that the samples come from the same distribution) in just a few cases, namely:
• the DT evolved on the video-level MSE outperforms all the other approaches in the setting with threshold 1;
• the DT evolved with the multi-objective approach outperforms the other approaches in three cases.
In all the other case, the methods result statistically equivalent. However, we should note that this lack of statistical significance is likely due to the small number of folds (i.e., samples for the T-test), even though this has been manually tuned to balance fold size (to reduce overfitting and assess better the generalization capabilities) and number of folds.
Finally, compared with the two methods described in [22] we observe that while the DTs evolved in the multi-objective setting have a comparable (usually better) minimum agreement than that of JASA L and JASA L+S, they perform substantially better when considering the mean agreement computed on the folds. Also in this case, due to the small number of folds, we can statistically confirm only the increase in performance w.r.t. JASA L+S in the exam-level mean agreement when using a threshold of 10.
Table 7 Descriptive statistics of the prognostic-level agreement on all the folds. “Th” stands for the (prognostic) threshold used for the evaluation of the results.
Method Th Min Mean Std Med Max
JASA L 24 63.13 78.12 7.96 80.64 85.00
JASA L+S 24 57.90 76.78 12.01 83.87 90.00
Video 24 57.89 79.63 11.68 85.00 90.32
Exam 24 61.29 76.25 12.76 77.42 95.00
Prognostic 24 73.68 81.89 5.27 81.25 90.00
3 objectives 24 68.42 82.11 7.51 84.38 90.00
Table 8 Statistical comparison of the different approaches on the prognostic-level agreement. “+” means “statistically better”, “=” means “statistically equivalent”, and “–” means “statistically worse” (Welch T-test, α=0.05).
Method Th JASA L JASA L+S Video Exam Prognostic 3 objectives
L 24 = = = = = =
L+S 24 = = = = = =
Video 24 = = = = = =
Exam 24 = = = = = =
Prognostic 24 = = = = = =
3 objectives 24 = = = = = =
4 Analysis of the decision trees
In this section, we show the best evolved DTs in each setting and interpret them to understand the relationships they captured on the predictions made by the DNNs. We consider as “best DT” the trees that satisfy the following properties (over the best solutions obtained in the 10 runs). For the setup considering only the video-level MSE, the best tree is the one that obtains the smallest MSE. When we only consider exam-level MSE, the best tree is the tree that achieves the smallest exam-level MSE. When we consider only the prognostic-level agreement, the best tree is the one with the highest prognostic-level agreement. Finally, when we consider all the objectives simultaneously, the best tree is the one that achieves the best prognostic-level agreement. In this case, if there are ties between two solutions, we choose the one that has the best trade-off between video- and exam-level MSE. In all the cases, the conditions of the DT are numbered as when doing a pre-order traversal of the DT.
4.1 Decision tree evolved on the video-level MSE
Fig. 5 shows the best DT obtained in this setting. While this DT performs worse than JASA L and JASA L+S when no tolerance is given to the prediction, it outperforms them significantly when a threshold of 1 is allowed. For this reason, we will interpret it considering that each prediction yˆ must be considered as a value ranging in [yˆ−1,yˆ+1] (constraining the values in [0,3]).
This DT checks very few things about the predictions made by the two DNNs. In fact, in the first split (lmax>smax) it captures a very simple pattern: when the maximum class of risk predicted by the labeling DNN is greater than the maximum risk class predicted by the segmentation DNN, then it assigns the video a risk varying in [0,2], i.e., it excludes the class 3. On the other hand, when the root condition is false, it checks whether the fraction of frames classified as maximum risk (by the labeling DNN) is bigger than the fraction of predictions made by the segmentation DNN in which the highest score is 2 (s2<l3). If so, it assigns the video a score varying in [2,3], i.e., high risk. Basically, this condition checks whether the video refers to a high-risk patient. In fact, the condition can be interpreted as: If the ratio of samples classified as maximum risk by the labeling DNN is bigger than the ratio of samples classified as risk 2 by the segmentation DNN, then give the priority to the labeling DNN and assign the maximum score to the video.
To confirm this hypothesis, in Fig. 6(c) we plot the histogram of the number of videos assigned to each class that fall in the case explained above (note that in the other sub-figures of Fig. 6 we do the same for all the other conditions in the DT). We observe that the number of videos belonging to class 3 is significantly higher w.r.t. the other classes. Finally, in the third condition (l2<0.15) the DT makes an extremely simple check: If the ratio of frame labeled with class 2 is low (i.e., under a threshold of 0.15), then probably the number of frames assigned to class 3 will be even lower, so assign a score ranging in[0,2]to the video. Otherwise, there is a high chance that the severity score is higher than 0, so assign a score ranging in[1,3]to the video.
From this DT, we infer that the labeling DNN may be “biased” towards high scores. The DT is then evolved to make use of the output of the segmentation DNN in order to reduce this bias.
Fig. 6 Class histograms for each of the branches of the DT shown in Fig. 5. Note that the nodes are counted as in a pre-order traversal of the DT.
4.2 Decision tree evolved on the exam-level MSE
This DT (shown in Fig. 7) achieves a better worst-case agreement with low thresholds (2 and 5). However, in this case (and the following ones), we cannot use the threshold as a tolerance value to be used on the output value of the DT. This is due to the fact that, in this case, the DT outputs the gravity for each video, but the tolerance is expressed at the exam level.
The first condition of this DT (l0<0.72) checks that the severity of the patient is high, by ensuring that the fraction of frames labeled as minimum risk is lower than an evolved threshold. If so, it then assesses the severity of the conditions by using the segmentation DNN, checking if the fraction of frames that are classified as maximum risk is more than the half (s3>0.51). If so, it assigns the maximum risk to the video.
If the first condition is false, then the DT performs additional checks. In fact, the right part of the DT is basically a decision list, i.e., an extremely unbalanced DT, which isolates one particular case at each split. While the first condition on the right (s2>lmax) may make no sense at a first sight, it is a simple trick that the DT uses to perform an and between two conditions. In fact, we know that s2∈[0,1] and lmax∈{0,1,2,3}. This means that the condition s2>lmax evaluates to true only in case s2>0 and lmax=0.
The second case isolated by the decision list checks for a particular case (s1=l3). By analyzing the training set, we observe that this case only happens when s1=l3=0. Moreover, in these cases s1 and l3 are the only variables that are always equal to zero. While this may seem a remote possibility, we found that this condition (conjoined with the conditions that are evaluated before it) evaluates to true for about a quarter of the samples in the training set. Hence, this condition exploits a bias of the two DNNs to detect cases in which the severity is likely to be low (45% of the cases with score 0, 31% of the cases with score 1, 17% with score 2, 7% with score 3).
The third condition on the right branch (smin=0) checks whether the patient has at least one frame with minimum risk (i.e., 0, opposed to a case in which no damaged tissue is detected, i.e., −1) by checking the outputs of the segmentation DNN. If so, it assigns the class 2 to the video.
Finally, the last condition (s0>0.37) checks the number of frames with minimum risk (detected by the segmentation model): if they consist of more than the 37% of the frames in the video, the risk assigned to the video is very high (3), otherwise it is assigned a lower score (1).
It is important to note that the video-level predictions performed by this model aim to reduce the worst-case exam-level MSE w.r.t. the physicians’ judgment.
Also in this case, for the sake of completeness we report in Fig. 8 the distribution of classes for each condition in the DT.
Fig. 7 Best DT evolved on the exam-level MSE.
Fig. 8 Class histograms for each of the branches of the DT shown in Fig. 7. Note that the nodes are counted as in a pre-order traversal of the DT.
4.3 Decision tree evolved on the prognostic-level agreement
This DT (shown in Fig. 9) achieves the best worst-case prognostic-level agreement among all the best evolved DTs (in this case, evolved with the single threshold value, 24).
In the root condition (l3<s1), this DT checks whether the confidence given to class 3 from the labeling DNN is smaller than the confidence given to class 1 by the segmentation DNN. This, intuitively, tries to filter out the cases where the probability of having the maximum risk is high. In fact, as shown in Fig. 10(a), the ratio of samples belonging to class 3 is not so high in this case (12.5%), see the other sub-figures in Fig. 10 for the distributions of classes corresponding to the other conditions in the DT.
The second condition (i.e., the left branch of the root, l0>0.75) naturally follows the first one: given that, as shown in Fig. 10(a), the distribution of the classes is skewed towards class 2, is there a way to filter out the samples belonging to class 2? While this condition does not filter perfectly the samples belonging to class 2, it is able to filter 67.9% of them (as shown in Fig. 10, Fig. 10).
The third condition (lmax=lmin) seeks for cases in which the maximum class and the minimum class predicted by the labeling DNN are equal. Of course, this condition is way more likely to happen in low-risk frames, as confirmed by Fig. 10(e). However, as we can see from Fig. 10(f), not all the samples with low risk are filtered out by this condition.
For this reason, the purpose of the fourth condition (s3<s0) is to separate the low-risk cases from the higher-risk ones. In fact, what it does is simply checking the predictions made by the segmentation DNN: if the ratio of samples assigned to class 0 is higher than the ratio of samples assigned to class 3, then it predicts 0, otherwise 2.
Note that this DT has been optimized to maximize the prognostic-level agreement. This explains why, in some cases, the outputs are not coherent with what a human expects when trying to predict the label for each video. In fact, we hypothesize that these counter-intuitive tests aim to soften the contributions of each video to the prognostic score. This can be seen in the fact that predictions for the class 3 never appear in this DT, and also that the predictions for class 1 are not frequent (even when they would minimize the video-level MSE, which is not taken into account when optimizing this DT).
Fig. 9 Best DT evolved on the prognostic-level agreement.
Fig. 10 Class histograms for each of the branches of the DT shown in Fig. 9. Note that the nodes are counted as in a pre-order traversal of the DT.
4.4 Decision tree evolved on three objectives
This DT (shown in Fig. 11) has comparable, but often better, performance with respect to all the other DTs evolved in the other settings. One interesting feature of this DT is that it uses the labeling DNN to make coarse-grained decisions, that are then refined by using the segmentation DNN.
In the first condition (l3>l0), this DT simply checks the outputs coming from the labeling DNN to address the gravity of the conditions. Surprisingly, only checking if the ratio of samples assigned to risk 3 is higher than the ratio of samples assigned to risk 0 is enough to discriminate very well the high-risk cases, as shown in Fig. 12(a) (see the other sub-figures in Fig. 12 for the distributions of classes corresponding to the other conditions in the DT).
In the second condition (l2>0.07), the DT checks the ratio of labels assigned to class 2 by the labeling DNN. If they are more than 7%, then it makes a simple refinement using the segmentation DNN: if the segmentation DNN classifies all the samples as class 3 (s3=1), then it assigns the video the maximum score, otherwise it assigns the video a score of 2.
If the second condition evaluates to false, then the DT checks whether lmax<0.97. Since lmax is an integer, this corresponds to checking whether lmax=0. If so, again, the DT makes use of the segmentation DNN to refine the decision: if the ratio of samples assigned to class 3 by the segmentation DNN is more than the 60% (s3>0.60), then it assigns the video a score of 2. Otherwise, the gravity of the condition is not high enough, so it assigns a score 0 to the video. This condition handles a bias of the labeling DNN, that happens when this DNN classifies all the frames with a severity of 0, but, instead, their actual score is very different from 0.
Finally, if lmax>0.97, it uses a similar check (s3>0.59) to assign the samples either to class 0 or 1. Surprisingly, when s3 is greater than 59%, the DT assigns the sample to class 0 while, as we can see from Fig. 12(k), assigning it a value equal to 1 would reduce the video-level MSE. However, this reasoning applies to the video-level predictions, but it may affect negatively the other two metrics. On the other hand, when s3≤0.59, we observe that the probability for class 3 is quite low, so the DT classifies the sample as belonging to class 1, probably to minimize the video-level MSE.
Fig. 11 Best DT evolved on the three objectives.
Fig. 12 Class histograms for each of the branches of the DT shown in Fig. 11. Note that the nodes are counted as in a pre-order traversal of the DT.
5 Conclusions
The use of LUS techniques to monitor the state of the lungs in COVID-19 patients is spreading, due to its numerous advantages compared to other techniques. Moreover, with COVID-19, the need for an automatic diagnosis emerged. For this purpose, several approaches have been proposed to perform automatic COVID-19 patients’ evaluation from LUS images [17], [18], [19], [20]. A previous approach [22] proposed a combination of two DNNs to increase the overall performance. This approach used an empiric threshold-based approach that, while performing well, did not give any insight on the biases of the DNNs used as input.
In recent years, a new need has emerged at the intersection between AI and healthcare: the need for interpretability [26]. In fact, especially in this domain, the users (in this case, the physicians) usually do not trust decisions suggested by a black box model (such as one base on DDNs). Instead, they want to be able to understand each decision made by the model, to ensure its correctness. Interpretable AI allows to have data-driven models that are inherently understandable and “simulatable” by humans, thus ensuring that a physician can actually understand the decisions made by the model.
In this work, we use two previously proposed DNNs as feature extractors, and then we use a DT for combining the two predictions. We use both single- and multi-objective evolutionary optimization to evolve the DT that takes in input the predictions made by the two DNNs aggregated at the video-level (i.e., a collection of frames). When evaluating our approach on three different levels of agreement with the physicians’ judgment, we find that the multi-objective optimization approach leads to DTs that, in general, perform in most cases comparably or better than the DTs evolved on single objectives. Moreover, our approach appears to perform better (in terms of descriptive statistics) than the approach presented in [22], even though, due to the small number of samples used for the comparison, we were able to quantify the statistical significance of the results only in a small number of cases.
This aspect, i.e., the fact that the limited number of samples, in some cases, does not allow a statistically significant comparison w.r.t. the baseline algorithm from [22], represents one of the main limitations of the current work. Another limitation of this study is that the DTs use orthogonal conditions (i.e., they compare a variable with a constant), or conditions that compare one variable to another variable. Since there are more expressive types of conditions (e.g., oblique conditions), better results may be achievable through the use of different types of conditions. Moreover, considering more complex conditions to describe the relationship between more than two variables may lead to better insights about the biases of the DNNs.
In the light of these limitations, future research directions will be aimed at collecting more data, in order to increase the size of the dataset, and evolving DTs by using different types of conditions, including oblique ones.
Finally, we highlight that we make our data publicly available for further development and reproducibility.2 Moreover, in a separate repository3 we release the scripts used to produce the results shown in this paper.
CRediT authorship contribution statement
Leonardo Lucio Custode: Methodology, Software, Data curation, Investigation, Formal analysis, Visualization, Validation, Writing – original draft. Federico Mento: Data curation, Investigation, Visualization, Writing – original draft. Francesco Tursi: Data curation; Investigation; Validation. Andrea Smargiassi: Data curation, Investigation, Validation. Riccardo Inchingolo: Data curation, Investigation, Validation. Tiziano Perrone: Data curation, Investigation, Validation. Libertario Demi: Conceptualization, Resources, Supervision, Writing – review & editing. Giovanni Iacca: Conceptualization, Resources, Supervision, Writing – review & editing.
Declaration of Competing Interest
Libertario Demi is a cofounder of UltraAI. The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
1 Since the most “important” metric is the prognostic-level agreement, an attentive reader may point out that using Pareto optimization may not be the ideal choice in this case. However, we cannot use a lexicographic selection, since this would require specifying a preference also between the exam- and the video- level MSE while, in this case, we do not have a clear preference. Moreover, using a weighted sum of the three objectives is also not feasible, since this would require assigning a specific weight to each of the three objectives. For these reasons, we optimize the objectives using Pareto optimization and, then, we select the best solution according to the prognostic-level agreement.
2 https://drive.google.com/drive/folders/1Or4dF2fAM23H5fd_yxtq1vyAS8b7pL0s.
3 https://gitlab.com/leocus/neurosymbolic-covid19-scoring.
==== Refs
References
1 Soldati G. Smargiassi A. Inchingolo R. Buonsenso D. Perrone T. Briganti D.F. Perlini S. Torri E. Mariani A. Mossolani E.E. Tursi F. Mento F. Demi L. Is there a role for lung ultrasound during the COVID-19 pandemic? J. Ultrasound Med. 39 7 2020 1459 1462 10.1002/jum.15284 32198775
2 Soldati G. Smargiassi A. Inchingolo R. Buonsenso D. Perrone T. Briganti D.F. Perlini S. Torri E. Mariani A. Mossolani E.E. Tursi F. Mento F. Demi L. Proposal for international standardization of the use of lung ultrasound for patients with COVID-19 J. Ultrasound Med. 39 7 2020 1413 1419 10.1002/jum.15285 32227492
3 Poggiali E. Dacrema A. Bastoni D. Tinelli V. Demichele E. Mateo Ramos P. Marcianò T. Silva M. Vercelli A. Magnacavallo A. Can lung US help critical care clinicians in the early diagnosis of novel coronavirus (COVID-19) pneumonia? Radiology 295 3 2020 E6 10.1148/radiol.2020200847 32167853
4 Lomoro P. Verde F. Zerboni F. Simonetti I. Borghi C. Fachinetti C. Natalizi A. Martegani A. COVID-19 pneumonia manifestations at the admission on chest ultrasound, radiographs, and CT: single-center study and comprehensive radiologic literature review Eur. J. Radiol. Open 7 2020 10.1016/j.ejro.2020.100231
5 Nouvenne A. Ticinesi A. Parise A. Prati B. Esposito M. Cocchi V. Crisafulli E. Volpi A. Rossi S. Bignami E.G. Baciarello M. Brianti E. Fabi M. Meschi T. Point-of-care chest ultrasonography as a diagnostic resource for COVID-19 outbreak in nursing homes J. Am. Med. Direct. Assoc. 21 7 2020 919 923 10.1016/j.jamda.2020.05.050
6 Yasukawa K. Minami T. Point-of-care lung ultrasound findings in patients with COVID-19 pneumonia Am. J. Trop. Med. Hygiene 102 6 2020 1198 1202 10.4269/ajtmh.20-0280 URL https://www.ajtmh.org/view/journals/tpmd/102/6/article-p1198.xml
7 Xing C. Li Q. Du H. Kang W. Lian J. Yuan L. Lung ultrasound findings in patients with COVID-19 pneumonia Crit. Care 24 1 2020 174 10.1186/s13054-020-02876-9 32345353
8 Peng Q.-Y. Wang X.-T. Zhang L.-N. (CCUSG), Chinese Critical Care Ultrasound Study Group Findings of lung ultrasonography of novel corona virus pneumonia during the 2019–2020 epidemic Intensive Care Med. 46 5 2020 849 850 10.1007/s00134-020-05996-6 32166346
9 Duclos G. Lopez A. Leone M. Zieleskiewicz L. “No dose” lung ultrasound correlation with “low dose” CT scan for early diagnosis of SARS-CoV-2 pneumonia Intensive Care Med. 46 6 2020 1103 1104 10.1007/s00134-020-06058-7 32367166
10 Demi L. Lung ultrasound: The future ahead and the lessons learned from COVID-19 J. Acoust. Soc. Am. 148 4 2020 2146 2150 10.1121/10.0002183 33138522
11 Allinovi M. Parise A. Giacalone M. Amerio A. Delsante M. Odone A. Franci A. Gigliotti F. Amadasi S. Delmonte D. Parri N. Mangia A. Lung ultrasound may support diagnosis and monitoring of COVID-19 pneumonia Ultrasound Med. Biol. 46 11 2020 2908 2917 10.1016/j.ultrasmedbio.2020.07.018 32807570
12 Mento F. Soldati G. Prediletto R. Demi M. Demi L. Quantitative lung ultrasound spectroscopy applied to the diagnosis of pulmonary fibrosis: The first clinical study IEEE Trans. Ultrason. Ferroelectr. Freq. Control 67 11 2020 2265 2273 10.1109/TUFFC.2020.3012289 32746228
13 Mento F. Demi L. On the influence of imaging parameters on lung ultrasound B-line artifacts, in vitro study J. Acoust. Soc. Am. 148 2 2020 975 983 10.1121/10.0001797 URL https://asa.scitation.org/doi/abs/10.1121/10.0001797 32873037
14 Soldati G. Demi M. The use of lung ultrasound images for the differential diagnosis of pulmonary and cardiac interstitial pathology J. Ultrasound 20 2017 10.1007/s40477-017-0244-7
15 Mohanty K. Blackwell J. Egan T. Muller M. Characterization of the lung parenchyma using ultrasound multiple scattering Ultrasound Med. Biol. 43 5 2017 993 1003 10.1016/j.ultrasmedbio.2017.01.011 28318888
16 Zhang G. Zhang J. Wang B. Zhu X. Wang Q. Qiu S. Analysis of clinical characteristics and laboratory findings of 95 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a retrospective analysis Respir. Res. 21 1 2020 74 10.1186/s12931-020-01338-8 URL https://pubmed.ncbi.nlm.nih.gov/32216803https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7099829/ 32216803
17 Roy S. Menapace W. Oei S. Luijten B. Fini E. Saltori C. Huijben I. Chennakeshava N. Mento F. Sentelli A. Peschiera E. Trevisan R. Maschietto G. Torri E. Inchingolo R. Smargiassi A. Soldati G. Rota P. Passerini A. Sloun R.J.G.V. Ricci E. Demi L. Deep learning for classification and localization of COVID-19 markers in point-of-care lung ultrasound IEEE Trans. Med. Imaging 39 8 2020 2676 2687 10.1109/TMI.2020.2994459 32406829
18 Carrer L. Donini E. Marinelli D. Zanetti M. Mento F. Torri E. Smargiassi A. Inchingolo R. Soldati G. Demi L. Bovolo F. Bruzzone L. Automatic pleural line extraction and COVID-19 scoring from lung ultrasound data IEEE Trans. Ultrason. Ferroelectr. Freq. Control 67 11 2020 2207 2217 10.1109/TUFFC.2020.3005512 32746195
19 Xue W. Cao C. Liu J. Duan Y. Cao H. Wang J. Tao X. Chen Z. Wu M. Zhang J. Sun H. Jin Y. Yang X. Huang R. Xiang F. Song Y. You M. Zhang W. Jiang L. Zhang Z. Kong S. Tian Y. Zhang L. Ni D. Xie M. Modality alignment contrastive learning for severity assessment of COVID-19 from lung ultrasound and clinical information Med. Image Anal. 69 2021 10.1016/j.media.2021.101975
20 Frank O. Schipper N. Vaturi M. Soldati G. Smargiassi A. Inchingolo R. Torri E. Perrone T. Mento F. Demi L. Galun M. Eldar Y.C. Bagon S. Integrating domain knowledge into deep networks for lung ultrasound with applications to COVID-19 IEEE Trans. Med. Imaging 2021 1 10.1109/TMI.2021.3117246
21 Perrone T. Soldati G. Padovini L. Fiengo A. Lettieri G. Sabatini U. Gori G. Lepore F. Garolfi M. Palumbo I. Inchingolo R. Smargiassi A. Demi L. Mossolani E.E. Tursi F. Klersy C. Di Sabatino A. A new lung ultrasound protocol able to predict worsening in patients affected by severe acute respiratory syndrome coronavirus 2 pneumonia J. Ultrasound Med. 2020 10.1002/jum.15548
22 Mento F. Perrone T. Fiengo A. Smargiassi A. Inchingolo R. Soldati G. Demi L. Deep learning applied to lung ultrasound videos for scoring COVID-19 patients: A multicenter study J. Acoust. Soc. Am. 149 5 2021 3626 3634 34241100
23 Ryan C. Collins J. Neill M.O. Grammatical evolution: Evolving programs for an arbitrary language European Conference on Genetic Programming 1998 Springer Berlin, Heidelberg 83 96 10.1007/BFb0055930 URL http://link.springer.com/10.1007/BFb0055930
24 Custode L.L. Iacca G. Evolutionary learning of interpretable decision trees 2021 arXiv:2012.07723
25 Deb K. Agrawal S. Pratap A. Meyarivan T. A fast elitist non-dominated sorting genetic algorithm for multi-objective optimization: NSGA-II International Conference on Parallel Problem Solving from Nature 2000 Springer 849 858
26 Vellido A. The importance of interpretability and visualization in machine learning for applications in medicine and health care Neural Comput. Appl. 32 24 2020 18069 18083
| 0 | PMC9746028 | NO-CC CODE | 2022-12-15 00:04:04 | no | Appl Soft Comput. 2023 Jan 13; 133:109926 | utf-8 | Appl Soft Comput | 2,022 | 10.1016/j.asoc.2022.109926 | oa_other |
==== Front
J Allergy Clin Immunol Glob
J Allergy Clin Immunol Glob
The Journal of Allergy and Clinical Immunology. Global
2772-8293
Elsevier Inc
S2772-8293(22)00095-9
10.1016/j.jacig.2022.10.003
Article
“The Conundrum of COVID-19 mRNA Vaccine-Induced Anaphylaxis”
Khalid Muhammad Bilal M.D. 1
Frischmeyer-Guerrerio Pamela A. M.D., Ph.D. 1∗
1 Laboratory of Allergic Diseases, Food Allergy Research Section, National Institute of Allergy and Infectious Diseases, National Institute of Health, Bethesda, MD, USA
∗ Corresponding Author: Pamela A Frischmeyer-Guerrerio, M.D., Ph.D. Laboratory of Allergic Diseases, Food Allergy Research Section, National Institute of Allergy and Infectious Diseases, National Institute of Health, Bethesda, MD, USA Phone: 301-402-9782, Mailing Address: MSC 1881, 10 Center drive, Bethesda, MD 20892
13 12 2022
13 12 2022
26 6 2022
19 10 2022
26 10 2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Novel mRNA vaccines have proven to be effective tools against coronavirus disease-19 (COVID-19) and changed the course of the pandemic. However, early reports of mRNA vaccine-induced anaphylaxis resulted in public alarm, contributing towards vaccine hesitancy. While initial reports were concerning for an unusually high rate of anaphylaxis to the mRNA vaccines, the true incidence is likely comparable to other vaccines. These reactions occurred predominantly in young to middle-aged females and many had prior history of allergies. Though initially thought to be triggered by polyethylene glycol (PEG), lack of reproducibility of these reactions with subsequent dosing and absent PEG sensitization points away from an IgE-mediated PEG allergy in most. PEG skin testing has poor post-test probability and should be reserved for evaluating non-vaccine related PEG allergy without influencing decisions for subsequent mRNA vaccination. Immunization stress-related response (ISRR) can closely mimic vaccine-induced anaphylaxis and warrants consideration as a potential etiology. Current evidence suggests that many individuals who developed anaphylaxis to the first dose of an mRNA vaccine can likely receive a subsequent dose after careful evaluation. The need to understand these reactions mechanistically remains critical as the mRNA platform is rapidly finding its way into other vaccinations and therapeutics.
Keywords
COVID-19
mRNA
vaccine
anaphylaxis
immunization stress-related response
ISRR
allergy
allergic reaction
polyethlene glycol
PEG
Abbreviations
AE, Adverse event
APC, Antigen presenting cell
BAT, Basophil activation test
CARPA, Complement activation related pseudoallergy
CDC, Centers for Disease Control and Prevention
COVID-19, Coronavirus Disease-2019
FDA, Food and Drug Administration
HaT, Hereditary alpha tryptasemia
Ig, Immunoglobulin
ISM, Indolent systemic mastocytosis
ISRR, Immunization stress-related response
LNP, Lipid nanoparticle
mRNA, messenger ribonucleic acid
PEG, Polyethylene glycol
SARS-CoV2, Severe Acute Respiratory Syndrome Coronavirus 2
TLR, Toll-like receptor
VAERS, Vaccine Adverse Event Reporting System
VSD, Vaccine Safety Datalink
==== Body
pmcFunding:
The work on this manuscript was supported by the Division of Intramural Research, National Institutes of Allergy and Infectious Diseases, National Institutes of Health.
Disclosures:
The authors declare no conflicts of interest.
| 0 | PMC9746073 | NO-CC CODE | 2022-12-15 00:03:23 | no | J Allergy Clin Immunol Glob. 2022 Dec 13; doi: 10.1016/j.jacig.2022.10.003 | utf-8 | J Allergy Clin Immunol Glob | 2,022 | 10.1016/j.jacig.2022.10.003 | oa_other |
==== Front
JAAD Case Rep
JAAD Case Rep
JAAD Case Reports
2352-5126
Published by Elsevier on behalf of the American Academy of Dermatology, Inc. This is an open access..
S2352-5126(22)00540-9
10.1016/j.jdcr.2022.12.001
Article
Indolent cutaneous lymphoma with gamma/delta expression after COVID-19 vaccination
Hobayan Catherine Grace BS 1
Chung Catherine G. MD 23∗
1 The Ohio State University College of Medicine, Columbus, OH
2 Departments of Pathology, The Ohio State University, Columbus, OH
3 Departments of Dermatology, The Ohio State University, Columbus, OH
∗ Corresponding author: Catherine G. Chung, MD, 2050 Kenny Road MMT930, Columbus, OH 43221
13 12 2022
13 12 2022
2 11 2022
6 12 2022
7 12 2022
© 2022 Published by Elsevier on behalf of the American Academy of Dermatology, Inc. This is an open access..
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Keywords
COVID-19 vaccine
general dermatology
injection site reaction
medical dermatology
oncology
pathology
viral vector vaccine
cutaneous lymphoma
gamma/delta T-cell lymphoma
==== Body
pmcFunding sources: None
Conflicts of Interest: None declared.
Consent for the publication of all patient photographs and medical information was provided by the authors at the time of article submission to the journal stating that all patients gave consent for their photographs and medical information to be published in print and online and with the understanding that this information may be publicly available.
| 0 | PMC9746077 | NO-CC CODE | 2022-12-15 00:03:23 | no | JAAD Case Rep. 2022 Dec 13; doi: 10.1016/j.jdcr.2022.12.001 | utf-8 | JAAD Case Rep | 2,022 | 10.1016/j.jdcr.2022.12.001 | oa_other |
==== Front
Infect Dis Now
Infect Dis Now
Infectious Diseases Now
2666-9919
Elsevier Masson SAS.
S2666-9919(22)00275-5
10.1016/j.idnow.2022.12.002
Review
Spread of viruses, which measures are the most apt to control COVID-19?
Tandjaoui-Lambiotte Yacine
Lomont Alexandra
Moenne-Locoz Pierre
Seytre Delphine
Ralph Zahar Jean ⁎
Service de Pneumologie-Infectiologie, CH Saint Denis, 2 rue Dr. Delafontaine, 93200
Unité de Prévention du Risque Infectieux, Service de microbiologie clinique, GHU Paris Seine Saint-Denis, Université Sorbonne Paris Nord
⁎ Corresponding author at: Unité de Préveion du Risque Infectieux
13 12 2022
13 12 2022
11 11 2022
22 11 2022
6 12 2022
© 2022 Elsevier Masson SAS. All rights reserved.
2022
Elsevier Masson SAS
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The persistent debate about the modes of transmission of SARS-CoV2 and preventive measures has illustrated the limits of our knowledge regarding the measures to be implemented in the face of viral risk. Past and present (pandemic-related) scientific data underline the complexity of the phenomenon and its variability over time. Several factors contribute to the risk of transmission, starting with incidence in the general population (i.e., colonization pressure) and herd immunity. Other major factors include intensity of symptoms, interactions with the reservoir (proximity and duration of contact), the specific characteristics of the virus(es) involved, and a number of unpredictable elements (humidity, temperature, ventilation…).
In this review, we will emphasize the difficulty of “standardizing” the situations that might explain the discrepancies found in the literature. We will show that the airborne route remains the main mode of transmission. Regarding preventive measures of prevention, while vaccination remains the cornerstone of the fight against viral outbreaks, we will remind the reader that wearing a mask is the main barrier measure and that the choice of type of mask depends on the risk situations.
Finally, we believe that the recent pandemic should induce us in the future to modify our recommendations by adapting our measures in hospitals, not to the pathogen concerned, which is currently the case, but rather to the type of at-risk situation.
Keywords
airborne transmission
COVID-19
preventive measures
hand hygiene
mask
==== Body
pmc1 Introduction:
The COVID-19 pandemic raised many questions about preventive measures in community and in-hospital settings. These questions persist more than 2 years after the onset of the pandemic. The contradictory data in different published studies have added to the confusion and caused concern. Awareness the modes of transmission of a pathogen is an important, if not essential step in controlling the risk of its spread. Like all respiratory viruses, SARS-CoV-2 is transmitted by the respiratory and airborne routes. However, it is important to specify not only the role of each transmission route, but also the so-called droplet or airborne character of the respiratory route. In this work, we will take a critical look at the data in the literature, and try to provide “food for thought” and aid to decision-making.
2 What we previously knew about respiratory viruses
Prior to the pandemic [1], [2], existing data could have allowed us to approach the initial phases more serenely in terms of preventive measures designed to decrease risk of transmission. The fear aroused by the lack of knowledge of SARS-CoV2, justified on account of clinical consequences, appears less appropriate as regards modes of transmission. Previous data concerning the modes of transmission of respiratory viruses made it possible to anticipate risks and to propose preventive measures [3], [4]. One of the major difficulties in understanding transmission phenomena lies in the need to take into account a multitude of confounding factors. At the individual level, the probability of being infected involves both the probability of being exposed to the pathogen and frequency of exposure. For example, a route with a low probability of acquisition by exposure but occurring frequently may contribute more to the total number of acquisitions than the ones with rare exposure but with a high probability of transmission by exposure.
At the collective level, several factors contribute to the risk of transmission: the collective reservoir or colonization pressure [i.e., disease prevalence] [5], the individual reservoir e [i.e., symptom intensity, proximity and duration of contact], and the immunity of exposed hosts [6]. In addition, the frequency of infected and asymptomatic individuals is variable for different viruses and makes it more difficult to understand the modes of transmission. In fact, the greater the number of asymptomatic individuals, the greater the need for systematic screening of contacts in order to define secondary acquisitions and transmission routes. These are some reasons why it seems even more difficult to define the modes of transmission during pandemic periods.
3 Modes of transmission of respiratory viruses:
As for other pathogens, different modes of transmission of respiratory viruses exist [7], [8]. For a very long time, we artificially divided transmission into 3 different modes, forgetting that for the same pathogen, there were multiple modes of transmission that varied over time. While one mode may dominate, it is now clear that different modes overlap.
Transmission can be indirect, through the contaminated hands of an individual, whether in a hospital environment [health care workers or visitor] or in a community environment [colleague or relatives]. This contamination can be the result of contact with a human reservoir [infected person] or with a contaminated environment [surfaces]. In the latter case, it will depend on the frequency of environmental contamination linked to a given reservoir, on the virus' ability to survive in this environment, and on the hand disinfection products used.
The second mode of transmission is direct, airborne, from the reservoir [9]. It requires direct contact with the reservoir and duration of exposure. This mode of transmission is divided into two distinct modes: droplet transmission and airborne transmission. This definition is meant to differentiate between short-range [droplet transmission] and long-range [airborne transmission] particles. It is generally based on the size of the exhaled droplets, the former being particles of more than 5µm that deposit very quickly on surfaces and the latter being particles of less than 5µm, therefore lighter and able to “float” longer in the air. Droplet transmission refers to direct inhalation of the virus exhaled by an infected person when a contact person is close to him/her. Airborne transmission refers to the inhalation of small droplets of aerosol exhaled by an infected person at a distance exceeding two meters.
Numerous works on respiratory viruses preceding the crisis had questioned this definition and underlined the complexity of the phenomenon [2], [10]. Several factors could explain the mixed character [droplets and air] of the transmission of respiratory viruses. They can be summarized as factors related to the type of reservoir [intensity and symptomatology of the emitting source], external factors [physio-chemical phenomena that explain the possibility of desiccation of the emitted particles], and conditions of care. For example, aerosol generating procedures [AGP] such as intubation or nebulizer therapy [3] generate a large quantity of particles with variable flow and speed. These factors may explain why the risk cannot be categorized as droplets or air. Several studies of patients infected by the influenza virus have highlighted the presence of particles of different sizes at a variable distance from the emitting source [2], depending on the intensity of the symptoms [2], [11] and the conditions under which the samples were taken. These findings suggest that particles >5µm [supposed to be responsible for droplet transmission] and particles <5µm [supposed to be responsible for airborne transmission] are emitted by the same patient at the same moment [12]. If there is a continuum in the size of the emitted particles, this cannot explain why some viruses are supposed to be highly infective in an airborne route, whereas others are associated with a droplet route. This being said, the risk should be considered in relation to viral load and the distance from the source. In fact, it has been shown that the small droplets produced by an infected individual have a low viral load [13]. They do not pose a serious threat of infection, whereas larger droplets have a higher viral load with higher risk for healthy individuals, but they settle quickly. This phenomenon could be more complex as these larger droplets partially evaporate during their ballistic trajectory before settling on the ground. In this situation, the droplets shrink due to friction forces and evaporation in atmospheric air. This could transform a large particle >5µm into various particles, of which some become suspended in the air but still retain their pre-evaporated viral load. These infectious droplets can travel longer distances in the air and have a high enough viral load to infect healthy individuals [13]. It is important to highlight that many factors can impact the behavior of respiratory droplets: temperature and humidity, [12], [13], [14], and air flows due to ventilation systems or human activity. Indeed, the co-presence of a ventilation system and of an infectious source is at the origin of airflow disturbances that could modify the trajectory and the range of the particles. [15].
In the community, the risk of transmission/contamination seems to be much higher than in a hospital environment. There are many reasons for this, including duration of contact with the potential reservoir, proximity of the reservoir, the absence of protective measures and the viral load in the patient’s airways [high at disease onset]. The risk is certainly different at home [where contact times are longer] compared to other living environments [including public transportation]. However, this difference is modifiable by the number of reservoirs present [i.e., colonization pressure], promiscuity, and level of air renewal. Similarly, the risk is related not only to direct but also to indirect transmission.
In the hospital, the risks differ, depending on the type of structure. In long-term care facilities, they seem greater than in acute care. There are many reasons for this, including colonization pressure, promiscuity and shared activities. In acute care, risks are variable and depend on colonization pressure [the higher it is, the greater the risks], the earliness of patient hospitalization [in relation to the onset of symptoms], and the extent of care load and the duration of contact with infected patients.
4 Benefits of wearing a mask
For more than a decade there has been a scientific debate about the usefulness of wearing a mask and its effectiveness during periods of viral circulation. While medical masks are designed to block droplets, N95 respirators are tailored to capture 95% of particles up to 0.3 µm in diameter. Numerous studies have suggested the importance of wearing a mask to prevent the risk of infection. A prospective cluster-randomized study conducted during the 2006 winter season found that compliance with mask use significantly reduced the risk of influenza-like illness [ILI] in households [16]. Furthermore, a blinded cluster randomized study [17] was aimed at determining whether hand hygiene and/or face mask use prevented influenza transmission in households. The authors found that hand hygiene, with or without wearing a face mask, appeared to reduce influenza transmission, but differences from the control group were not significant. That much said, infection appeared to be reduced when interventions were implemented within 36 hours of symptom onset, a finding suggesting that hand hygiene and masks may prevent transmission of influenza virus at home when implemented earlyt after symptom onset in the index patient.
In a meta-analysis including 6 randomized and 23 observational studies, Offedu and colleagues [18] suggested a protective effect of masks and N95 respirators against clinical respiratory illness (CRI) and Influenza-like illness (ILI) in healthcare settings. Moreover, the authors suggested that N95 respirators conferred superior protection against CRI and laboratory-confirmed bacterial infections, but surprisingly not against viral infections or ILI.
However, a previous study evaluating continuous or targeted N95 respirators during high-risk procedures involving 1669 hospital-based healthcare workers in Beijing (China) during the winter of 2009-2010 found a higher clinical respiratory illness rate in the medical mask group, followed by the group wearing targeted N95 respirators and the group continuously wearing them. Furthermore, after adjustment for confounders, compared with targeted N95 respirator use, only continuous N95 respirator use remained significant against CRI [19].
Most guidelines recommend masks for droplet-transmitted infections. As “airborne precautions,” N95 masks offer superior protection against droplet-transmitted infections. To ensure the health and safety of healthcare workers, the superiority of N95 respirators in preventing respiratory infections should be highlighted in infection control guidelines.
5 Why should prevention measures for SARS-CoV2 be different ?
5.1 The virus, from the prevention standpoint
Concerning viruses, several characteristics specific to the virus suggest a variable risk of secondary transmission. Among these characteristics, the ability to survive in the environment and resistance to detergents and disinfectants appear primordial. Indeed, adding an environmental reservoir to the human reservoir would increase the risk of transmission. At the beginning of the pandemic, Van Doremalen et al. evaluated the stability and desiccation rate of the SARS-CoV2 virus compared to SARS-CoV1 on different surfaces, suggesting 3h survival of the virus in aerosols and viability on surfaces lasting up to 3 days[20]. Survival in the environment was found to vary, depending on initial concentration, types of surfaces and temperature and humidity conditions [20], [21]. Many studies with contradictory results have analyzed survival under different conditions. For example, iphysico-chemical conditions act as an amplifier of risk. Experimental data show that SARS-CoV-2 activity decreases with increasing temperature. SARS-CoV-2 can survive for 14 days at 4°C, 1 day at 37°C and only 30 minutes at 56°C. Similarly, a person is more likely to be infected when the relative humidity is low, between 10 and 20%. In addition, experiments conducted over the past 60 years have indicated that the activity of viral pathogen-carrying droplets is negatively related to ambient humidity and that dryness promotes virus spread. SARS-CoV-2 is sensitive to disinfection[22] and exhibits stability over a wide range of pH values (pH 3-10) at room temperature [21].
In fact, SARS-CoV-2 is sensitive to a wide variety of disinfectants. Lipid solvents, including ethanol (>75%), formaldehyde (>0.7%), isopropanol (>70%), povidone-iodine (>0.23%), sodium hypochlorite (>0.21%), or hydrogen peroxide (H2O2; >0.5%), can be used to inactivate SARS-CoV-2 [23]
6 What do we know about SARS-CoV2 transmission?
Recent data on SARS-CoV2 suggest that the risk of transmission begins during the period of asymptomatic infection, 48-72 hours before the onset of symptoms, and progressively decreases until in most situations it disappears by the 8th day after the first symptoms. Respiratory samples taken from infected symptomatic and asymptomatic patients, as well as epidemiological and clinical studies, have demonstrated viral shedding beginning 48 to 72 hours before the first symptoms and persisting up to 8 days after. Transmissionwould begin 2 to 3 days before the onset of symptoms, reach its peak approximately 1 day before the onset of symptoms [24], and rapidly decrease within the first 7 days [24], [25], [26].
SARS-CoV-2 viral loads in the respiratory tract decline rapidly after symptom onset. The highest loads shift from the upper to the lower respiratory tract [27], [28] While patients with severe disease have higher respiratory viral loads than those with mild disease, all loads decline over time [29]. Researchers in China have estimated the duration of RNA shedding from various sites based on detailed analysis of samples from 49 patients with COVID-19 and reported a median duration of shedding from the nasopharynx of 22 days for mild cases and 33 days for severe cases. The median duration of excretion from the nasopharynx was likewise 22 days for mild cases and 33 days for severe cases, with some individuals excreting for more than 2 months[30]. It should be noted that the period of infectiousness is much shorter than the duration of detectable RNA. For mild to moderate cases, infectious virus can be isolated from specimens only until approximately the 8th day of symptoms. Multiple studies have found virtually no viable virus in patients with mild to moderate disease after 10 days of symptoms, despite continuous and frequent RNA shedding[27], [31], [32]. In a study that included patients from 0 to 21 days after symptom onset, viable virus was isolated in 26 out of 90 samples, but no viral growth was found when the cycle threshold was greater than 24 or the patient had more than 8 days of symptoms [33]. A group from the Netherlands evaluated 129 hospitalized patients, 89 of whom required intensive care, and collected upper and lower tract samples [34]. Isolation of infectious virus occurred on average 8 days after the onset of symptoms. The probability of isolating infectious virus was less than 5% at 15.2 days, and decreased with increasing time since symptom onset, lower viral loads, and higher neutralizing antibody titers. The “tardiest” isolation of infectious virus occurred 20 days after symptom onset. Despite the late isolation of infectious virus, no late transmission was documented, including in health care settings. Some field observations have confirmed that after 6 days of symptoms, the risk of transmission from an index case to HCW is low[35]
To summarize, the period of greatest risk seems to be between Day 2 and Day 3 of the first symptoms, which would explain a greater risk of contamination in the community than in the hospital, since patients are rarely hospitalized as soon as the first symptoms appear, in a situation where the risk of transmission is decreasing. As with many other viruses, pre-symptomatic patients as well as symptomatic patients, are at the origin of transmission. However, the risk of transmission seems to be lower in the former than in the latter, and the period during which pre-symptomatic patients are infectious remains unknown to this day, as do the dynamics of excretion in asymptomatic patients. Among those who develop symptoms, the secondary attack rate may increase with the intensity of the source case's symptoms. However, many authors have suggested the presence of super spreaders [36].
7 Transmission and Animal Model:
Numerous animal models [37] have been used to understand the modes of transmission and pathogenicity of SARS-Cov2. Transmission by direct contact remains the most frequently identified mode of transmission regardless of the animal model used [38]. However, in ferrets, authors have emphasized the importance of airborne transmission over short distances [animals separated by a maximum of 10 cm] more frequently than over long distances [animals separated by more than one meter][39]. Similarly, numerous studies carried out in the hamster model and testing the different variants of SARS-CoV2[40] have shown a risk of transmission by contact [40] and by air, with risks varying according to the different variants. In one study, the authors demonstrated[41]that infected hamsters produced aerosolized SARS-CoV-2 infectious particles both before and during the onset of a mild disease. The average emission rate in this study was 25 infectious virions/hour on days 1 and 2 after inoculation, with average levels of viral RNA 200-fold higher than infectious virus in the aerosol particles. In this work, the majority of the virus was contained in particles.
8 Human data
In a recent study, the authors demonstrated a correlation between the viral load of index cases and the occurrence of secondary cases [42]. In this study including 212 index cases and 365 contacts, 19% tested positive after their exposure, and after adjustment for cough, time between test and exposure risk of transmission to a close contact was significantly associated with the viral load. In a home setting, an original study based on the detection of exhaled mRNA showed a correlation between the threshold of mRNA positivity and secondary transmission[43]. Frequency of home transmission was associated with viral load.
Among community activities with high transmission risk, it seems that private gatherings (family meals) or densely populated public places lead to higher viral diffusion [44]. Indeed, in a study including 32 [45] studies involving 68206 participants, the authors highlighted lower attack rates in the context of extra-familial gatherings. For example, in this work, transportation (RR 10.55, 95% confidence interval (CI) 1.43-77.85), medical care (RR 11.68, 95% CI 1.58-86.61), and work or study locations (RR 10.15, 95% CI 1.40-73.38) had lower attack rates. All in all, attack rates were highest at home (95.3%), meals or gatherings (81.4%), public places (58.9%), daily conversation (50.1%), transportation (30.8%), medical care (18.2%), and work or study sites (15.3%). In another study of home attack rates [46], including 276 index cases and 644 contact cases, the authors found 200 cases of secondary SARS-CoV-2 infection, representing an attack rate of 45.7% (95% CI: 39.7-51.7%) per household. Asymptomatic or mildly symptomatic index cases had a lower risk of transmission. This work also showed that the majority of transmissions occurred early after the introduction of SARS-CoV-2 into a household.
9 Is transmission related to surfaces or to the air?
During the pandemic, numerous studies sought to identify the risk factors associated with environmental contamination and to distinguish between direct transmission through the air and indirect transmission through contaminated surfaces [47].
Environmental contamination depends not only on virus characteristics and patient-related factors but also on several other factors such as temperature, exposure to UV, moisture or surface characteristics [22]. The results of the various studies carried out during the pandemic are contradictory, in terms of both air sampling and surface sampling. In addition, many limitations to the different studies must be underlined before considering the frequency of contamination of the different types of samples. Firstly, most published studies have initially searched by RT-PCR for viral RNA and not for the presence of viable virus, neglecting the possibility that the results obtained do not reflect actual infectivity. Secondly, when interpreting their results few studies have taken into account certain confounding factors, such as the intensity and precocity of symptoms and the duration of exposure of the environments to the reservoir. Thirdly, the heterogeneity of the populations studied make it difficult to interpret the results. Finally, during the pandemic, several sampling devices and methods were used for surfaces or for air, and they were not standardized (Table 1 ).Table 2. Table 1 Confounding factors accounting for the discrepancies in the literature .
Patient-related Intensity of symptoms (coughing, sneezing)Sampling time (compared to the onset of symptoms)Colonization pressure
Physico-chemical data HumidityTemperature
Ward-related Surface and volume of the roomsVentilation sytems (Air change rate, Air filtration)
Sampling method Sampling techniques, Positioning of sampling systemsDuration and quantity of air intakeMicrobiological methods (RT-PCR vs Viral isolation)
Table 2 Preventive measures according to different situations.
Setting Mode of transmission Preventive measures Comments
Surface Air Vaccination Hand Hygien Mask (type)
Community Home ++ ++ ++ ++ - At home, airborne transmission occurs early (even before symptoms). Wearing a surgical mask is interesting for symptomatic patients, less for contact patients.
Gathering + ++ ++ + + (surgical) The risk depends on the colonization pressure and the volume and ventilation level of the room
Public transport + ++ ++ + +(cloth) The risk is low compared to the situations described above. In ventilated spaces, virus survival on surfaces is low.
Hospital Emergency department ++ ++ ++ ++ ++ (surgical) At risk because of the number of patients managed, the proximity of care and the earliness of management in relation to the onset of symptoms. N95 respirators if AGP
Medical Ward ++ ++ ++ ++ ++ (surgical) Less risk than ED, due to patient “tracking” and isolation. However, the delay in management from the onset of symptoms is the major risk factor. N95 respirators if AGP
ICU ward + + ++ ++ ++ Management of ICU patients is usually delayed relative to the onset of symptoms. Patients are generally less shedding. N95 respirators if AGP
10 Frequency and risk factors of surface contamination
The data in the literature are variable and contradictory suggesting contamination on between 5% and 50% of contaminated surfaces. In an initial study carried out in China [48], the authors highlighted surface contamination in 57% of situations. In this study, they noted high surface contamination in ten (66.7%) of the 15 patients during the first week of illness, and in only three (20%) after the first week of illness, suggesting like others[48] a decrease in the frequency of environmental contamination. Similarly [49], in another prospective study the authors correlated the cycle threshold (Ct) for reverse transcription polymerase chain reaction (RT-qPCR) of severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) on a nasopharyngeal swab (test taken closest to the day of collection) with the percentage of RT-qPCR-positive surfaces [Spearman rank correlation coefficient of -0.48 (P<0.001)]. In a more recent study [50], the authors assessed environmental contamination of inpatient rooms in a dedicated COVID-19 unit. Among the 347 samples, they found a contamination rate of 5.5% based on RT-PCR results and only 0.3% based on cell culture. They confirmed that despite the frequent identification of viral RNA by RT-PCR, the viable nature of the virus in surface samples remains rare [51]. Finally, a meta-analysis including about fifty in-hospital studies suggested that contamination was more frequent in the ICU than outside and, specifically, in patient rooms compared to non-patient areas[52]. Numerous factors explain these discrepancies, and analysis of the data in the literature underscores several points concerning the contamination of surfaces, which depends on (1) the earliness of sampling compared to the onset of symptoms, (2) the intensity of symptoms noted at the time of sampling and (3) the type of surface sampled [53]. Also, the longer the time from the onset of symptoms, the more negative the results of environmental sampling will be. All in all, environmental contamination is less frequent in asymptomatic patients compared to symptomatic ones. However, recent study suggested a low risk of viral transmission, in most instances by fomites [54].
11 Air contamination
Aerosol transmission was better understood during the pandemic. Several studies have tried to detect SARS-CoV-2 in the air. As mentioned above, one of the difficulties of interpretation is related to the different sampling methods applied. In a systematic review including 25 studies [55], while the authors identified 15 studies with at least one positive sample, only one of them suggested the presence of viable viruses in the air [56]. In a more recent prospective study conducted in an acute care setting between March and May 2020, the authors suggested that air contamination was rarely identified (2% among 146 samples) [51]. In a meta-analysis of in-hospital data, the authors suggested the presence of SARS-CoV2 in 16% of air sampling [52]. In this meta-analysis, SARS-CoV2 RNA was more frequently detected in patient areas, and the highest prevalence was found in ICU patients. However, the viral concentrations were lower in ICU patients’ rooms when given in copies and did not differ from non-ICU patients in CT-values. Finally, the most recent systematic review[57], including 73 published papers, suggested lower air contamination outdoors compared to hospitals and care facilities, suggesting a lower acquisition risk.
An issue widely debated in the literature is that of the risk associated with ventilation support. To date, several studies have suggested that there is no increased risk with CPAP or HFO [58], [59], whereas, as suggested in the aforementioned meta-analysis suggested a higher risk of detecting surface or airborne SARS contamination when AGPs are performed.
Despite some discrepancies, surface contamination has been more frequently identified than air contamination. While many authors have emphasized the frequency of surface contamination, few have proven the viability of the virus. Indeed, most of the studies are limited to the identification of viral RNA, and do not deal with the viability of the virus by cell culture[60].
Few studies have attempted to identify the proportion of surface transmission compared to airborne transmission. Despite debated data in the literature, there are some indirect arguments to suggest a low risk of transmission through contaminated surfaces. Some authors have even suggested, with the help of video recordings, that there is no transmission from surfaces in the community [61]. Similarly, although the frequency of positivity of surface samples in RT-PCR is high, the thresholds obtained suggest a low viral concentration and the cytopathic effect has rarely been demonstrated. However, it has become increasingly clear that the fomite route is less important than previously thought[62].
Evidence has been accumulated, especially since the outbreak of COVID-19, of the primacy of the airborne transmission route. The most important evidence of airborne transmission concerns long-range airborne transmission, especially in poorly ventilated indoor environments[15], [63]. Relevant studies include analyses of COVID-19 outbreaks on a bus, a cruise ship, a gym, and a restaurant. It has become clear that poor ventilation increases long-range airborne transmission. On the other hand, in the majority of the articles included, this scoping review shows that the risk of SARS-CoV-2 infection via contaminated surfaces is low [61].
12 Air or droplets?
This debate seems outdated. In fact, the pandemic has reminded us that there exist several modes of transmission; the question is to know what is the most appropriate mode of prevention according to the different situations. In a very specific moment, a human being infected by SARS-CoV-2 would exhale a very high number of droplets not all of which would be of the same size. During a single act of exhaling, the size of droplets emitted by a human being follows a normal distribution around a mean. Depending on vesico-elastic properties of the respiratory mucus and saliva, both of which are modified by the natural history of the respiratory infection, the mean of the normal distribution changes. A single patient can exhale a continuum of particles from aerosols to droplets, and that the combination can change along time. Numerous models and reviews published during the pandemic pointed out that the transmission of respiratory viruses in general and SARS-Cov2 in particular is mixed, including both droplet and airborne transmission [64]. Similarly, many authors have cited factors and variables associated with the different modes of transmission. In a recent systematic review including 18 community-based studies conducted during the pandemic [65], the authors concluded that long-distance airborne transmission was possible in indoor environments such as restaurants, workplaces, and choral settings. They identified factors such as inadequate air exchange as likely contributors to transmission. These findings reinforce the need for mitigation measures in indoor environments, particularly adequate ventilation. Numerous studies have sought to identify the presence of SARS-Cov2 in air. Despite conflicting data, the presence of SARS-CoV2 has rarely been demonstrated in RT-PCR or cell culture samples. However, these contradictory data are essentially related to the different sampling techniques and patient profiles[66].
13 Which mask for which risk?
The debate about the effectiveness of surgical masks or N95 respirators was one of the highlights of the pandemic. While all laboratory studies have demonstrated the expected superiority of filtration by N95 respirators, their superiority in clinical practice remains debated. One disadvantage of the N95 respirators is the increased risk of leakage if the mask is not fitted, and lower adherence has been reported due to higher rates of adverse events. Despite numerous methodological biases, many studies have tried to answer the question of the effectiveness of different masks in terms of prevention for health care workers. A meta-analysis including 4 randomized studies [67] suggested, with a low level of certainty, no difference for either clinical or microbiologically proven infections. Similarly [68], a second meta-analysis including 6 studies suggested that there are insufficient data to definitively determine whether N95 respirators are superior to medical masks in protecting against acute respiratory infections. Further randomized trials are needed to compare the above respiratory protection methods in the context of COVID-19 incidence. The meta-analysis by Maccintyre et al. included the 2 randomized controlled studies performed by the same team in China and including 3591 subjects while comparing 4 interventions: (i) continuous use of an N95 respirator, (ii) targeted use of an N95 respirator, (iii) use of a medical mask, and (iv) control arm. In the adjusted analysis, this study highlighted laboratory-confirmed bacterial colonization rates (RR 0.33, 95% CI 0.21-0.51), laboratory-confirmed viral infections (RR 0.46, 95% CI 0.23-0.91), and droplet-transmitted infections (RR 0.26, 95% CI 0.16-0.42) were significantly lower in the continuous N95 group. Rates of laboratory-confirmed influenza were also lowest in the continuous N95 group (RR 0.34, 95% CI 0.10-1.11), but the difference was not statistically significant. Rates of laboratory-confirmed bacterial colonization (RR 0.54, 95% CI 0.33-0.87) and droplet-transmitted infections (RR 0.43, 95% CI 0.25-0.72) were also lower in the targeted N95 group, but not in the medical mask group.
14 Discussion-Conclusion :
Despite initial uncertainty, the aerosol route is now recognized as the principal path of transmission of COVID-19. Several arguments favor airborne transmission: the effect of ventilation, the difference between indoor and outdoor transmission, the possibility of transmission despite the use of masks and goggles, animal data, and airflow simulations[69]. Several works have suggested more frequent contamination of surfaces compared to airborne viruses. However, it is difficult if not impossible to standardize different clinical situations and many factors, which vary over time, contribute to the risk of transmission. Concerning the clinical elements, it is important to note that most of the studies were carried out on a given day, even though the infection and the risks of contamination continue to evolve. In several studies, the time lapse between sample taking and onset of symptoms remains a first limiting factor. In fact, the longer the time elapsed, the lower the probability of finding the virus in the environment. That said, there exists a window during which the possibility of finding the virus seems higher (D-2 to D+8). Secondly, the fact that patient profiles are not taken into account makes it difficult to interpret the studies and explains some of the discrepancies. Furthermore, whatever the time of sampling, the intensity of respiratory symptoms is associated with environmental contamination. The third study limitation is related to the unit in which sampling is carried out. For example, intubated and ventilated patients in intensive care units are less at risk of aerosolization than those hospitalized in medical units. Fourthly, the non-standardization of sampling methods further complicates the interpretation of literature data.
Like clinical studies, fundamental studies encounter numerous limitations: the lack of standardization of the diffusion methods, the viral concentrations emitted, and the difficulty of taking into account the physicochemical data associated with the risks of diffusion in the air. Despite the contradictory data during the pandemic [70], there are several clinical and theoretical arguments in favor of N95 respirators in contact with infected patients. Studies during the pandemic suggested the superiority in terms of filtration of N95 respirators compared to surgical masks, and in some cases[71], a lower risk of contamination of health care workers. It is now clear that the presence of fine particles exhaled by infected patients and the risk situations in hospitals [3] are two factors that argue in favor of N95 respirators. However, N95 respirators require a seal[72], [73], and previous work suggested that in terms of the risk of penetration, peripheral leaks had a greater impact than level of filtration.
In conclusion, while airborne transmission of respiratory viruses is a certainty, air vs. droplet is in our opinion an outdated debate. Conditions associated with patient status, ventilation data or therapeutic procedures, can transform droplet risk into air risk. If transmission is mainly airborne, it is difficult to measure the environmental part linked to surface contamination. However, studies suggest that the further one moves away from the onset of symptoms, the more this risk decreases. Future work should discuss not so much preventive measures (mask and hand hygiene) as the situations in which the systematic wearing of a mask is useful and the type of mask which, in our opinion, should be adapted not to types of virus, but rather to the care situations that may expose a HCW to greater or lesser risk.
==== Refs
References
1 Rule A.M. Apau O. Ahrenholz S.H. Brueck S.E. Lindsley W.G. de Perio M.A. Healthcare personnel exposure in an emergency department during influenza season PLoS One. 13 8 2018 e0203223 30169507
2 Yan J. Grantham M. Pantelic J. Bueno de Mesquita P.J. Albert B. Liu F. Infectious virus in exhaled breath of symptomatic seasonal influenza cases from a college community Proc Natl Acad Sci U S A. 115 5 2018 Jan 30 1081 1086 29348203
3 Tran K. Cimon K. Severn M. Pessoa-Silva C.L. Conly J. Aerosol generating procedures and risk of transmission of acute respiratory infections to healthcare workers: a systematic review PLoS One. 7 4 2012 e35797 22563403
4 Thomas R.E. Do we have enough evidence how seasonal influenza is transmitted and can be prevented in hospitals to implement a comprehensive policy? Vaccine. 34 27 2016 Jun 8 3014 3021 27171752
5 Salgado C.D. Farr B.M. Hall K.K. Hayden F.G. Influenza in the acute hospital setting Lancet Infect Dis. 2 3 2002 Mar 145 155 11944184
6 Salgado C.D. Giannetta E.T. Hayden F.G. Farr B.M. Preventing nosocomial influenza by improving the vaccine acceptance rate of clinicians Infect Control Hosp Epidemiol. 25 11 2004 Nov 923 928 15566025
7 Cowling B.J. Ip D.K.M. Fang V.J. Suntarattiwong P. Olsen S.J. Levy J. Aerosol transmission is an important mode of influenza A virus spread Nat Commun. 4 2013 1935 23736803
8 Gautret P. Gray G.C. Charrel R.N. Odezulu N.G. Al-Tawfiq J.A. Zumla A. Emerging viral respiratory tract infections–environmental risk factors and transmission Lancet Infect Dis. 14 11 2014 Nov 1113 1122 25189350
9 Cole E.C. Cook C.E. Characterization of infectious aerosols in health care facilities: an aid to effective engineering controls and preventive strategies Am J Infect Control. 26 4 1998 Aug 453 464 9721404
10 Canini L. Woolhouse M.E.J. Maines T.R. Carrat F. Heterogeneous shedding of influenza by human subjects and its implications for epidemiology and control Sci Rep. 14 6 2016 Dec 38749
11 Fennelly K.P. Particle sizes of infectious aerosols: implications for infection control Lancet Respir Med. 8 9 2020 Sep 914 924 32717211
12 Chamseddine A. Soudani N. Kanafani Z. Alameddine I. Dbaibo G. Zaraket H. Detection of influenza virus in air samples of patient rooms J Hosp Infect. 108 2021 Feb 33 42 33152397
13 Liu L. Wei J. Li Y. Ooi A. Evaporation and dispersion of respiratory droplets from coughing Indoor Air. 27 1 2017 Jan 179 190 26945674
14 Asadi S. Wexler A.S. Cappa C.D. Barreda S. Bouvier N.M. Ristenpart W.D. Aerosol emission and superemission during human speech increase with voice loudness Sci Rep. 9 1 2019 Feb 20 2348 30787335
15 Luongo J.C. Fennelly K.P. Keen J.A. Zhai Z.J. Jones B.W. Miller S.L. Role of mechanical ventilation in the airborne transmission of infectious agents in buildings Indoor Air. 26 5 2016 Oct 666 678 26562748
16 MacIntyre C.R. Cauchemez S. Dwyer D.E. Seale H. Cheung P. Browne G. Face mask use and control of respiratory virus transmission in households Emerg Infect Dis. 15 2 2009 Feb 233 241 19193267
17 Cowling B.J. Chan K.H. Fang V.J. Cheng C.K.Y. Fung R.O.P. Wai W. Facemasks and hand hygiene to prevent influenza transmission in households: a cluster randomized trial Ann Intern Med. 151 7 2009 Oct 6 437 446 19652172
18 Offeddu V. Yung C.F. Low M.S.F. Tam C.C. Effectiveness of Masks and Respirators Against Respiratory Infections in Healthcare Workers: A Systematic Review and Meta-Analysis Clin Infect Dis. 65 11 2017 Nov 13 1934 1942 29140516
19 MacIntyre C.R. Wang Q. Seale H. Yang P. Shi W. Gao Z. A randomized clinical trial of three options for N95 respirators and medical masks in health workers Am J Respir Crit Care Med. 187 9 2013 May 1 960 966 23413265
20 van Doremalen N. Bushmaker T. Morris D.H. Holbrook M.G. Gamble A. Williamson B.N. Aerosol and Surface Stability of SARS-CoV-2 as Compared with SARS-CoV-1 N Engl J Med. 382 16 2020 Apr 16 1564 1567 32182409
21 Chin A.W.H. Chu J.T.S. Perera M.R.A. Hui K.P.Y. Yen H.L. Chan M.C.W. Stability of SARS-CoV-2 in different environmental conditions Lancet Microbe. 1 1 2020 May e10 32835322
22 Paton S. Spencer A. Garratt I. Thompson K.A. Dinesh I. Aranega-Bou P. Persistence of Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) Virus and Viral RNA in Relation to Surface Type and Contamination Concentration Appl Environ Microbiol. 87 14 2021 Jun 25 e0052621 33962986
23 Viana Martins C.P. Xavier C.S.F. Cobrado L. Disinfection methods against SARS-CoV-2: a systematic review J Hosp Infect. 119 2022 Jan 84 117 34673114
24 He X. Lau E.H.Y. Wu P. Deng X. Wang J. Hao X. Temporal dynamics in viral shedding and transmissibility of COVID-19 Nat Med. 15 2020 Apr 1 4
25 Lauer S.A. Grantz K.H. Bi Q. Jones F.K. Zheng Q. Meredith H.R. The Incubation Period of Coronavirus Disease 2019 (COVID-19) From Publicly Reported Confirmed Cases: Estimation and Application Ann Intern Med. 172 9 2020 May 5 577 582 32150748
26 Yang L. Dai J. Zhao J. Wang Y. Deng P. Wang J. Estimation of incubation period and serial interval of COVID-19: analysis of 178 cases and 131 transmission chains in Hubei province, China Epidemiol Infect. 19 148 2020 Jun e117
27 Wölfel R. Corman V.M. Guggemos W. Seilmaier M. Zange S. Müller M.A. Virological assessment of hospitalized patients with COVID-2019 Nature 2020 Apr 1
28 To K.K.W. Tsang O.T.Y. Leung W.S. Tam A.R. Wu T.C. Lung D.C. Temporal profiles of viral load in posterior oropharyngeal saliva samples and serum antibody responses during infection by SARS-CoV-2: an observational cohort study The Lancet Infectious Diseases. 20 5 2020 May 1 565 574 32213337
29 Liu Y. Yan L.M. Wan L. Xiang T.X. Le A. Liu J.M. Viral dynamics in mild and severe cases of COVID-19 Lancet Infect Dis. 20 6 2020 Jun 656 657 32199493
30 Sun J. Xiao J. Sun R. Tang X. Liang C. Lin H. Prolonged Persistence of SARS-CoV-2 RNA in Body Fluids Emerg Infect Dis. 26 8 2020 Aug 1834 1838 32383638
31 Perera R.A.P.M. Tso E. Tsang O.T.Y. Tsang D.N.C. Fung K. Leung Y.W.Y. SARS-CoV-2 Virus Culture and Subgenomic RNA for Respiratory Specimens from Patients with Mild Coronavirus Disease Emerg Infect Dis. 26 11 2020 Nov 2701 2704 32749957
32 Singanayagam A, Patel M, Charlett A, Lopez Bernal J, Saliba V, Ellis J, et al. Duration of infectiousness and correlation with RT-PCR cycle threshold values in cases of COVID-19, England, January to May 2020. Euro Surveill. 2020 Aug;25(32).
33 Bullard J. Dust K. Funk D. Strong J.E. Alexander D. Garnett L. Predicting Infectious Severe Acute Respiratory Syndrome Coronavirus 2 From Diagnostic Samples Clin Infect Dis. 71 10 2020 Dec 17 2663 2666 32442256
34 van Kampen J.J.A. van de Vijver D.A.M.C. Fraaij P.L.A. Haagmans B.L. Lamers M.M. Okba N. Duration and key determinants of infectious virus shedding in hospitalized patients with coronavirus disease-2019 (COVID-19) Nat Commun. 12 1 2021 Jan 11 267 33431879
35 Cheng H.Y. Jian S.W. Liu D.P. Ng T.C. Huang W.T. Lin H.H. Contact Tracing Assessment of COVID-19 Transmission Dynamics in Taiwan and Risk at Different Exposure Periods Before and After Symptom Onset. JAMA Intern Med. 2020 May 1
36 Luo L. Liu D. Liao X. Wu X. Jing Q. Zheng J. Contact Settings and Risk for Transmission in 3410 Close Contacts of Patients With COVID-19 in Guangzhou, China : A Prospective Cohort Study Ann Intern Med. 173 11 2020 Dec 1 879 887 32790510
37 Saravanan U.B. Namachivayam M. Jeewon R. Huang J.D. Durairajan S.S.K. Animal models for SARS-CoV-2 and SARS-CoV-1 pathogenesis, transmission and therapeutic evaluation World J Virol. 11 1 2022 Jan 25 40 56 35117970
38 de Vries R.D. Rockx B. Haagmans B.L. Herfst S. Koopmans M.P. de Swart R.L. Animal models of SARS-CoV-2 transmission Curr Opin Virol. 50 2021 Oct 8 16 34256352
39 Kutter J.S. de Meulder D. Bestebroer T.M. Lexmond P. Mulders A. Richard M. SARS-CoV and SARS-CoV-2 are transmitted through the air between ferrets over more than one meter distance Nat Commun. 12 1 2021 Mar 12 1653 33712573
40 Yuan S. Ye Z.W. Liang R. Tang K. Zhang A.J. Lu G. Pathogenicity, transmissibility, and fitness of SARS-CoV-2 Omicron in Syrian hamsters Science. 377 6604 2022 Jul 22 428 433 35737809
41 Hawks S.A. Prussin A.J. Kuchinsky S.C. Pan J. Marr L.C. Duggal N.K. Infectious SARS-CoV-2 Is Emitted in Aerosol Particles mBio. 26;12(5):e0252721 2021 Oct
42 Bhavnani D. James E.R. Johnson K.E. Beaudenon-Huibregtse S. Chang P. Rathouz P.J. SARS-CoV-2 viral load is associated with risk of transmission to household and community contacts BMC Infect Dis. 22 1 2022 Aug 5 672 35931971
43 Pan D, Williams CM, Decker J, Fletcher E, Sze S, Assadi S, et al. Long title: Exhaled SARS-CoV-2 RNA viral load kinetics measured by facemask sampling associates with household transmission. Clin Microbiol Infect. 2022 Jul 14;S1198-743X(22)00369-X.
44 Liu C. Huang J. Chen S. Wang D. Zhang L. Liu X. The impact of crowd gatherings on the spread of COVID-19 Environ Res. 9 213 2022 Jun 113604
45 Zhao X. Shen Z. Sun L. Cheng L. Wang M. Zhang X. A network meta-analysis of secondary attack rates of COVID-19 in different contact environments Epidemiol Infect. 5 149 2021 Oct e219
46 Verberk J.D.M. de Hoog M.L.A. Westerhof I. van Goethem S. Lammens C. Ieven G. Transmission of SARS-CoV-2 within households: a remote prospective cohort study in European countries Eur J Epidemiol. 37 5 2022 May 549 561 35644003
47 Cai J. Sun W. Huang J. Gamber M. Wu J. He G. Indirect Virus Transmission in Cluster of COVID-19 Cases, Wenzhou, China, 2020 Emerg Infect Dis. 26 6 2020 Jun 1343 1345 32163030
48 Chia P.Y. Coleman K.K. Tan Y.K. Ong S.W.X. Gum M. Lau S.K. Detection of air and surface contamination by SARS-CoV-2 in hospital rooms of infected patients Nat Commun. 11 1 2020 May 29 2800 32472043
49 Nagle S. Tandjaoui-Lambiotte Y. Boubaya M. Athenaïs G. Alloui C. Bloch-Queyrat C. Environmental SARS-CoV-2 contamination in hospital rooms of patients with acute COVID-19 J Hosp Infect. 126 2022 Aug 116 122 35569577
50 Warren B.G. Nelson A. Barrett A. Addison B. Graves A. Binder R. Severe acute respiratory syndrome coronavirus 2 environmental contamination in hospital rooms is uncommon using viral culture techniques Clin Infect Dis. 75 1 2022 Aug 24 e307 e309 35023553
51 Kotwa J.D. Jamal A.J. Mbareche H. Yip L. Aftanas P. Barati S. Surface and Air Contamination With Severe Acute Respiratory Syndrome Coronavirus 2 From Hospitalized Coronavirus Disease 2019 Patients in Toronto, Canada, March-May 2020 J Infect Dis. 225 5 2022 Mar 2 768 776 34850051
52 Ribaric N.L. Vincent C. Jonitz G. Hellinger A. Ribaric G. Hidden hazards of SARS-CoV-2 transmission in hospitals: A systematic review Indoor Air. 32 1 2022 Jan e12968 34862811
53 Belluco S. Mancin M. Marzoli F. Bortolami A. Mazzetto E. Pezzuto A. Prevalence of SARS-CoV-2 RNA on inanimate surfaces: a systematic review and meta-analysis Eur J Epidemiol. 36 7 2021 Jul 685 707 34313896
54 Meister T.L. Dreismeier M. Blanco E.V. Brüggemann Y. Heinen N. Kampf G. Low risk of SARS-CoV-2 transmission by fomites - a clinical observational study in highly infectious COVID-19 patients J Infect Dis. 2022 May 5;jiac170.
55 Borges J.T. Nakada L.Y.K. Maniero M.G. Guimarães J.R. SARS-CoV-2: a systematic review of indoor air sampling for virus detection Environ Sci Pollut Res Int. 28 30 2021 Aug 40460 40473 33630259
56 Lednicky JA, Lauzardo M, Fan ZH, Jutla A, Tilly TB, Gangwar M, et al. Viable SARS-CoV-2 in the air of a hospital room with COVID-19 patients. medRxiv. 2020 Aug 4;2020.08.03.20167395.
57 Dinoi A. Feltracco M. Chirizzi D. Trabucco S. Conte M. Gregoris E. A review on measurements of SARS-CoV-2 genetic material in air in outdoor and indoor environments: Implication for airborne transmission Sci Total Environ. 25 809 2022 Feb 151137
58 Lomont A. Boubaya M. Khamis W. Deslandes A. Cordel H. Seytre D. Environmental contamination related to SARS-CoV-2 in ICU patients ERJ Open Res. 6 4 2020 Oct 00595 02020 33257911
59 Winslow R.L. Zhou J. Windle E.F. Nur I. Lall R. Ji C. SARS-CoV-2 environmental contamination from hospitalised patients with COVID-19 receiving aerosol-generating procedures Thorax. 77 3 2022 Mar 259 267 34737194
60 Onakpoya I.J. Heneghan C.J. Spencer E.A. Brassey J. Plüddemann A. Evans D.H. SARS-CoV-2 and the role of fomite transmission: a systematic review F1000Res. 10 2021 233 34136133
61 Mohamadi M. Babington-Ashaye A. Lefort A. Flahault A. Risks of Infection with SARS-CoV-2 Due to Contaminated Surfaces: A Scoping Review Int J Environ Res Public Health. 18 21 2021 Oct 20 11019 34769538
62 Goldman E. Exaggerated risk of transmission of COVID-19 by fomites The Lancet Infectious Diseases. 20 8 2020 Aug 892 893 32628907
63 Stockwell R.E. Ballard E.L. O’Rourke P. Knibbs L.D. Morawska L. Bell S.C. Indoor hospital air and the impact of ventilation on bioaerosols: a systematic review J Hosp Infect. 103 2 2019 Oct 175 184 31279762
64 Meyerowitz E.A. Richterman A. Gandhi R.T. Sax P.E. Transmission of SARS-CoV-2 Ann Intern Med. 174 7 2021 Jul 1037
65 Duval D. Palmer J.C. Tudge I. Pearce-Smith N. O’Connell E. Bennett A. Long distance airborne transmission of SARS-CoV-2: rapid systematic review BMJ. 29 377 2022 Jun e068743
66 Noorimotlagh Z. Jaafarzadeh N. Martínez S.S. Mirzaee S.A. A systematic review of possible airborne transmission of the COVID-19 virus (SARS-CoV-2) in the indoor air environment Environ Res. 193 2021 Feb 110612
67 Bartoszko J.J. Farooqi M.A.M. Alhazzani W. Loeb M. Medical masks vs N95 respirators for preventing COVID-19 in healthcare workers: A systematic review and meta-analysis of randomized trials Influenza Other Respir Viruses. 14 4 2020 Jul 365 373 32246890
68 Barycka K. Szarpak L. Filipiak K.J. Jaguszewski M. Smereka J. Ladny J.R. Comparative effectiveness of N95 respirators and surgical/face masks in preventing airborne infections in the era of SARS-CoV2 pandemic: A meta-analysis of randomized trials PLoS One. 15 12 2020 e0242901 33320847
69 Wang C.C. Prather K.A. Sznitman J. Jimenez J.L. Lakdawala S.S. Tufekci Z. Airborne transmission of respiratory viruses Science. 27;373(6558):eabd9149 2021 Aug
70 Szajek K. Fleisch F. Hutter S. Risch M. Bechmann T. Luyckx V.A. Healthcare institutions’ recommendation regarding the use of FFP-2 masks and SARS-CoV-2 seropositivity among healthcare workers: a multicenter longitudinal cohort study Antimicrob Resist Infect Control. 11 1 2022 Jan 10 6 35012679
71 Haller S. Güsewell S. Egger T. Scanferla G. Thoma R. Leal-Neto O.B. Impact of respirator versus surgical masks on SARS-CoV-2 acquisition in healthcare workers: a prospective multicentre cohort Antimicrob Resist Infect Control. 11 1 2022 Feb 5 27 35123572
72 Landry S.A. Subedi D. Barr J.J. MacDonald M.I. Dix S. Kutey D.M. Fit-Tested N95 Masks Combined With Portable High-Efficiency Particulate Air Filtration Can Protect Against High Aerosolized Viral Loads Over Prolonged Periods at Close Range J Infect Dis. 226 2 2022 Aug 24 199 207 35535021
73 Ingabire J. McKenney H. Sebesta C. Badhiwala K. Kemere C. Kapur S. Evaluation of Aerosol Particle Leak and Standard Surgical Mask Fit With 3 Elastomeric Harness Designs JAMA Netw Open. 5 1 2022 Jan 4 e2145811 35099550
| 0 | PMC9746078 | NO-CC CODE | 2022-12-15 00:03:23 | no | Infect Dis Now. 2022 Dec 13; doi: 10.1016/j.idnow.2022.12.002 | utf-8 | Infect Dis Now | 2,022 | 10.1016/j.idnow.2022.12.002 | oa_other |
==== Front
Neuroscience
Neuroscience
Neuroscience
0306-4522
1873-7544
IBRO. Published by Elsevier Ltd.
S0306-4522(22)00615-7
10.1016/j.neuroscience.2022.12.007
Neuroscience Forefront Review
Action of the purinergic and cholinergic anti-inflammatory pathways on oxidative stress in patients with Alzheimer's disease in the context of the COVID-19 pandemic
Simões Júlia L.B. a
Sobierai Leilane D. b
Leal Inayá F. c
Dos Santos Miriam V.R. d
Victor Coiado João e
Bagatini Margarete D. f⁎
a Medical School, Federal University of Fronteira Sul, Chapecó, SC, Brazil
b Medical School, Federal University of Fronteira Sul, Chapecó, SC, Brazil
c Medical School, Federal University of Fronteira Sul, Chapecó, SC, Brazil
d Medical School, Federal University of Fronteira Sul, Chapecó, SC, Brazil
e Medical School, Federal University of Fronteira Sul, Chapecó, SC, Brazil
f Graduate Program in Biomedical Sciences, Federal University of Fronteira Sul, Chapecó, SC, Brazil
⁎ Corresponding author at: Graduate Program in Medical Sciences, Federal University of Fronteira Sul, Chapecó, SC, Brazil
13 12 2022
13 12 2022
22 4 2022
2 12 2022
5 12 2022
© 2022 IBRO. Published by Elsevier Ltd. All rights reserved.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the etiologic agent of the 2019 coronavirus disease (COVID-19), has affected more than 20 million people in Brazil and caused a global health emergency. This virus has the potential to affect various parts of the body and compromise metabolic functions. The virus-mediated neural inflammation of the nervous system is due to a storm of cytokines and oxidative stress, which are the clinical features of Alzheimer's disease (AD). This neurodegenerative disease is aggravated in cases involving SARS-CoV-2 and its inflammatory biomarkers, accelerating accumulation of β-amyloid peptide, hyperphosphorylation of tau protein, and production of reactive oxygen species, which lead to homeostasis imbalance. The cholinergic system, through neurons and the neurotransmitter acetylcholine (ACh), modulates various physiological pathways, such as the response to stress, sleep and wakefulness, sensory information, and the cognitive system. Patients with AD have low concentrations of ACh; hence, therapeutic methods are aimed at adjusting the ACh titers available to the body for maintaining functionality. Herein, we focused on acetylcholinesterase inhibitors, responsible for the degradation of ACh in the synaptic cleft, and muscarinic and nicotinic receptor agonists of the cholinergic system owing to the therapeutic potential of the cholinergic anti-inflammatory pathway in AD associated with SARS-CoV-2 infection.
Keywords
SARS-CoV-2
Alzheimer's disease
purinergic
cholinergic
oxidative stress
==== Body
pmcIntroduction
AD is a progressive neurodegenerative disease, which is the most common cause of dementia in the elderly population worldwide. In the scenario of COVID-19, it has emerged as a key comorbidity, given the increased morbidity and mortality of COVID-19 in patients with AD due to multiple pathological changes. Among them, we can mention: the high expression of the viral angiotensin-converting enzyme 2 (ACE2) receptor, the increase in pro-inflammatory cytokines and several complications of AD (diabetes, changes in lifestyle, use of medication). Adding to the picture, the direct attack of the virus to the central nervous system (CNS) results in neurological symptoms, cognitive impairment, neuronal inflammation, hospitalization and post-COVID-19 delirium and syndrome (Xia et al., 2021). Furthermore, the COVID-19 crisis also worsens behavioral symptoms in uninfected AD patients and presents new challenges for preventing AD and controlling its progression.
Clinical signs have an insidious onset of memory deficits, with an emphasis on recent memory impairments, learning difficulties, mood, and behavioral changes, which progress to variations in personality, language, calculation, visuospatial orientation, severe retrograde amnesia, and aphasia. In more advanced cases, in addition to psychotic symptoms, individuals present alterations in the sleep-wake cycle, irritability and aggressiveness, speech, gait and self-care difficulties (Ferreira, Nordberg and Westman, 2020). Among the risk factors, studies list more than 20 variants associated with AD including age, family heritage, exposure to aluminum, traumatic brain injury (TBI) and comorbidities such as vascular disease and infections.
Genetic mutations stand out in rare forms of early-onset familial AD (EO-FAD), such as alterations in the amyloid precursor protein (APP) and presenilin genes (PSEN1/2), while late-onset sporadic AD (LO-SAD) presents a compilation of genetic and environmental factors. Thus, it may be related to age, genetic risk factors such as allelic variation in apolipoprotein E (Apo E) and many other genes, as well as infections, mitochondrial function, metal exposure, immune system defects, vascular disease, TBI, and risk factors associated with diet. From this perspective, the variables are reorganized to the point of acting together to: a) increase the concentration of oxygen free radicals, b) external factors acting on early and late regulatory genes (the 'double whammy' hypothesis) or c) increasing the 'cumulative allostatic’ load on the body over a lifetime (Armstrong, 2019). Thus, the ways to modulate this pathogenesis are diverse, ranging from changes in habits to genetic management.
Among the hypotheses for the development of AD, it is believed that the protagonism of the β-amyloid protein. The Aβ hypothesis selects the imbalance between production and clearance of Aβ42 and Aβ peptides as an EO-FAD factor. Thus, knowing that the catalytic site of γ-secretase is presenilin, he oriented the causative factor of AD from the beginning on the alteration of the substrate, that is, on the APP or on the protease (presenilin) of the reaction that generates Aβ (Selkoe and Hardy, 2016). Thus, it is believed that the deposition of Aβ plaques results in neurodegeneration of brain tissue and this accumulation is the main influence on pathogenesis, followed by the formation of neurofibrillary tangles containing tau protein (Hardy and Selkoe, 2002). In addition, recent studies indicate that low Aβ42 concentration in the cerebrospinal fluid (CSF) and amyloid-PET positivity precede other manifestations of AD by many years, in addition to new results with three different Aβ antibodies (solanezumab, crenezumab and aducanumab) (Selkoe and Hardy, 2016). These suggested a slowing of cognitive decline in post hoc analysis of individuals with mild AD, leading to several factors that contribute to the development and progression of the disease (Selkoe and Hardy, 2016).
Thus, with mutations in presenilins implying alteration of the enzymatic activity of γ-secretase and the production of Aβ, the pathogenesis of hereditary AD in middle age results, although it is indistinguishable from sporadic and late AD. Mutations within and immediately flanking the Aβ region of APP cause aggressive forms of FAD. An increase in Aβ42-to-Aβ40 is understood to occur by mutations in presenilin, driven by a reduction in carboxypeptidase clipping and the increase in longer forms of Aβ, including membrane-associated forms Aβ45-Aβ49 (Wolfe, 2019). Relative increases in the production of Aβ42/43 peptides lead to a profound deposition of Aβ in middle age because they are hydrophobic species that self-aggregate easily (Selkoe and Hardy, 2016).
Certain biomarkers, such as interleukin (IL)-1 and IL-6, are also detected in the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. The new coronavirus is an enveloped virus with a single-stranded RNA, belonging to the family Coronaviridae, which has seven subtypes that affect humans and six other strains (Weiss and Navas-Martin, 2005; Rahman et al., 2021a). This virus can cause neural tissue inflammation, and in patients with AD, it can aggravate the clinical picture, given that it uses the olfactory bulb to permeate the nervous tissue, inducing the activation of defense cells, reactive astrogliosis, and the necroinflammatory cascade (Zotova, et al., 2013, Wu, et al., 2020). Concomitantly, there is an increase in the levels of the plasma protein Galectin-3 (GAL-3), which is closely linked to the oligomerization of Aβ. Thus, there is a close association between SARS-CoV-2 infection and the poor prognosis of AD.
Another factor that boosts AD progression is the production of reactive oxygen species (ROS) and reactive nitrogen species (RNS), which at high concentrations trigger an imbalance in the homeostasis of antioxidant substances and is responsible for neutralizing and reducing damage caused by free radicals. Oxidative stress in the neuronal environment from the exacerbated production of free radicals by the mitochondria, a by-product of respiration, generates brain aging. This scenario is present both in aging and in the pathogenesis of AD, the so-called mitochondrial cascade hypothesis, in which damaged mitochondria produce beta-amyloid, a toxic oxidative stressor. Free radicals are capable of breaking unprotected mitochondrial DNA double strands without histones, as well as manipulating the epigenome and have far-reaching effects, from repression of cognitive genes by H3K9me3 to hypomethylation and activation of the APP promoter. Studies with mitochondrial antioxidants in aged animal models reduce oxidative stress and DNA damage, while restoration of mitophagy in AD models may minimize pathology (Ionescu-Tucker and Cotman, 2021). Thus, acting on different lines of attack against the progression of the disease can contribute to the reduction of neurodegeneration, although several pathways are interconnected to this process.
Such alterations are related to the development of neurodegenerative diseases, which can lead to the accumulation of Aβ and the generation of tau neurofibrillary tangles (NFTs), compromising axonal transport, synaptic processing, and neuronal death (Gauthier et al., 2016). Furthermore, oxidative stress can be amplified by the activation of the N-methyl-D-aspartate receptor (NMDAR) in the stress signaling pathways in neurons and by cytoplasmic calcium deposition induced by Aβ plaques, which also stimulates the extracellular accumulation of glutamate. Upon activation of the central apoptotic pathway, cytochrome C is released to mediate the breakdown of synaptic proteins (Kamat et al., 2016).
There are different biomarkers considered for the biochemical diagnosis of AD and, based on different studies and much progress in this area, it was possible to validate three standard variables in the analysis of cerebrospinal fluid (CSF). Among them, the 42 amino acid long beta-amyloid peptide (Aβ 1-42), total tau protein (T-tau) and phosphorylated tau on threonine 181 (P-tau 181). They have been incorporated into research diagnostic criteria for AD and have added value both in diagnosis and exclusion in case of ambiguous clinical diagnoses of dementia (Bjerke and Engelborghs, 2018). Future researchers are leaning their studies on the implementation of the ratio of CSF Aβ 1-42 /Aβ 1-40A, acting directly on the analytical variability of the biomarker and early and differential diagnosis of AD. Based on many human biomarker studies, low CSF Aβ42 and positive amyloid PET precede other AD-related changes (increased CSF tau, decreased cerebral glucose metabolism, brain atrophy, clinical dementia) by years (Selkoe and Hardy, 2016). Furthermore, other studies analyze other pathological features of AD, such as incorrect amyloid metabolism, tau pathology or synaptic or neuronal degeneration, as well as neurodegenerative, vascular or inflammatory markers unrelated to AD (Bjerke and Engelborghs, 2018).
The neurotransmitter acetylcholine (ACh) is present in fundamental physiological pathways, such as the stress response in the cognitive system, sensory information, and sleep and wakefulness (Ferreira-Vieira et al., 2016a); therefore, neurochemical changes in the cholinergic system have been linked to AD. The formation of Aβ in hyperphosphorylated tau plaques and tangles causes wear at the synapses and machinery, including the ACh pathways and cholinergic system, leading to a reduction in the concentration of this neurotransmitter in the synaptic cleft. ACh promotes neuron plasticity and is responsible, along with the serotonergic system, for cognitive, behavioral, and emotional regimes; therefore, a low concentration of ACh is considered one of the main characteristics of AD (Machado et al., 2020).
Damage to these chemical systems and mechanisms induces the development of AD (Rasch et al., 2006, Kuo et al., 2007) via presynaptic cholinergic lesions in nucleus basalis of Meynert neurons and axonal projections from the cerebral cortex, affecting local muscarinic and nicotinic receptors. Thus, maintaining adequate levels of ACh in the synapses can be used as a treatment method for AD. This can be achieved by administrating drugs that block the action of acetylcholinesterase (AChE), as they are capable of modulating the supply of ACh in the synaptic cleft (Rasch et al., 2006, Kuo et al., 2007).
Thus, the relationship between AD and COVID-19 is related to the neuroinflammatory process and neuronal oxidative state, and the association between these pathologies and its effect on the prognosis is evidenced by more than 50 million individuals with some type of dementia and the global reality of a highly infectious virus. Hence, it is possible to establish a close association between the cholinergic disorders present in both diseases, and modulation of the cholinergic anti-inflammatory pathway stands out for its therapeutic potential and must be focused on and applied in further research to enable the development of alternative therapies (Fig. 1 ).Figure 1 Alzheimer's disease (AD) is caused either by genetic factors or by oxidative stress, inflammation, mitochondrial dysfunction, among others, which cause the accumulation of β-amyloid peptide (Aβ) and tau protein neurofibrillary tangles in the cleft synaptic, forming senile plaques. Neurochemical alterations in the cholinergic system are present in AD. The β-amyloid plaques erode the synapses and the mechanism of the acetylcholine (Ach) pathways, reducing the concentration of this neurotransmitter in the synaptic space, which modifies the stress response. SARS-CoV-2 can aggravate neurotissue inflammation by inducing reactive astrogliosis, defense cell activation, and the neuroinflammatory cascade, which is closely linked to Aβ oligomerization. β-amyloid peptide - Aβ, reactive oxygen species - ROS, acetylcholine - ACh, acetylcholinesterase - AChE, acetylcholinesterase enzyme inhibitors - IAChE.
Neurochemical Cross-talk Between Coronavirus disease 2019 and Alzheimer's disease
COVID-19 was initially detected as an outbreak with its epicenter located in Hubei Province, China, but quickly spread to other countries (Gómez-Mesa et al., 2021). The causal agent for COVID-19 is the SARS-CoV-2 virus, representative of the coronavirus family, which was first described in humans in 1966 when Tyrell and Bynoe observed the cultivation of viruses related to common cold (Chazal, 2021). Coronaviruses are named for their appearance, i.e., spherical-shaped with a central shell and surface projections that surround it like a crown (Velavan and Meyer, 2020). After the first reports of infection in 2019, it was identified that the responsible virus could be transmitted between humans (Baig et al., 2020). Based on the exponential increase in case notification rates across international, on March 11, 2020, the Emergency Committee of the World Health Organization declared that the outbreak of “Corona Virus 2019,” first diagnosed in a Chinese province, had become a global health emergency (Asselah et al., 2021).
Coronaviruses are enveloped, single-stranded, positive-sense RNA viruses ranging from approximately 26–32 kilobases in size (the largest genome known for an RNA virus) (Weiss and Navas-Martin, 2005). Four subfamilies are currently recognized and categorized as alpha, beta, gamma, and delta-coronaviruses. Alpha and beta coronaviruses are strongly associated with mammals in the order Chiroptera. In contrast, gamma and delta viruses originate from pigs and birds (Shu and McCauley, 2017; Zhou et al., 2020). Phylogenetically, the etiological agent of COVID-19 belongs to the subgenus Sarbecovirus, genus Betacoronavirus, subfamily Orthocoronavirinae, and family Coronaviridae (Cornidovirinea: Nidovirales) (Su et al., 2016). Among the seven subtypes of coronaviruses that infect humans, beta-coronaviruses are associated with serious diseases and fatalities (Shu and McCauley, 2017; Zhou et al., 2020).
All coronaviruses have similar characteristics in their organization and expression, in which 16 non-structural proteins (nsp1 to nsp16) are encoded by the open reading frame (ORF) 1a/b at the 5' end, followed by the peak expression of structural proteins (S), envelope (E), membrane (M), and nucleocapsid (N), which are encoded by other ORFs at the 3' end (Kin et al., 2015). Currently, seven human coronaviruses (HCoVs) have been identified, namely HCoV-229E, HCoV-OC43, SARS-CoV, HCoV-NL63, HCoV-HKU1, Middle East respiratory syndrome coronavirus, and SARS-CoV-2 (Kin et al., 2015).
SARS-CoV-2 primarily enters the host cell via the ACE2 receptor (Beacon et al., 2021). ACE2 is similar to ACE and acts in the regulation of blood pressure and electrolyte homeostasis in healthy individuals. ACE converts angiotensin I into angiotensin II, leading to vasoconstriction, renal sodium reabsorption and potassium excretion, increased aldosterone synthesis, and induction of inflammation and pro-fibrotic pathways. ACE2 cleaves angiotensin II into angiotensin (1-7), culminating in anti-inflammatory and vasodilatory effects (Bourgonje et al., 2020). Furthermore, ACE2 mediates the metabolism of bradykinins in the lungs, reducing pro-inflammatory effects such as vasodilation and increased vascular permeability (Bryce-Moncloa et al., 2021).
ACE2 is strongly expressed in alveolar and small intestinal epithelial cells, which are the sites frequently affected by SARS-CoV-2. Furthermore, the presence of this enzyme has been observed in vascular endothelial and smooth muscle cells (Hamming et al., 2004). SARS-CoV expresses an S1 peak protein for virion and cell membrane binding and interaction with the host cell's ACE2 receptor (Wrapp et al., 2020). Regarding the peak proteins between coronaviruses, the subtle differences in codon pair sequence alignments may explain the higher binding affinity between the spike protein of COVID-19 and the ACE2 receptor (Wrapp et al., 2020). According to the studies by Baig et al. (2020) proximal tubular cells of the urinary tract and neuronal and glial cells present in the CNS were also sites where the ACE2 receptor was found. Similar to the expression in pulmonary epithelial cells, the expression of ACE2 in blood-brain barrier (BBB) endothelial cells may allow HCoV to bind to pulmonary epithelial cells, thus causing viral dissemination in the CNS (Fu et al., 2020).
In addition to the ACE2 receptor, coronaviruses can use the olfactory bulb to enter neural tissue, taking advantage of the trans-synaptic pathway existing at the site (Steardo et al., 2020). Once in the CNS, SARS-CoV-2 induces reactive astrogliosis (inflammation of neural tissue), activation of immune defense cells (microglia), and a neuroinflammatory cascade. Simultaneously, viral spread in lung epithelial cells can cause a systemic inflammatory response, producing increased levels of pro-inflammatory cytokines, which also interfere with the CNS (Wu et al., 2020). In addition, the ventilatory function in lungs altered by infection can lead to respiratory failure and intense hypoxia, with consequent cerebral vasodilation and risk of evolution to edema and cerebral ischemia (Wu et al., 2020).
Rahman et al. (2021) revealed that biomarkers can be identified to track the progression of COVID-19 owing its significant role in the CNS. Some of these biological markers have also been identified in AD. IL-6 is a mediator with several effects on different cell groups (pleiotropic effects), acting in the inflammatory process, immune response, and hematopoiesis. At the onset of inflammation, IL-6 is synthesized at the injured site, travels to the liver via the bloodstream, and induces the production of positive acute phase proteins, such as C-reactive protein (CRP), serum amyloid A, fibrinogen, haptoglobin, and α1-antichymotrypsin. Simultaneously, IL-6 also acts by decreasing the production of fibronectin, albumin, and transferrin (negative acute phase proteins)(Zotova, et al., 2013, Tanaka et al., 2014). The 11 family groups of antagonist ligands and receptors (IL-1α, IL-1β, IL-18, IL-33, IL-36α, IL-36β, IL-36g, IL-1Ra, IL-36Ra, IL-38, and IL-37) for IL-1 are responsible for independently mediating local or generalized inflammatory processes or inducing anti-inflammatory responses (Mendiola and Cardona, 2018).
SARS-CoV-2 activates IL-1 at the beginning of infection, which stimulates the secretion of TNF, IL-6, and other pro-inflammatory cytokines, causing a cytokine storm (Conti et al., 2020). Zhao, et al., 2020, Zhao et al., 2020 concluded that certain markers, including C-C motif chemokine ligand 5 levels and IL-1RA and IL-10 in the blood (individually and in combination), may be useful in prognosis and to guide treatment strategies, (Zhao, et al., 2020, Zhao et al., 2020). In patients with AD, the increase of pro-inflammatory ILs in plasma has been analyzed owing to its inflammatory response (Culjak et al., 2020). The changes in IL-1α, IL-10, and TNF-α concentration in patients with AD partially confirmed its association with the neuroinflammatory response in AD (Culjak et al. 2020). Therefore, monitoring these levels may aid in our understanding of various AD stages and the effect of SARS-CoV-2 infection on the prognosis, given that elevated serum pro-inflammatory cytokines are associated with an increased rate of decline in cognitive impairment in AD (Ide et al., 2016, Lin et al., 2022).
However, from a broader perspective, in a study that included the elderly as a control group, the variables IL-6 (p = 0.138), TNF-α (p = 0.451), and CRP (p = 0.07) were not significant, indicating that patients with AD do not have higher differentiation markers than other elderly patients (Ng et al., 2018).
Clinical evidence has linked the cascade of cytokine release to the presence of viral infection and is a prominent cause of mortality in patients with -19 (Culjak et al., 2020). In patients compromised by SARS-CoV-2, the marked elevation of plasma GAL-3 levels (belongs to the galectin family of proteins) stands out compared to that of healthy controls. These lectins can interact with other proteins through the carbohydrate recognition domain of a galectin with a β-galactoside conjugate on a specific protein, which mediates several physiological effects (Barondes et al., 1994, MacHado et al., 2020). GAL-3 is a carbohydrate-binding protein expressed in lung cells, such as macrophages, epithelial cells, and alveolar cells (Reyfman et al., 2019). An increase in this substance is observed in patients affected by COVID-19, which is precisely explained by the association of galectin with the immune system and the activation of pro-inflammatory macrophages (McGonagle et al., 2020).
Tao et al. (2020) observed that GAL-3 promoted Aβ oligomerization and Aβ toxicity in animal models of AD. Evidence indicates that Aβ oligomers are responsible for altering the integrity of the bilipid membranes (membrane lipid bilayers), causing an increase in the influx of sodium (Na+) and Ca2+ ions, causing a deficit in synaptic transmission and contributing to the pathology of AD (Fändrich, , 2012, Salahuddin et al., 2016). Thus, GAL-3 can be considered a biomarker of inflammation related to both COVID-19 and AD, and the modulation of bioavailability seems to be promising in the treatment of these manifestations.
Cytoskeleton-associated protein 4 (CKAP4) is another protein of equal importance as a potential biomarker for both conditions. CKAP4, also known as CLIMP-63 and ERGIC-63, is a non-glycosylated type II transmembrane protein present in the endoplasmic reticulum (ER) of all tissues. In the ER, the cytoplasmic region of CKAP4 binds to the microtubules, creating a link between microtubules and the ER. (Osugi et al., 2019). Furthermore, CKAP4 is a Dicer-binding protein that regulates the microRNA (miRNA) pathway and mRNA translation by anchoring Dicer to the ER (Pépin et al., 2012). A study by Cancino (2013) suggested that CKAP4 is essential for maintaining the appropriate number of neuronal precursor cells (NPCs) in neurons born in the adult hippocampus. Furthermore, the removal (ablation) of P63 from the site leads to an increase in the apoptosis of NPCs via the pro-apoptotic p53-PUMA pathway, which culminates in neuronal memory deficits and decreased learning capacity (Cancino et al., 2013). Analysis of the serum proteomic profile of patients affected by SARS-CoV-2 identified 6 proteins related to disease severity, including CKAP4 (Poyiadji et al., 2020).
Apolipoprotein E (APOE), the main cholesterol transporter in the CNS, is another protein involved in the maintenance of neuronal function and is a constituent of very low-density lipoproteins (Hultman et al., 2013). Among its three alleles (ε2, ε3 and ε4), the ε4 allele (allele of the APOE gene) is the greatest risk factor for the development of AD (Serrano-Pozo et al., 2016) because the APOE ɛ4/4 genotype induces an increase in fibrinogenesis in the brains of these individuals (Hultman et al., 2013). Normally, astrocytes (and to a lesser degree, microglia) are primarily responsible for expressing and secreting APOE (Serrano-Pozo et al., 2016). Among the variables, the APOE ε4 allele is the best-established risk factor for AD. ApoE4 carriers have already been included in typical LO-SAD. This allele has been found to markedly increase the risk of AD and decrease cerebral Aβ clearance, leading to excess Aβ aggregation and downstream AD-typical neuropathology (Selkoe and Hardy, 2016). Whitwell et al (2021) evaluated the relationship between APOE genotype, age of onset, Aβ deposition, and typical versus atypical clinical presentations in AD. As a result, it highlighted the heterogeneous nature of AD and that the APOE genotype varies according to the studied variables. It can be mentioned that its frequency increased with the age of onset in atypical AD, although it presented a bell-shaped curve in typical AD, with higher frequencies between 65 and 70 years. Comparing typical and atypical AD, the former presented higher APOE ε4 frequencies only between the ages of 57 and 69 years. Finally, the overall proportions of standard Aβ absorption values did not differ according to APOE e4 status in either group (Whitwell et al., 2021).
APOE is one of the genes co-expressed with the ACE2 receptor gene in type II alveolar cells, which is the route of coronavirus entry (Zhao, et al., 2020, Zhao et al., 2020). Thus, individuals with AD carrying the APOE4 allele are at a higher risk of developing severe COVID-19. ACE2 is primarily responsible for the invasion of SARS-CoV-2 into the host cells (Chaudhry et al., 2020). Thus, the level of ACE2 receptor expression is a crucial determinant of viral replication and pathogenesis. Notably, ACE2 is not expressed or localized in all human cells, and its main site of concentration is the surface of type II alveolar epithelial cells (AT2 pneumonocytes) (Hamming et al., 2004). In view of the difference in tissue concentrations, the expression of the ACE2 gene increased tenfold in the nerve cells of individuals with AD compared to those without the disease (Lim et al., 2020). Thus, individuals with AD have a higher risk of contracting COVID-19.
One of the most prominent manifestations of COVID-19 is anosmia, which can be defined as the loss of the ability to detect odors (Moein et al., 2020). Simultaneously, several neurodegenerative diseases such as Parkinson's disease and AD present olfactory dysfunction as one of the initial symptoms prior to the appearance of motor symptoms and cognitive decline (Marin et al., 2018). Being the second most abundant metal in the human body, zinc is an essential micronutrient. Zinc deficiency (zincopenia) is related to the manifestation of anosmia and ageusia (taste dysfunction) because carbonic anhydrase, an enzyme responsible for maintaining the function of smell and taste, is a zinc-dependent metalloenzyme (Equils et al., 2021). Given the anti-inflammatory properties of zinc, zincopenia induces an increase in the expression of IL-6 and IL-1β, pro-inflammatory cytokines, and intercellular adhesion molecule 1, which are important for leukocyte extravasation (Dhama et al., 2020). According to Equils et al. (2021), SARS-CoV-2 infection induces an immune response in the nasopharyngeal mucosa that can lead to local zinc deficiency. Such a decrease in zinc (Zn) levels is also observed in patients with AD, contributing to the progression and severity of the disease (Sensi et al., 2018).
Nitric oxide (NO) is a gaseous molecule that can easily diffuse into body tissues and is produced by the enzymatic activity of the NO synthase (NOS) family. In neural tissues, neurons, glial cells, and vascular cells can express NOS and are potential sources of local NO (Tajes et al., 2013). Activation of NMDAR in the hippocampus induces a cascade of reactions involving NO production. The diffusion of gas from the postsynaptic to the presynaptic termination stimulates the release of vesicles via a mechanism independent of guanylyl cyclase, forming an activation cycle called long-term potentiation (LTP), which is attributed to the physiological mechanisms of learning and memory (Picón-Pagès et al., 2019). SARS-CoV-2 can decrease neuronal NO production; thus, patients with AD who have contracted COVID-19 may show increased behavioral and cognitive decline owing to the low concentration of this neurotransmitter (Alkeridy et al., 2020) (Fig. 2 ).Figure 2 In patients with COVID-19, there is an indirect reduction of critical factors, which also decrease in Alzheimer's disease (AD). The reduction of nitric oxide (NO), a molecule involved in the learning and memory process, and the cytoskeleton-associated protein 4 (CKAP4 protein), which acts in the maintenance of the appropriate number of neuronal precursor cells, and reduction in the level of this protein culminates in a deficit in the memorization process. One of the roles of SARS-CoV-2 in the host organism is linked to the low expression of galectin-3 (GAL-3), with a consequent increase in β-amyloid peptide (Aβ) oligomerization, a condition that characterizes the pathogenesis of AD. Apolipoprotein E (APOE 4) is another molecule that is associated with pathological Aβ deposition, which also acts to increase the activation of circulating pro-inflammatory macrophages, a factor that increases the risk of severe COVID-19 in an individual. Finally, the reduction of zinc in patients with AD and infected with COVID-19, together with profound alterations in the functioning of the cholinergic system in these patients, proved to be a stimulatory condition for the activation of already activated pro-inflammatory cells owing to the SARS-CoV-2 infection. In this scenario, patients with AD may have a worse prognosis of COVID-19 compared to age-matched controls. Zn - Zinc; GAL-3 - protein Galectin-3; APOE4 - apolipoprotein E; NO - nitric oxide; CKAP4 - cytoskeleton-associated protein 4; LTP - long-term potentiation; ACE 2 - angiotensin-converting enzyme 2;
ACh is another neurotransmitter involved in both the manifestations. It is a fast-acting point-to-point neurotransmitter in the neuromuscular junction and autonomic ganglia. Its main function is to control interneuronal and muscular communications. It also acts in the maintenance of movement, heart rate, digestion, breathing, and other autonomic functions (Picciotto et al., 2012). In addition to these well-characterized processes, ACh is involved in other activities such as vasodilation and action on the immune system (Cox et al., 2020).
According to the cholinergic theory, the decline in cognitive function in patients with AD is largely related to structural changes in cholinergic synapses, the loss of specific subtypes of ACh receptors, and the death of ACh-generating neurons, resulting in decreased cholinergic neurotransmission (Stanciu et al., 2019). Depending on the calcium ion influx, ACh release occurs via exocytosis of synaptic vesicles. They fuse with the presynaptic membrane, eliminating their neurotransmitter content in the synaptic cleft, where they activate muscarinic and nicotinic receptors (Stanciu et al., 2020). Effect of ACh stimulation on nicotinic receptors in macrophages was observed, which resulted in a concentration-dependent inhibition of the synthesis and release of pro-inflammatory cytokines, such as IL-1β, TNF-α, IL- 6 and HMGB1, without changing the concentration of anti-inflammatory cytokines, such as IL-10 (Hoover, 2017). A cytokine storm is characterized by the accentuated expression of IL-2, IFN-γ, IL-4, and IL-13. SARS-CoV-2 can rapidly activate pathogenic Th1 cells to secrete pro-inflammatory cytokines (Petrone et al., 2021). Therefore, treatment with nicotinic substances and the cholinergic system could reduce the generation of the inflammatory storm observed in patients infected with the new HCoV, while simultaneously being useful in the maintenance of neuronal functions in patients with AD (Farsalinos et al., 2020) (Fig 2).
In addition to AD biomarkers concomitantly related to COVID-19, some AD-specific features should be emphasized. Among them, AD is mainly characterized by the formation of extracellular aggregates of Aβ outside neurons: the Aβ protein, the APP 21q21 gene and the long arm of chromosome 21 (Xie et al., 2020). Aβ plaques begin to develop in the basal, temporal, and orbitofrontal regions of the brain's neocortex and then progress to other regions such as the neocortex, hippocampus, amygdala, diencephalon, and basal ganglia. In severe AD, Aβ plaques are observed throughout the midbrain, lower brainstem, and cerebellar cortex (Goedert, 2015). In the pathogenesis of AD, the deposition of misfolded amyloid fibrils into plaques causes the activation of kinases, hyperphosphorylation of tau protein associated with microtubules, and its polymerization into NFTs. The formation of extracellular Aβ plaques and NFT tangles formed by hyperphosphorylated tau protein act directly on the recruitment and activation of microglia, with a local inflammatory response and increased neurotoxicity (Pinheiro and Faustino, 2019, Tiwari, et al., 2019) (Fig.2).
However, tau is another factor that contributes to the possible development of AD. Tau is a microtubule-associated protein that plays a key role in microtubule stability (Gauthier et al., 2016). Furthermore, several functions of tau, such as maintenance of genomic DNA integrity, regulation of neuronal activity, neurogenesis, and iron export, have been elucidated in recent years (Gao et al., 2018, Peña-Bautista et al., 2019). The Tau protein is encoded on chromosome seventeen of MAPT gene. Alternative splicing of eight of the 16 exons of the MAPT gene allows the expression of six Tau isoforms in the CNS and six additional isoforms in the peripheral nervous system (PNS), ranging from 58 kDa to 66 kDa and one isoform of 110 kDa (Naseri et al., 2019). Post-translational modifications may occur during AD progression, and tau phosphorylation at various sites is the main characteristic of AD progression (Neddens et al., 2018).
Phosphorylation of serine and threonine residues near or within the microtubule-binding domain causes changes in the conformation of microtubules. This alteration releases stored tau, leading to its accumulation in the somatodendritic compartment of a pair of helical filaments and other abnormal conformations (Wesseling et al., 2020). Shigemoto et al. (2018) suggested that low levels of tau deposition, together with lower amyloid deposition, induce compensatory responses against early neuronal damage or chronic inflammation due to aging (Gauthier et al., 2016). In contrast, the coexistence of amyloid deposition and increased tau concentration induces a decrease in neuronal connectivity, contributing to the progression of AD (Shigemoto et al., 2018).
Presenilin 1 (PS1) is a ubiquitous transmembrane protein with several biological roles, such as cell adhesion, apoptosis, calcium homeostasis, and synaptic plasticity. Despite being rare and representing only 0.5% of all AD cases, FAD has attracted growing interest in the scientific community. This autosomal dominant inherited condition is characterized by highly penetrating mutations in three genes (Canevelli et al., 2014): (a) the APP gene on chromosome 21 (Goate et al., 1991); (b) the presenilin 1 gene (PSEN1) on chromosome 14 (Sherrington et al., 1995); and (c) the presenilin 2 gene (PSEN2) on chromosome 1 (Levy-Lahad et al., 1995). Changes in these variables are associated with increased production and/or deposition of β-amyloid (Aβ) (Selkoe, 1997, Walker, et al., 2005, Canevelli et al., 2014). As a clinical presentation, humans with trisomy 21 (Down syndrome), who harbor 3 copies of APP, invariably have the neuropathologically typical AD. Furthermore, on biopsy in early adolescence, abundant diffuse Aβ plaques are found without neuritic dystrophy, microgliosis, astrocytosis, and tangle formation, all of which gradually accumulate in such individuals in late adolescence and beyond (Selkoe and Hardy, 2016).
Another factor that causes neuronal effects and AD manifestation is the occurrence of mitochondrial defects (Macdonald et al., 2018), as normal synaptic function requires a high energy demand (Velavan and Meyer, 2020). In addition, neuroimaging studies using fluorodeoxyglucose positron emission tomography (FDG-PET) have demonstrated that the brains of individuals with AD absorb less glucose than those of cognitively normal controls. Therefore, a reduction in the cerebral metabolic rate of glucose, as measured by FDG-PET, is now considered a hallmark of AD (Zhou et al., 2018).
Genetics plays a vital role in AD risk and pathogenesis (Shao et al., 2017). miRNAs are a 19–23 nucleotide class of single-stranded non-coding RNA involved in post-transcriptional epigenetic regulation of mRNA (Bradley-Whitman and Lovell, 2013). Features such as size, amphipathic nature, and high solubility make miRNAs extremely mobile in the brain. The pathogenic miRNA gene families in the neocortex, hippocampus, limbic system, and CNS in general make miRNAs prime candidates for modulating the expression of many mRNA targets in complex, progressive, and ultimately lethal CNS neurological disorders, including AD (Takousis, et al., 2019, Nunomura and Perry, 2020). Four miRNAs, miR-31, miR-93, miR-143, and miR-146a, are decreased in the serum of patients with AD and, hence, are biomarkers for AD pathology (Dong et al., 2015).
The transcription factor, p53, is essential for the maintenance of several cellular functions related to the integrity of the genome, which includes cell cycle control, response to DNA damage, and apoptosis (Peña-Bautista et al., 2019, Farmer et al., 2020). Furthermore, the roles of p53 in controlling synaptic function, regulating the inflammatory process, and reducing Aβ are observed as factors of great importance in preventing the manifestation and progression of neurodegenerative diseases (Abate et al., 2020). Thus, in AD, the highest concentration of Aβ peptides mediates the degradation of HIPK-2 protein, affecting the conformation of p53 (Mantzavinos and Alexiou, 2017). Consequently, the unfolded p53 protein forms oligomers and fibrils that are associated with AD pathology (Farmer et al., 2020). In addition, severe stress caused by AD can trigger the induction, by P53, of increased cellular oxidative stress by increasing ROS production, which induces neuronal apoptosis and contributes to the development of AD (Farmer et al., 2020; Beacon et al., 2021).
Oxidative Stress, Synaptic Dysfunction and Alzheimer's disease
ROS are chemical compounds that contain oxygen with reactive characteristics, which can be produced by enzymatic means (such as in macrophages) to inactivate invading agents or by non-enzymatic means (such as oxidative phosphorylation). The respiratory chain is the largest generator of reactive species, as it has oxygen as its final acceptor. This process is known as “aerobic cell metabolism,” which continuously synthesizes reactive species, such as superoxide anions, hydrogen peroxide, and hydroxyl radicals. Physiologically, ROS levels are controlled and preserved in lower concentrations in the body; however, in AD, there is an irregularity in the cytochrome C oxidase enzyme, one of the mitochondrial electron transport enzymes/complexes, which causes greater production of ROS, making the cells more prone to apoptosis (Kamat et al., 2016). Thus, significant accumulation of these substances is known as “oxidative stress.”
The energy necessary for the maintenance of the human body is acquired through reduction reactions (in the transfer of electrons) via the electron transport chain (ETC) located in mitochondrial crests. Ingested food and its processed products generate electrons, which are donated to the mitochondrial complexes of the ETC to be finally accepted by oxygen, which has a high reduction capacity, and consequently, prominent reaction energy. In these reactions, oxygen normally receives four protons and four electrons to generate two water molecules (Cheignon et al., 2018). However, owing to genetic alterations and aging, the enzymatic complexes of the respiratory chain become less effective and alter the redox activity, leading to a high formation of ROS and RNS, which cause mitochondrial dysfunction that is linked to the progression of neurodegenerative diseases, such as AD, and cognitive impairment (Cheignon et al., 2018).
Neurons, which are the essential operating unit of the brain, have a high metabolic rate compared to that of other cells in the body, thus showing greater susceptibility and damage to oxidative stress. Considered as a common pathological feature in AD, oxidative stress has not yet been established, as it acts in pathophysiology. However, it has been suggested that, in addition to the destruction or insufficiency of the constituents of the antioxidant system - such as superoxide dismutase (SOD), catalase (CAT,) and glutathione peroxidase (GPx) - in the mitochondria and cytosol, mitochondrial dysfunction, tau hyperphosphorylation, Aβ accumulation, inflammation, and metal accumulation are implicit means of its induction (Chen and Zhong, 2014). A contributing factor to the progression of oxidative stress is that 700 mL of blood flow per minute in the brain and, hence, is responsible for the consumption of approximately 20% of the oxygen generated owing to the gas exchange occurring in the respiratory system (Gallucci Neto et al., 2005).
In addition, according to Chen and Zhong (Chen and Zhong, 2014), phospholipids in the brain are essential for the basis and process of neurotransmission and cognition and have a high proportion of polyunsaturated fatty acids. However, in AD, an increase in ROS and free radicals causes a decrease in polyunsaturated fatty acids, an increase in malondialdehyde and 4-hydroxynonenal, synthesis of isoprostanes (F2-IsoPs and F4-IsoPs), and slight cognitive impairment (Shinto et al., 2014). The oxidation of proteins, mainly tyrosine, by ROS and RNS gives rise to di-tyrosine and 3-nitrotyrosine, which are unfavorably related to scores on the Mini Mental State Examination; in addition, protein nitration is considered an early milestone in the pathogenesis of AD. Concomitantly, DNA oxidation results in the generation of 8-hydroxydeoxyguanosine and 8-hydroxyguanosine (8-OHG), which increase in the parietal, temporal, and frontal cortices of individuals with AD(Kawamoto et al., 2005, Weimann et al., 2018).
8-OHG is related to the precedence of characteristic features of the disease, such as the accumulation of Aβ plaques and NFTs, usually years before the onset of clinical signs (Chen and Zhong, 2014). In early-onset AD, missense mutations in presenilin 1 or 2 are indicated in the literature as the major causes, resulting in relative increases in the production of Aβ42/43 peptides. These peptides have hydrophobic characteristics, facilitating their self-aggregation and exacerbated Aβ deposition in middle age. As a result, there are two cleavage lines from Aβ48/49 or Aβ49/50 ε. Overall, presenilin mutations directly affect C- to N-terminal cleavage and, consequently, the relative production of longer Aβ peptides increases, which are more hydrophobic and self-aggregating (Selkoe and Hardy, 2016).
Thus, there is an increase in the Aβ42/Aβ40 ratio in humans, which are highly auto-aggregating, while Aβ40 may be anti-amyloidogenic (Kim et al., 2007). Studies in mice indicate that human Aβ42 oligomers induce tau hyperphosphorylation on AD-relevant epitopes and cause neuritic dystrophy in neurons, and when modulated by co-administration of Aβ antibodies, this scenario was completely avoided. From another perspective, studies demonstrate that the inheritance of a missense mutation in APP that decreases Aβ production and aggregation throughout life protects against AD and age-related cognitive decline (Selkoe and Hardy, 2016).
In AD, most cases are due to the influence of external and environmental factors, and only 1% of cases are related to familial mutations in genes encoding APP or preselins (PS1 and PS2). The deposition of intracellular Aβ residues occurs before the accumulation of extracellular Aβ, which is present in the mitochondria, ER, trans-Golgi network, and lysosomal and endosomal membranes, that damages synaptic activity, leading to synaptic dysfunction and memory deficits (Tönnies and Trushina, 2017). Although the function of APP is still unknown, it is closely related to the regulation of intracellular calcium, cell adhesion and growth, metal ion homeostasis, and axonal transport of vesicles (Smith et al., 2007).
According to a study by Paula et al. (Paula et al., 2009), the tau protein, located in the axons, stabilizes microtubules through the incorporation of tubulin and can present in soluble and insoluble forms; the latter configuration is detected in the basic element of NFTs, defined as paired helical filaments. Hyperphosphorylation can occur because of the increase in the functioning of kinases, such as taukinases, and the sub-impact of phosphatases or both. However, in line with a study by Tönnies and Trushina (2017), the formation of Aβ causes a translocation in the tau protein, causing it to be hyperphosphorylated, causing it to lose its stabilization capacity, displacing the microtubules, and suspending the neuronal circulation mechanism. Tau hyperphosphorylation is not very well understood but is known to affect the functionality of mitochondrial complex I, and its NFTs are found in other neurological diseases, causing morphological and biological dysfunctions in neurons.
The accumulation of Aβ, generation of NFTs, and presence of hyperphosphorylated tau compromise axonal transport and synapse processing, which can lead to the suppression of cell viability, degradation of the microtubular cytoskeleton, and neuronal death (Paula, Guimarães and Forlenza, 2009). These events may be related to oxidative stress (intensified through signaling stress in neurons), activation of the NMDAR (a neurotransmitter mediated by the cationic channel glutamate and an essential element of excitatory synaptic transmission), excitotoxicity, and neuronal plasticity. Activation of postsynaptic NMDARs induces an abundant influx of calcium ions (Ca2+) into postsynaptic cells, resulting in synaptic dysfunction, tau phosphorylation, mitochondrial operational deficiencies, activation of permeability transition pores in the inner mitochondrial membrane, loss of ATP, cytochrome C release, and continuous synthesis of ROS from intracellular cascades owing to the high concentration of cytoplasmic Ca2+ (Kamat et al., 2016).
In addition to Aβ plaques inducing cytoplasmic deposition of Ca2+, they boost the extracellular accumulation of glutamate, which corroborates the increased synthesis of ROS and oxidative stress. Glutamate receptor over-induction can lead to apoptosis, with neuronal cell death triggered by a set of toxic events. Excitotoxicity, defined by a sustained induction of excitatory amino acid receptors (NMDARs), occurs owing to toxic events such as the alteration of Ca2+ homeostasis, harmful over-regulation of the signaling means, RNS and ROS, which cause more nitrosative and oxidative stress, resulting in activation of apoptotic processes. The activation of caspases (caspase-3 and caspase-9) via central apoptosis is determined by a mitochondrial disorder that releases cytochrome C, which mediates the breakdown of synaptic proteins, such as the degradation of α-amino-acid receptor subunits. 3-hydroxy-5-methyl-4-isoxazolpropionic combined with the glutamate dose sample causes a decrease in the influx of Ca2+, leading to excitotoxicity (Kamat et al., 2016).
According to a study by Kamat et al. (2016), high levels of caspase-3 in the postsynaptic density portion of the AD brain can be linked to excitatory synaptic transmission variation, reduction in size and spinal density, memory impairment, and long-term depression (Snigdha et al., 2012). Therefore, the suppression of caspase-3 by pharmacological intervention may improve synapse transmission, memory impairment, and spine size and is considered opportune to reduce cognitive decline (Snigdha, et al., 2012, Wessels, et al., 2020). Another triggering factor for neuronal apoptosis is the relationship between age and mitochondrial oxidative stress, in which mitochondria influence the course of cell aging owing to damage caused to their structures by oxidative stress in pyramidal neurons and, hence, are subjected to neurodegeneration (Vassar, 2007). The main causes of neurodegeneration are damage to mitochondrial respiration, decreased synthesis and/or depletion of ATP, and deterioration of energy metabolism through the inhibition of important enzymes such as cytochrome C oxidase (Vassar, 2007).
AD neurodegeneration is also linked to several metals such as zinc and copper, which can substantially aggravate neuronal toxicity because the high concentration of Aβ plaques forms insoluble amyloid fibers that adhere to metals. This results in the development of high-affinity complexes with neurotoxicity depending on the type of metal connected to the senile plates, with copper and zinc being able to solubilize them (Sereniki et al., 2008). Tönnies and Trushina (2017) pointed out that soluble Aβ residues can be even more toxic to the body, influencing several molecular mechanisms that induce synapse abnormalities. In addition to the action on amyloid plaques, the altered homeostasis of bioactive metals can interfere with the synthesis of oxidative stress and free radicals and the deposition of tau protein. As a result, altered neuronal metal homeostasis in AD generates accumulation of a certain metal to compensate for the deficiency.
Zinc occasionally affects APP processing by binding to the membrane along with copper, iron, and aluminum binding directly to Aβ, leading to its agglomeration. The binding of iron and copper to Aβ plates can produce even more hydrogen peroxide, which reveals that metals and Aβ plates act synergistically for the elaboration of extra-mitochondrial ROS and oxidative stress (Tönnies and Trushina, 2017). Hydrogen peroxide is harmful in the presence of active redox metals, as it produces hydroxyl radicals from the Haber-Weiss or Fenton reaction. Enzymes such as GPx and CAT, which regulate the concentrations of hydrogen peroxide and act as antioxidant components of the protective mechanism, are found in smaller amounts in the brains of patients with neurodegenerative diseases such as AD. This contributes to an increased imbalance between pro-oxidants and antioxidants (Cheignon et al., 2018).
The human body has an antioxidant defense system that aims to reduce the damage caused by oxidative stress and free radicals and is composed of the following elements: non-enzymatic - can be endogenous (such as glutathione) or exogenous (acquired from the intake of vitamins C and E, minerals, carotenoids, organosulfur, and cofactors) (Awad et al., 2018) - and enzymatic- are endogenously synthesized, neutralize free radicals, restore structures, and associate peroxidized lipids and xenobiotics. Enzymatic antioxidants are composed of SOD, CAT, glutathione S-transferase, γ-glutamylcysteine synthase, GPx, and glutathione reductase and are responsible for oxidative damage and cell death (Mantzavinos and Alexiou, 2017) (Fig. 3 ).Figure 3 The formation of amyloid plaques can occur through increased levels of reactive oxygen species (ROS), mitochondrial dysfunction, inflammation, genetic mutations, etc., which interfere with intramembrane proteolysis, mediated by β-secretase and γ-secretase of amyloid precursor protein (APP), leading to the accumulation of β-amyloid peptide (Aβ) in the synaptic cleft and tau protein neurofibrillary tangles. The location of senile plaques occurs mainly in the cerebellar tonsils, hippocampus and entorhinal cortex of the temporal lobe, resulting in neural and synaptic dysfunction, loss of neurons and synapses, atrophy of different brain areas and, consequently, dementia and others characteristic clinical signs of Alzheimer's disease (AD).
Under physiological conditions, antioxidant enzymes such as SOD, GPx, CAT, glutaredoxins, and thioredoxins operate as a free radical excluding machinery and regulate the degree of ROS. This protective system is aided by the activation of nuclear factor 2 related to erythroid-2 (Nrf2), a transcription factor that undergoes ubiquitination by the E3 ubiquitin ligase, that is negatively mediated by its binding to the analogous ECH-associated protein to Kelch from stress sensor 1 (KEAP 1) and to the cytoplasmic repressor, passing as a substrate adapter (Kawamoto et al., 2005). When chemical reactives are present, such as metals and toxic reactive species, KEAP 1 releases Nrf2 into the cell nucleus, which is responsible for activating the transcription of cytoprotective genes through promoter segments that have conserved antioxidant response constituents. This increases the concentration of antioxidant enzymes and induces replacement of damaged organelles. However, in AD, the levels of Nrf2 and antioxidant response element (AREs) may be increased or decreased by mechanisms that are not yet known but are influenced by the organization of the disease and aging (Tönnies and Trushina, 2017) (Fig. 3).
Basic neurochemical changes in the cholinergic system are reported in AD indicating that cholinergic dysfunction is linked to changes in learning, memory, attention, and cognitive processing. Cholinergic dysfunction can be caused by the fraction of muscarinic receptors, intracellular signaling stimulated by these receptors, or by the influence of excessive amounts of endogenous low-molecular-weight inhibitor protein. This induces an endogenous divergence of muscarinic cholinergic receptors, causing a decrease in the brain levels of ACh and, thus, revealing the typical clinical manifestations of this disease (Ventura et al., 2010).
Cholinergic system and treatment of Alzheimer's disease
The triggering of events that lead to late AD (one that is not derived from the autosomal dominant form) presents a complex polygenic form, with interactions between several molecular cascades. These late manifestations are related to several factors, such as age, APOE ε4, possible cardiovascular conditions, and lifestyle (Rasch et al., 2006, Kuo et al., 2007). Furthermore, the implication of the neurotransmitter ACh in important physiological pathways, such as the response to stress, memory, learning, sleep and wakefulness, and sensory information, was highlighted by Ferreira-Vieira et al. (2016). The disease presents itself as a chronic syndrome that affects the CNS and has some factors that may contribute to its manifestation, such as Aβ aggregates and increased phosphorylation of tau protein, which collaborate with the progressive degeneration of the disease, causing harmful effects on the cognitive system, altering language, memory, judgment, orientation, learning, and deficits in decision-making (Sharma, 2019a).
Tau protein plays an important role in the proper functionality of this system and is hyperphosphorylated in the pathophysiology of AD, culminating in brain damage (Iqbal et al., 2005, Gauthier et al., 2016). Tau, which is associated with microtubules of normal neurons, plays a major role in the clinical manifestations and neurodegenerative signs of the disease as well as links APOE ε4 to the development of pathogenesis, as this allele has the capacity to influence Aβ deposition (Shi et al., 2017). All AD patients show progressive Aβ deposition followed by neuritic and surrounding glial cytopathology in brain regions that serve memory and cognition (Selkoe and Hardy, 2016). Currently, the main approach to cognitive and behavioral symptoms of early and late stages of AD has been the restoration of the cholinergic system. In age-associated neurodegenerative diseases, cholinergic atrophy and cognitive decline are accelerated (Giacobini, Cuello and Fisher, 2022). Furthermore, abnormal central cholinergic changes can also induce abnormal tau protein phosphorylation, nerve cell inflammation, cell apoptosis, and other pathological phenomena (Chen et al., 2022). Thus, reduced synthesis of this neurotransmitter is associated with loss of cholinergic neurons, as highlighted in the cholinergic theory of AD. New investigations indicate that memory loss related to both AD and other pathologies may be triggered by alterations in APP processing or ACh-mediated neuronal function, or both, which in turn trigger amyloid beta overexpression, synaptic malfunction and loss of trophic factor in targeted regions, eventually leading to synaptic and dendritic loss with age (Ferreira-Vieira et al., 2016b; Sharma, 2019b).
ACh is produced through choline acetyltransferase, with choline and acetate as precursor substrates and is subsequently stored in synaptic vesicles that are transferred to two classes of postsynaptic receptors: nicotinic (ionotropic) and muscarinic (metabotropic) receptors. Each of them has its specificities: nicotinic are oriented by ionic ligands, and muscarinic are associated with the G protein, which has five subclasses, each of which is found in smaller or larger amounts in the organism. M1 receptors, one of the subtypes of G protein receptors, are found abundantly in some brain regions, such as the cerebral cortex, hippocampus and striatum and, as inferred from their location, are involved in learning and memory mechanisms (Renard and Jean, 2017). Furthermore, cholinergic structures are found in both parasympathetic (pre- and post-ganglionic neurons) and sympathetic systems (pre-ganglionic neurons) (Ferreira-Vieira et al., 2016a).
The five subtypes of muscarinic receptors can act on both excitability and cell rest, depending on the characteristics of the activated cell and may inhibit either potassium (K+) or Ca2+ channels to promote excitability or cell rest, respectively. M1 and M3 receptors are mostly found at the postsynaptic level, whereas M2 and M4 receptors are usually located presynaptically. Although these receptors have the capacity for depolarization and repolarization, they often involve excitation, particularly in the cortical region, in addition to modulating other molecular pathways such as GABAergic inhibition and glutamatergic stimulation via secondary messengers, resulting in brain excitation (Ferreira-Vieira et al., 2016a).
Nicotinic receptors are selective for cations, such as K+, Na+, and Ca2+ , and are formed by five subunits: α, β, δ, γ (fetal), and ε (adult). The distinct combinations of these subunits form nine subclasses of nicotinic receptors, which are manifested in different structures that may occasionally have different specificities and purposes. Ganglia and muscles consist of receptors with five subunits, with alpha being doubled, while neurons comprise receptors with only two subunits, α 2-10 and β 2-4. The CNS has an accentuated α7 subunit with a particularly modulatory function, whereas the receptor reflects synaptic transmission in the PNS. Another relevant contrast between these receptors in different systems is that in the CNS, they are usually found in the presynaptic membrane and can modulate several neurotransmitters, such as glutamate, GABA, dopamine, serotonin, norepinephrine, and ACh, which are usually located in the postsynaptic membrane (Ferreira-Vieira et al., 2016a).
ACh and cholinergic neurons are also present in the CNS (at the neuromuscular junction) and autonomic nervous system (sympathetic and parasympathetic), presenting specific nicotinic and muscarinic receptors in each of these structures. After its biosynthesis and exocytosis to the synaptic cleft, ACh is degraded by AChE, which functions to decrease its concentration, consequently culminating in the depression of its action on postsynaptic receptors. Anticholinergic denomination is also attributed to processes that act in the reduction of ACh, which may occur in different ways, such as the reduction in the synthesis or release of ACh, increase in the functioning of AChE, or inhibition of receptors. In addition, the brain expresses five subclasses of muscarinic receptors; however, M1 plays a significant role in anticholinergic manifestations, such as cognitive impairment, confusion, sedation, delirium, and dizziness (Volpicelli-Daley et al., 2003).
The cholinergic system widely innervates brain regions; has memory, learning, and aspects of cognition involved in its regulation; has ACh as a neurotransmitter; and has the capacity to promote the plasticity of neurons; hence, it is involved in the central and peripheral modulation of the nervous system. Notably, the Meynert basal nucleus and its cholinergic composition play a fundamental role in the memory mechanism, and the degradation of neurons in this area leads to loss of functionality in the region, which is characteristic of AD (Rasch, Born and Gais, 2006a; Kuo et al., 2007a). The cholinergic and serotonergic systems are responsible for cognitive, emotional, and behavioral processes. Damage to these chemical structures and mechanisms reflects the AD process. This is because the brain regions of the basal forebrain, thalamus, and hippocampus are the main areas affected by the disease, and with the neural tissue destroyed, there is a depletion of both cholinergic nuclei in the basal area of the forebrain and cholinergic innervation directed to the cerebral cortex (Klaassens et al., 2019).
The forebrain is characterized by cholinergic aggregates and has a basal Meynert nucleus, which contains ramifications of neurons in the cortex and amygdala, with modulatory physiological functions acting on cortical activities. Under the effects of AD, this nucleus manifests as neuronal degeneration (Ferreira-Vieira et al., 2016a). These aspects of Meynert nucleus appear as one of the three hallmarks of the cholinergic hypothesis, the other two being presynaptic cholinergic markers in scarcity in the cerebral cortex, confirming that cholinergic antagonists impair memory. Hence, cholinergic depletion highlighted in AD neuropathology was observed, since there is a significant reduction in the action of the AChE enzyme in the cerebral cortex and limbic system in individuals with AD (Rasch et al., 2006, Kuo et al., 2007).
The cholinergic lesions that arise in AD are presynaptic and appear in the prodromal state of the disease; that is, they are asymptomatic, resulting in the depletion of neurons in the Meynert nucleus and axonal projections of the cerebral cortex, affecting nicotinic and muscarinic receptors. However, this highlights the involvement of cortical postsynaptic nicotinic receptors, given that there is a partial preservation of M1 that eventually becomes dysfunctional. The action of cholinergic agonists can also be observed, which have a positive modulation in neurochemical cascades in the neuronal configuration of AChE in the cerebral cortex. ACh has other brain implications such as neuroplasticity and the hemodynamic system owing to its capacity for cortical remodeling through the synchronization and connectivity of the neuronal network as well as the promotion of vasodilation and perfusion of the brain (Rasch et al., 2006, Kuo et al., 2007).
AChE is an important modulator for ACh supply in the synaptic cleft and has a significantly accentuated expression in the muscles and brain; therefore, AChE inhibitors are a method of therapy, as in the case of anesthetics, when promoting muscle block, and selective AChE inhibitors are used in the treatment of AD (Ferreira-Vieira et al., 2016a). Furthermore, AChE inhibitors can help in the treatment of AD, as they improve the cognitive symptoms of AD by preventing degradation of ACh. In addition, the serotonergic theory suggests that reduced functionality in AD contributes to the worsening of cognitive, behavioral, and mood changes, which are already compromised by the depletion of the cholinergic system (Klaassens et al., 2019).
The cholinergic theory, based on the adequate maintenance of acetylcholine levels in synapses, is of fundamental importance for the therapeutic method in AD because the signs and symptoms observed in this disease result, among other factors, from the decay of this neurotransmitter. Therefore, compounds that have the potential to block AChE can be used in the management of the disease to promote a higher concentration of ACh for metabolism. Drugs approved by the Food and Drug Administration for the treatment of AD include substances capable of inhibiting AChE, such as donepezil, rivastigmine, and galantamine (Sangaleti et al., 2021). These drugs have shown improvement in cognition and some behavioral areas and are administered in symptomatic treatments (Rasch et al., 2006b; Rosengarten et al., 2006; Kuo et al., 2007b).
Traditional acetylcholinesterase enzyme inhibitor (IAChE) inhibitors include physostigmine, tacrine, donepezil, rivastigmine, galantamine, and metrifonates (Rosengarten et al., 2006). Physostigmine, also known as eserian, has a short half-life and marked side effects; therefore, although it can cross the BBB, it has little therapeutic use. Tacrine has been approved for the treatment of AD but has been discontinued owing to its impact on metabolism (eg., hepatotoxicity). Donepezil, an IAChE, is the most prescribed drug (∼68%) for the treatment of the mild-moderate stage of the disease (Rosengarten et al., 2006, Cornelli, 2010). By modulating neurotransmitter levels, it is able to act at the molecular and cellular levels at various stages of the pathogenesis of the disease, such as, for example, depressing the expression of inflammatory cytokines and reducing the impact of oxidative stress (Hampel et al., 2018; Sharma, 2019a). Rivastigmine has slow activation kinetics, inactivates AChE for a limited time, and causes significant side effects. Metrifonate is a long-acting organophosphate inhibitor, and although it shows considerable improvement in people with mild to moderate AD, its development was discontinued owing to its adverse repercussions, one of them being respiratory paralysis. Galantamine binds to nicotinic cholinergic receptors and is effective in the treatment of cognitive symptoms of AD, and the tolerability of this drug is associated with a gradual increase in dosage (Sharma, 2019a; Sangaleti et al., 2021). Galantamine can also reduce the mortality of individuals with AD and improve cognitive depression (Ferreira-Vieira et al., 2016a; Sangaleti et al., 2021).
There is a discrepancy between the proper uses of cholinesterase inhibitors as these drugs are prescribed for less than half of the consultations for individuals with AD. Certain barriers prevent the proper use of these drugs such as the lack of medical knowledge and experience in the use of cholinesterase inhibitors in primary care and judicious requirements on the part of professionals regarding the clinical effectiveness of these inhibitors; furthermore, neurologists and psychiatrists are more inclined to prescribe these drugs (Rosengarten et al., 2006, Cornelli, 2010, Sangaleti et al., 2021). Thus, although this therapy has the potential to attenuate the signs of cognitive impairment commonly seen in all stages of AD, there are different and complex clinical contexts in the fickle use of these cholinesterase blockers (Rasch et al., 2006, Kuo et al., 2007).
The uncoupled muscarinic M1 receptor is associated with the worsening of AD, and the impaired functionality of this receptor is characteristically present in AD. Another property of this receptor is its ability to transfer the APP processing pathway in a non-amyloidogenic direction and to depress tau hyperphosphorylation. These muscarinic singularities have become instruments of disease therapies through M1 ACh agonists, such as M1 AF267B, which can recover the cognitive impairment representative of neurodegeneration and prevent the manufacture of Aβ peptide by increasing the α enzyme secretase. Furthermore, allosteric M1 modulators may contribute to Aβ depletion, and selective muscarinic agonists for M1 and M4 can, depending on the dose, decrease memory deficits, mood disorders, agitation, and hallucinations (Ferreira-Vieira et al., 2016a). Overall, M1 agonists act as negative regulators of both the amyloidogenic processes and neural protection that is provided by the excitation of nicotinic receptor 7 against the toxicity produced by Aβ, in addition to its connection with the anti-inflammatory pathways provided by this receptor (Hampel et al., 2018). Simultaneously, positive allosteric modulators and agonist have the ability to increase cholinergic function in neuronal degeneration processes (Renard and Jean, 2017).
The α 7 nicotinic acetylcholine receptor (α7nAChR) is encoded by a gene with variability and polymorphism and has high Ca2+ permeability, with rapid desensitization and, consequently, activation, in addition to being located in several brain regions, particularly the prefrontal cortex and hippocampus. In addition, this receptor plays a significant role in anti-inflammatory signaling owing to its ability to activate one of the modulators responsible for regulating oxidative stress (Nrf2). Thus, selective α7 agonists are analyzed with the aim of developing a drug route for the treatment of AD, such as encenicline, which showed improvement on cognitive impairments, pozanicline, ABT-418, etc. However, they have considerable side effects; hence, several studies on these drugs have been discontinued (Hoskin et al., 2019). The importance of the α 7 receptor is also because of its manifestation in cells as a precursor of oligodendrocytes, endothelial cells, astrocytes, and microglia, which play roles in immunity, inflammation, and neurological protection. Thus, this nicotinic receptor has cholinergic anti-inflammatory potential and acts in several different pathological contexts, including neurodegenerative processes, such as AD (Bouzat et al., 2018) (Fig 4 )(Table 1 ).Figure 4 The α 7 nicotinic receptors (α7nAChR) found in various locations in the body are one of the synaptic cleft acetylcholine scavengers. Additionally, they have the ability to regulate the concentrations of reactive oxygen species via the activation of an oxidative stress modulator nuclear factor 2 related to erythroid-2 (Nrf2) and, concomitantly, act in the signaling of the anti-inflammatory pathway. Thus, the use of α7nAChR agonists is one of the drug alternatives for the adjuvant treatment of Alzheimer's disease (AD) with the aim of promoting neurological protection. Muscarinic M1 receptors also act to capture the neurotransmitter acetylcholine to increase the concentration of the enzyme α secretase, which depresses the manufacture of β-amyloid peptide (Aβ) and decreases the hyperphosphorylation of tau protein. Thus, M1 receptor agonists act as negative modulators of amyloidogenic processes, as they allow the processing of amyloid precursor protein (APP) to a non-amyloidogenic pathway, which, consequently, increases cholinergic functions. Thus, appropriate levels of acetylcholine (ACh) are important in the maintenance of cognitive processes; therefore, the use of acetylcholinesterase (AChE) inhibitors promotes greater availability of this neurotransmitter in the synaptic cleft and provides an improvement in cognitive symptoms of Alzheimer's.
Table 1 Role of cholinergic receptors in Alzheimer's disease
Receptor Modulation Application Reference
M1 receptors Agonists Learning and memory Renard and Jean 2017
M1-M4 receptors GABAergic inhibition and glutamatergic stimulation Brain excitement Ferreira-Vieira et al. 2016
α7nAChR Modulation of several neurotransmitters, such as: glutamate, GABA, dopamine, serotonin, norepinephrine, and acetylcholine in the CNS CNS modulation Hoskin, Al-Hasan, and Sabbagh 2019
M1 receptors Modulation of processes that act to reduce acetylcholine Cognitive impairment, confusion, sedation, delirium, and dizziness Migirov and Datta 2019
Cholinergic agonist receptors Positive modulation in the neuronal configuration of choline acetyltransferase from the cerebral cortex Neuroplasticity, vasodilation, and brain perfusion Hampel et al. 2018
M1 and M4 receptors Modulation of β-amyloid peptide manufacturing Decreased memory deficits, mood disorder, agitation, and hallucinations Ferreira-Vieira et al. 2016
α7nAChR Modulation of Nrf2 activation (oxidative stress regulation) Improved anti-inflammatory signaling Hoskin, Al-Hasan, and Sabbagh 2019
α7nAChR Modulation of cells as the precursors of oligodendrocytes, endothelials, astrocytes and microglia Enhanced immunity, anti-inflammatory potential and neurological protection Bouzat et al. 2018
Purinergic therapeutic potential in Alzheimer's neuroinflammation
The purinergic system comprises an organization of specific receptors, enzymes, and signaling components and is responsible for promoting essential organic functions for body balance (Burnstock, 2020). Among these components, it is extremely important to mention adenosines, UTP, UDP and ATP, which act on different purinergic receptors, such as P1 (subdivided into A1, A2, A3 and A4) and P2 (subdivided into ionotropic P2X, containing seven subunits, and the metabotropic P2Y, containing eight subunits) (Burnstock, 2020, Simões and Bagatini, 2021). Under normal homeostasis conditions, the balance of these nucleotides is unaltered; however, under stress, inflammation, and apoptosis conditions, cells release extracellular ATP that activates some of the purinergic receptors, which may lead to imbalance and negative outcomes in homeostasis (Gratal et al., 2020). P2X and P2Y receptors (P2XR and P2YR) are found in astrocytes, oligodendrocytes, neurons, microglia, and endothelial cells in the CNS (Di Virgilio et al., 2009). Adenosine P1 receptors (P1R) are found in astrocytes, microglia, and neurons (Cieślak and Wojtczak, 2018). Furthermore, ATP is associated with neurotransmission, neurosecretion, and neuromodulation, particularly in the short term. In the long term, it plays an important role in the signaling involved in cell proliferation, differentiation, and death (Burnstock, 2020).
The main pathophysiology of AD is the accumulation of Aβ, followed by the deposition of NFTs, as previously discussed. This process may activate an inflammatory response, with an emphasis on the action of microglia, astrocytes, and neurons that may release ATP into the extracellular environment when exposed to Aβ (Orellana et al., 2011). This characteristic can play an important role in disease progression, particularly in more advanced cases. However, microglial activation is not the underlying cause of the pathological process but a synergistic epiphenomenon that occurs along with the negative outcomes brought about by long-term exposure to Aβ.
In initial and acute neuroinflammatory situations, microglial activity can contribute to the balance of neuronal homeostasis by increasing the clearance of Aβ and dead neurons as they phagocytose these components, reducing their neurotoxicity (Kim et al., 2012a; Heppner et al., 2015, Erb et al., 2019). This glial cell has two phenotypes with important roles in neuromodulation, namely M1, which is considered the classic form, characterized by the production of ROS and pro-inflammatory cytokines, such as IL-1β, IL-12, IL-23, STAT3, and TNF-α. M2 is characterized by its remodeling, repair, and angiogenesis owing to the release of anti-inflammatory cytokines, such as TGF-β, IL-10, IL-4, and IL-13. During advanced stages of the disease, microglia undergo a change in their phenotype—from the neuroprotective M2 to the classic M1 form (Wang et al., 2015), which helps to explain the pathophysiological changes according to the stage of the disease.
Purinergic receptors may be associated with this special feature of microglial cells, since under control conditions, with low levels of ATP and consequently low activation of the P2X7 receptor (P2X7R), they act as auxiliary receptors that phagocytize extracellular sediments. However, under temporary stimulation, autophagy is stimulated, material is degraded, and microglia expresses a mixed M1/M2 phenotype. Finally, in the case of prolonged stimulation of P2X7R, there was an increase in lysosomal pH with the extracellular release of autophagosomal content. Microglia expresses genes and proteins associated with M1 activation state (Campagno and Mitchell, 2021). Thus, in the pathogenesis of AD, an acute inflammatory response would be able to promote a protective effect in the CNS, while chronic neuroinflammation could contribute to neurodegenerative outcomes due to the increased release of pro-inflammatory cytokines that contribute to neuronal loss (Heppner et al., 2015, Erb et al., 2019, Thawkar and Kaur, 2019). The notion that different treatments should be targeted according to the temporal phase of the disease and that the purinergic system can play important roles in this cascade can help in its management and in the search for new therapeutic targets.
Continuing on the effects of microglia on the pathogenesis of the disease (Orellana, Froger, et al., 2011a; Orellana, Shoji, et al., 2011b), microglia releases cytokines, such as TNF-α and IL-1β, under stress conditions after exposure to Aβ, resulting in the increase of ATP and glutamate release via Cx43 hemichannels in astrocytes, which activate Pannexin 1 (PANX-1) in neurons through NMDAR and P2XR. The PANX-1 opening further increases ATP release, creating a vicious cycle by activating more P2XR, which may have a role in intracellular neurotoxic cascades. The blockage of receptors involved in the opening of this hemichannel, such as P2X7R, may contribute to reduced pro-inflammatory activity at the CNS level, reducing Aβ-induced neuronal death (Ryu and McLarnon, 2008, Orellana et al., 2011b). P2X7R is one of the most studied purinergic receptors in neurodegenerative diseases. It plays a role in AD neuroinflammation by activating nuclear factor-kappa B and the NLRP3 inflammasome, which are essential molecules in inflammatory cascade initiation (Thawkar and Kaur, 2019). P2X7R upregulation is also related to an increase in ROS production by microglia, which occurs mainly in chronic conditions and can lead to synaptic loss and death of cortical cells (Parvathenani et al., 2003). Munoz et al. (2017) and Bartlett and colleagues (2013) showed that the activation of P2X7R by ATP induces the uptake of organic cations, ROS formation, and cell death, and these events may be inhibited by the use of AZ10606120 and A438079 (P2X7R antagonists). Thus, the purinergic system is associated with oxidative stress, which, as described in this article, participates in the pathophysiology of AD.
Other mechanisms involving this receptor have been described by Sanz et al. (Sanz et al., 2009), who showed that wild-type rats exposed to Aβ had greater ATP release, IL-1β secretion and accumulation, and microglial plasma membrane permeability than that of P2X7R knockout rats. In addition, this last group of the study showed a reduction in the number of amyloid plaques in the hippocampal region due to an increase in α-secretase activity (Diaz-Hernandez et al., 2012). Coincidentally, α-secretase activity decreases under conditions of enhanced P2X7R expression (Cieślak and Wojtczak, 2018). McLarnon et al. (2006) demonstrated that AD patients had increased expression of P2X7R compared to those without the disease, and that cells exposed to Aβ also had higher expression of the P2X7R than those without contact with this agent. The use of P2X7R antagonists is also related to neuroprotection by reducing the inflammatory response and improving cognition and memory in Aβ-induced AD rat models (Ryu and McLarnon, 2008, Chen et al., 2014). Therefore, a correlation between the blockage of these receptors and neuroprotection in the pathophysiology of AD is plausible.
P2X4R and P2Y6R are other receptors that may also play a role in microglia-mediated neuroinflammation and contribute to AD neurodegeneration (Di Virgilio, et al., 2009, Godoy et al., 2019). P2X4R expression increases after nerve tissue damage (Varma et al., 2009) and its upregulation can be induced in AD cases by ATP release by microglia via Aβ induction, leading to neuronal death (Woods et al., 2016). This receptor drives microglial motility via phosphatidylinositol-3-kinase/Akt, in addition to being involved in the production of prostaglandin E2 (PGE2), brain-derived neurotrophic factor, and TNF-α, which may be related to the pathophysiology of the disease. P2X4R has a synergistic relationship with P2X7R (the absence of expression of one can further activate the other, acting as a compensation and complementation mechanism) (Suurväli, et al., 2017, Castillo et al., 2022). Finally, it is highly expressed in cases of spinal cord diseases, inflammatory pain, and pre-term hypoxia-ischemia and may also act on LTP and plasticity at CA1 synapses in the rat hippocampus, suggesting that its overexpression may be related to synaptic dysfunction and microglial phagocytic function in patients with AD (Castillo et al., 2022). P2Y6R is selectively activated by UDP, is upregulated after neuronal damage, and is associated with the release of cytokines, such as CCL2 and CCL3, in microglia and astrocytes (Kim, et al., 2011, Morioka et al., 2013, Woods, et al., 2016). Koizumi and collaborators (Koizumi et al., 2007) showed that kainic acid exposure in the hippocampus of rats increased extracellular uridine nucleotide levels and phagocytic activity of microglia. This phagocytic activation may increase Aβ clearance; however, as a consequence, it can also result in phagocytosis of viable neurons, which is associated with neurodegeneration (Brown and Neher, 2014, Anwar et al., 2020). The use of P2Y6R antagonists reduces neuronal death after LPS exposure in rats (Neher et al., 2014) and reduces neuronal loss and memory deficits in knockout mice induced by Aβ or tau oligomers (Puigdellívol et al., 2021).
Receptors associated with neuroprotection, such as the metabotropics P2Y1 and P2Y2, have also been studied (Peterson et al., 2010, Cieślak and Wojtczak, 2018). Mishra et al. (Mishra et al., 2006) showed that the activation of these receptors could increase the proliferation of neural stem cells and act on neurogenesis, which may be a new therapy to regenerate damaged hippocampal neurons in patients with AD or even attenuate its progression. In addition, P2Y1R protects neuronal tissue against oxidative stress, with IL-6 being the main signaling molecule for this event (Fujita et al., 2009), which may have dual activity in neural inflammation (Rothaug, Becker-Pauly and Rose-John, 2016). However, recent studies have shown that its action can worsen AD dysfunction because it is involved in astrocytic hyperactivity (Reichenbach et al., 2018), which is associated with the worsening of AD dysfunction. Delekate et al. (2014) demonstrated that P2Y1R antagonism returned astroglial network dysfunction to its normal state. Further studies are required to confirm its protective effects. P2Y2R is present in several cells in the CNS and its activation may occur owing to the release of ATP in the presence of Aβ (Cieślak and Wojtczak, 2018).
Deletion of P2Y2R in mice increased Aβ, causing neurological deficits, and decreased the expression of the microglial marker CD11b, which is associated with the severity of microglial activation (Ajit et al., 2014); hence, it acts on the recruitment of microglial cells and Aβ clearance. Aβ clearance may occur by the upregulation of P2Y2R, which plays a role in Aβ degradation by the phagocytic action of microglial cells (Kim et al., 2012b). Kong et al. (2009a) reinforced that this role can be mediated by P2X2R via IL-1β through the IκB-α/NF-κB signaling pathway, which stimulates α-secretase activity mediated by ADAM10/17 metalloproteinases (Kong et al., 2009b). An important finding is that the activity of P2X7R is related to the reduction of α-secretase, whereas P2Y2R has the opposite effect, increasing the release of α-secretase (León-Otegui et al., 2011). Based on this finding, it is suggested that P2Y2R activation concomitant with P2X7R antagonism leads to an efficient activation of α secretase; thus, they are promising candidates as new therapeutic targets in AD.
Finally, the adenosine-activated receptor P1 family, such as A1 and A2A, are most studied in relation to neurological diseases and can affect the glutamatergic, dopaminergic, and cholinergic signaling pathways and alter synaptic plasticity in regions responsible for learning and memory development (Ribeiro et al., 1996, Gomes et al., 2011, Cieślak and Wojtczak, 2018). Antagonism of these receptors is related to cognitive improvement and a reduced risk of dementia (Woods et al., 2016). Caffeine is an antagonist of these receptors and reduces the neurotoxicity induced by Aβ, which was verified by blocking A2AR but not A1R (Huang et al., 2011, Wang, et al., 2014). In a recent in vivo study, Faivre et al. (2018) showed that long-term treatment for 6 months with the A2AR antagonist MSX-3 in a mouse model of AD prevented memory deficits and reduced Aβ1-42 levels in the cortex, and Stazi et al. (2022) showed that treatment for 4 months reduced the loss of neurons in the hippocampus and improved neurogenesis, memory, and learning ability, which may reflect A2AR antagonism. Similar results in cognitive abilities were found in mouse models with tauopathy and those with transgenic lack of A2AR gene encoding (Laurent et al., 2016). Finally, A1R and A2AR modulate cortical release of ACh in the prefrontal cortex, acting on behavioral excitation and sleep (Baghdoyan and Lydic, 2009). A1R antagonism is also associated with the transmission of ACh in the cortical region and an increase in its extracellular levels in cholinergic terminals (Rahman, 2009), which provides opportunities for further studies to elucidate its behavior and therapy in AD.
SARS-CoV-2 exacerbates the immune system, promoting the worsening of inflammation, which contributes to the advancement of the disease and activation of purinergic receptors negatively involved in the pathophysiology of AD. Thus, therapeutic measures aimed at the analysis of these receptors are essential to prevent the neural progression of the disease in these patients, not only for those who had contact with the new coronavirus but also for the uninfected ones that can benefit from treatment with these therapeutic targets (Table 2 ).Table 2 Purinergic receptors and main roles in Alzheimer Disease
Receptor Agonist Cell activation Main role Reference
P2X7 ATP, ADP Microglia - Acute response scavenger receptor, promote M2 microglia phenotype- Chronic response promote M1 microglia phenotype- Linked to Panx-1 hemichannel opening- Activation of nuclear factor-kappa B (NF-kB) and NLRP3 inflammasome- ROS production- Decrease α-secretase activity Bartlett, Yerbury and Sluyter, 2013; Hernandez et al., 2012; Munoz et al, 2017; Parvathenani et al. , 2003; Orellana et al., 2011a; Sans et al., 2009; Thawkar; Kaur, 2019;Wang et al., 2015;
P2X4 ATP Microglia - Synergic to P2X7R response Castillo et al., 2022; Suurvali, et al., 2017
P2Y6 UDP Microglia and astrocytes - Cytokines release- Phagocytosis of neurons Aβ induced Anwar; Pons; Rivest, 2020; Brown; Neher, 2014; Erb et al., 2015; Kim, et al., 2011, Morioka et al., 2013;
P2Y2 UTP Microglia - Increase α-secretase release and promotes Aβ clearence Ajit et al., 2015; Kim et al., 2012
P2Y1 ADP, ATP Microglia and astrocytes - Astrocytic hyperactivity- Neuroprotection via microglia activation Delekate et al. 2014; Fujita; Tozaki-Saitoh; Inoue, 2009; Reichenbach et al., 2018;
A2A Adenosine Microglia - Reduces the neurotoxicity induced by Aβ- Neurogenesis Dall'Igna et al. 2003; Faivre et al., 2018; Stazi et al., 2022
Future proposals
Neurodegeneration as a result of AD mainly affects cholinergic pathways, such as forebrain neurons, affecting memory deficits. Thus, pharmacology focused on this area is a promising segment for the treatment of the disease (Richter et al., 2018), since abnormal cholinergic pathways result in inefficiency of distribution and function in the body, for eg., cognitive deficits (Rasch et al., 2006, Kuo et al., 2007). In addition, cholinesterase inhibitors are relevant tools in therapies aimed at improving AD, given their ability to increase the availability of ACh in the synaptic cleft and, consequently, to the body (Ferreira-Vieira et al., 2016a) as well as anti-amyloid and anti-tau interventions (Rasch et al., 2006, Kuo et al., 2007). Furthermore, studies aimed at understanding the complexity of AD pathophysiology can improve treatment by analyzing triggering and aggravating factors, considering oxidative stress, aggregation of Aβ and tau protein, genetic predisposition, inflammatory mechanisms, and mitochondrial disturbances (Hampel et al., 2018; Sharma, 2019a). Considering these involvements, multifactorial diseases can be better interpreted and oriented, contributing to the development of personalized drug therapies and interventions (Rasch et al., 2006, Kuo et al., 2007) (Fig. 4).
Active investigation into anti-inflammatory methods triggered by some receptors, such as nicotinic 7, is significantly relevant, as they can be used and directed to improve cognition in several pathologies, including AD. Understanding their functionality and impact on organisms may point to useful modulation pathways for treatments and as targets for pharmacological tools (Bouzat et al., 2018). Because of the wide location of brain nicotinic receptors, their impact on AD pathophysiology, and their effects and means of regulation, cholinergic receptors have become potential therapeutic targets, enabling the use of agonists that stimulate the cholinergic system to play a role in the organism as this system is compromised in AD. Thus, the study of the cholinergic pathway is essential to contribute to the improvement of brain impairment caused by AD (Hoskin et al., 2019) (Fig. 4).
Other receptors that comprise the purinergic system are also essential tools for AD pharmacotherapy. The P2X7R is one of the most closely related receptors to the pathophysiology of AD. Its activation, depending on the stage of the disease, is related to negative outcomes and affects the inflammatory response, with a role in microglial modulation, release of ROS, and reduction in the levels of α-secretase (Cieślak and Wojtczak, 2018, Campagno and Mitchell, 2021). P2X4R has a synergistic role with P2X7R in inflammation, and its antagonism may be an adjuvant target in treatment as well (Castillo et al., 2022). Other receptors such as P2Y6 and P2Y1 have setbacks in this field: although the former is associated with an increase in α-secretase, it is linked with neuronal damage and the latter has a protective effect against ROS but is associated with possible negative outcomes by worsening astrocyte activity (Fujita et al., 2009, Anwar et al., 2020). Hence, these receptors need further studies to elucidate their pathophysiology of neurodegenerative diseases. The activity of P2Y2 is closely related to neuroprotection, and agonism of P2Y2 promotes an increase in α-secretase and Aβ clearance (Cieślak and Wojtczak, 2018). The antagonism of P1 family receptors, such as A2A, is also associated with positive outcomes in neurodegenerative models, as it improves aspects of cognition, such as memory, and protects neural tissue against Aβ-induced damage (Woods et al., 2016). Therefore, purinergic receptors are promising therapeutic targets against the progression of AD, and further studies on their roles in the course of this neurodegenerative disease are essential to find a feasible elucidation of their use in clinical practice (Fig. 5 ).Figure 5 Adenosine triphosphate (ATP) is released via neuronal damage/tissue inflammation or microglial cell exposure to β-amyloid peptide (Aβ), which liberate tumor necrosis factor alpha (TNF-α) and Interleukin 1 beta (IL-1B), acting on astrocyte and releasing additional ATP. It activates P2X7 receptor (P2X7R) in macroglia and neuron, causing α-secretase reduction and activating PANX-1, which is involved in ATP secretion. P2X7R is also involved in reactive oxygen species (ROS) liberation. P2X4R has effects similar to P2X7 and is involved in TNF-α and prostaglandin E2 (PGE2) secretion.
Conclusion
The global pandemic triggered by SARS-CoV-2 has led to several analyses regarding the impact of COVID-19 on the body. Its involvement in the neural network can aggravate other pre-existing dysfunctions, as is the case in individuals with AD who already have a neurodegenerative condition and owing to the hyperinflammation produced by the virus along with the oxidative stress process, result in prominent nerve damage. Thus, because of the involvement and impairment of the nervous system, studies on anti-inflammatory modulation pathways may contribute to attenuating the effects of COVID-19 and neural degeneration processes of AD. The cholinergic system, acting through the neurotransmitter ACh and cholinergic neurons, contributes to the anti-inflammatory potential through muscarinic and nicotinic receptor agonists, which have protective effects against neurodegenerative processes. The cholinergic system could be beneficial in preventing these neurological disorders; additionally, the purinergic system is a promising target. The antagonism of purinergic receptors, such as P2X7, P2X4, P2Y6, and A2A, and agonism of P2Y2 showed improvement in inflammation, oxidative stress, and cognition. Therefore, the use of cholinergic and purinergic therapies is a tool for modulating the nervous system and providing an alternative drug for the treatment of some physiological disorders.
Ethics approval and consent to participate
Not applicable
Consent for publication
Informed consent and consent for publication were obtained from each participant.
Availability of data and materials
Not applicable
Competing interests
The authors declare no conflicts of interest.
Funding
CNPq (Proj. N. 404256/2021-0 and Proj. N. 310606/2021-7)
Authors' contributions
JLBS conceptualized the paper. JLBS, LDS, IFL, MVRS, JVC, and MDB performed the literature search and data analysis as well as drafted and critically revised the work.
Uncited references
Beacon and Davie, , 2021, Van Dort et al., 2009, Kamat et al., 2016, Kamat et al., 2016, de Paula et al., 2009, Picciotto and Mineur, , 2012, Shao and Wang, 2017, Sharma, 2019, Sharma, 2019, Lastname, et al., 2020a, Xie, 2020, Lastname, et al., 2020b.
Acknowledgements
JLBS is grateful to the Federal University of Fronteira Sul for the research grant that promoted the production of this publication.
Disclosure of potential conflicts of interest
Not applicable
Research involving Human Participants and/or Animals
Not applicable
Informed consent
Not applicable
==== Refs
References
Abate G. ‘The pleiotropic role of p53 in functional/dysfunctional neurons: focus on pathogenesis and diagnosis of Alzheimer’s disease’, Alzheimer’s Research & Therapy, 12(1) Available at: 2020 10.1186/S13195-020-00732-0
Ajit D. Loss of P2Y2 Nucleotide Receptors Enhances Early Pathology in the TgCRND8 Mouse Model of Alzheimer’s Disease Molecular neurobiology 49 2 2014 1031 10.1007/S12035-013-8577-5 Available at: 24193664
Alkeridy A Unique Presentation of Delirium in a Patient with Otherwise Asymptomatic COVID-19 Journal of the American Geriatrics Society 68 7 2020 1382 1384 10.1111/JGS.16536 Available at: 32383778
Anwar S. Pons V. Rivest S. ‘Microglia Purinoceptor P2Y6: An Emerging Therapeutic Target in CNS Diseases’, Cells, 9(7) Available at: 2020 10.3390/CELLS9071595
Armstrong R.A. Risk factors for Alzheimer’s disease Folia neuropathologica 57 2 2019 87 105 10.5114/FN.2019.85929 Available at: 31556570
Asselah T. COVID-19: Discovery, diagnostics and drug development Journal of Hepatology 74 1 2021 168 10.1016/J.JHEP.2020.09.031 Available at: 33038433
Awad M.A. Aldosari S.R. Abid M.R. Genetic Alterations in Oxidant and Anti-Oxidant Enzymes in the Vascular System Frontiers in Cardiovascular Medicine 2018 107 10.3389/FCVM.2018.00107 Available at: 30140678
Baig A.M. Evidence of the COVID-19 Virus Targeting the CNS: Tissue Distribution, Host-Virus Interaction, and Proposed Neurotropic Mechanisms ACS Chemical Neuroscience 11 7 2020 995 10.1021/ACSCHEMNEURO.0C00122 Available at: 32167747
Barondes Galectins: a family of animal beta-galactoside-binding lectins Cell 76 4 1994 597 598 10.1016/0092-8674(94)90498-7 Available at: 8124704
Bartlett R. Yerbury J.J. Sluyter R. ‘P2X7 Receptor Activation Induces Reactive Oxygen Species Formation and Cell Death in Murine EOC13 Microglia’, Mediators of Inflammation, 2013 Available at: 2013 10.1155/2013/271813
Beacon D. Davie, Epigenetic regulation of ACE2, the receptor of the SARS-CoV-2 virus 1 Genome 64 4 2021 386 399 10.1139/GEN-2020-0124 Available at: 33086021
Bjerke M. Engelborghs S. Cerebrospinal Fluid Biomarkers for Early and Differential Alzheimer’s Disease Diagnosis Journal of Alzheimer’s disease : JAD 62 3 2018 1199 1209 10.3233/JAD-170680 Available at: 29562530
Bourgonje A.R. Angiotensin-converting enzyme-2 (ACE2), SARS-CoV-2 and pathophysiology of coronavirus disease 2019 (COVID-19) The Journal of Pathology 251 3 2020 228 248 10.1002/PATH.5471 Available at: 32418199
Bouzat C. Molecular function of α7 nicotinic receptors as drug targets The Journal of Physiology 596 10 2018 1847 10.1113/JP275101 Available at: 29131336
Bradley-Whitman M.A. Lovell M.A. Epigenetic changes in the progression of Alzheimer’s disease Mechanisms of ageing and development 134 10 2013 486 495 10.1016/J.MAD.2013.08.005 Available at: 24012631
Brown G.C. Neher J.J. Microglial phagocytosis of live neurons Nature Reviews Neuroscience 15 4 2014 209 216 10.1038/NRN3710 Available at: 24646669
Bryce-Moncloa A. COVID-19 from a cardiovascular perspective Available at: Archivos de cardiologia de Mexico 91 Suplemento COVID 2021 086 094 10.24875/ACM.20000304 33871468
Burnstock, G. (2020) ‘Introduction to Purinergic Signaling’, Methods in Molecular Biology, 2041, pp. 1–15. Available at: https://doi.org/10.1007/978-1-4939-9717-6_1.
Campagno K.E. Mitchell C.H. The P2X7 Receptor in Microglial Cells Modulates the Endolysosomal Axis, Autophagy, and Phagocytosis Frontiers in Cellular Neuroscience 15 2021 66 10.3389/FNCEL.2021.645244/BIBTEX Available at:
Cancino G.I. p63 Regulates Adult Neural Precursor and Newly Born Neuron Survival to Control Hippocampal-Dependent Behavior The Journal of Neuroscience 33 31 2013 12569 10.1523/JNEUROSCI.1251-13.2013 Available at: 23904595
Canevelli M. Familial Alzheimer’s disease sustained by presenilin 2 mutations: systematic review of literature and genotype-phenotype correlation Neuroscience and biobehavioral reviews 42 2014 170 179 10.1016/J.NEUBIOREV.2014.02.010 Available at: 24594196
Castillo C. ‘Microglial Activation Modulated by P2X4R in Ischemia and Repercussions in Alzheimer’s Disease’, Frontiers in Physiology, 13 Available at: 2022 10.3389/FPHYS.2022.814999
Chaudhry ‘Manipulation of ACE2 expression in COVID-19’, Open heart, 7(2) Available at: 2020 10.1136/OPENHRT-2020-001424
Chazal N. Coronavirus, the King Who Wanted More Than a Crown: From Common to the Highly Pathogenic SARS-CoV-2, Is the Key in the Accessory Genes? Frontiers in Microbiology 12 2021 682603 10.3389/FMICB.2021.682603 Available at:
Cheignon Oxidative stress and the amyloid beta peptide in Alzheimer’s disease Redox biology 14 2018 450 464 10.1016/J.REDOX.2017.10.014 Available at: 29080524
Chen X. Brilliant Blue G improves cognition in an animal model of Alzheimer’s disease and inhibits amyloid-β-induced loss of filopodia and dendrite spines in hippocampal neurons Neuroscience 279 2014 94 101 10.1016/J.NEUROSCIENCE.2014.08.036 Available at: 25193238
Chen and Zhong Oxidative stress in Alzheimer’s disease Neuroscience bulletin 30 2 2014 271 281 10.1007/S12264-013-1423-Y Available at: 24664866
Chen Z.R. ‘Role of Cholinergic Signaling in Alzheimer’s Disease’, Molecules (Basel, Switzerland), 27(6) Available at: 2022 10.3390/MOLECULES27061816
Cieślak M. Wojtczak A. Role of purinergic receptors in the Alzheimer’s disease Purinergic Signalling 14 4 2018 331 10.1007/S11302-018-9629-0 Available at: 30362042
Conti Coronavirus-19 (SARS-CoV-2) induces acute severe lung inflammation via IL-1 causing cytokine storm in COVID-19: a promising inhibitory strategy Available at: Journal of biological regulators and homeostatic agents 34 6 2020 1971 1975 10.23812/20-1-E 33016027
Cornelli U. Treatment of Alzheimer’s disease with a cholinesterase inhibitor combined with antioxidants Neuro-degenerative diseases 7 1–3 2010 193 202 10.1159/000295663 Available at: 20224285
Cox M.A. Beyond neurotransmission: acetylcholine in immunity and inflammation Journal of Internal Medicine 287 2 2020 120 133 10.1111/JOIM.13006 Available at: 31710126
Culjak The Association between TNF-alpha, IL-1 alpha and IL-10 with Alzheimer’s Disease Current Alzheimer research 17 11 2020 972 984 10.2174/1567205017666201130092427 Available at: 33256580
Delekate, A. et al. (2014) ‘Metabotropic P2Y1 receptor signalling mediates astrocytic hyperactivity in vivo in an Alzheimer’s disease mouse model’, Nature Communications 2014 5:1, 5(1), pp. 1–14. Available at: https://doi.org/10.1038/ncomms6422.
Dhama K. Coronavirus disease 2019–COVID-19 Clinical Microbiology Reviews 33 4 2020 1 48 10.1128/CMR.00028-20 Available at:
Diaz-Hernandez J.I. In vivo P2X7 inhibition reduces amyloid plaques in Alzheimer’s disease through GSK3β and secretases Neurobiology of aging 33 8 2012 1816 1828 10.1016/J.NEUROBIOLAGING.2011.09.040 Available at: 22048123
Dong ‘Serum MicroRNA Profiles Serve as Novel Biomarkers for the Diagnosis of Alzheimer’s Disease’, Disease markers, 2015 Available at: 2015 10.1155/2015/625659
Van Dort C.J. Baghdoyan H.A. Lydic R. ‘Adenosine A(1) and A(2A) receptors in mouse prefrontal cortex modulate acetylcholine release and behavioral arousal’, The Journal of neuroscience : the official journal of the Society for Neuroscience 29 3 2009 871 881 10.1523/JNEUROSCI.4111-08.2009 Available at: 19158311
Equils O. Proposed mechanism for anosmia during COVID-19: The role of local zinc distribution Available at: Journal of Translational Science 7 1 2021 10.15761/JTS.1000397
Erb L. Purinergic signaling in Alzheimer’s disease Brain Research Bulletin 151 2019 25 37 10.1016/J.BRAINRESBULL.2018.10.014 Available at: 30472151
Faivre E. ‘Beneficial effect of a selective adenosine A2A receptor antagonist in the APPswe/PS1dE9 mouse model of Alzheimer’s disease’, Frontiers in Molecular Neuroscience, 11 Available at: 2018 10.3389/FNMOL.2018.00235/FULL
Fändrich, Oligomeric intermediates in amyloid formation: structure determination and mechanisms of toxicity Journal of molecular biology 421 4–5 2012 427 440 10.1016/J.JMB.2012.01.006 Available at: 22248587
Farmer K.M. ‘P53 aggregation, interactions with tau, and impaired DNA damage response in Alzheimer’s disease’, Acta Neuropathologica Communications, 8(1) Available at: 2020 10.1186/S40478-020-01012-6
Farsalinos K. Editorial: Nicotine and SARS-CoV-2: COVID-19 may be a disease of the nicotinic cholinergic system Toxicology Reports 7 2020 658 10.1016/J.TOXREP.2020.04.012 Available at: 32355638
Ferreira D. Nordberg A. Westman E. Biological subtypes of Alzheimer disease: A systematic review and meta-analysis Neurology 94 10 2020 436 10.1212/WNL.0000000000009058 Available at: 32047067
Ferreira-Vieira T. Alzheimer’s disease: Targeting the Cholinergic System Current Neuropharmacology 14 1 2016 101 115 10.2174/1570159x13666150716165726 Available at: 26813123
Fu Y. Cheng Y. Wu Y. Understanding SARS-CoV-2-Mediated Inflammatory Responses: From Mechanisms to Potential Therapeutic Tools Virologica Sinica 35 3 2020 266 10.1007/S12250-020-00207-4 Available at: 32125642
Fujita T. Tozaki-Saitoh H. Inoue K. P2Y1 receptor signaling enhances neuroprotection by astrocytes against oxidative stress via IL-6 release in hippocampal cultures Glia 57 3 2009 244 257 10.1002/GLIA.20749 Available at: 18756525
Gallucci Neto J. Tamelini M.G. Forlenza O.V. Diagnóstico diferencial das demências Archives of Clinical Psychiatry (São Paulo) 32 3 2005 119 130 10.1590/S0101-60832005000300004 Available at:
Gao, Y.-L. et al. (2018) ‘Tau in neurodegenerative disease’, Annals of Translational Medicine, 6(10), pp. 175–175. Available at: https://doi.org/10.21037/ATM.2018.04.23.
Gauthier S. Efficacy and safety of tau-aggregation inhibitor therapy in patients with mild or moderate Alzheimer’s disease: a randomised, controlled, double-blind, parallel-arm, phase 3 trial Lancet (London, England) 388 10062 2016 2873 2884 10.1016/S0140-6736(16)31275-2 Available at: 27863809
Giacobini E. Cuello A.C. Fisher A. Reimagining cholinergic therapy for Alzheimer’s disease Brain : a journal of neurology 145 7 2022 2250 2275 10.1093/BRAIN/AWAC096 Available at: 35289363
Goate A. Segregation of a missense mutation in the amyloid precursor protein gene with familial Alzheimer’s disease Nature 349 6311 1991 704 706 10.1038/349704A0 Available at: 1671712
Godoy P.A. Ramírez-Molina O. Fuentealba J. ‘Exploring the Role of P2X Receptors in Alzheimer’s Disease’, Frontiers in Pharmacology, 10 Available at: 2019 10.3389/FPHAR.2019.01330
Goedert (2015) ‘NEURODEGENERATION. Alzheimer’s and Parkinson’s diseases: The prion concept in relation to assembled Aβ, tau, and α-synuclein’, Science (New York, N.Y.), 349(6248). Available at: https://doi.org/10.1126/SCIENCE.1255555.
Gomes C.V. ‘Adenosine receptors and brain diseases: Neuroprotection and neurodegeneration’, Biochimica et Biophysica Acta (BBA) - Biomembranes 1808 5 2011 1380 1399 10.1016/J.BBAMEM.2010.12.001 Available at:
Gómez-Mesa ‘Thrombosis and Coagulopathy in COVID-19’, Current problems in cardiology, 46(3) Available at: 2021 10.1016/J.CPCARDIOL.2020.100742
Gratal P. Purinergic System Signaling in Metainflammation-Associated Osteoarthritis Frontiers in Medicine 7 2020 506 10.3389/FMED.2020.00506/BIBTEX Available at: 32984382
Hamming, I. et al. (2004) ‘Tissue distribution of ACE2 protein, the functional receptor for SARS coronavirus. A first step in understanding SARS pathogenesis’, The Journal of Pathology, 203(2), p. 631. Available at: https://doi.org/10.1002/PATH.1570.
Hampel H. ‘The cholinergic system in the pathophysiology and treatment of Alzheimer’s disease’, Brain 2018 Oxford University Press 1917 1933 Available at:
Hardy J. Selkoe D.J. ‘The amyloid hypothesis of Alzheimer’s disease: progress and problems on the road to therapeutics’, Science 297 2002 353 356 Available at: 12130773
Heppner F.L. Ransohoff R.M. Becher B. Immune attack: The role of inflammation in Alzheimer disease Nature Reviews Neuroscience 16 6 2015 358 372 10.1038/NRN3880 Available at: 25991443
Hoover D.B. Cholinergic Modulation of the Immune System Presents New Approaches for Treating Inflammation Pharmacology & therapeutics 179 2017 1 10.1016/J.PHARMTHERA.2017.05.002 Available at: 28529069
Hoskin J.L. Al-Hasan Y. Sabbagh M.N. Nicotinic Acetylcholine Receptor Agonists for the Treatment of Alzheimer’s Dementia: An Update Nicotine & Tobacco Research 21 3 2019 370 10.1093/NTR/NTY116 Available at: 30137524
Huang Z.-L. Urade Y. Hayaishi O. The role of adenosine in the regulation of sleep Current topics in medicinal chemistry 11 8 2011 1047 1057 10.2174/156802611795347654 Available at: 21401496
Hultman K. Strickland S. Norris E.H. The APOE ɛ4/ɛ4 genotype potentiates vascular fibrin(ogen) deposition in amyloid-laden vessels in the brains of Alzheimer’s disease patients Journal of Cerebral Blood Flow & Metabolism 33 8 2013 1251 10.1038/JCBFM.2013.76 Available at: 23652625
Ide M. ‘Periodontitis and Cognitive Decline in Alzheimer’s Disease’, PloS one, 11(3) Available at: 2016 10.1371/JOURNAL.PONE.0151081
Ionescu-Tucker A. Cotman C.W. Emerging roles of oxidative stress in brain aging and Alzheimer’s disease Neurobiology of Aging 107 2021 86 95 10.1016/J.NEUROBIOLAGING.2021.07.014 Available at: 34416493
Iqbal K. ‘Tau pathology in Alzheimer disease and other tauopathies’, Biochimica et Biophysica Acta - Molecular Basis of Disease Elsevier 2005 198 210 10.1016/j.bbadis.2004.09.008 Available at:
Kamat Mechanism of Oxidative Stress and Synapse Dysfunction in the Pathogenesis of Alzheimer’s Disease: Understanding the Therapeutics Strategies Molecular neurobiology 53 1 2016 648 661 10.1007/S12035-014-9053-6 Available at: 25511446
Kawamoto E.M. Oxidative state in platelets and erythrocytes in aging and Alzheimer’s disease Neurobiology of aging 26 6 2005 857 864 10.1016/J.NEUROBIOLAGING.2004.08.011 Available at: 15718044
Kim, B. et al. (2011) ‘Uridine 5’-diphosphate induces chemokine expression in microglia and astrocytes through activation of the P2Y6 receptor’, Journal of immunology (Baltimore, Md. : 1950), 186(6), pp. 3701–3709. Available at: https://doi.org/10.4049/JIMMUNOL.1000212.
Kim H.J. Nucleotides released from Aβ1-42 -treated microglial cells increase cell migration and Aβ1-₄₂ uptake through P2Y₂ receptor activation Journal of neurochemistry 121 2 2012 228 238 10.1111/J.1471-4159.2012.07700.X 22353164
Kim J. ‘Abeta40 inhibits amyloid deposition in vivo’, The Journal of neuroscience : the official journal of the Society for Neuroscience 27 3 2007 627 633 10.1523/JNEUROSCI.4849-06.2007 Available at: 17234594
Kin Genomic Analysis of 15 Human Coronaviruses OC43 (HCoV-OC43s) Circulating in France from 2001 to 2013 Reveals a High Intra-Specific Diversity with New Recombinant Genotypes Viruses 7 5 2015 2358 2377 10.3390/V7052358 Available at: 26008694
Klaassens B.L. Cholinergic and serotonergic modulation of resting state functional brain connectivity in Alzheimer’s disease NeuroImage 199 2019 143 152 10.1016/j.neuroimage.2019.05.044 Available at: 31112788
Koizumi S. UDP acting at P2Y6 receptors is a mediator of microglial phagocytosis Nature 446 7139 2007 1091 1095 10.1038/NATURE05704 Available at: 17410128
Kong Q. IL-1β ENHANCES NUCLEOTIDE-INDUCED AND α-SECRETASE-DEPENDENT APP PROCESSING IN RAT PRIMARY CORTICAL NEURONS VIA UPREGULATION OF THE P2Y2 RECEPTOR Journal of neurochemistry 109 5 2009 1300 10.1111/J.1471-4159.2009.06048.X Available at: 19317852
Kong Q. Interleukin-1beta enhances nucleotide-induced and alpha-secretase-dependent amyloid precursor protein processing in rat primary cortical neurons via up-regulation of the P2Y(2) receptor Journal of neurochemistry 109 5 2009 1300 1310 10.1111/J.1471-4159.2009.06048.X Available at: 19317852
Kuo M.F. ‘Focusing effect of acetylcholine on neuroplasticity in the human motor cortex’, The Journal of neuroscience : the official journal of the Society for Neuroscience 27 52 2007 14442 14447 10.1523/JNEUROSCI.4104-07.2007 Available at: 18160652
Laurent C. A2A adenosine receptor deletion is protective in a mouse model of Tauopathy Molecular psychiatry 21 1 2016 97 107 10.1038/MP.2014.151 Available at: 25450226
León-Otegui M. Opposite effects of P2X7 and P2Y2 nucleotide receptors on α-secretase-dependent APP processing in Neuro-2a cells FEBS letters 585 14 2011 2255 2262 10.1016/J.FEBSLET.2011.05.048 Available at: 21651910
Levy-Lahad E. ‘Candidate gene for the chromosome 1 familial Alzheimer’s disease locus’, Science 269 1995 973 977 Available at: 7638622
Lim K.-H. Elevation of ACE2 as a SARS-CoV-2 entry receptor gene expression in Alzheimer’s disease The Journal of Infection 81 3 2020 e33 Available at:
Lin P.Y. Omega-3 fatty acids and blood-based biomarkers in Alzheimer’s disease and mild cognitive impairment: A randomized placebo-controlled trial Brain, behavior, and immunity 99 2022 289 298 10.1016/J.BBI.2021.10.014 Available at: 34755655
Macdonald R. Mitochondrial abnormalities in Parkinson’s disease and Alzheimer’s disease: can mitochondria be targeted therapeutically? Biochemical Society transactions 46 4 2018 891 909 10.1042/BST20170501 Available at: 30026371
MacHado A. ‘The cholinergic system in subtypes of Alzheimer’s disease: an in vivo longitudinal MRI study’, Alzheimer’s research & therapy, 12(1) Available at: 2020 10.1186/S13195-020-00620-7
Mantzavinos and Alexiou ‘Biomarkers for Alzheimer’s Disease Diagnosis’, Current Alzheimer research, 14(11) Available at: 2017 10.2174/1567205014666170203125942
Marin, C. et al. (2018) ‘Olfactory Dysfunction in Neurodegenerative Diseases’, Current Allergy and Asthma Reports 2018 18:8, 18(8), pp. 1–19. Available at: https://doi.org/10.1007/S11882-018-0796-4.
McGonagle D. The Role of Cytokines including Interleukin-6 in COVID-19 induced Pneumonia and Macrophage Activation Syndrome-Like Disease Autoimmunity Reviews 19 6 2020 102537 10.1016/J.AUTREV.2020.102537 Available at:
McLarnon J.G. Upregulated expression of purinergic P2X(7) receptor in Alzheimer disease and amyloid-beta peptide-treated microglia and in peptide-injected rat hippocampus Journal of neuropathology and experimental neurology 65 11 2006 1090 1097 10.1097/01.JNEN.0000240470.97295.D3 Available at: 17086106
Mendiola, A.S. and Cardona, A.E. (2018) ‘The IL-1β phenomena in neuroinflammatory diseases’, Journal of neural transmission (Vienna, Austria : 1996), 125(5), p. 781. Available at: https://doi.org/10.1007/S00702-017-1732-9.
Mishra S.K. Extracellular nucleotide signaling in adult neural stem cells: synergism with growth factor-mediated cellular proliferation Development (Cambridge, England) 133 4 2006 675 684 10.1242/DEV.02233 Available at: 16436623
Moein S.T. Smell dysfunction: a biomarker for COVID-19 International Forum of Allergy & Rhinology 10 8 2020 944 950 10.1002/ALR.22587 Available at: 32301284
Morioka N. The activation of P2Y6 receptor in cultured spinal microglia induces the production of CCL2 through the MAP kinases-NF-κB pathway Neuropharmacology 75 2013 116 125 10.1016/J.NEUROPHARM.2013.07.017 Available at: 23916475
Munoz, F.M. et al. (2017) ‘Neuronal P2X7 receptor-induced reactive oxygen species production contributes to nociceptive behavior in mice’, Scientific Reports 2017 7:1, 7(1), pp. 1–12. Available at: https://doi.org/10.1038/s41598-017-03813-7.
Naseri N.N. The complexity of tau in Alzheimer’s disease Neuroscience letters 705 2019 183 10.1016/J.NEULET.2019.04.022 Available at: 31028844
Neddens J. ‘Phosphorylation of different tau sites during progression of Alzheimer’s disease’, Acta Neuropathologica Communications 6 1 2018 52 10.1186/S40478-018-0557-6 Available at:
Neher J.J. Inhibition of UDP/P2Y6 purinergic signaling prevents phagocytosis of viable neurons by activated microglia in vitro and in vivo Glia 62 9 2014 1463 1475 10.1002/GLIA.22693 Available at: 24838858
Ng A. ‘IL-1β, IL-6, TNF- α and CRP in Elderly Patients with Depression or Alzheimer’s disease: Systematic Review and Meta-Analysis’, Scientific reports, 8(1) Available at: 2018 10.1038/S41598-018-30487-6
Nunomura A. Perry G. ‘RNA and Oxidative Stress in Alzheimer’s Disease: Focus on microRNAs’, Oxidative medicine and cellular longevity, 2020 Available at: 2020 10.1155/2020/2638130
Orellana J.A. Shoji K.F. Amyloid β-Induced Death in Neurons Involves Glial and Neuronal Hemichannels The Journal of Neuroscience 31 13 2011 4962 10.1523/JNEUROSCI.6417-10.2011 Available at: 21451035
Orellana J.A. Froger N. ATP and glutamate released via astroglial connexin43 hemichannels mediate neuronal death through activation of pannexin 1 hemichannels Journal of neurochemistry 118 5 2011 826 10.1111/J.1471-4159.2011.07210.X Available at: 21294731
Osugi Y. Fumoto K. Kikuchi A. ‘CKAP4 Regulates Cell Migration via the Interaction with and Recycling of Integrin’, Molecular and Cellular Biology, 39(16) Available at: 2019 10.1128/MCB.00073-19
Parvathenani L.K. P2X7 Mediates Superoxide Production in Primary Microglia and Is Up-regulated in a Transgenic Mouse Model of Alzheimer’s Disease Journal of Biological Chemistry 278 15 2003 13309 13317 10.1074/JBC.M209478200 Available at: 12551918
de Paula V.J.R. Guimarães F.M. Forlenza O.V. Papel da proteína Tau na fisiopatologia da demência frontotemporal Archives of Clinical Psychiatry (São Paulo) 36 5 2009 197 202 10.1590/S0101-60832009000500004 Available at:
Peña-Bautista C. ‘Oxidative Damage of DNA as Early Marker of Alzheimer’s Disease’, International journal of molecular sciences, 20(24) Available at: 2019 10.3390/IJMS20246136
Pépin G. Perron M.P. Provost P. Regulation of human Dicer by the resident ER membrane protein CLIMP-63 Nucleic Acids Research 40 22 2012 11603 10.1093/NAR/GKS903 Available at: 23047949
Peterson T.S. P2Y2 nucleotide receptor-mediated responses in brain cells Molecular neurobiology 41 2–3 2010 356 366 10.1007/S12035-010-8115-7 Available at: 20387013
Petrone L. A whole blood test to measure SARS-CoV-2-specific response in COVID-19 patients Clinical microbiology and infection : the official publication of the European Society of Clinical Microbiology and Infectious Diseases 27 2 2021 286.e7 286.e13 10.1016/J.CMI.2020.09.051 Available at:
Picciotto H. Mineur, Acetylcholine as a neuromodulator: cholinergic signaling shapes nervous system function and behavior Neuron 76 1 2012 116 129 10.1016/J.NEURON.2012.08.036 Available at: 23040810
Picón-Pagès P. Garcia-Buendia J. Muñoz F.J. Functions and dysfunctions of nitric oxide in brain Biochimica et Biophysica Acta (BBA) - Molecular Basis of Disease 1865 8 2019 1949 1967 10.1016/J.BBADIS.2018.11.007 Available at: 30500433
Pinheiro L. Faustino C. Therapeutic Strategies Targeting Amyloid-β in Alzheimer’s Disease Current Alzheimer research 16 5 2019 418 452 10.2174/1567205016666190321163438 Available at: 30907320
Poyiadji N. COVID-19–associated Acute Hemorrhagic Necrotizing Encephalopathy: CT and MRI Features Radiology 296 2 2020 E119 E120 10.1148/RADIOL.2020201187 Available at: 32228363
Puigdellívol M. The microglial P2Y6 receptor mediates neuronal loss and memory deficits in neurodegeneration Cell Reports 37 13 2021 110148 10.1016/J.CELREP.2021.110148 Available at:
Rahman A. The Role of Adenosine in Alzheimer’s Disease Current Neuropharmacology 7 3 2009 207 10.2174/157015909789152119 Available at: 20190962
Rahman M.A. Neurobiochemical Cross-talk Between COVID-19 and Alzheimer’s Disease Molecular Neurobiology 58 3 2021 1017 10.1007/S12035-020-02177-W Available at: 33078369
Rasch B.H. Born J. Gais S. Combined blockade of cholinergic receptors shifts the brain from stimulus encoding to memory consolidation Journal of cognitive neuroscience 18 5 2006 793 802 10.1162/JOCN.2006.18.5.793 Available at: 16768378
Reichenbach N. P2Y1 receptor blockade normalizes network dysfunction and cognition in an Alzheimer’s disease model The Journal of experimental medicine 215 6 2018 1649 1663 10.1084/JEM.20171487 Available at: 29724785
Renard P.Y. Jean L. Probing the cholinergic system to understand neurodegenerative diseases Future Medicinal Chemistry. Future Medicine Ltd. 2017 131 133 10.4155/fmc-2016-0213 Available at:
Reyfman P.A. Single-Cell Transcriptomic Analysis of Human Lung Provides Insights into the Pathobiology of Pulmonary Fibrosis American Journal of Respiratory and Critical Care Medicine 199 12 2019 1517 10.1164/RCCM.201712-2410OC Available at: 30554520
Ribeiro J.A. Purinergic regulation of acetylcholine release Progress in brain research 109 1996 231 241 10.1016/S0079-6123(08)62107-X Available at: 9009712
Richter N. Effect of cholinergic treatment depends on cholinergic integrity in early Alzheimer’s disease Brain 141 3 2018 903 915 10.1093/BRAIN/AWX356 Available at: 29309600
Rosengarten B. Acetylcholine esterase inhibitor donepezil improves dynamic cerebrovascular regulation in Alzheimer patients Journal of neurology 253 1 2006 58 64 10.1007/S00415-005-0926-5 Available at: 16096820
Rothaug M. Becker-Pauly C. Rose-John S. The role of interleukin-6 signaling in nervous tissue Biochimica et biophysica acta 1863 6 Pt A 2016 1218 1227 10.1016/J.BBAMCR.2016.03.018 Available at: 27016501
Ryu J.K. McLarnon J.G. Block of purinergic P2X(7) receptor is neuroprotective in an animal model of Alzheimer’s disease Neuroreport 19 17 2008 1715 1719 10.1097/WNR.0B013E3283179333 Available at: 18852683
Salahuddin Structure of amyloid oligomers and their mechanisms of toxicities: Targeting amyloid oligomers using novel therapeutic approaches European journal of medicinal chemistry 114 2016 41 58 10.1016/J.EJMECH.2016.02.065 Available at: 26974374
Sangaleti C.T. ‘The Cholinergic Drug Galantamine Alleviates Oxidative Stress Alongside Anti-inflammatory and Cardio-Metabolic Effects in Subjects With the Metabolic Syndrome in a Randomized Trial’, Frontiers in immunology, 12 Available at: 2021 10.3389/FIMMU.2021.613979
Sanz, J.M. et al. (2009) ‘Activation of microglia by amyloid {beta} requires P2X7 receptor expression’, Journal of immunology (Baltimore, Md. : 1950), 182(7), pp. 4378–4385. Available at: https://doi.org/10.4049/JIMMUNOL.0803612.
Selkoe D.J. ‘Alzheimer’s disease: genotypes, phenotypes, and treatments’, Science 275 1997 630 631 Available at: 9019820
Selkoe D.J. Hardy J. The amyloid hypothesis of Alzheimer’s disease at 25 years Available at: EMBO molecular medicine 8 6 2016 595 608 10.15252/EMMM.201606210 27025652
Sensi S.L. Copper and Zinc Dysregulation in Alzheimer’s Disease Trends in Pharmacological Sciences 39 12 2018 1049 1063 10.1016/J.TIPS.2018.10.001 Available at: 30352697
Sereniki, A. et al. (2008) ‘A doença de Alzheimer: aspectos fisiopatológicos e farmacológicos Alzheimer’s disease: pathophysiological and pharmacological features’.
Serrano-Pozo Plaque-Associated Local Toxicity Increases over the Clinical Course of Alzheimer Disease The American journal of pathology 186 2 2016 375 384 10.1016/J.AJPATH.2015.10.010 Available at: 26687817
Shao P. Wang, Genetics of Alzheimer’s disease: From pathogenesis to clinical usage Journal of clinical neuroscience : official journal of the Neurosurgical Society of Australasia 45 2017 1 8 10.1016/J.JOCN.2017.06.074 28869135
Sharma K. ‘Cholinesterase inhibitors as Alzheimer’s therapeutics (Review)’, Molecular Medicine Reports 2019 Spandidos Publications 1479 1487 Available at:
Sherrington R. Cloning of a gene bearing missense mutations in early-onset familial Alzheimer’s disease Nature 375 6534 1995 754 760 10.1038/375754A0 Available at: 7596406
Shi Y. ApoE4 markedly exacerbates tau-mediated neurodegeneration in a mouse model of tauopathy Nature 549 7673 2017 523 527 10.1038/nature24016 Available at: 28959956
Shigemoto Y. Association of deposition of tau and amyloid-β proteins with structural connectivity changes in cognitively normal older adults and Alzheimer’s disease spectrum patients Brain and Behavior 8 12 2018 1145 10.1002/BRB3.1145 Available at:
Shinto L. A randomized placebo-controlled pilot trial of omega-3 fatty acids and alpha lipoic acid in Alzheimer’s disease Journal of Alzheimer’s disease : JAD 38 1 2014 111 120 10.3233/JAD-130722 Available at: 24077434
Shu and McCauley ‘GISAID: Global initiative on sharing all influenza data - from vision to reality’, Euro surveillance : bulletin Europeen sur les maladies transmissibles = European communicable disease bulletin, 22(13) Available at: 2017 10.2807/1560-7917.ES.2017.22.13.30494
Simões J.L.B. Bagatini M.D. Purinergic Signaling of ATP in COVID-19 Associated Guillain-Barré Syndrome Journal of Neuroimmune Pharmacology 16 1 2021 48 58 10.1007/S11481-020-09980-1/FIGURES/4 Available at: 33462776
Smith D.G. Cappai R. Barnham K.J. ‘The redox chemistry of the Alzheimer’s disease amyloid β peptide’, Biochimica et Biophysica Acta (BBA) - Biomembranes 1768 8 2007 1976 1990 10.1016/J.BBAMEM.2007.02.002 Available at:
Snigdha, S. et al. (2012) ‘Caspase-3 activation as a bifurcation point between plasticity and cell death’, Neuroscience bulletin, 28(1), pp. 14–24. Available at: https://doi.org/10.1007/S12264-012-1057-5.
Stanciu ‘Alzheimer’s Disease Pharmacotherapy in Relation to Cholinergic System Involvement’, Biomolecules, 10(1) Available at: 2019 10.3390/BIOM10010040
Stanciu G.D. ‘Alzheimer’s Disease Pharmacotherapy in Relation to Cholinergic System Involvement’, Biomolecules, 10(1) Available at: 2020 10.3390/BIOM10010040
Steardo ‘Neuroinfection may contribute to pathophysiology and clinical manifestations of COVID-19’, Acta physiologica (Oxford, England), 229(3) Available at: 2020 10.1111/APHA.13473
Su, S. et al. (2016) ‘Epidemiology, Genetic Recombination, and Pathogenesis of Coronaviruses’, Trends in Microbiology, 24(6), p. 490. Available at: https://doi.org/10.1016/J.TIM.2016.03.003.
Suurväli, J. et al. (2017) ‘P2X4: A fast and sensitive purinergic receptor’, Biomedical Journal, 40(5), p. 245. Available at: https://doi.org/10.1016/J.BJ.2017.06.010.
Tajes M. ‘Nitro-oxidative stress after neuronal ischemia induces protein nitrotyrosination and cell death’, Oxidative Medicine and Cellular Longevity [Preprint] Available at: 2013 10.1155/2013/826143
Takousis, P. et al. (2019) ‘Differential expression of microRNAs in Alzheimer’s disease brain, blood, and cerebrospinal fluid’, Alzheimer’s & dementia : the journal of the Alzheimer’s Association, 15(11), pp. 1468–1477. Available at: https://doi.org/10.1016/J.JALZ.2019.06.4952.
Tanaka, T., Narazaki, M. and Kishimoto, T. (2014) ‘IL-6 in Inflammation, Immunity, and Disease’, Cold Spring Harbor Perspectives in Biology, 6(10), pp. 16295–16296. Available at: https://doi.org/10.1101/CSHPERSPECT.A016295.
Tao, C.-C. et al. (2020) ‘Galectin-3 promotes Aβ oligomerization and Aβ toxicity in a mouse model of Alzheimer’s disease’, Cell Death and Differentiation, 27(1), p. 192. Available at: https://doi.org/10.1038/S41418-019-0348-Z.
Thawkar, B.S. and Kaur, G. (2019) ‘Inhibitors of NF-κB and P2X7/NLRP3/Caspase 1 pathway in microglia: Novel therapeutic opportunities in neuroinflammation induced early-stage Alzheimer’s disease’, Journal of Neuroimmunology, 326, pp. 62–74. Available at: https://doi.org/10.1016/J.JNEUROIM.2018.11.010.
Tiwari, S. et al. (2019) ‘Alzheimer’s disease: pathogenesis, diagnostics, and therapeutics’, International Journal of Nanomedicine, 14, p. 5541. Available at: https://doi.org/10.2147/IJN.S200490.
Tönnies, E. and Trushina, E. (2017) ‘Oxidative Stress, Synaptic Dysfunction, and Alzheimer’s Disease’, Journal of Alzheimer’s Disease, 57(4), p. 1105. Available at: https://doi.org/10.3233/JAD-161088.
Varma, R. et al. (2009) ‘Amyloid-β induces a caspase-mediated cleavage of P2X4 to promote purinotoxicity’, NeuroMolecular Medicine, 11(2), pp. 63–75. Available at: https://doi.org/10.1007/S12017-009-8073-2.
Vassar, R. (2007) ‘Caspase-3 cleavage of GGA3 stabilizes BACE: implications for Alzheimer’s disease’, Neuron, 54(5), pp. 671–673. Available at: https://doi.org/10.1016/J.NEURON.2007.05.018.
Velavan, T.P. and Meyer, C.G. (2020) ‘The COVID‐19 epidemic’, Tropical Medicine & International Health, 25(3), p. 278. Available at: https://doi.org/10.1111/TMI.13383.
Ventura, A.L.M. et al. (2010) ‘Sistema colinérgico: revisitando receptores, regulação e a relação com a doença de Alzheimer, esquizofrenia, epilepsia e tabagismo’, Archives of Clinical Psychiatry (São Paulo), 37(2), pp. 66–72. Available at: https://doi.org/10.1590/S0101-60832010000200007.
Di Virgilio, F. et al. (2009) ‘Purinergic signalling in inflammation of the central nervous system’, Trends in neurosciences, 32(2), pp. 79–87. Available at: https://doi.org/10.1016/J.TINS.2008.11.003.
Volpicelli-Daley, L.A. et al. (2003) ‘Altered striatal function and muscarinic cholinergic receptors in acetylcholinesterase knockout mice’, Molecular pharmacology, 64(6), pp. 1309–1316. Available at: https://doi.org/10.1124/MOL.64.6.1309.
Walker, L.C. et al. (2005) ‘Emerging prospects for the disease-modifying treatment of Alzheimer’s disease’, Biochemical pharmacology, 69(7), pp. 1001–1008. Available at: https://doi.org/10.1016/J.BCP.2004.12.015.
Wang, T. et al. (2014) ‘Chronic caffeine treatment protects against experimental autoimmune encephalomyelitis in mice: therapeutic window and receptor subtype mechanism’, Neuropharmacology, 86, pp. 203–211. Available at: https://doi.org/10.1016/J.NEUROPHARM.2014.06.029.
Wang, W.Y. et al. (2015) ‘Role of pro-inflammatory cytokines released from microglia in Alzheimer’s disease’, Annals of Translational Medicine, 3(10), p. 136. Available at: https://doi.org/10.3978/J.ISSN.2305-5839.2015.03.49.
Weimann, A., Simonsen, A.H. and Poulsen, H.E. (2018) ‘Measurement of 8-oxo-7,8-dihydro-2’-deoxyguanosine and 8-oxo-7,8-dihydro-guanosine in cerebrospinal fluid by ultra performance liquid chromatography-tandem mass spectrometry’, Journal of chromatography. B, Analytical technologies in the biomedical and life sciences, 1073, pp. 110–117. Available at: https://doi.org/10.1016/J.JCHROMB.2017.12.011.
Weiss and Navas-Martin (2005) ‘Coronavirus pathogenesis and the emerging pathogen severe acute respiratory syndrome coronavirus’, Microbiology and molecular biology reviews : MMBR, 69(4), pp. 635–664. Available at: https://doi.org/10.1128/MMBR.69.4.635-664.2005.
Wesseling, H. et al. (2020) ‘Tau PTM Profiles Identify Patient Heterogeneity and Stages of Alzheimer’s Disease’, Cell, 183(6), pp. 1699-1713.e13. Available at: https://doi.org/10.1016/J.CELL.2020.10.029.
Wessels, A.M. et al. (2020) ‘Cognitive outcomes in trials of two BACE inhibitors in Alzheimer’s disease’, Alzheimer’s & dementia : the journal of the Alzheimer’s Association, 16(11), pp. 1483–1492. Available at: https://doi.org/10.1002/ALZ.12164.
Whitwell, J.L. et al. (2021) ‘Relationship of APOE, age at onset, amyloid and clinical phenotype in Alzheimer disease’, Neurobiology of Aging, 108, pp. 90–98. Available at: https://doi.org/10.1016/J.NEUROBIOLAGING.2021.08.012.
Wolfe, M.S. (2019) ‘Dysfunctional γ-secretase in familial Alzheimer’s disease’, Neurochemical research, 44(1), p. 5. Available at: https://doi.org/10.1007/S11064-018-2511-1.
Woods, L.T. et al. (2016) ‘Purinergic receptors as potential therapeutic targets in Alzheimer’s disease’, Neuropharmacology, 104, pp. 169–179. Available at: https://doi.org/10.1016/J.NEUROPHARM.2015.10.031.
Wrapp et al. (2020) ‘Cryo-EM structure of the 2019-nCoV spike in the prefusion conformation’, Science (New York, N.Y.), 367(6483), pp. 1255–1260. Available at: https://doi.org/10.1126/SCIENCE.ABB2507.
Wu, Y. et al. (2020) ‘Nervous system involvement after infection with COVID-19 and other coronaviruses’, Brain, Behavior, and Immunity, 87, p. 18. Available at: https://doi.org/10.1016/J.BBI.2020.03.031.
Xia X. Wang Y. Zheng J. ‘COVID-19 and Alzheimer’s disease: how one crisis worsens the other’, Translational neurodegeneration, 10(1) Available at: 2021 10.1186/S40035-021-00237-2
Xie, Wu and Zhao (2020) ‘Multifunctional roles of zinc in Alzheimer’s disease’, Neurotoxicology, 80, pp. 112–123. Available at: https://doi.org/10.1016/J.NEURO.2020.07.003.
Zhao Y. ‘Longitudinal COVID-19 profiling associates IL-1RA and IL-10 with disease severity and RANTES with mild disease’, JCI insight, 5(13) Available at: 2020 10.1172/JCI.INSIGHT.139834
Zhao, Yu et al. (2020) ‘Single-cell RNA expression profiling of ACE2, the putative receptor of Wuhan 2019-nCov’, bioRxiv, p. 2020.01.26.919985. Available at: https://doi.org/10.1101/2020.01.26.919985.
Zhou et al. (2020) ‘A pneumonia outbreak associated with a new coronavirus of probable bat origin’, Nature, 579(7798), pp. 270–273. Available at: https://doi.org/10.1038/S41586-020-2012-7.
Zhou, Z. et al. (2018) ‘Mitochondrial Metabolism in Major Neurological Diseases’, Cells, 7(12), p. 229. Available at: https://doi.org/10.3390/CELLS7120229.
Zotova, E. et al. (2013) ‘Inflammatory components in human Alzheimer’s disease and after active amyloid-β42 immunization’, Brain : a journal of neurology, 136(Pt 9), pp. 2677–2696. Available at: https://doi.org/10.1093/BRAIN/AWT210.
| 0 | PMC9746135 | NO-CC CODE | 2022-12-15 00:03:57 | no | Neuroscience. 2022 Dec 13; doi: 10.1016/j.neuroscience.2022.12.007 | utf-8 | Neuroscience | 2,022 | 10.1016/j.neuroscience.2022.12.007 | oa_other |
==== Front
Am J Ophthalmol
Am J Ophthalmol
American Journal of Ophthalmology
0002-9394
1879-1891
Elsevier Inc.
S0002-9394(20)30331-7
10.1016/j.ajo.2020.06.035
Correspondence
Comment on: Is this a 737 Max Moment for Brolucizumab
Kayath Marcia
Sauer Dirk
Basel, Switzerland
15 9 2020
3 2021
15 9 2020
223 446446
© 2020 Elsevier Inc. All rights reserved.
2020
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcEditor:
We read with interest the editorial titled “Is this a 737 Max Moment for Brolucizumab.”1 At Novartis, providing safe and effective treatments for patients is our highest priority. Working closely with health authorities around the world, including FDA, we continuously monitor the benefit-risk profile of our medicines. Although other anti–vascular endothelial growth factor (anti-VEGF) agents are available, there are current unmet needs with neovascular AMD (nAMD) treatment that we believe brolucizumab addresses. Moreover, we believe the choice of treatment should ultimately be left to individual treating physicians and their patients, after appropriate evaluation of the benefit-risk profile of the product.
As a greater number of patients were exposed to brolucizumab following FDA approval, Novartis received reports of retinal vasculitis, including retinal occlusive vasculitis. Novartis initiated its own internal review of these postmarketing safety case reports, including the establishment of an external safety review committee (SRC) to provide an independent review of these cases and compare them to events seen in the brolucizumab Phase III trials. Using the terminology defined by the SRC, Novartis concluded a confirmed safety signal of rare adverse events termed “retinal vasculitis” and/or “retinal vascular occlusion” that may result in severe vision loss.
Additionally, Novartis has established a fully dedicated research, drug development, and medical task force who are working with top external global specialists with the goal of examining the following key questions: (1) root cause; (2) identifying at-risk patient characteristics; (3) risk mitigation strategies; and (4) treatment algorithms for these rare events.
Since the launch of brolucizumab, transparency and communication with the retina community have been first and foremost in our minds. In addition to the commissioning of the SRC and the task force, Novartis worked closely with the American Society of Retina Specialists (ASRS) ReST Committee to provide access to postmarketing data to ensure physicians and patients fully understood the risks and benefits associated with brolucizumab. We have also created a global safety website, brolucizumab.info, to provide the latest information and guidance. Other actions included (1) working with health authorities to update the prescribing information worldwide; (2) informing investigators of ongoing clinical trials and asking them to reconsent patients; (3) amending the protocols, informed consent forms, and investigator brochures of all Novartis-sponsored trials; and (4) informing all physicians who request brolucizumab through our Managed Access Program.
Physicians are advised to carefully monitor each patient treated with brolucizumab for evidence of inflammation or other adverse events. It is advised they follow recommendations set forth in/by the brolucizumab label, and specialty societies and organizations, such as the ASRS, regarding management and timing of repeated administrations of anti-VEGF agents. Brolucizumab is contraindicated in patients with ocular or periocular infections, active intraocular inflammation, or known hypersensitivity to brolucizumab.
Brolucizumab represents an important treatment option for patients with nAMD. At Novartis, we support individual physicians, who we believe, whether or not they choose to use brolucizumab, are able to make the best treatment choices for their patients.
Funding/Support: This study received no funding. Financial Disclosures: Marcia Kayath is an employee of Novartis Pharmaceuticals Corporation. Dirk Sauer is an employee of Novartis AG. All authors attest that they meet the current ICMJE criteria for authorship.
==== Refs
Reference
1 Rosenfeld P.J. Browning D.J. Is this a 737 Max moment for brolucizumab? Am J Ophthalmol 216 2020 A7 A8 https://www.ajo.com/article/S0002-9394(20)30242-7/fulltext 32505363
| 32948293 | PMC9746201 | NO-CC CODE | 2022-12-15 00:03:35 | no | Am J Ophthalmol. 2021 Mar 15; 223:446 | utf-8 | Am J Ophthalmol | 2,020 | 10.1016/j.ajo.2020.06.035 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Published by Elsevier Inc.
S0002-9343(21)00166-2
10.1016/j.amjmed.2021.02.010
Commentary
Resilience for Frontline Health Care Workers: Evidence-Based Recommendations
Southwick Steven M. MD a⁎
Charney Dennis S. MD b
a Department of Psychiatry, Yale University School of Medicine, New Haven, Conn
b Department of Psychiatry, Icahn School of Medicine at Mount Sinai, New York, NY
⁎ Requests for reprints should be addressed to Steven Southwick, MD, 27 Castle Rock, Branford, CT 06371.
25 3 2021
7 2021
25 3 2021
134 7 829830
© 2021 Published by Elsevier Inc.
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcIn the battle against COVID-19, frontline health care workers are dealing with enormous levels of stress and uncertainty. They are facing fears about infecting their patients, colleagues, and family, being re-assigned to practice in settings where they feel insufficiently prepared, having to make life-and-death decisions based on limited availability of resources, knowing that some of their colleagues, or they themselves, will become ill and perhaps die, and watching patients die alone without being able to say good-bye to their loved ones.
In this Commentary, we share evidence-based strategies known to enhance resilience and decrease the likelihood of developing burnout, depression, and post-traumatic stress disorder during times of adversity. These strategies come from research with stress-hardy civilian men and women, including physicians, as well as military personnel, including former prisoners of war and Special Forces instructors. The strategies and recommendations are strongly supported by a large body of peer-reviewed psychosocial and neurobiological research.1
Face Fear
Primary drivers of fear include loss of a sense of control and lack of predictability, so it is natural to feel afraid in the current medical climate. Moderate levels of anxiety and fear enhance performance. However, when left unattended, fear can escalate into panic. When this happens, the prefrontal cortex becomes flooded with catecholamines and loses its capacity to inhibit the limbic system's fight or flight response.2 While this may be advantageous during acute danger, it can be harmful in medical settings where rational decision-making is required. Methods to help maintain fear and the fight or flight response within an adaptive range include: acknowledge and monitor your reactions to fear; obtain accurate information about what you fear from reliable sources; manage inflow of negative information whenever possible; follow evidence-based procedures to protect against the virus; share your concerns and fears with trusted colleagues and support one another; practice relaxation techniques such as deep belly breathing and mindfulness meditation even for brief periods; try to reframe fearful situations as a challenge in addition to being a threat; train for likely scenarios whenever possible. Courage is not the absence of fear but the ability to act despite the fear. If you currently are a health care worker on the front lines, you are acting courageously, even if you feel afraid.
Realistic Optimism
Realistic optimism and positive emotions repeatedly have been associated with resilience, and good mental and physical health. Methods to increase realistic optimism include: train sufficiently to feel prepared for the specific challenges that you will face; cognitively reappraise overly negative thoughts by challenging them and searching for more constructive ways to view the situation; embrace humor when appropriate; increase behaviors commonly associated with positive emotions because behaviors affect mood and are under more direct control of the will than emotions; and try to associate with optimistic colleagues because emotions can be contagious.
Social Support
Positive social support consistently is among the strongest psychosocial predictors of resilience and well-being. On the other hand, social isolation negatively impacts mental and physical health, with effects on longevity equal to those of cigarette smoking, obesity, and sedentary lifestyle.3 With the current need for social distancing, it is critical for health care workers to find creative solutions to stay socially connected, such as virtual group meetings where physicians and health care teams use video conferencing to discuss what they are experiencing and suggestions for effective coping.
For leaders of frontline caregivers in the battle against COVID-19, it is imperative to create an atmosphere of camaraderie, respect, and psychological safety where fear, shame, and guilt are acknowledged and discussed; where responsibilities and resources are shared; where peer and mentor support are fostered; and where team members feel valued and understood.4 Navy Seals and Special Forces soldiers often attribute their own courage and resilience to the power of team members who “have each other's back” and will even risk their life for one another. The message from healthcare leaders should be clear: “Team, Team, Team. You are your brother and sister's keeper. We are all in this together, fighting for a common noble cause. It is a privilege to be working alongside such a remarkable group of colleagues.”
Businesses, communities, and government agencies can also support frontline workers by providing medical supplies (eg, personal protective equipment, ventilators), financial resources (eg, hazard pay, free lunches), and emotional support. Early in the pandemic, at 7 PM every evening, New Yorkers would scream, shout, clap, bang pots and pans, and play musical instruments for 2 minutes as a way to honor and say “thank you” to the thousands of frontline workers who ‘walk into the fire’ and risk their lives every day.
Coping
Resilience is associated with active and flexible coping (eg, gathering information, acquiring skills, problem solving, seeking out social support) rather than passive coping (denial there is a problem, avoiding, procrastinating or withdrawing, using substances, repetitive negative venting, blaming someone else). Active coping includes self-care through exercise, good nutrition, relaxation strategies, and restorative sleep. Strategies to modulate sympathetic and parasympathetic tone, such as mindful meditation, deep belly breathing, and yoga are also effective, because regularly bringing the stress response back to baseline is an essential component of resilience.
Cognitive Strategies
Acceptance involves acknowledging the reality of a situation even if painful or frightening, accepting what you cannot change, abandoning goals that are no longer feasible, and intentionally redirecting efforts toward goals that can be achieved. Acceptance is not a passive giving up but, instead, a conscious decision not to waste effort on pointless attempts to change the unchangeable. Acceptance forms the basis of the well-known Serenity prayer.
Positive reappraisal refers to the intentional effort to reassess the meaning of an event from negative to more positive, leading to more adaptive emotional and behavioral responses, partly by increasing activation of the executive region of the brain (prefrontal cortex) and attenuating limbic system alarm reactions. You might ask, “Is there a more constructive way I can think about this situation?” Trauma may force us to learn something new or possibly even grow as a person.
Religious and Spiritual Practices
Religious and spiritual practices have repeatedly been associated with physical and emotional well-being, partly through a focus on altruism, gratitude, supporting one another, prayer and meditation, meaning and purpose, and connection with a higher power or with something far larger than the self. Faith-based beliefs and practices are especially powerful during times of crisis and provide great strength for many health care workers and their families.
Meaning and Purpose
Having a clear and valued purpose, and committing to a worthy mission, can markedly strengthen one's resilience, even if that mission entails danger and suffering. In the words of psychiatrist and Holocaust survivor Viktor Frankl, “There is nothing in the world I would venture to say that would so effectively help one to survive even the worst conditions as the knowledge that there is a meaning to life.”5 This is especially true at a time of great need, like the present pandemic, and when meaning is connected to a calling, like medicine.
It is our sincere hope that one or more of these strategies will prove helpful to the courageous health care workers who are fighting in our hospitals to save lives.
Funding: None.
Conflicts of Interest: Both authors receive an honorarium from Cambridge University Press for the book Resilience: The Science of Mastering Life's Greatest Challenges, 2018.
Authorship: Both authors have participated in the preparation of the manuscript.
==== Refs
References
1 Southwick SM Charney DS Resilience: The Science of Mastering Life's Greatest Challenges 2nd ed 2018 Cambridge University Press Cambridge, England
2 Arnsten AFT Stress signaling pathways that impair prefrontal cortex structure and function Nature 10 6 2009 410 422
3 Holt-Lunstad J Smith TB Baker M Harris T Stephenson D Loneliness and social isolation as risk factors for mortality: a meta-analytic review Perspect Psychol Sci 10 2 2015 227 237 25910392
4 Southwick SM Southwick FS The loss of a sense of social connectedness as a major contributor to physician burnout: applying organizational and teamwork principles for prevention and recovery JAMA Psychiatry 77 5 2020 449 450 32074385
5 Frankl VE Man's Search for Meaning: An Introduction to Logotherapy 1984 Pocket Books, Simon and Schuster New York
| 33773971 | PMC9746223 | NO-CC CODE | 2022-12-15 23:21:55 | no | Am J Med. 2021 Jul 25; 134(7):829-830 | utf-8 | Am J Med | 2,021 | 10.1016/j.amjmed.2021.02.010 | oa_other |
==== Front
Accid Anal Prev
Accid Anal Prev
Accident; Analysis and Prevention
0001-4575
1879-2057
Elsevier Ltd.
S0001-4575(21)00256-6
10.1016/j.aap.2021.106225
106225
Article
Speeding through the pandemic: Perceptual and psychological factors associated with speeding during the COVID-19 stay-at-home period
Tucker A. a⁎
Marsh K.L. b
a Conecticut Transportation Safety Research Center, Storrs Mansfield, CT, USA
b University of Connecticut Psychological Sciences Department, USA
⁎ Corresponding author at: 270 Middle Turnpike, Unit 5202 Longley Building, Storrs Mansfield, CT 06269, USA.
12 6 2021
9 2021
12 6 2021
159 106225106225
17 8 2020
7 5 2021
1 6 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
During the COVID-19 stay-at-home period there were observed increases in both the percentage of cars engaged in extreme speeding, and the percentage of cars traveling below the speed limit. These changes have been attributed to unusually low traffic volume during the stay-at-home period. We develop a novel theoretical account, based on existing empirical research, of perceptual and psychological processes that may account for changes in speeding behavior under low traffic volume conditions. These include impaired ability to accurately perceive and control speed due to change in visual information, decreased salience of certain norms about socially appropriate speeds, lower perceived risk of speeding, and increased boredom leading to risk-taking behaviors. Further, we consider that individual attitude functions may account for the observed split in speeding behavior.
Keywords
COVID-19
Speeding
Motivation
Perception
==== Body
pmc1 Introduction
Motor vehicle crashes and fatalities typically decrease during periods of economic hardship (Afroz et al., 2012, Blower et al., 2019). However, the COVID-19 pandemic and associated stay-at-home order has been associated with increases in the incidence of certain crash types, most notably fatal, single car crashes (Doucette et al., 2021). Further, changes in speeding behavior have been observed: the COVID-19 stay-at-home order in Connecticut was associated with increases in the percentage of cars engaged in extreme speeding (i.e. more than 20mph over the posted speed limit) and the percentage of cars traveling below the speed limit, but a decrease in the percentage of cars traveling fewer than 20mph over the speed limit (e.g., 66-84mph in a 65mph zone) (Shapiro et al., 2021). These changes are attributable to the decrease in traffic volume during the COVID-19 stay-at-home period, rather than seasonality; early reports indicate a 40 to 60% reduction in VMT (vehicle miles traveled) associated with the COVID-19 pandemic (Agrawal et al., 2020, Doucette et al., 2021, Shapiro et al., 2021).
Changes in observed speeding pose a unique challenge: it appears that a split in the driving population exists, such that some drivers became more cautious during the stay-at-home period, accounting for the increase in the percentage of cars traveling below the speed limit, whereas others drove more recklessly, accounting for the increase in the percentage of cars engaged in extreme speeding. However, it is not well understood how or why the behavior of any individual driver might change under the low-volume conditions associated with the COVID-19 stay-at-home order. Provisional answers may be gleaned from research on the perceptual and psychological factors that influence driver behavior under normal driving conditions. It is commonly claimed that 90% of the information drivers use to control vehicle speed and heading is visual; while the specific “90%” claim is difficult to verify, is it widely accepted that drivers guide their behavior through the detection of visual information (Sivak, 1996). Thus, it may be possible to gain insight into expected or observed changes in driver behavior during a pandemic by assessing changes in the visual information available to drivers. To this end, we draw on seminal work by Gibson and Crooks, 1938, Gibson, 1966; Gibson’s (1979) later theories of ecological optics and affordances, and more recent work in the Gibsonian tradition. Foundational to the Gibsonian approach is the notion that organisms directly perceive and act on environmental affordances - loosely, opportunities for action that exist by virtue of specific organism-environment relations – that are specified by visual and other sensory information (Gibson, 1966). Because the COVID-19 pandemic has been associated with steep declines in traffic volume (Doucette et al., 2021, Shapiro et al., 2021), we primarily consider the impacts of decreased traffic volume on the available visual information, and relatedly, the affordances available to drivers. With atypically low traffic volume, the opportunities for action available drivers are expanded; how drivers behave in light of these changed conditions may depend on a number of social and motivational factors. On the basis of existing research, it appears that changes in the available visual information concomitant with lower traffic volume may be associated with impairments in drivers’ ability to perceive and control their speed, absence of information about socially appropriate speeds (i.e., how others are driving), lower perceived risk of driving faster, and increased boredom leading to risk-taking behaviors.
2 Perceptual factors
Psychological models of driver behavior have historically been influenced by, and reciprocally influence, models of human locomotion (Gibson and Crooks, 1938). Whereas locomotion is guided by visual and haptic (or proprioceptive) information, driving relies almost entirely on visual information. This is of particular importance as it is known that, during locomotion, haptic or proprioceptive information can play a role in the perception of speed and distance traveled (Turvey et al., 2009, Turvey et al., 2012, Chrastil and Warren, 2014). Lacking these, drivers are almost entirely dependent on visual information for speed regulation; vestibular information (i.e. that from the inner ear) plays only a minor role in perception of speed during driving (Lappe et al., 1999). Of course, information regarding speed is available through a vehicle’s speedometer. In everyday driving, however, drivers may not regularly check their speedometer, but instead rely on what feels safe or appropriate for prevailing conditions; experimental evidence suggests that speedometer glances are relatively short and infrequent, especially among more experienced drivers (Lee et al., 2006, Lehtonen et al., 2020). In on-road experiments, it has been observed that drivers spend 0.7% to 3% of the time looking at the speedometer or instrument panel (Lansdown, 2003, Recarte and Nunes, 2003), with a frequency of 0.36 to 1.49 glances per minute (Lansdown, 2003). At 65mph, a driver might therefore travel 0.72 to 3 miles between instrument panel glances. Perception of speed, safety, and appropriateness, then, are derived primarily through variables present in optic flow (Gibson, 1979, Lappe et al., 1999)—the transformations of the visual field that distinguish self-motion from the motion of external objects.
Two key variables present in optic flow are particularly important for control of vehicle speed: optical edge rate, and tau, specifying time-to-collision. The first, optical edge rate, refers to “the rate at which local discontinuities cross a fixed point of reference in the observer’s field of view” (Larish and Flach, 1990; p 295), where local discontinuities may be objects in the environment, other vehicles, or elements of the ground texture. Optical edge rate may be contrasted with global flow rate, which scales with the velocity of forward motion independent of ground texture or objects in the environment (Larish and Flach, 1990). Put quite broadly, lower optical edge rate and lower global flow rate are both commonly associated with slower speeds. However, lower optical edge rate also occurs in very sparse environments. Decreased traffic volume could therefore lead to reduced optical edge rate. Further, optical edge rate, but not global flow rate, is a primary source of information by which the speed of self-motion is perceived (Larish and Flach, 1990, Andersen et al., 1999). If it is in fact the case that decreased traffic volume is associated with lower optical edge rate, this may result in systematic underestimation of speed by drivers. This effect has been demonstrated in both lab and field settings (Denton, 1980, Warren, 1982). Despite this, it is not currently known if or how drivers may adapt and come to use other sources of visual information to compensate for impacts of decreased traffic volume on optical edge rate. With low traffic volume, optical edge rate information generated by the ground texture and road markings may still be used (Lidestam et al., 2019). Further, there is some evidence that global flow rate may substitute for optical edge rate in situations where edge rate is deemed unreliable (Dyre, 1997, Ballard et al., 1998). However, decreased traffic volume could also reduce the magnitude of motion parallax, leading to underestimation of speed (Dyre et al., 2006). Given this, with fewer other vehicles on the road, drivers may be missing important external referents for accurately perceiving their own speed. On the whole, we would expect these changes in optical information to result in underestimation of speed by drivers, potentially leading to speeding.
The second key variable, tau, is the inverse of the rate of optical expansion of angles in the visual field, and specifies time-to-collision (Lee, 1976). Objects in an environment subtend a certain optical angle. The change in that optical angle over time, as occurs through motion of the object or self, is termed optical expansion or contraction as appropriate. The rate of an object’s optical expansion (i.e., the expansion rate divided by optical angle), in particular, plays a key role in speed management and braking; drivers are able to perceive with quite high accuracy whether or not they will collide with an object or vehicle ahead of them, and make time-to-collision estimates, on the basis of optical expansion rate (Lee, 1976, Rio et al., 2014). Like optical edge rate, use of optical expansion rate reveals the importance of external referents in the perception of one’s own speed. However, where other vehicles passed by a driver provide an external referent used for speed perception through optical edge rate, the presence of a lead vehicle provides a referent through optical expansion or time-to-collision. In normal driving, a lead vehicle necessarily constrains speed to some extent, as the following vehicle cannot driver faster than it without risking a collision. In addition to this, it also creates visual information – optical expansion or contraction – that reflects the relation between the speeds of the leading and following vehicles. The environmental affordances for, or constraints on, a vehicle’s movement and associated visual information are therefore concomitant; visual information is generated by real relations among objects in an environment, and specifies to a perceiver what actions may be taken in that environment. When many vehicles share a road, each vehicle creates visual information that supports other drivers’ accurate speed perception and management. With fewer vehicles on the road than normal, optical information is changed in such a way that drivers perceive the road as affording faster travel.
Additionally, the changes in perceptual variables associated with lower traffic volume during the COVID-19 pandemic parallel some differences between urban and rural driving. Namely, rural driving is typically associated with relatively lower visual demand, affecting driving behaviors and drivers’ eye movements (Lansdown, 2003, Engström et al., 2005, Chapman and Underwood, 1998). In comparisons of urban and rural driving it has been found that higher visual demand in driving is associated with lower speeds (Engström et al., 2005), and that lower visual demand is associated with higher speeds (Antonson et al., 2009). Similarly, drivers in tunnels, which are characterized by very low visual demand, have been shown to underestimate their speed, leading to increased speeding (Wan et al., 2018). Tunnel sidewall markings, which may increase optical edge rate, have been shown to reduce this effect (Wan et al., 2018). We may then expect that lower visual demand caused by decreased traffic volume would be associated with higher speeds.
3 Risk and boredom
Perceptions of risk might also be challenging in low traffic volume conditions. Early field-theoretic models of driver behavior suggested that drivers can perceive, on the basis of visual information similar to that identified above, a “field of safe travel” – an area in or through which the vehicle may be safely driven (Gibson and Crooks, 1938). The size of the field of safe travel necessarily depends on environmental layout, including the nature and quality of the road, presence of other vehicles and obstacles, and prevailing weather conditions. Relatedly, drivers identify a “minimal stopping zone”, which represents the area within which they could stop the vehicle. The perceived risk of a driver’s current actions could reflect the relation between the sizes of the field of safe travel and minimal stopping zone: as size of the minimal stopping zone approaches that of the field of safe travel, greater risk is perceived. Notably, risk is therefore not perceived in reference to absolute units, such as distance or miles per hour, but rather in reference to environmental and behavioral factors; it corresponds to the relation between what actions the environment affords to the driver, and the driver’s capabilities to act. Further, although the concepts field of safe travel and minimal stopping zone predate Gibson’s later ecological theory and related research, such as that on tau, we expect that they are compatible. A driver might identify, based on tau and other visual information, affordances that are functionally similar to the field of safe travel and minimal stopping zone.
More recent motivational accounts of driver behavior have suggested that drivers seek homeostasis of one or more related variables, such as risk, task difficulty, or challenge (Ranney, 1994, Fuller, 2005). That is to say, drivers adjust their behavior so as to maintain a more-or-less constant and acceptable level of perceived risk, challenge, or some other variable. There is not broad agreement on which specific variable drivers prioritize, and the preference for or attention given to one over others may depend on individual factors. However, because psychological variables such as risk are managed, and not physical variables like speed in itself, these models account for some of the flexibility drivers exhibit in different environments or situations. A combination of field-theoretic and motivational models may help explain changes in driver behavior under low traffic volume conditions. When fewer cars are on the road, a driver’s field of safe travel is expanded. Given that perceived risk is a function of the relation between the field of safe travel and the minimal stopping zone, maintaining a constant level of perceived risk would entail expanding the minimal stopping zone - that is, relaxing the constraint on speed and driving faster. Changes in the environment are thereby met with changes in behavior in order to achieve homeostasis of a psychological variable. There are several disadvantages to this response. Speed is, in general, positively related to crash risk and severity (Pei et al., 2012, Aarts and van Schagen, 2006). Additionally, there may be secondary perceptual or psychological effects. It has been found that adaptation to moving at a high speed causes systematic underestimation of speed when it is decreased (Hietanen et al., 2008). A driver who acclimates to a higher speed on a highway, for example, may then tend to underestimate their speed when transferring onto an off ramp or local road. Similar effects occur with changes in gaze fixation distance: when one switches from a distant point of fixation to a nearer one while in motion, speed is typically underestimated (Yotsutsuji and Kita, 2010).
Finally, lower traffic volume may cause increased boredom among drivers, potentially resulting in increased speeding or other unsafe behaviors as coping strategies. Boredom here is defined as “the aversive experience of having an unfulfilled desire to be engaged in satisfying activity” (Fahlman et al., 2013, p. 69), and is understood to occur in situations where stimulation or information is sparse, monotonous, or meaningless. In driving, it is associated with low traffic volume, routine trips, slower speeds, and use of cruise control or other kinds of automation (Steinberger, 2018). To cope with boredom drivers may adopt any of a number of strategies, which are characterized as either “approach strategies” or “avoidance strategies” (Steinberger et al., 2017). Approach strategies are those which re-engage the bored driver in the driving task, and can include changing lanes, aggressive driving, and speeding. Avoidance strategies tend to reduce task engagement further, and may include phone use, day-dreaming, and listening to music; these are more typically associated with distraction and taking one’s eyes off the road (Steinberger et al., 2017). Secondary task engagement while driving is associated with reduced speed (Young and Regan, 2007). Specifically, phone use is associated with both reduced speed (Iio et al., 2021) and increased time to recover speed lost by braking (Jamson et al., 2004). Boredom may therefore lead to either increased or decreased speed, contingent on individual drivers’ preference for approach or avoidance strategies. However, it is not presently known why or when drivers might adopt one coping strategy over others. This too may be associated with individual factors; Zuckerman (2007) suggests that individuals high on sensation-seeking are more likely to compensate for boredom with risky driving. Nonetheless, both approach and avoidance strategies are likely to increase risk. Approach strategies such as speeding are directly associated with increased crash risk (Aarts and van Schagen, 2006), while avoidance strategies which involve distraction are more likely to result in heading errors, leading to lane keeping errors (Engström et al., 2005).
4 Social and motivational factors
The information that other moving vehicles convey not only provides perceptual support for monitoring one’s own speed adequately, but it also provides social support for maintaining a given speed; it allows drivers to ground their actions in the context of what is socially acceptable behavior. Descriptive norms convey what speed other people are driving, whereas injunctive norms convey “ought” information about what other people, societal regulations, or laws imply drivers should do (Cialdini et al., 1991). Research suggests that descriptive norms in particular can affect driving speed. For example, young drivers’ beliefs about how fast their friends drive have been found to strongly predict their own speeding (Møller and Haustein, 2014). Moreover, interventions that lower the believed speed of other drivers using dynamic signage can reduce driver speeds (Van Houten and Nau, 1983). But in most daily circumstances, the perception that one’s own speed is within a normal range and is unlikely to lead to negative societal consequences (a speeding ticket, or angry honks) comes from the direct influence of a heuristic to drive within the normal flow of other drivers.
When normative information is unavailable to guide behavior, driving speed will likely be more strongly influenced by individual differences in the motives inherently underlying people’s driving behavior, and the attitude functions which these behaviors satisfy (Katz, 1960). Attitude function approaches to predicting behavior suggest that individuals’ attitudes towards different objects or behaviors serve different attitude functions (Katz, 1960, Shavitt, 1990), and that attitude functions may vary across different situations (Marsh and Julka, 2000, Julka and Marsh, 2005). For example, an individual may hold positive attitudes about speeding insofar as it serves the attitude function of utility. However, certain attitude functions may become more or less relevant under conditions of substantially reduced traffic volume; in a driving environment that evokes few social pressures, provides less information about others’ behavior, and has no “audience” of other drivers, motivation to drive slower or faster to convey a certain impression on others, or to avoid negative consequences from others’ reactions, is likely minimal.
Other intrinsic motives might instead play a role. Research suggests that the same attitude topic, such as attitude toward personally speeding, can serve different functions for given individuals (DeBono, 1987, Katz, 1960, Maio and Olson, 2000, Shavitt, 1990, Smith et al., 1956). For some individuals, attitudes surrounding driving may meet certain needs for status and superiority—getting ahead of, cutting in front of, and showing disdain for other drivers. Under low density situations, such motives are likely to be less evident than other functional bases of driving attitudes. For some individuals, utilitarian and hedonic functions of speeding would involve practical benefits and costs of driving slower versus faster, including the potentially pleasurable experience of going fast, or the boredom of driving slowly. For other individuals, driving behaviors could serve the function of expressing one’s personal values, such as about energy conservation; a driver of a hybrid vehicle might strive to get positive feedback from his/her energy display that suggests optimal driving, or consciously try to conserve gas by minimizing hard braking and rapid accelerations. In extreme cases, risky driving behavior may reflect depression or suicidality, which may be increased during a pandemic. Specific to the COVID-19 pandemic, slower driving may reflect a general concern for others’ safety, or respect for and adherence to rules that are intended to promote safety. It may also reflect a concern for one’s own health and well-being, insofar as slower driving decreases the risk of interacting with others (e.g., police at a traffic stop, EMTs at a crash). Some support for the notion that drivers tend to vary in the functions that underlie their attitudes toward speeding comes from research finding that drivers do vary in driving styles—aggressive, cautious, or defensive for instance (Sagberg et al., 2015). Accordingly, a motive-based account for driving tendencies would suggest that under an unusual drop in traffic volume, driver responses may fairly diverge as the relative weight or relevance of different individual attitude functions change.
5 Conclusions and limitations
On balance, existing empirical research suggests that decreased traffic volume is likely to result in impairments in drivers’ ability to accurately perceive, and thus control, their own speed. Three primary processes are likely at play. First, in low traffic volume conditions drivers lack visual information that they would typically use to detect their own speed. Systematic underestimation of speed, and thus speeding, are possible results. Second, it has been suggested that drivers seek homeostasis of perceived risk. With fewer cars on the road, speeding is likely to be perceived as less risky, as any given car’s field of safe travel is expanded. If this is in fact the case, it would entail more speeding in order to maintain the desired or acceptable level of risk. Finally, low traffic volume is associated with increased boredom. Drivers may adopt a number of strategies to deal with boredom. Many of these, such as speeding, aggressive driving, and phone use, are associated with increased risk. A summary of factors considered and their anticipated effects on speeding is given in Table 1 .Table 1 Factors considered and anticipated effects on motor vehicle speeds.
Factor Effect on Speed
Decreased optical edge rate Increase
Absence of time-to-collision information Increase
Decreased motion parallax magnitude Increase
Lower visual demand Increase
Lower perceived risk Increase
Increased boredom Increase OR Decrease, dependent on coping strategy
Lack of social normative information Increase OR Decrease, dependent on individual attitudes or motivations
It should be noted that, while this assessment draws on the work of Gibson (1966); Gibson (1979), which lead to the discovery of optical variables such as optical edge rate and time-to-collision, our use of terminology is not wholly consistent with Gibson. For example, we have described certain variables as being involved in speed perception. However, speed, at least in terms of metrical units such as miles per hour, is not in itself perceptible (Gibson, 1979). Organisms regulate their behavior with regard to the affordances, or possibilities for action, in an environment, which are perceived directly. Perception is then of action-relevant and action-scaled properties (Gibson, 1966, Gibson, 1979). A driver might directly perceive, for example, that they will or won’t make a traffic light, and not that they are so many meters away from the light, and traveling at so many miles per hour. These perceptions of affordances – of what actions may be taken in a given situation – necessarily hinge not only on the environment but also on the driver’s capabilities for action: their driving aptitude, the acceleration or braking performance of their vehicle, and so on. In this way, affordances inhere in the relation between driver and environment. In driving, however, behavior must be regulated both with regard to affordances and with regard to more abstract rules, such as speed limits, that set certain constraints. Regulation of behavior according to an abstract rule or variable requires either a tool, such as a speedometer, or the individual calibration over time of that which is immediately perceived (i.e., affordances) with that which must be ascertained (i.e., speed in miles per hour). Thus, while we have described factors such as lower optical edge rate as affecting speed perception, it may be more accurate to say that they affect the mapping of perceptions of environmental affordances onto more abstract variables. Further, we would not suggest that changes in perceptual information result in drivers perceiving that a road affords speeding, as speeding exists only in relation to traffic rules that are not present in optic flow in the manner of optical edge rate and time-to-collision. Rather, it may be that drivers perceive a road as affording safe travel at such-and-such a rate in the way that a pedestrian might perceive a slippery patch of sidewalk as affording running over, walking over, or neither. That is, the perception occurs first in terms of “What actions can I take?”, and for the driver only later in terms of “How many MPH/KMPH?” and “Is it legal?”.
The above perceptual factors suggest that decreased traffic volume ought to be associated with increased speeding. Indeed, some research has found increases in speeding in conjunction with decreased traffic volume in the early months of the COVID-19 pandemic (Katrakazas et al., 2020, Lee et al., 2020, Shapiro et al., 2021). Absent unique effects from the COVID-19 pandemic, we would expect lower traffic volume to lead to increased speeding for most drivers. However, while there was an observed increase in extreme speeding during the COVID-19 stay-at-home period, there was also an increase in the percentage of cars traveling below the speed limit (Shapiro et al., 2021). This divergence likely reflects differences in individual attitudes or motivations that affect the impact of any of the above-mentioned perceptual factors on actual driving behavior. It may be the case that, lacking perceptual information that supports accurate speed perception and normative information that creates situational pressures, drivers’ different motivational tendencies are revealed, or their effects are amplified. For example, drivers prioritizing utility, in the sense of reaching their destination faster, may be more likely to engage in speeding. Drivers prioritizing value expression, which may, for example, include expression of concern for the well-being of others, might instead drive more cautiously. Others prioritizing health and safety may also drive slower to avoid potential interactions with police that do not allow for social distancing. Further empirical research is required to examine the mechanisms by which conditions unique to the COVID-19 period affect speeding.
These conclusions naturally warrant empirical evaluation through simulator, road video, or on-road studies. This is especially relevant as many of the theoretical constructs applied were not developed with driving in mind, and their application here is novel. For example, the role of optical edge rate in speed perception has primarily been explored with reference to ground texture. Controlled studies are therefore required to determine whether changes in traffic volume can produce change in optical edge rate sufficient to affect speed perception, and whether drivers in low volume conditions instead use other visual information to determine their speed. These may be conducted in driving simulators where driver control is required, or else with road video where increased verisimilitude is desirable. Multiple experiments are likely required to establish the specific roles of optic flow variables in driver speed perception under a variety of traffic volume, visibility, road type, and other conditions; through these, it may be possible to develop the present theoretical account into a robust mathematical model. While we expect that changes in perceptual information resulting from lower traffic volume during the COVID-19 pandemic should result in underestimation of speed and increased speeding, it is not clear if these changes can account for the observed increase in extreme speeding. Therefore, determining the magnitude of perceptual effects on speed perception and speeding behavior is of great concern. Additionally, the role of attitudes and motivations in particular warrants further empirical investigation. Extensive evidence suggests that individual-level factors, such as sensation-seeking and boredom proneness, affect driving behavior under normal circumstances (Jonah, 1997, Dahlen et al., 2005). However, it is not presently known how these effects might differ under unusually low traffic volume. Unique circumstances associated with the COVID-19 epidemic might also introduce new motivations that would not normally impact driver behavior. Some motivations, such as those related to social distancing, may persist even as states reopen and traffic volumes return to normal levels.
Through empirical investigation, it is hoped that this theoretical account leads to the development of interventions to reduce speeding that target either drivers’ individual motivations, or the social and environmental context in which those motivations play out. Interventions affecting the perceptual information available to drivers have been validated in some simulator studies. These include transverse strips or herringbone patterns on curves (Ariën et al., 2017), and sidewall markings in tunnels (Wan et al., 2018). Similar approaches, including road markings or physical objects like delineator posts, might be temporarily employed to reduce speeding by increasing optical edge rate while traffic volume remains below normal levels. Other interventions might provide additional information or incentives regarding speed to drivers, as has been attempted with some gamified driving apps (Steinberger et al., 2016). However, the effectiveness of such approaches is uncertain. We anticipate that considerable basic empirical research is required to develop interventions based on the current theoretical account, but are optimistic for the potential to improve roadway safety.
CRediT authorship contribution statement
K.L. Marsh: Conceptualization, Writing - original draft, and Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
We wish to thank Claire Michaels for comments on an earlier version of this manuscript.
==== Refs
References
Aarts L. van Schagen I. Driving speed and the risk of road crashes: a review Accid. Anal. Prev. 38 2006 215 224 16256932
Afroz B. Moniruzzaman S. Stark Ekman D. Andersson R. The impact of economic crisis on injury mortality: the case of the ‘Asian crisis’ Public Health 126 10 2012 836 838 22884860
Agrawal A.W. King H. Wachs M. Marks J. The Impact of the COVID-19 Recovery on California Transportation Revenue: A Scenario Analysis through 2040 2020 Mineta Transportation Institute
Andersen G.J. Cisneros J. Atchley P. Saidpour A. Speed, size, and edge-rate information for the detection of collision events J. Exp. Psychol. Hum. Percept. Perform. 25 1 1999 256 10069034
Antonson H. Mårdh S. Wiklund M. Blomqvist G. Effect of surrounding landscape on driving behaviour: a driving simulator study J. Environ. Psychol. 29 4 2009 493 502
Ariën C. Brijs K. Vanroelen G. Ceulemans W. Jongen E.M.M. Daniels S. Brijs T. Wets G. The effect of pavement markings on driving behaviour in curves: a simulator study Ergonomics 60 5 2017 701 713 27376174
Ballard T.G. Roach T. Dyre B.P. Use of global optical flow rate and discontinuity rate depends on their validity as determinants of egospeed 42 1998 1540 1544
Blower D. Flannagan C. Geedipally S. Lord D. Wunderlich R. Identification of Factors Contributing to the Decline of Traffic Fatalities in the United States No. Project Number 2019 17 67 from 2008 to 2012
Cialdini R.B. Kallgren C.A. Reno R.R. A focus theory of normative conduct: a theoretical refinement and reevaluation of the role of norms in human behavior Zanna M.P. Advances in Experimental Social Psychology Vol. 24 1991 Academic Press 201 234
Chapman P.R. Underwood G. Visual search of driving situations: danger and experience Perception 27 8 1998 951 964 10209634
Chrastil E.R. Warren W.H. Does the human odometer use an extrinsic or intrinsic metric? Attention Percept. Psychophys. 76 1 2014 230 246
Dahlen E.R. Martin R.C. Ragan K. Kuhlman M.M. Driving anger, sensation seeking, impulsiveness, and boredom proneness in the prediction of unsafe driving Accid. Anal. Prev. 37 2 2005 341 348 15667821
DeBono K.G. Investigating the social-adjustive and value-expressive functions of attitudes: Implications for persuasion processes J. Personal. Soc. Psychol. 52 2 1987 279
Denton G.G. The influence of visual pattern on perceived speed Perception 9 4 1980 393 402 7422457
Doucette M.L. Tucker A. Auguste M.E. Watkins A. Green C. Pereira F.E. Borrup K.T. Shapiro D. Lapidus G. Initial impact of COVID-19’s stay-at-home order on motor vehicle traffic and crash patterns in Connecticut: an interrupted time series analysis Injury Prevention 27 1 2021 3 9 10.1136/injuryprev-2020-043945 33115707
Dyre, B.P., 1997. Perception of accelerating self-motion: Global optical flow rate dominates discontinuity rate. In: Proceedings of the Human Factors and Ergonomics Society Annual Meeting, vol. 41(2). SAGE Publications, Los Angeles, pp. 1333–1337.
Dyre B.P. Cooper S. Lew R. Wotring B. The magnitude of motion parallax affects control of egospeed Proceedings of the Human Factors and Ergonomics Society Annual Meeting 50 16 2006 1666 1669
Engström J. Johansson E. Östlund J. Effects of visual and cognitive load in real and simulated motorway driving Transp. Res. Part F: Traffic Psychol. Behav. 8 2 2005 97 120
Fahlman S.A. Mercer-Lynn K.B. Flora D.B. Eastwood J.D. Development and validation of the multidimensional state boredom scale Assessment 20 1 2013 68 85 21903683
Fuller R. Towards a general theory of driver behaviour Accid. Anal. Prev. 37 3 2005 461 472 15784200
Gibson J.J. The senses considered as perceptual systems 1966 Houghton Mifflin
Gibson J.J. The ecological approach to visual perception 1979 Houghton Mifflin
Gibson J.J. Crooks L.E. A theoretical field-analysis of automobile-driving Am. J. Psychol. 51 3 1938 453 471
Hietanen M.A. Crowder N.A. Ibbotson M.R. Differential changes in human perception of speed due to motion adaptation J. Vision 8 11 2008 6-6
Iio K. Guo X. Lord D. Examining driver distraction in the context of driving speed: An observational study using disruptive technology and naturalistic data Accid. Anal. Prev. 153 2021 105983 33618100
Jamson A.M. Strayer D.L. Drews F.A. Profiles in driver distraction: effects of cell phone conversations on younger and older drivers Hum. Factors 46 4 2004
Jonah B.A. Sensation seeking and risky driving: a review and synthesis of the literature Accid. Anal. Prev. 29 5 1997 651 665 9316713
Julka D.L. Marsh K.L. An attitude functions approach to increasing organ-donation participation J. Appl. Soc. Psychol. 35 4 2005 821 849
Katrakazas C. Michelaraki E. Sekadakis M. Yannis G. A descriptive analysis of the effect of the COVID-19 pandemic on driving behavior and road safety Transp. Res. Interdiscip. Perspect. 7 2020 100186 34173462
Katz D. The functional approach to the study of attitudes Public Opinion Quarterly 24 2 1960 163 204
Lansdown T.C. Visual demands associated with rural, urban, & motorway driving McCabe P.T. Contemporary Ergonomics 2003 2003 Taylor & Francis New York 331 336
Lappe M. Bremmer F. Van den Berg A.V. Perception of self-motion from visual flow Trends Cogn. Sci. 3 9 1999 329 336 10461195
Larish J.F. Flach J.M. Sources of optical information useful for perception of speed of rectilinear self-motion J. Exp. Psychol. Hum. Percept. Perform. 16 2 1990 295 2142200
Lee D.N. A theory of visual control of braking based on information about time-to-collision Perception 5 4 1976 437 459 1005020
Lee S.E. Olsen E.C.B. Simons-Morton B. Eyeglance behavior of novice teen and experienced adult drivers Transp. Res. Rec. 1980 1 2006 57 64
Lee J. Porr A. Miller H. Evidence of increased vehicle speeding in Ohio’s major cities during the COVID-19 pandemic Transport Findings, June 2020
Lehtonen E. Malhotra N. Starkey N.J. Charlton S.G. Speedometer monitoring when driving with a speed warning system Eur. Transp. Res. Rev. 12 1 2020 1 12
Lidestam B. Eriksson L. Eriksson O. Speed perception affected by field of view: Energy-based versus rhythm-based processing Transp. Res. Part F Psychol. Behav. 65 2019 227 241
Maio G.R. Olson J.M. Why We Evaluate: Functions of Attitudes 2000 Lawrence Erlbaum Associates Publishers
Marsh K.L. Julka D.L. A motivational approach to experimental tests of attitude functions theory Gregory M.R. Olson J.M. Why we evaluate: Functions of attitudes (271–294) 2000 Lawrence Erlbaum Associates Publishers
Møller M. Haustein S. Peer influence on speeding behaviour among male drivers aged 18 and 28 Accid. Anal. Prev. 64 2014 92 99 24355559
Pei X. Wong S.C. Sze N.N. The roles of exposure and speed in road safety analysis Accid. Anal. Prev. 48 2012 464 471 22664713
Ranney T.A. Models of driving behavior: a review of their evolution Accid. Anal. Prev. 26 6 1994 733 750 7857489
Recarte M.A. Nunes L.M. Mental workload while driving: effects on visual search, discrimination, and decision making J. Experiment. Psychol. Appl. 9 2 2003 119 137
Rio K.W. Rhea C.K. Warren W.H. Follow the leader: visual control of speed in pedestrian following J. Vision 14 2 2014 4-4
Sagberg, F., Selpi, Bianchi Piccinini, G.F., Engström, J., 2015. A review of research on driving styles and road safety. Hum. Factors 57(7), 1248–1275.
Shapiro D. Tucker A. O’Brien B. Doucette M.L. Changes in motor vehicle driving behaviors and crashes in Connecticut related to COVID-19: A descriptive analysis 2021 Under Review
Shavitt S. The role of attitude objects in attitude functions J. Exp. Soc. Psychol. 26 2 1990 124 148
Sivak M. The information that drivers use: is it indeed 90% visual? Perception 25 1996 1081 1089 8983048
Smith M.B. Bruner J.S. White R.W. Opinions and Personality 1956 Wiley New York
Steinberger F. Risky gadgets to the rescue: Reframing in-car technology use as task engagement (Doctoral dissertation) 2018 Queensland University of Technology
Steinberger, F., Proppe, P., Schroeter, R., Alt, F., 2016, October. CoastMaster: an ambient speedometer to gamify safe driving. In: Bolls, S., Green, P., Burnett, G., Gabbard, J., Osswald, S. (Eds.) Proceedings of the 8th International Conference on Automotive User Interfaces and Interactive Vehicular Applications, Association for Computing Machinery, pp. 83–90.
Steinberger F. Schroeter R. Watling C.N. From road distraction to safe driving: evaluating the effects of boredom and gamification on driving behaviour, physiological arousal, and subjective experience Comput. Hum. Behav. 75 2017 714 726
Turvey M.T. Harrison S.J. Frank T.D. Carello C. Human odometry verifies the symmetry perspective on bipedal gaits J. Exp. Psychol. Hum. Percept. Perform. 38 4 2012 1014 1025 22506786
Turvey M.T. Romaniak-Gross C. Isenhower R.W. Arzamarski R. Harrison S. Carello C. Human odometer is gait-symmetry specific Proc. Royal Soc. B: Biol. Sci. 276 1677 2009 4309 4314
Van Houten R. Nau P.A. Feedback interventions and driving speed: a parametric and comparative analysis J. Appl. Behav. Anal. 16 3 1983 253 281 16795666
Wan H. Du Z. Yan Q. Chen X. Evaluating the effectiveness of speed reduction markings in highway tunnels Transport 33 3 2018 647 656
Warren, R., 1982. Optical transformation during movement: Review of the optical concomitants of egomotion. Ohio State Univ Research Foundation Columbus.
Yotsutsuji H. Kita H. Self-motion speed perception of visual information in driving: modeling and validation Traffic Transp. Stud. 2010 2010 1029 1045
Young K. Regan M. Driver distraction: a review of the literature Faulks I.J. Regan M. Stevenson M. Brown J. Porter A. Irwin J.D. Distracted driving 2007 Australasian College of Road Safety Sydney, NSW 379 405
Zuckerman M. Sensation seeking and risky driving, sports, and vocations Zuckerman M. Sensation seeking and risky behavior 2007 American Psychological Association 73 106
| 34130056 | PMC9746225 | NO-CC CODE | 2022-12-15 00:03:38 | no | Accid Anal Prev. 2021 Sep 12; 159:106225 | utf-8 | Accid Anal Prev | 2,021 | 10.1016/j.aap.2021.106225 | oa_other |
==== Front
Transplantation
Transplantation
TP
Transplantation
0041-1337
1534-6080
Lippincott Williams & Wilkins Hagerstown, MD
36329002
00006
10.1097/TP.0000000000004387
3
Commentaries
COVID-19 Vaccines and Solid Organ Transplantation: More Doses, More Protection
Safa Kassem MD 12
Kotton Camille Nelson MD 34
1 Transplant Center, Massachusetts General Hospital, Harvard Medical School, Boston, MA.
2 Division of Nephrology, Massachusetts General Hospital, Harvard Medical School, Boston, MA.
3 Transplant and Immunocompromised Host Infectious Diseases, Transplant Center, Massachusetts General Hospital, Harvard Medical School, Boston, MA.
4 Infectious Diseases Division, Massachusetts General Hospital, Harvard Medical School, Boston, MA.
Correspondence: Camille Nelson Kotton, MD, Transplant and Immunocompromised Host Infectious Diseases, Infectious Diseases Division, Massachusetts General Hospital, Harvard Medical School, 55 Fruit St, Cox 5, Boston, MA 02114. ([email protected]).
04 11 2022
1 2023
04 11 2022
107 1 2122
5 8 2022
10 8 2022
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
2022
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
==== Body
pmcMore than 2 y into the COVID-19 pandemic, solid organ transplant (SOT) recipients remain vulnerable to infection. Better testing, therapeutics, and vaccination have improved their overall mortality from infection and led to less disruption in performing kidney transplants.1 Nonetheless, SOT recipients face hurdles in returning to regular life activities given diminished vaccine efficacy in this population, relating to their immune suppression and baseline comorbid conditions.
Early after vaccine implementation, it became apparent that immunocompromised hosts have an attenuated response after 1 and 2 vaccine doses, as was highlighted in multiple interventional and observational studies and was consistent regardless of the vaccine platform used. The use of a third dose as part of the primary vaccine series was recommended by the United States Centers for Disease Control and Prevention for immunocompromised hosts in August 2021.2 In this issue, Bailey et al3 synthesized a meta-analysis to examine the response to a third dose of mRNA SARS-CoV vaccine in immunocompromised transplant recipients and showed an increased humoral response to a third dose in solid organ transplant recipients; almost half of those who did not have a serologic response after 2 vaccine doses developed a response after a third one. The cellular immune response was similarly augmented with a third dose. The study included >1200 patients from a dozen studies that met preset inclusion criteria. Their work provides compelling evidence for the need of additional vaccines in immunocompromised hosts. This supports the current guidance from the Centers for Disease Control and Prevention, which recommends that SOT recipients receive a primary vaccination series of 3 doses of the mRNA vaccine, followed by a first booster (fourth dose) and second booster (fifth dose) 3 and 4 mo after the third dose, respectively as well as another booster with the bivalent mRNA vaccine. ref: https://www.cdc.gov/vaccines/covid-19/clinical-considerations/interim-considerations-us.html#immunocompromised, accessed 6 October 2022.4 The World Health Organization currently recommends 1 booster shot in addition to the triple vaccine primary series for immunocompromised patients.5
The Omicron variant caused an intense and somewhat lethal surge in COVID-19 cases worldwide near the end of 2021. New subvariants continue to evolve and sublineages are emerging with increased immune evasion from both disease and vaccine-acquired immunity. Receipt of additional mRNA vaccine doses decreased the risk of Omicron variants COVID-19 associated hospitalization rates and disease severity in immunocompetent hosts with the risk further attenuated after 4 doses compared with 3 doses.4 Among SOT recipients, case series have shown an improved neutralizing antibody level after 4 vaccine doses. In 1 series, the improvement in humoral response to a fourth vaccine dose was seen in 41.9% of those who did not respond to 3 doses.6 In another series, 89% of 128 SOT recipient developed a serologic response after a fourth vaccine dose, with 61% of previously seronegative patients developing detectable immunity.7 Although the meta-analysis by Bailey et al supports 3 doses as the primary series, additional doses are clearly needed, especially for longer-term protection.
Given the blunted immune response to vaccination in SOTR, preexposure prophylaxis with tixagevimab/cilgavimab is being increasingly used, and it has shown effectiveness in reducing breakthrough infections with the Omicron variant.8 The in vitro activity of the monoclonal antibodies cocktail was maintained against the BA.5 sublineage in a recent study.9,10 Immune escape, however, via selective pressure from monoclonal antibody exposure could lead to the emergence of new variants in immunocompromised hosts, especially those with prolonged COVID-19 infections.11
Serious adverse events after additional mRNA vaccine booster continue to be rare, and according to the Vaccine Adverse Event Reporting System, a US safety surveillance system to monitor adverse events after vaccination, studying adverse events between January and March 2022, a time period in which about 500 000 immunocompromised hosts received a fourth shot, 4015 subjects registered through the system, and only 17 serious adverse events were reported.12 From the immunological adverse events standpoint, vaccines are very well tolerated by transplant recipients, and in general, events such as rejection or clinically significant de novo donor-specific antibodies formation13 are not associated with vaccination. Specific to COVID-19 vaccination, except for case reports describing allograft rejection in isolated cases, no widespread or significantly increased risk in allograft rejection has been described. Several studies looking at the immunogenicity of a third COVID shot in immunocompromised host showed no increased risk in allograft rejection. The work by Bailey et al in this issue highlights this as no acute rejection episodes were reported in the studies examined.
Multipronged approaches will continue to be needed in SOT recipients to prevent or diminish COVID-19 infection. When possible, the optimal strategy starts with primary vaccination before organ transplant, when vaccine efficacy is at its highest, followed by timely vaccine booster doses posttransplant. Such an approach supports the pretransplant COVID-19 vaccine mandates that were implemented at many transplant centers, as a best method to prevent severe COVID-19 infections. For those SOT recipients who underwent transplant before the vaccines were available, utilization of multiple doses of vaccine for the primary series followed by periodic boosters, along with preexposure prophylaxis monoclonal antibody, provides the safest and most effective method to prevent severe, life-threatening COVID-19. Moving forward, SOT recipients will likely need additional doses of vaccine, optimally tailored to actively circulating strains, such as the bivalent vaccines. We encourage the transplant community to keep SOT recipients up to date with their vaccines and preexposure prophylaxis monoclonal antibodies, along with methods for optimal infection control, to provide the best protection possible against this ongoing scourge.
C.N.K. serves on the Advisory Committee for Immunization Practice at The United States Centers for Disease Control and Prevention. The other author declares no conflicts of interest.
K.S. and C.N.K. participated in the writing of the paper.
==== Refs
REFERENCES
1. Nimmo A Gardiner D Ushiro-Lumb I . The global impact of COVID-19 on solid organ transplantation: two years into a pandemic. Transplantation. 2022;106 :1312–1329.35404911
2. Mbaeyi S Oliver SE Collins JP . The advisory committee on immunization practices’ interim recommendations for additional primary and booster doses of covid-19 vaccines—United States, 2021. MMWR Morb Mortal Wkly Rep. 2021;70 :1545–1552.34735422
3. Bailey AJM, Maganti HB, Cheng W, et al . Humoral and cellular response of transplant recipients to a third dose of mRNA SARSCoV-2 vaccine: a systematic review and meta-analysis. Transplantation. 2023;107 :204–215.36398334
4. Centers for Disease Control and Prevention. Interim clinical considerations for use of COVID-19 vaccines currently approved or authorized in the United States. July 20, 2022. Available at https://www.cdc.gov/vaccines/covid-19/clinical-considerations/interim-considerations-us.html#immunocompromised. Accessed October 6, 2022.
5. World Health Organization. Interim statement on the use of additional booster doses of Emergency Use Listed mRNA vaccines against COVID-19. May 17, 2022. Available at https://www.who.int/news/item/17-05-2022-interim-statement-on-the-use-of-additional-booster-doses-of-emergency-use-listed-mrna-vaccines-against-covid-19. Accessed August 1, 2022.
6. Kamar N Abravanel F Marion O . Assessment of 4 doses of SARS-CoV-2 Messenger RNA-based vaccine in recipients of a solid organ transplant. JAMA Netw Open. 2021;4 :e2136030.34817587
7. Mitchell J Alejo JL Chiang TPY . Antibody response to a fourth dose of SARS-CoV-2 Vaccine in solid organ transplant recipients: an update. Transplantation. 2022;106 :e338–e340.35426888
8. Al Jurdi A Morena L Cote M . Tixagevimab/cilgavimab pre-exposure prophylaxis is associated with lower breakthrough infection risk in vaccinated solid organ transplant recipients during the omicron wave [Online ahead of print]. Am J Transplant 2022. doi: 10.1111/ajt.17128.
9. Touret F Baronti C Pastorino B . In vitro activity of therapeutic antibodies against SARS-CoV-2 Omicron BA.1, BA.2 and BA.5. Sci Rep. 2022;12 :12609.35871089
10. Takashita E Yamayoshi S Simon V . Efficacy of antibodies and antiviral drugs against omicron BA.2.12.1, BA.4, and BA.5 subvariants. N Engl J Med. 2022;387 :468–470.35857646
11. Scherer EM Babiker A Adelman MW . SARS-CoV-2 evolution and immune escape in immunocompromised patients. N Engl J Med. 2022;386 :2436–2438.35675197
12. Hause AM Baggs J Marquez P . Safety monitoring of COVID-19 mrna vaccine first booster doses among persons aged ≥12 years with presumed immunocompromise status—United States, January 12, 2022-March 28, 2022. MMWR Morb Mortal Wkly Rep. 2022;71 :899–903.35834416
13. Mulley WR Dendle C Ling JEH . Does vaccination in solid-organ transplant recipients result in adverse immunologic sequelae? A systematic review and meta-analysis. J Heart Lung Transplant. 2018;37 :844–852.29609844
| 36329002 | PMC9746227 | NO-CC CODE | 2022-12-15 23:21:55 | no | Transplantation. 2023 Jan 4; 107(1):21-22 | utf-8 | Transplantation | 2,022 | 10.1097/TP.0000000000004387 | oa_other |
==== Front
Transplantation
Transplantation
TP
Transplantation
0041-1337
1534-6080
Lippincott Williams & Wilkins Hagerstown, MD
00038
10.1097/TP.0000000000004400
3
Letters to the Editor
Single-center Experience on Nonlung Solid Organ Transplantation From SARS-CoV-2–positive Donors
Covarrubias Karina MD 1
Brubaker Aleah L. MD, PhD 1
Mekeel Kristin MD 1
Pretorius Victor MD 2
Shah Mita MD 3
Adler Eric MD 4
Ajmera Veeral MD 5
Schnickel Gabriel T. MD 1
Aslam Saima MD, MS 6
1 Department of Surgery, Division of Transplant and Hepatobiliary Surgery, UC San Diego, San Diego, CA.
2 Department of Surgery, Division of Cardiothoracic Surgery, UC San Diego, San Diego, CA.
3 Department of Medicine, Division of Nephrology, UC San Diego, San Diego, CA.
4 Department of Medicine, Division of Cardiology, UC San Diego, San Diego, CA.
5 Department of Medicine, Division of Hepatology, UC San Diego, San Diego, CA.
6 Department of Medicine, Division of Infectious Diseases and Global Public Health, UC San Diego, San Diego, CA.
Correspondence: Saima Aslam, MD, MS, Department of Medicine, Division of Infectious Diseases and Global Public Health, UC San Diego, 4510 Executive Dr, MC 7745, San Diego, CA 92121. ([email protected]).
13 10 2022
1 2023
13 10 2022
107 1 e41e42
19 8 2022
6 9 2022
08 9 2022
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
2022
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
STATUSONLINE-ONLY
SDCT
==== Body
pmcEarly in the coronavirus disease 2019 (COVID-19) pandemic, limited testing availability and knowledge about disease transmission raised concern about the safety of using organ donors with positive severe acute respiratory distress coronavirus 2 (SARS-CoV-2) testing.1 Recent registry reports suggest the safety of this approach, but granular data on such transplants remain limited.2-5 We report our single-center experience of 36 nonlung solid organ transplant (SOT) allografts from polymerase chain reaction (PCR) SARS-CoV-2+ donors.
We conducted a retrospective review of all nonlung SOT recipients at our center who underwent transplantation between July 2020 and July 2022 from deceased donors with SARS-CoV-2+ PCR testing within 28 d of donation (institutional review board no. 805160) (Figure 1). Donor SARS-CoV-2 PCR testing was obtained from nasopharyngeal (NP) swabs, tracheal aspirates, or bronchoalveolar lavage (BAL). The lowest cycle threshold (Ct) value was reported when available. Data analysis was performed by Stata 14.2 for Mac (College Station, TX). We report continuous variables as medians with associated range or interquartile range and discrete variables as percentages.
Thirty-six transplanted allografts from 29 deceased donors with SARS-CoV-2+ PCR within 28 d of donation were identified. None of the recipients developed donor-derived COVID-19. Twenty of 29 donors (69%) had a SARS-CoV-2+ test within 72 h of donation, 11 (38%) were symptomatic within 28 d of donation, and 4 (14%) had died from COVID-19 (Table 1). Of the 20 donors who were SARS-CoV-2+ within 72 h of donation, 6 donors had NP+ and BAL+, 5 had NP+ and BAL–, and 1 donor had NP– but BAL+ SARS-CoV-2 test. Eight donors had NP SARS-CoV-2+ test without lower respiratory tract testing. Ct values were available for 16 of 20 donors; the median Ct value was 33 (range, 18.0–38.2).
TABLE 1. Donor and recipient characteristics
Donor characteristics (N = 29)
Age, y, median (IQR) 33 (27–47)
Cause of death, n (%)
Trauma 4 (14)
Anoxia 16 (55)
Cerebrovascular accident 5 (17)
COVID-19 4 (14)
Symptomatic COVID-19 within 28 d,a n (%) 11 (38)
SARS-CoV-2 PCR+ within 72 h,b n (%) 20 (69)
SARS-CoV-2 PCR+ typec
NP+, BAL+, n (%) 6 (30)
NP+, BAL–, n (%) 5 (25)
NP+, no BAL, n (%) 8 (40)
NP–, BAL+, n (%) 1 (5)
Cycle threshold value, median (range) (N = 16) 33.0 (18.0–38.4)
Recipient characteristics (N = 34)
Age, y, median (IQR) 57 (42–62)
Organ, n (%)
Kidney 17 (50)
cPRA, median (range) 0 (0–86)
KAS, median (range) 9.60 (5.39–14.14)
Liver 8 (24)
MELD,d median (range) 31 (17–40)
Heart 7 (20)
Status, median (range) 2 (1–3)
Heart-kidney 1 (3)
Status 2
Liver-kidney 1 (3)
MELD 29
Induction agent, N (%)
Antithymocyte globulin 20 (59)
Basiliximab 1 (3)
Steroid 13 (38)
Pretransplant vaccine,e n (%) 26 (76)
Doses of COVID-19 vaccine, median (range) 2 (1–5)
Prior SARS-CoV-2 infection, n (%) 7 (21)
a Symptomatic COVID-19 was defined as symptoms compatible with COVID-19 within 28 d of procurement.
b Positive SARS-CoV-2 PCR test within 72 h of procurement.
cDonors who were PCR+ within 72 h of donation underwent all listed combinations of NP or BAL SARS-CoV-2 testing.
dAllocation MELD score.
eReceipt of at least 1 dose of any COVID-19 vaccine before transplant.
BAL, bronchoalveolar lavage; COVID-19, coronavirus disease 2019; cPRA, calculated panel-reactive antibody; IQR, interquartile range; KAS, kidney allocation score; MELD, model for end-stage liver disease score; NP, nasopharyngeal; PCR, polymerase chain reaction; SARS-CoV-2, severe acute respiratory coronavirus 2.
Thirty-four transplant recipients including 17 kidney (50%), 8 liver (24%), 7 heart (20%), 1 heart-kidney (3%), and 1 liver-kidney (3%) recipients were identified (Table 1). Twenty-six recipients (76%) received at least 1 dose of any COVID-19 vaccine before transplant. Seven recipients (21%) had COVID-19 before transplant; median time from COVID-19 to transplant was 31 d (range, 6–637). The first 3 recipients from donors with SARS-CoV-2+ PCRs within 72 h of donation received prophylactic remdesivir for 3 d per our protocol although this practice was later discontinued.
Overall mortality related to COVID-19 has reduced considerably owing to improved therapeutics and vaccination.6 As COVID-19 is now endemic, we suggest that SARS-CoV-2+ allografts be considered in the same way centers consider other respiratory viral infections in donors for nonlung SOT, such as influenza and respiratory syncytial virus, for which risk of transmission is low and vaccination and/or antivirals are available.
FIGURE 1. Organ acceptance criteria for all non-lung solid organs from donors with a history of COVID-19 at our center. COVID-19, coronavirus disease 2019; PCR, polymerase chain reaction; SARS-CoV-2, severe acute respiratory distress coronavirus 2.
To date, this is one of the largest single-center series of SARS-CoV-2+ donor allografts transplanted into nonlung SOT recipients and is consistent with other studies demonstrating a lack of transmission in this setting.2-5,7 We have extended prior work by including donors with SARS-CoV-2+ PCRs within 72 h of donation, adding granularity to a growing body of evidence supporting the safety of using acutely SARS-CoV-2+ allografts for nonlung SOT recipients.
Supplementary Material
K.C. and A.L.B. are cofirst authors.
Funding for this study was provided in part by the National Library of Medicine. K.C. is supported by T15LM011271 from the National Library of Medicine.
S.A. received honoraria from Merck and Gilead; consultant fee from BioMx, Phico, and Pherecydes; grant funding for research from Contrafect, Cystic Fibrosis Foundation, and National Institutes of Health (unrelated to current study). The other authors declare no conflicts of interest.
A.L.B. and S.A. conceived this study. K.C., A.L.B., and G.T.S. performed the chart review. K.C. performed the data analysis. K.C., A.L.B., and S.A. wrote the first draft of the article. Every author has read, edited, and approved the final article.
Supplemental Visual Abstract; http://links.lww.com/TP/C613.
Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com).
==== Refs
REFERENCES
1. American Society of Transplantation. Recommendations and Guidance for Organ Donor Testing. American Society of Transplantation; 2021.
2. Dhand A Okumura K Nabors C . Solid organ transplantation from COVID positive donors in the United States: analysis of United Network for Organ Sharing database. Transpl Infect Dis. [Epub ahead of print. August 9, 2022]. doi:10.1111/tid.13925.
3. Schold JD Koval CE Wee A . Utilization and outcomes of deceased donor SARS-CoV-2-positive organs for solid organ transplantation in the United States. Am J Transplant. 2022;22 :2217–2227.35730252
4. Ushiro-Lumb I Callaghan CJ Pettigrew GJ ; NHSBT Organ and Tissue Donation and Transplantation Clinical Team. Transplantation of organs from SARS-CoV-2 RNA positive deceased donors: the UK experience so far. Transplantation. 2022;106 :e418–e419.35581692
5. Boan P Marinelli T Opdam H . Solid organ transplantation from donors with COVID-19 infection. Transplantation. 2022;106 :693–695.35238852
6. Dong E Du H Gardner L . An interactive web-based dashboard to track COVID-19 in real time. Lancet Infect Dis. 2020;20 :533–534.32087114
7. La Hoz RM Mufti AR Vagefi PA . Short-term liver transplant outcomes from SARS-CoV-2 lower respiratory tract NAT positive donors. Transpl Infect Dis. 2022;24 :e13757.34741572
| 36240442 | PMC9746228 | NO-CC CODE | 2022-12-15 23:21:55 | no | Transplantation. 2023 Jan 13; 107(1):e41-e42 | utf-8 | Transplantation | 2,022 | 10.1097/TP.0000000000004400 | oa_other |
==== Front
Transplantation
Transplantation
TP
Transplantation
0041-1337
1534-6080
Lippincott Williams & Wilkins Hagerstown, MD
36398334
00027
10.1097/TP.0000000000004386
3
Original Clinical Science—General
Humoral and Cellular Response of Transplant Recipients to a Third Dose of mRNA SARS-CoV-2 Vaccine: A Systematic Review and Meta-analysis
Bailey Adrian J.M. BSc 12
Maganti Harinad B. PhD 23
Cheng Wei PhD 4
Shorr Risa MLIS 5
Arianne Buchan C. MD, FRCPC 16
Allan David S. MD, FRCPC 1237
1 Faculty of Medicine, University of Ottawa, Ottawa, ON, Canada.
2 Canadian Blood Services, Stem Cells and Centre for Innovation, Ottawa, ON, Canada.
3 Clinical Epidemiology and Regenerative Medicine, Ottawa Hospital Research Institute, Ottawa, ON, Canada.
4 Department of Biostatistics, Yale School of Public Health, New Haven, CT.
5 Library and Information Services, The Ottawa Hospital, Ottawa, ON, Canada.
6 Division of Infectious Diseases, Department of Medicine, The Ottawa Hospital, Ottawa, ON, Canada.
7 Blood and Marrow Transplant Program, Division of Hematology, Department of Medicine, The Ottawa Hospital, Ottawa, ON, Canada.
Correspondence: David S. Allan, BSc, Ottawa Hospital Research Institute, 501 Smyth Rd, Box 704 12, Ottawa, ON K1H 8L6, Canada. ([email protected]).
04 11 2022
1 2023
04 11 2022
107 1 204215
14 2 2022
31 8 2022
01 9 2022
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
2022
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
Background.
High rates of nonresponse to 2 doses of mRNA severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) vaccine have been reported in transplant recipients. Several studies have investigated the efficacy of a third dose in this population. However, efficacy remains unclear, as response rates vary across studies. Therefore, we conducted a systematic review and meta-analysis to determine the efficacy of a third dose of any mRNA SARS-CoV-2 vaccine in transplant recipients.
Methods.
Preferred Reporting Items for Systematic Review and Meta-Analysis reporting guidelines (PROSPERO:CRD42021281498) were followed. Medline, Embase, and CENTRAL were searched from inception to December 2, 2021, without restrictions. All full-text studies reporting on the efficacy of a third dose of any mRNA SARS-CoV-2 vaccine in pediatric and adult transplant recipients were included. The National Institutes of Health quality assessment tool for case series and the Cochrane risk of bias tool determined study quality. Meta-analysis was performed via the DerSimonian-Laird random-effect model.
Results.
Of 84 records, 12 studies totaling 1257 patients met inclusion criteria. One study was a randomized controlled trial, whereas all other studies were observational. Across 7 studies (801 patients), humoral response after 3 doses was observed in 66.1% (95% confidence interval, 62.8%-69.4%; I2 = 0%) of transplant recipients. Triple immunosuppression, mycophenolate, antiproliferatives, and belatacept use were associated with reduced odds of humoral response in studies reporting multivariate analyses. Transplant recipients receiving a third dose displayed higher levels of neutralizing antibodies to SARS-CoV-2 variants (Alpha, Beta, and Delta) compared with placebo.
Conclusions.
A third dose SARS-CoV-2 mRNA vaccine should be strongly considered in transplant recipients. Limitations included lack of controlled studies and clinically relevant thresholds to determine response to vaccination.
SDCT
==== Body
pmcINTRODUCTION
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), responsible for coronavirus disease 2019 (COVID-19), is a zoonotic viral pathogen that has become a global health concern. Low seroconversion rates (54%) after the second dose of mRNA SARS-CoV-2 vaccine have been reported in transplant recipients,1 and severe cases of COVID-19 have been described in transplant recipients who have received 2 doses of vaccine.2 Given the high mortality rate of COVID-19 in transplant recipients,3,4 health authorities have issued recommendations to administer a third dose in this population.1 However, the efficacy of a third dose of mRNA SARS-CoV-2 vaccine in transplant recipients varies across studies,5-7 with some studies reporting humoral response after 3 doses as low as 6.4%.5 Additionally, the relative efficacy of a third dose according to transplant type, mRNA vaccine, and factors associated with immune response remains unknown. The literature is also highly reliant on humoral response and lacks other indicators that are more suggestive of immunity, specifically cellular response and neutralizing antibody assays.8 As neutralization level is highly predictive of immune protection9 and T cells reduce viral loads and disease even when neutralizing antibody levels are low,10 these outcomes are of vital importance. Finally, these vaccines have demonstrated decreased efficacy against variants of concern (VOC),11,12 which are the dominant circulating strains in the community,13 and data on vaccine efficacy with respect to VOC in transplant populations remain unknown. Therefore, a systematic review and meta-analysis is required to determine the humoral and cellular response of transplant recipients to 3 doses of mRNA SARS-CoV-2 vaccine, according to viral strain, and factors associated with immunity in this population.
MATERIALS AND METHODS
This review was reported in accordance with the Preferred Reporting Items for Systematic Review and Meta-Analysis14 (Tables S1 and S2, SDC, http://links.lww.com/TP/C610). Our protocol was registered on PROSPERO (CRD42021281498; September 24, 2021).
Search Strategy and Study Selection
Medline, Embase, and Cochrane Central Register of Controlled Trials (CENTRAL) were searched from inception to December 2, 2021, without language or geographic restrictions, using a search strategy developed by a librarian (R.S.) specializing in systematic searches (Table S3, SDC, http://links.lww.com/TP/C610) according to the Peer Review of Electronic Search Strategies guidelines.15
Two independent reviewers (A.J.M.B. and H.B.M.) screened all records in Rayyan software,16 with supervising author providing consensus (D.S.A.). All studies reporting on the efficacy of a third dose of any mRNA SARS-CoV-2 vaccine in pediatric and adult transplant recipients were included. Studies reporting on nontransplant recipients (eg, Car-T-cell recipients) or studies reporting other doses (ie, first or second doses) were excluded. Studies were not required to be comparative. Randomized trials, case-control studies, cohort studies, cross-sectional studies, and single-arm studies were included, whereas commentaries, editorials, conference abstracts, reviews, gray literature, preprints, and guidelines were excluded. Manual searches through reference lists of included articles were conducted to capture any other relevant literature not captured by our search strategy.
Outcomes
Our primary outcomes consisted of the humoral and cellular response of transplant recipients of any age and transplant type to a third dose of any mRNA SARS-CoV-2 vaccine. Humoral response was considered positive if antibody titers exceeded the study’s threshold. Cellular response was considered positive if the ratio of SARS-CoV-2–specific T cells to T cells, respectively, exceeded the study threshold. If a study threshold was not provided, a threshold from another study using the same T-cell population was applied. Secondary outcomes included factors associated with immunity, adverse events, and safety.
Data Extraction
Two reviewers (A.J.M.B. and H.B.M.) independently extracted data using data extraction forms for all primary and secondary outcomes, with supervising author providing consensus (D.S.A.). Extracted data included study characteristics (authors, year, study design, patient age and sex, transplant type, time since transplant, vaccine type received, time since second dose, time of assessment of outcomes), details regarding humoral and cellular assays (antibody assay, measured antibody, threshold, cellular type, markers analyzed), and outcomes (prevalence of humoral and cellular response after 3 doses, prevalence of humoral and cellular response after 3 doses in patients without response to 2 doses, acute rejection, factors associated with humoral response, side effects, adverse events, and acute rejection episodes).
Risk of Bias Assessment
Two reviewers (A.J.M.B. and H.B.M.) independently assessed the risk of bias for all primary and secondary outcomes, with supervising author providing consensus (D.S.A.). For the synthesis of noncomparative data, such as humoral response in a single group, the National Institutes of Health quality assessment tool for case series was used (https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools, accessed October 2, 2021). For comparative data, version 2.0 of the Cochrane Risk of Bias Assessment Tool for randomized trials was applied (low risk, some concerns, or high risk of bias).17 All poor-quality studies or studies with a high risk of bias were excluded from the qualitative or quantitative synthesis. Sensitivity analysis was performed according to study quality. Funnel plots were used to assess for publication bias.
Certainty of the Evidence
The Grading of Recommendations Assessment, Development, and Evaluation (GRADE)18 was used to assess the certainty of the evidence for all outcomes. Although there is no formal guidance for GRADE in systematic reviews of prevalence,19 a recommended framework was applied.20 Two reviewers (A.J.M.B. and W.C.) independently assessed certainty of the evidence, with supervising author providing consensus (D.S.A.).
Statistical Analysis
Meta-analysis of the prevalence of humoral and cellular response of transplant recipients after 3 doses of mRNA SARS-CoV-2 vaccine was performed in OpenMetaAnalyst21 using the DerSimonian-Laird random-effect model.22 I2 >50% represented substantial heterogeneity,23 and heterogeneity was evaluated using the Cochrane Q-statistics with a significance level of P < 0.10.24 P <0.05 was considered as significant for all other analyses. Prespecified subgroup analyses were performed according to transplant (ie, heart, lung, kidney) and vaccine type (ie, mRNA-1273 versus BNT162b2). Subgroup analyses were also performed according to study threshold and according to correlates of protection according to SARS-CoV-2 variant (wild type, alpha, delta).25 The presence of publication bias was only assessed if >10 studies were available for each outcome.24 Solid organ transplant recipients were analyzed separately from hematopoietic stem cell transplant (HSCT) recipients.
For analysis of associations, the odds ratios (ORs) of predictive factors for humoral response could not be pooled as the multivariable logistic regression performed across studies accounted/adjusted for different covariates. Thus, associations were synthesized qualitatively.
All data are presented with 95% confidence intervals (CIs). All tests were 2-sided and statistical significance was based on the 95% CIs excluding the null.
RESULTS
Study Selection
Of 84 records, 12 studies5-7,25-33 were included encompassing 1257 patients (Figure S1, SDC, http://links.lww.com/TP/C610).34 Along with editorials and commentaries, 3 studies were excluded in the full-text phase. One study reported prevalence of humoral response in a single cohort of transplant recipients receiving 2 or 3 doses,35 another study did not adequately characterize its study populations (ie, various data such as time since transplant and immunosuppressive regimens were not reported) and did not meet threshold for quality,36 and another study applied inconsistent thresholds in their data reporting.37
Study Characteristics
Study and patient characteristics are summarized in Table 1. All patients were aged >18 y, and the range of medians/means was 54.8 to 66.6 y. In most (8/10) studies, the third dose was administered 2 mo after the second dose (range of medians/means, 21–168 d), with outcome measurement occurring 1 mo afterward (range of medians/means, 14–52 d). All studies used defined thresholds to determine humoral and cellular response, although assays and thresholds varied across studies (Table 2). Transplant types included 878 kidney (69.8%), 147 heart (11.7%), 101 liver (8.0%), 42 allogenic HSCT (aHSCT; 3.3%), 30 lung (2.4%), and 59 other transplants (4.7%), such as pancreas and combined transplants.
TABLE 1. Characteristics of included studies
Author (location) Study design Age, y (%M) Transplant Time since transplant (y) Vaccine Time since second dose Time of assessment Humoral assay (threshold)
Hall et al7(Canada) and Kumar et al27 (Canada) RCT 66.6 [63.3–71.4] (74%) 29 Kidney29 Lung20 Liver18 Heart24 Other 3.16 [1.71–6.12] mRNA-1273 60 d 30 d Antispike-RBD IgG(≥100 U/mL)
Kamar et al28(France) Single arm 57.6 ± 17.2 (69.7%) 76 Kidney12 Liver7 Other 8.1 ± 1.0 BNT162b2 61 ± 1 d 30 d Anti–SARS-CoV-2(S/Co>1.1)
Del Bello et al29(France) Single arm 54.8 ± 12.8 (65.4%) 277 Kidney69 Liver33 Heart17 Other 6.8 ± 8.1 BNT162b2 59 d [47–67] 30 d Anti–SARS-CoV-2(S/Co>1.1)
Chavarot et al5 (France) Single arm 63.5 [51–72] (58%) 62 Kidney 4.0 [2.1–6.6] BNT162b2 69.5 d [40–84] 28 [28–33] SARS-CoV-2 spike protein(≥50 AU/mL)
Peled et al30(Israel) Single arm 61.0 [49.8–68.0] (71%) 96 Heart 6.3 [3.5–13.6] BNT162b2 168 ± 18 d 18 d Antispike-RBD IgG(>12.6 geometric titer)
Westhoff et al31 (Germany) Single arm 59.5 ± 11.6 (80%) 10 Kidney 5.2 ± 4.7 mRNA-1273 64 ± 15 d 14 d SARS-CoV-2 spike and nucleocapsid protein(≥0.75 U/mL)
Benotmane et al6(France) Single arm 57.6 [49.6–66.1] (61.6%) 159 Kidney 5.3 [1.9–11.1] mRNA-1273 51 d [48–59] 28 d [28–32] Antispike-RBD IgG(>50 AU/mL)
Redjoul et al32 (France) Single arm 59 [50–64] (65%) 42 Allogenic HSCT <1 y (22)>1 y (20) BNT162b2 51 ± 22 d 26 ± 6 d Antispike-RBD IgG(≥4160 AU/mL)
Masset et al33(France) Single arm 63.7 ± 11.7 (63.2%) 124 Kidney12 Other 9.5 ± 8.1 BNT162b2 50 d ± NR 28 d [27–33] Antispike-RBD IgG(>250 UI/L)
Bertrand et al34 (France) Single arm 63.6 ± 15.7 (60%) 80 Kidney 7.3 [3.4–14.1] BNT162b2 67.5 d[57–70] At least 28 d Antispike-RBD IgG(>50 AU/mL)
Massa et al26 (France) Single arm 58 [47.1–66.1] (72.1%) 61 Kidney 4.5 [1.8–11.3] BNT162b2 21 d 28 d Antispike-RBD IgG(>50 AU/mL)
Interquartile range is presented in square brackets.
AU, Arbitrary units; HSCT, hematopoietic stem cell transplantation; IgG, immunoglobulin G; IU, international units; %M, % male; NR, not reported; RBD, receptor-binding domain; RCT, randomized controlled trial; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; S/Co, signal to cutoff; U, units.
TABLE 2. Detailed description of humoral and cellular assays across studies
Study Cellular assay(s) Threshold (s) Humoral assay (s) Antibody measuring Threshold (s) positive
Del Bello et al29 Not reported Not reported Viral Ab titer: Wantai microplate ELISA Anti–SARS-CoV-2 antibodies S/Co>1.1
Hall et al7 andKumar et al27 Intracellularcytokine stainingMarkers analyzed: IFN-gamma/IL-2Cells analyzed: CD4+, CD8+ cells <0.01(absolute values were reported)a Viral Ab titer: anti–SARS-CoV-2 spike enzyme immunoassay (Roche Elecsys) Antispike-RBD IgG ≥100 U/mL
Neutralizing antibody test:(i) SARS-CoV-2 surrogate virus neutralization test(ii) SARS-CoV-2 pseudovirus neutralization assay (Kumar et al27 only) (i) Antispike-RBD IgG(ii) N/A (i) ≥30%(ii) Absence of 50% neutralization with undiluted serum
Kamar et al28 Not performed Not reported Viral Ab titer: Wantai microplate ELISA Anti–SARS-Cov-2 antibodies S/Co>1.1
Redjoul et al32 Not performed Not reported Viral Ab titer:Abbott Architect SARS-CoV-2 IgG Quant II assay (Abbott, Sligo, Ireland) Antispike-RBD IgG ≥4160 AU/mL
Westhoff et al31 Intracellularcytokine stainingMarkers analyzed: IL-2/IL-4/IFN-γ/TNF-α/GrzBCell analyzed: CD4+CD154+CD127+and CD8+ CD137+ >0% Viral Ab titer: measure using Elecsys Anti–SARS-CoV-2-S (Roche, Mannheim, Germany) SARS-CoV-2 spike and nucleocapsid protein ≥0.75 U/mL
Neutralizing antibody test: VSVG (FLuc) pseudovirus system bearing the SARS-CoV-2 spike protein SARS-CoV-2 spike protein ≥0 AU
Benotmane et al6 Not reported Not reported Viral Ab titer: ARCHITECT IgG IIQuant test (Abbott) Antispike-RBD IgG >50 AU/mL
Chavarot et al5 Not reported Not reported Viral Ab titer: SARS-CoV-2 IgG II Quant antibody test (Abbott, USA) SARS-CoV-2 spike protein ≥50 AU/mL
Masset et al33 Not performed Not reported Viral Ab titer:(i) Chemiluminescent microparticle immunoassay (Abbott Architect)(ii) Chemiluminescence immunoassay (Siemens Atellica)(iii) Electrochemiluminescence immunoassay (Roche Elecsys) Antispike-RBD IgG >250 IU/L
Peled et al30 T lymphocytes(T-cell markers are not described) Not reported Viral Ab titer: in-house ELISA Antispike-RBD IgG >12.6 geometric titer
Neutralizing antibody test: GFP reporter-based VSVG backbone coated with the SARS-CoV-2 spike protein SARS-CoV-2 spike protein >1024 geometric titer
Bertrand et al34 Intracellularcytokine stainingMarkers analyzed: IFN-gammaCells analyzed: CD3+ T cells S1 > 25 SFC/106 CD3+ T cellsS2 > 40 SFC/106CD3+ T cells Viral Ab titer: ARCHITECT IgG II Quant test (Abbott) Antispike-RBD IgG >50 AU/mL
Massa et al26 ELISpot Path Human IFN-g (SARS-CoV-2, S1 scan+S2N+SNMO) ALP assay (MABTECH) measured IFN-gamma–secreting SARS-CoV-2–specific T cells per 106 PBMCs None Viral Ab titer: Abbott SARS-CoV-2 IgG assay Antispike-RBD IgG >50 AU/mL
Neutralizing antibody test: V-PLEX! SARS-CoV-2 Panel 7 (ACE2) Kit (MSD) was used to measure binding of the SARS-CoV-2 full-length Spike protein to soluble ACE2 receptor Anti–full-length SARS-CoV-2 spike protein antibodies Noneb
a Threshold from Westhoff et al31 was applied in meta-analysis.
b Threshold from Kumar et al27 was applied in meta-analysis.
Ab, antibody; ACE2, angiotensin-converting enzyme 2; ALP, Alkaline phosphatase; AU, Arbitrary units; ELISA, enzyme-linked immunosorbent assay; GFP, Green fluorescent protein; IFN, interferon; IgG, immunoglobulin G; IL, interleukin; IU, international units; MSD, Meso scale diagnostics; N/A, not available; PBMC, peripheral blood mononuclear cells; RBD, receptor-binding domain; S1, 15-mer peptide pools spanning the sequence of SARS-CoV-2 spike protein S spanning the N-terminal; S2, 15-mer peptide pools spanning the sequence of SARS-CoV-2 spike protein S spanning the C-terminal; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; SFC, spot-forming cells; S/Co, signal to cutoff; TNF, tumor necrosis factor; U, units; VSVG, vesicular stomatitis virus glycoprotein.
Humoral Response
Although all studies measured response against SARS-CoV-2 spike protein, assays differed across studies—depending on the manufacturer and the kit sensitivity, the threshold cutoff varied. When the prevalence of humoral response after 3 doses was pooled, high heterogeneity (I2 = 97.6%) was observed (Figure S2, SDC, http://links.lww.com/TP/C610). This heterogeneity was found to be because of one study,5 which reported 6.4% response in a cohort of kidney transplant recipients (KTRs) receiving belatacept, and was removed as an outlier. Across 7 studies7,26,28-30,33,34 (801 patients), the prevalence of humoral response after 3 doses was 66.1% (95% CI, 62.8%-69.4%, I2 = 0%; Figure 1A), with similar rates found across studies of kidney (61.7%; 95% CI, 53.7%-69.7%; I2 = 0%)26,34 and heart (66.7%; 95% CI, 57.2%-76.1%)30 transplant recipients (test for subgroup differences, P = 0.56; Figure 1B). Prevalence of humoral response according to mRNA vaccine was higher in transplant recipients who received 3 doses of BNT162b2 (66.9%; 95% CI, 63.5%-70.3%; I2 = 0%)26,28-30,33,34 versus 3 doses of mRNA-1273 (55.0%; 95% CI, 42.4%-67.6%; I2 = N/A)7 (test for subgroup differences, P = 0.07; Figure 1C). Across 9 studies6,7,26,28-31,33,34 (747 patients), humoral response to a third dose after humoral nonresponse to 2 doses was observed in 45.7% of patients (95% CI, 41.8%-49.7%; I2 = 13.1%; Figure 2A), with similar rates found across studies of kidney (43.2%; 95% CI, 33.6%-52.9%; I2 = 45.5%)6,26,31,34 and heart (54.3%; 95% CI, 42.6%-66.0%)30 (test for subgroup differences, P = 0.30; Figure 2B). In aHSCT recipients, prevalence was 47.6% (95% CI, 32.5%-62.7%).32 In one randomized controlled trial (RCT),7 a greater number of patients in the mRNA-1273 group (55%) demonstrated humoral response compared with the placebo group (18%; risk ratio = 3.1; 95% CI, 1.7-5.8; P < 0.001). No differences in humoral response after humoral nonresponse to 2 doses were observed according to mRNA vaccine (mRNA-12736,7,31 49.6%; 95% CI, 43.0%-56.1%; I2 = 0% versus BNT162b226,28-30,32-34 43.9%; 95% CI, 38.6%-49.2%; I2 = 15.6%; test for subgroup differences, P = 0.19; Figure 2C). Notably, humoral response after 3 doses was observed in 52 of 64 patients (81.2%; 95% CI, 71.7%-90.8%) who had a weak humoral response and 26 of 95 patients (27.4%; 95% CI, 18.4%-36.3%) who had no humoral response to 2 doses.6 Prevalence of humoral response after 3 doses of any mRNA SARS-CoV-2 vaccine did not vary according to study threshold (test for subgroup differences, P = 0.43; Figure S3, SDC, http://links.lww.com/TP/C610). Prevalence of humoral response after 3 doses also did not vary according to correlates of protection for the wild type, alpha variant, and delta variant (test for subgroup difference P = 0.17; Figure S4, SDC, http://links.lww.com/TP/C610).
FIGURE 1. Prevalence of humoral response after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients (A, top panel), also according to transplant type (B, middle panel), and mRNA vaccine (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
FIGURE 2. Prevalence of humoral response after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients that did not display a humoral response to 2 doses of an mRNA SARS-CoV-2 vaccine (A, top panel), also according to transplant type (B, middle panel), and mRNA vaccine (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
Cellular Response
Five studies reported cellular response.7,26,30,31,34 Intracellular cytokine staining was used as a marker to measure cellular response, and markers analyzed within the studies largely varied. One study26 was not able to be included in the meta-analysis because it reported the number of SARS-CoV-2–specific T cells per peripheral blood mononuclear cells. The percentage of patients achieving cellular response after 3 doses (75.3%; 95% CI, 66.6%-83.9%; I2 = 24.5%)7,30,31,34 (Figure 3A) was significantly higher (P = 0.0001) compared with after 2 doses (49.3%; 95% CI, 39.5%-59.1%; I2 = 0%)30,31,34 (Figure 3B). Among patients with cellular nonresponse to 2 doses, 57.8% (95% CI, 30.0%-85.6%; I2 = 64.5%)30,31,34 displayed cellular response after 3 doses (Figure 3C).30,31,34 In one RCT,7 median SARS-CoV-2–specific T-cell counts after 3 doses of mRNA-1273 were greater than placebo (432 versus 67 cells per 106 CD4+ T cells, 95% CI for the between-group difference, 46–986).
FIGURE 3. Prevalence of cellular response after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients (A, top panel), after 2 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients (B, middle panel), and after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients that did not display a cellular response to 2 doses (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
Neutralizing Antibody Response
Four studies reported neutralizing antibodies.7,26,30,31 Although all studies measured response against SARS-CoV-2 spike protein, detected spike protein variants varied on the basis of the manufacturer. Furthermore, although 1 study27 used a standard enzyme-linked immunosorbent assay strategy to measure neutralizing antibody response, 2 studies30,31 used vesicular stomatitis virus glycoprotein pseudovirus expressed luciferase or Green fluorescent protein to detect neutralizing antibody responses. One study26 used a high throughput alternative to the traditional neutralization assay that identified over 8 variants of receptor-binding domain antigens as supposed to the traditional 3 variants. When the percentage of patients demonstrating a neutralizing antibody response above study threshold after 3 doses was pooled, high heterogeneity (I2 = 95.7%) was observed (Figure 4A). This heterogeneity was found to be because of one study,26 which used an alternative to the traditional neutralization assay, and was removed as an outlier. With this study removed, the percentage of patients demonstrating a neutralizing antibody response above study threshold after 3 doses was 60.9% (95% CI, 53.2%-68.6%; I2 = 0%; Figure 4B). In one RCT,7 neutralizing antibody positivity was observed in 60% of patients after 3 doses of mRNA-1273 versus 25% receiving placebo, respectively (risk ratio = 2.4; 95% CI, 1.5-4.0).
FIGURE 4. Prevalence of neutralizing antibody response above study threshold after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients with outlier study (A, top panel) and without outlier study (B, middle panel) and prevalence of neutralizing antibody response above study threshold after 3 doses of any mRNA SARS-CoV-2 vaccine in transplant recipients according to SARS-CoV-2 variant and study (C, bottom panel). CI, confidence interval; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
Two studies26,27 quantified the percentage of transplant recipients mounting a neutralizing antibody response to SARS-CoV-2 VOC, such as the Alpha,26,27 Beta,26,27 Delta,26,27 and Gamma26 variants. Although the prevalence across studies varied widely (I2 = 99.1%), the percentage of patients mounting a neutralizing antibody response above study threshold according to SARS-CoV-2 variant was similar within each study (Figure 4C). One RCT27 found a smaller percentage of transplant recipients displayed a neutralizing antibody response above threshold after 2 doses of mRNA-1273 to all 3 variants versus wild-type virus (differences in proportions: Alpha, 19% [95% CI, 8%-31%]; Beta, 21% [95% CI, 10%-32%]; and Delta, 16% [95% CI, 4% to 28%]). After 3 doses of mRNA-1273 vaccine, the number of transplant recipients displaying a neutralizing antibody response above threshold was greater for each variant compared with placebo (differences in proportions: Alpha variant 45% [95% CI, 28%-61%], Beta variant 38% [95% CI, 21%-55%], and Delta variant 37% (95% CI, 20%-55%] via surrogate virus neutralization test and wild-type virus 35% [95% CI, 17%-51%], Alpha variant 36% [95% CI, 17%-51%], Beta variant 31% [95% CI, 15%-46%], and Delta variant 37% [95% CI, 19%-53%] via pseudovirus neutralization assay, with P < 0.001 for all comparisons).
Factors Associated With Humoral Response
Five studies investigated factors associated with humoral response.5,29,30,32,33 These studies could not be pooled as the multivariable logistic regression performed across these studies accounted and adjusted for different covariates. In 2 multivariable analyses,29,30 mycophenolate use was associated with reduced odds of response across a cohort of kidney, liver, heart, and lung transplant recipients (OR = 0.28; 95% CI, 0.14-0.54; P < 0.001)29 and in heart transplant recipients alone (OR = 0.1; 95% CI, 0.01-0.49; P = 0.01).30 Additionally, in another multivariable analysis,29 belatacept was associated with reduced odds of humoral response across a cohort of kidney, liver, heart, and lung transplant recipients (OR = 0.14; 95% CI, 0.04-0.46; P = 0.001), and, in one study5 of 62 belatacept-treated KTRs, only 4 patients (6.4%) developed anti–SARS-CoV-2 immunoglobulin G with low antibody titers (median 209, interquartile range 20–409 AU/mL).
Across multivariable analyses, lymphocyte count <1500 mm3 (OR = 0.31; 95% CI, 0.10-0.63; P = 0.004)33 and lymphocyte count <1000 mm3 (OR = 0.20; 95% CI, 0.04-0.65; P = 0.02),26 antiproliferatives (OR = 0.06; 95% CI, 0.01-0.41; P = 0.01),26 triple immunosuppression (OR = 0.42; 95% CI, 0.21–0.86; P = 0.02),29 and higher CRP (OR = 0.88; 95% CI, 0.79-0.96; P = 0.02)30 were associated with nonhumoral response, whereas allograft function (OR = 1.03; 95% CI, 1.01-1.06; P = 0.02)33 and higher eGFR (OR = 1.03; 95% CI, 1.01-1.06; P = 0.04)30 were associated with an increased likelihood of achieving a positive antibody response in solid organ transplant recipients. Conflicting results were found with respect to age29,30 and sex.30,33 Finally, in a multivariable analysis of aHSCT recipients,32 only a B-cell count >0.25 g/L was associated with a humoral response (OR = 7.1; 95% CI, 1.5-34.1; P < 0.001).
Adverse Events
No serious (grade 3–4) adverse events or acute rejection episodes were reported across studies. Most common side effects included pain at injection site (62.4%), fatigue (18.5%), and headache (9.6%) with BNT162b226,30 and pain at injection site (76.7%), fatigue (43.3%), and myalgias (28.3%) with mRNA-1273 vaccine.7 One RCT7 found local and systematic events were more common after mRNA-1273 versus placebo.
Risk of Bias
All included noncomparative data, such as humoral response in a single group, were of good quality5-7,25-33 (Table S4, SDC, http://links.lww.com/TP/C610), and all comparative data were of low risk of bias7,27 (Table S5, SDC, http://links.lww.com/TP/C610). Funnel plot asymmetry was not detected (Figure S5, SDC, http://links.lww.com/TP/C610). When poor-quality studies36 were included in the analysis, significant heterogeneity was detected (Figure S6, SDC, http://links.lww.com/TP/C610).5-7,26-34
GRADE Assessments
Table S6 (SDC, http://links.lww.com/TP/C610) summarizes the certainty levels of all outcomes. Most outcomes had a low level (67%) or moderate level of certainty (17%). Most common reasons for decreasing certainty of the estimate included varying thresholds across studies and imprecision of the estimate.
DISCUSSION
Across solid organ and hematopoietic transplant recipients, this meta-analysis confirms 66.1% of transplant recipients overall and 45.9% of recipients without humoral response to 2 doses display a humoral response after 3 doses of mRNA SARS-CoV-2 vaccine, which did not significantly vary across transplant types. As humoral response has been shown to correlate with neutralizing antibody titers38 and a relationship between neutralization level after SARS-CoV-2 vaccination and protection against COVID-19 has been demonstrated,9 this suggests humoral response to vaccination may be clinically relevant. Given approximately half of transplant recipients do not display humoral response after 2 doses,1 a third dose of mRNA SARS-CoV-2 vaccine should be strongly considered. However, despite this, roughly a third of transplant recipients do not display humoral response after 3 doses, which suggests that this patient population remains at risk for infection.
Notably, humoral response after 3 doses was markedly attenuated with immunosuppression, as patients receiving triple immunosuppression, antiproliferatives, or belatacept had significantly reduced odds of developing a humoral response. Specifically, with belatacept use, only 6.4% of KTRs developed immunity. Thus, these patients seem to be at particular risk. Similarly, low lymphocyte counts were associated with markedly reduced response. Given the increased risk for progression to severe COVID-19 in high-risk patients who are vaccinated but are not expected to mount an adequate immune response, current guidelines recommend measures in addition to vaccination.39 These measures include primary prevention, such as masking and social distancing, and secondary prevention, such as consideration of the administration of anti–SARS-CoV-2 monoclonal antibodies upon confirmation of SARS-CoV-2 infection.39 Furthermore, in nonresponders and weak responders to 3 doses of vaccine, a fourth dose of SARS-CoV-2 vaccine may lead to an improved serological response vaccine.40
Prevalence of humoral response was found to be lower in transplant recipients after 3 doses of mRNA-1273 versus BNT162b2. However, analysis of 1647 healthcare workers demonstrated higher antibody titers in participants vaccinated with 2 doses of mRNA-1273 compared with those vaccinated with BNT162b2.41 Given only 1 study was included in the mRNA-1273 group in our analysis and other analyses did not demonstrate a difference in efficacy between the 2 vaccines, it is likely that this result will be altered by future investigation.
Prevalence of humoral response did not significantly vary according to study thresholds. Given the wide variety of thresholds reported across studies, this result may change because more investigations are completed using similar thresholds to those already present in the literature. It is important to note that correlate of protection likely exists on a spectrum42 and that higher antibody levels are needed to achieve the same level of protection after vaccination compared with unvaccinated individuals with prior infection.43 Thus, correlates of protection must be taken from vaccinated populations. Unfortunately, none of the studies in this review used thresholds that were based on clinical correlates of protection in vaccinated populations.43 Individual studies7 used thresholds that conferred 50% protective neutralization in previous clinical studies of SARS-CoV-2–infected individuals. Other studies6,32 used thresholds that correlated with in vitro neutralization of SARS-CoV-2. Therefore, the use of common, established thresholds that are proven to correlate with clinical protection in vaccinated individuals are required in future investigations of SARS-CoV-2 vaccine efficacy in transplant populations.
Prevalence of humoral response after 3 doses did not vary according to published correlates of protection for the wild type, alpha variant, and delta variant. This result suggests that mRNA SARS-CoV-2 vaccines provided transplant recipients protection from VOC, which is significant considering the higher risk of severe disease and hospitalization with certain variants, such as the Delta variant.44 Future research is needed to establish correlates for protection for other VOC to guide vaccine administration and policy.42
A third dose was associated with increased cellular and neutralizing antibody response. Specifically, the percentage of transplant recipients displaying a cellular response above study threshold after 3 doses was significantly higher compared with after 2 doses, with higher SARS-CoV-2–specific T-cell counts compared with placebo. Neutralizing antibody levels were also significantly increased after 3 doses of mRNA SARS-CoV-2 vaccine. As neutralization level is highly predictive of immune protection9 and T cells reduce viral loads and disease even when neutralizing antibody levels are low,10 these data suggest that a third dose may provide at least partial protection and reduce the likelihood of severe COVID-19 and infection in transplant recipients. Notably, the proportion of transplant recipients demonstrating a neutralizing antibody response above threshold to VOCs was higher after 3 doses compared with 2 doses, with a greater proportion of patients demonstrating a neutralizing antibody response to VOCs compared with placebo. Given neutralizing antibody positivity after 2 doses of mRNA-1273 vaccine was low for all 3 SARS-CoV-2 VOCs (Alpha, Beta, and Delta), these data suggest that a third dose may at least provide partial protection of transplant recipients to VOCs.
Across select cellular and neutralizing antibody outcomes, heterogeneity was observed. With respect to cellular response, this may be because of the use of different assays across studies. For example, the study34 reporting the lowest prevalence measured a single cytokine whereas other studies30,31 measured multiple cytokines in their cellular assays. In addition, heterogeneity was also observed with neutralizing antibody response to 3 doses. This heterogeneity was found to be because of one study,26 which only allowed for 21 d between second and third dose administration versus >60 d for the other studies.7,30 Given previous investigations have demonstrated that neutralizing antibody titers peak around 1 mo after infection45,46 and decrease significantly thereafter,45,47 high neutralizing antibody titers at the time of third dose administration (because of a short interval between second and third doses) may explain this study’s reduced neutralizing antibody titers after 3 doses. Additionally, this study22 used a commercially available high throughput alternative to the traditional neutralization assay that identified over 8 variants of receptor-binding domain antigens as supposed to the traditional 3 variants.23 These differences could largely contribute to the heterogeneity in the results observed. Thus, standardization of cellular response assays, particularly with respect to measured cytokines, and neutralizing antibody assays could be considered. Future investigation of the cellular and neutralizing antibody response is required.
Interestingly, no consensus in relation to the tests used for cellular assays, neutralizing assays, or humoral assays was observed among the studies included in our analysis. Considering the assays differed according to measured cytokines (cellular assays) and had variable cutoffs (humoral and neutralizing assays), emphasis should be made toward standardizing the tools used to measure cellular and humoral responses. Alternatively, information should be provided on how the results from 1 assay correlate with other standard assays used. Such an effort would enable direct comparison and contrast of results across studies, which would minimize heterogeneity and allow for more robust recommendations.
A paucity of literature exists regarding humoral and cellular response in other populations after 3 doses of mRNA SARS-CoV-2 vaccine. Nonetheless, when compared with existing literature on other patient populations, transplant recipients seem to have a lower prevalence of humoral response. For example, in a cohort of patients on maintenance hemodialysis,48 92.4% of patients displayed humoral response after 3 doses compared with 66.1% of transplant recipients. Additionally, similar to transplant recipients with absent or minimal humoral response to 2 doses, small case series have demonstrated that a third dose of vaccine augments humoral response in patients with rheumatoid arthritis with absent or minimal humoral response to 2 doses.49 Finally, meta-analyses50,51 on the efficacy of 2 doses in patients with immune-mediated inflammatory diseases have demonstrated an association between attenuation of humoral response and anti-CD20 (rituximab) and anticytotoxic T lymphocyte–associated antigen therapies. This attenuation of humoral response has also been noted in transplant recipients receiving antimetabolites after 2 doses of vaccine.1 As seen in our review, some immunosuppressive therapies were associated with reduced humoral response after 3 doses. Future investigation is required to determine interventions that can promote serological response in these populations, with some investigations already planned.52
Limitations of this review include (1) the lack of literature reporting on cellular response and neutralizing antibodies and (2) the various assays and thresholds used across studies. Further studies are needed for standardization of SARS-CoV-2 serology and to correlate clinical protection with humoral and cellular response in this population.53 Additionally, the inclusion of the poor-quality study resulted in significant heterogeneity. This study was judged to be of poor quality as, among other factors, this study did not report important data such as time since transplant and immunosuppressive regimens. As seen from our analyses, immunosuppressive regimens seem to greatly influence humoral response after 3 doses of vaccine. Thus, unreported factors, such as immunosuppressive regimen, may lead to the reduced humoral response reported in this poor-quality study. Finally, in this review, the duration of response and the effect of previous immunization, such as viral vector, were not able to be assessed.
In conclusion, a third dose SARS-CoV-2 mRNA vaccine should be strongly considered in transplant recipients. Future study and establishment of standardized assays and clinically relevant thresholds for humoral and cellular response are required. As this population seems to remain at risk for infection post 3 doses, investigations regarding the efficacy of future doses should be considered, and patients and healthcare providers should remain vigilant regarding exposure to infection.
Supplementary Material
The authors declare no funding or conflicts of interest.
All authors met the International Committee of Journal Editors criteria for authorship. A.J.M.B., H.B.M., W.C., and D.S.A. gave substantial contributions to the conception or design of the work, the acquisition, analysis, and interpretation of data for the work; drafting the work and revising it critically for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. R.S. and C.A.B. gave substantial contributions to the conception or design of the work; drafting the work and revising it critically for important intellectual content; final approval of the version to be published; and agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
This study met the definition of Institutional Review Board exempt research.
All additional data can be obtained by contacting the corresponding author.
Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com).
==== Refs
REFERENCES
1. Boyarsky BJ Werbel WA Avery RK . Antibody response to 2-dose SARS-CoV-2 mRNA vaccine series in solid organ transplant recipients. JAMA. 2021;325 :2204–2206.33950155
2. Wadei HM Gonwa TA Leoni JC . COVID-19 infection in solid organ transplant recipients after SARS-CoV-2 vaccination. Am J Transplant. 2021;21 :3496–3499.33890410
3. Ljungman P de la Camara R Mikulska M . COVID-19 and stem cell transplantation; results from an EBMT and GETH multicenter prospective survey. Leukemia. 2021;35 :2885–2894.34079042
4. Muñoz Serrano A Arias A Moreno-Torres V . Coronavirus disease 2019 (COVID-19) in solid organ transplant recipients: a case-control study. Ann Transplant. 2021;26 :e933152.34764235
5. Chavarot N Morel A Leruez-Ville M . Weak antibody response to three doses of mRNA vaccine in kidney transplant recipients treated with belatacept. Am J Transplant. 2021;21 :4043–4051.34431207
6. Benotmane I Gautier G Perrin P . Antibody response after a third dose of the mRNA-1273 SARS-CoV-2 vaccine in kidney transplant recipients with minimal serologic response to 2 doses. JAMA. 2021;326 :1063–1065.34297036
7. Hall VG Ferreira VH Ku T . Randomized trial of a third dose of mRNA-1273 vaccine in transplant recipients. N Engl J Med. 2021;385 :1244–1246.34379917
8. Nixon DF Ndhlovu LC . Vaccine breakthrough infections with SARS-CoV-2 variants. N Engl J Med. 2021;385 :e7.
9. Khoury DS Cromer D Reynaldi A . Neutralizing antibody levels are highly predictive of immune protection from symptomatic SARS-CoV-2 infection. Nat Med. 2021;27 :1205–1211.34002089
10. Shroff RT Chalasani P Wei R . Immune responses to two and three doses of the BNT162b2 mRNA vaccine in adults with solid tumors. Nat Med. 2021;27 :2002–2011.34594036
11. Lopez Bernal J Andrews N Gower C . Effectiveness of Covid-19 vaccines against the B.1.617.2 (Delta) variant. N Engl J Med. 2021;385 :585–594.34289274
12. Tenforde MW Self WH Naioti EA ; IVY Network Investigators; IVY Network. Sustained effectiveness of Pfizer-BioNTech and Moderna vaccines against COVID-19 associated hospitalizations among adults—United States, March-July 2021. MMWR Morb Mortal Wkly Rep. 2021;70 :1156–1162.34437524
13. Callaway E . Delta coronavirus variant: scientists brace for impact. Nature. 2021;595 :17–18.34158664
14. Page MJ McKenzie JE Bossuyt PM . The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372 :n71.33782057
15. McGowan J Sampson M Salzwedel DM . PRESS peer review of electronic search strategies: 2015 guideline statement. J Clin Epidemiol. 2016;75 :40–46.27005575
16. Ouzzani M Hammady H Fedorowicz Z . Rayyan—a web and mobile app for systematic reviews. Syst Rev. 2016;5 :210.27919275
17. Sterne JAC Savović J Page MJ . RoB 2: a revised tool for assessing risk of bias in randomised trials. BMJ. 2019;366 :l4898.31462531
18. Guyatt G Oxman AD Akl EA . GRADE guidelines: 1. Introduction-GRADE evidence profiles and summary of findings tables. J Clin Epidemiol. 2011;64 :383–394.21195583
19. Borges Migliavaca C Stein C Colpani V ; Prevalence Estimates Reviews – Systematic Review Methodology Group (PERSyst). How are systematic reviews of prevalence conducted? A methodological study. BMC Med Res Methodol. 2020;20 :96.32336279
20. Iorio A Spencer FA Falavigna M . Use of GRADE for assessment of evidence about prognosis: rating confidence in estimates of event rates in broad categories of patients. BMJ. 2015;350 :h870.25775931
21. Wallace B Dahabreh I Trikalinos T . Closing the gap between methodologists and end-users: R as a computational back-end. J Stat Softw. 2012;49 :1–15.
22. DerSimonian R Laird N . Meta-analysis in clinical trials. Control Clin Trials. 1986;7 :177–188.3802833
23. Higgins JP Thompson SG Deeks JJ . Measuring inconsistency in meta-analyses. BMJ. 2003;327 :557–560.12958120
24. Higgins JPT Thomas J Chandler J . Cochrane Handbook for Systematic Reviews of Interventions Version 6.0. Cochrane; 2019.
25. Goldblatt D Fiore-Gartland A Johnson M . Towards a population-based threshold of protection for COVID-19 vaccines. Vaccine. 2022;40 :306–315.34933765
26. Massa F Cremoni M Gérard A . Safety and cross-variant immunogenicity of a three-dose COVID-19 mRNA vaccine regimen in kidney transplant recipients. EBioMedicine. 2021;73 :103679.34763205
27. Kumar D Ferreira VH Hall VG . Neutralization of SARS-CoV-2 variants in transplant recipients after two and three doses of mRNA-1273 vaccine: secondary analysis of a randomized trial. Ann Intern Med. 2022;175 :226–233.34807716
28. Kamar N Abravanel F Marion O . Three doses of an mRNA COVID-19 vaccine in solid-organ transplant recipients. N Engl J Med. 2021;385 :661–662.34161700
29. Del Bello A Abravanel F Marion O . Efficiency of a boost with a third dose of anti-SARS-CoV-2 messenger RNA-based vaccines in solid organ transplant recipients. Am J Transplant. 2022;22 :322–323.34331842
30. Peled Y Ram E Lavee J . Third dose of the BNT162b2 vaccine in heart transplant recipients: immunogenicity and clinical experience. J Heart Lung Transplant. 2022;41 :148–157.34565682
31. Westhoff TH Seibert FS Anft M . A third vaccine dose substantially improves humoral and cellular SARS-CoV-2 immunity in renal transplant recipients with primary humoral nonresponse. Kidney Int. 2021;100 :1135–1136.34509489
32. Redjoul R Le Bouter A Parinet V . Antibody response after third BNT162b2 dose in recipients of allogeneic HSCT. Lancet Haematol. 2021;8 :e681–e683.34487683
33. Masset C Kerleau C Garandeau C . A third injection of the BNT162b2 mRNA COVID-19 vaccine in kidney transplant recipients improves the humoral immune response. Kidney Int. 2021;100 :1132–1135.34474075
34. Bertrand D Hamzaoui M Lemée V . Antibody and T cell response to SARS-CoV-2 messenger RNA BNT162b2 vaccine in kidney transplant recipients and hemodialysis patients. J Am Soc Nephrol. 2021;32 :2147–2152.34112706
35. Stumpf J Tonnus W Paliege A . Cellular and humoral immune responses after 3 doses of BNT162b2 mRNA SARS-CoV-2 vaccine in kidney transplant. Transplantation. 2021;105 :e267–e269.34342963
36. Marlet J Gatault P Maakaroun Z . Antibody responses after a third dose of covid-19 vaccine in kidney transplant recipients and patients treated for chronic lymphocytic leukemia. Vaccines (Basel). 2021;9 :1055.34696163
37. Werbel WA Boyarsky BJ Ou MT . Safety and immunogenicity of a third dose of SARS-COV-2 vaccine in solid organ transplant recipients: a case series. Ann Intern Med. 2021;174 :1330–1332.34125572
38. Rubio-Acero R Castelletti N Fingerle V ; KoCo19 study team. In search of the SARS-CoV-2 protection correlate: head-to-head comparison of two quantitative S1 assays in pre-characterized oligo-/asymptomatic patients. Infect Dis Ther. 2021;10 :1505–1518.34137000
39. National Institutes of Health. Anti-SARS-CoV-2 monoclonal antibodies. 2022. Available at https://www.covid19treatmentguidelines.nih.gov/therapies/anti-sars-cov-2-antibody-products/anti-sars-cov-2-monoclonal-antibodies. Accessed June 1, 2022.
40. Midtvedt K Vaage JT Heldal K . Fourth dose of the SARS-CoV-2 vaccine in kidney transplant recipients with previously impaired humoral antibody response. Am J Transplant. 2022;22:2704–2706.
41. Steensels D Pierlet N Penders J . Comparison of SARS-CoV-2 antibody response following vaccination with BNT162b2 and mRNA-1273. JAMA. 2021;326 :1533–1535.34459863
42. Perry J Osman S Wright J . Does a humoral correlate of protection exist for SARS-CoV-2? A systematic review. PLoS One. 2022;17 :e0266852.35395052
43. Wei J Pouwels KB Stoesser N ; COVID-19 Infection Survey team. Antibody responses and correlates of protection in the general population after two doses of the ChAdOx1 or BNT162b2 vaccines. Nat Med. 2022;28 :1072–1082.35165453
44. Twohig KA Nyberg T Zaidi A ; COVID-19 Genomics UK (COG-UK) consortium. Hospital admission and emergency care attendance risk for SARS-CoV-2 delta (B.1.617.2) compared with alpha (B.1.1.7) variants of concern: a cohort study. Lancet Infect Dis. 2022;22 :35–42.34461056
45. Wang K Long QX Deng HJ . Longitudinal dynamics of the neutralizing antibody response to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection. Clin Infect Dis. 2021;73 :e531–e539.32745196
46. Karuna S Li SS Grant S ; HVTN 405/HPTN 1901 Study Team. Neutralizing antibody responses over time in demographically and clinically diverse individuals recovered from SARS-CoV-2 infection in the United States and Peru: a cohort study. PLOS Med. 2021;18 :e1003868.34871308
47. Crawford KHD Dingens AS Eguia R . Dynamics of neutralizing antibody titers in the months after severe acute respiratory syndrome coronavirus 2 infection. J Infect Dis. 2021;223 :197–205.33535236
48. Dekervel M Henry N Torreggiani M . Humoral response to a third injection of BNT162b2 vaccine in patients on maintenance haemodialysis. Clin Kidney J. 2021;14 :2349–2355.34754430
49. Schmiedeberg K Vuilleumier N Pagano S . Efficacy and tolerability of a third dose of an mRNA anti-SARS-CoV-2 vaccine in patients with rheumatoid arthritis with absent or minimal serological response to two previous doses. Lancet Rheumatol. 2022;4 :e11–e13.34723209
50. Jena A Mishra S Deepak P . Response to SARS-CoV-2 vaccination in immune mediated inflammatory diseases: systematic review and meta-analysis. Autoimmun Rev. 2022;21 :102927.34474172
51. Sakuraba A Luna A Micic D . Serologic response to coronavirus disease 2019 (COVID-19) vaccination in patients with immune-mediated inflammatory diseases: a systematic review and meta-analysis. Gastroenterology. 2022;162 :88–108.e9.34599933
52. Yahav D Rozen-Zvi B Mashraki T . Immunosuppression reduction when administering a booster dose of the BNT162b2 mRNA SARS-CoV-2 vaccine in kidney transplant recipients without adequate humoral response following two vaccine doses: protocol for a randomised controlled trial (BECAME study). BMJ Open. 2021;11 :e055611.
53. Perkmann T Perkmann-Nagele N Koller T . Anti-spike protein assays to determine sars-cov-2 antibody levels: a head-to-head comparison of five quantitative assays. Microbiol Spectr. 2021;9 :e0024721.34190591
| 36398334 | PMC9746229 | NO-CC CODE | 2022-12-15 23:21:55 | no | Transplantation. 2023 Jan 4; 107(1):204-215 | utf-8 | Transplantation | 2,022 | 10.1097/TP.0000000000004386 | oa_other |
==== Front
Transplantation
Transplantation
TP
Transplantation
0041-1337
1534-6080
Lippincott Williams & Wilkins Hagerstown, MD
36367927
00026
10.1097/TP.0000000000004383
3
Original Clinical Science—General
Humoral Response to the Fourth BNT162b2 Vaccination and Link Between the Fourth Dose, Omicron Infection, and Disease Severity in Renal Transplant Recipients
Hod Tammy MD 123
Ben-David Aharon MD 123
Mor Eytan MD 12
Olmer Liraz MSc 4
Halperin Rebecca RN, BA 5
Indenbaum Victoria PhD 26
Beckerman Pazit MD 23
Doolman Ram PhD 27
Asraf Keren PhD 27
Atari Nofar PhD 6
Benjamini Ohad MD 28
Lustig Yaniv PhD 26
Grossman Ehud MD 29
Mandelboim Michal PhD 26
Rahav Galia MD, PhD 25
1 Renal Transplant Center, Sheba Medical Center, Tel Hashomer, Israel.
2 Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel.
3 Nephrology Department, Sheba Medical Center, Tel Hashomer, Israel
4 Bio-statistical and Bio-mathematical Unit, The Gertner Institute of Epidemiology and Health Policy Research, Sheba Medical Center, Tel Hashomer, Israel.
5 The Infectious Diseases Unit, Sheba Medical Center, Tel Hashomer, Israel.
6 Central Virology Laboratory, Public Health Services, Ministry of Health and Sheba Medical Center, Tel Hashomer, Israel.
7 The A. Dworman Automated Mega-Laboratory, Sheba Medical Center, Tel Hashomer, Israel
8 Hematology Division, Sheba Medical Center, Tel Hashomer, Israel.
9 Internal Medicine Wing, Sheba Medical Center, Tel Hashomer, Israel.
Correspondence: Tammy Hod, Renal Transplant Center, Sheba Medical Center, Israel 52621. ([email protected]).
08 12 2022
1 2023
08 12 2022
107 1 192203
30 6 2022
29 8 2022
30 8 2022
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
2022
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
Background.
The effectiveness of the fourth BNT162b2 vaccination in reducing the rate and severity of coronavirus disease 2019 (COVID-19) caused by the Omicron variant in renal transplant recipients (RTRs) is unknown.
Methods.
Interviews were conducted with 447 RTRs regarding the status and timing of the fourth vaccination, prior vaccinations, and preceding COVID-19 infection. RTRs with polymerase chain reaction–confirmed COVID-19 infection from December 1, 2021, to the end of March 2022 were considered to have been infected with the Omicron variant and were interviewed to determine their disease severity. In a subgroup of 74 RTRs, the humoral response to the fourth dose was analyzed. In 30 RTRs, microneutralization assays were performed to reveal the humoral response to wild-type, Delta, and Omicron variant isolates before and after the fourth dose.
Results.
Of 447 RTRs, 144 (32.2%) were infected with the Omicron variant, with 71 (49.3%) of the infected RTRs having received the fourth vaccine dose. RTRs who did not receive the fourth dose before the infection had more serious illness. In a subgroup of 74 RTRs, the fourth dose elicited a positive humoral response in 94.6% (70/74), with a significant increase in geometric mean titer for receptor-binding domain immunoglobulin G and neutralizing antibodies (P < 0.001). The humoral responses to the Omicron variant before and after the fourth dose were significantly lower than the responses to the wild-type and the Delta variants.
Conclusions.
Overall, the fourth BNT162b2 dose was effective in reducing the rate and severity of Omicron disease in RTRs, despite the reduced humoral response to the variant.
SDCT
==== Body
pmcINTRODUCTION
Following the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) waves of the Wuhan, Alpha, and then Delta variants, populations in many parts of the world had acquired immunity from previous infection and vaccination. The subsequently appearing Omicron variant nonetheless infected people having this acquired immunity, albeit with a significantly less severe clinical picture compared with that with the previous variants.1 Whether the clinical presentation observed upon infection with Omicron versus previous variants is indeed a result of preexisting host immunity or whether the variant is less virulent is still unknown. For example, in a recent study conducted in South Africa, which included 5144 patients from wave 4 and 11 609 patients from prior waves, severe coronavirus disease of 2019 (COVID-19) outcomes were indeed found to be reduced, but the intrinsically reduced virulence of the Omicron variant only accounted for about a 25% reduced risk of severe hospitalization or death compared with the Delta variant.2 Findings such as these are of particular concern to nonimmune or immunocompromised people with an impaired response to vaccination, who may be severely affected by the Omicron variant.
A few studies have reported a poor humoral response in renal transplant recipients (RTRs) after SARS-CoV-2 mRNA vaccinations.3-6 A study in the Czech Republic found that as many as 80% of RTRs failed to generate a positive humoral response after 2 doses of the BNT162b2 mRNA vaccine and that 16% of vaccinated kidney recipients became infected within 3 mo, with a high proportion of severe cases and significant mortality compared with the course of COVID-19 in unvaccinated RTRs.7 Similarly, we observed a humoral response rate of as little as 35% in 120 RTRs after 2 doses of the BNT162b2 mRNA vaccine.8 However, it was shown that administration of a third dose of an mRNA vaccine enhanced the immune response in immunocompromised patients.9 In our patients, 85.9% of RTRs developed neutralizing antibodies (NAs) after a third BNT162b2 dose.10 In immunocompetent populations in Israel, rates of confirmed COVID-19 infection and severe illness were substantially lower after a third dose than for individuals vaccinated with only 2 doses for adults aged ≥60 y11 and for younger healthcare workers.12
Following the success of the third vaccine and in light of the rapid spread of the Omicron variant and concerns regarding a significant increase in severe cases and hospitalizations, the State of Israel decided, on January 2, 2022, that a fourth dose of the BNT162b2 vaccine should be given to healthcare workers, adults aged >60 y, and immunosuppressed individuals for whom at least 4 mo had passed since the administration of the third vaccine dose.
In this study, we aimed to elucidate the response to the fourth vaccine dose (specifically, the response to the different SARS-CoV-2 variants, including the Omicron variant) and the correlation between the fourth vaccine dose and the rate and severity of morbidity because of the Omicron variant in RTRs. We also monitored the adverse events (AEs) after the fourth booster dose in our population.
MATERIALS AND METHODS
Study Population
The study cohort consisted of 447 RTRs followed up at the outpatient RTR clinic at the Sheba Medical Center. In the first part of the study, all 447 transplant recipients were interviewed by telephone after giving their consent to participate in the study. The questionnaire included questions regarding the status and timing of the fourth vaccination, prior vaccinations, and whether and when infected with SARS-CoV-2 (information was also collected for patients who had contracted the disease more than once). RTRs with confirmed COVID-19 (by a polymerase chain reaction [PCR]) from December 1, 2021, until the end of March 2022 were considered to have been infected with the Omicron variant because almost all sequenced infections during this period were Omicron validated. These patients were interviewed about the self-reported severity of their disease. Patients were asked whether they had been hospitalized during the disease course and whether they had experienced any of the symptoms listed in Table 4 (fever, chills, cough, etc). In addition, they were requested to rate their disease severity as mild, moderate, or severe and to report antiviral therapy (specifically molnupiravir and nirmatrelvir/ritonavir) given during the course of the disease.
In addition, we conducted a prospective study of 79 of 447 RTRs who had previously received 3 doses of the BNT162b2 vaccine and had subsequently been vaccinated with a homologous fourth dose of the vaccine. Patients with a positive SARS-CoV-2 PCR test before or after the full 3-dose vaccination were excluded from the study, as were patients vaccinated before transplant (given the stronger response to the BNT162b2 vaccine in patients vaccinated before kidney transplant). Five of these RTRs were subsequently excluded from the study because they were infected with COVID-19 before the post fourth vaccine serology assessment, leaving a cohort of 74 RTRs for further assessment. Vaccination was avoided during the first 3 mo after transplantation and during active treatment for rejection. On the day of the fourth vaccination, blood was drawn, before administration of the booster dose, for baseline serology assessment of receptor-binding domain (RBD) immunoglobulin (IgG) and NAs. Three to 4 wk after the fourth booster dose, testing for RBD IgG and NA was repeated to assess the humoral response to the vaccine.
For 56 of the subgroup of 74 RTRs, NA had been determined before and 3 wk after the third vaccine, which enabled a comparison between the antibody response to the fourth versus the third dose. In addition, to study the antibody response to the third vaccination over time, blood was drawn from those 56 patients 3 mo after the third vaccination. Furthermore, in 30 of 74 RTRs, blood was drawn on the day of the fourth vaccination and 3 wk after that vaccination for assessment of NA to 3 different SARS-CoV-2 variants (wild-type, Delta, and Omicron) with the aim to evaluate the humoral response to the different SARS-CoV-2 variants before and after the fourth vaccine dose. Written informed consent was obtained from all 79 participants. The protocol and informed consent were approved by our institutional review board (8314-21-SMC).
Immunosuppression
In our medical center, the standard maintenance immunosuppression regimen for RTRs comprises a calcineurin inhibitor (usually tacrolimus), an antimetabolite (usually a mycophenolate-based drug, mainly mycophenolic acid [MPA]), and prednisone, as described previously.10 For RTRs with a low immunological risk of rejection, early steroid withdrawal is implemented 5 to 8 d after transplant, and the maintenance regimen thus consists of tacrolimus and MPA. Conversion to a mammalian target of rapamycin inhibitor (sirolimus or everolimus) is instituted according to the patient’s risk of malignancy and lack of tolerance to calcineurin inhibitors.
Primary Outcome
For the retrospective study of 447 RTRs, the rate and severity of Omicron infection were evaluated in those who received the fourth vaccine versus those nonvaccinated with the fourth vaccine. For the prospective study of 74 RTRs, a positive response to the fourth booster dose of the BNT162b2 vaccine was defined as the presence of NA capable of reducing viral replication by at least 50% at a ≥16-fold dilution.
Data Extraction and Study Assessments
Patient information was obtained from the electronic patient records in the MDClone data acquisition system of the Sheba Medical Center, as described previously (Table 1).8 This system facilitated retrieval of relevant biochemical and clinical information, as described previously.8 For 10 patients, total daily mycophenolate dose was converted to the equivalent MPA dose by dividing the mycophenolate dose by 1.388. The use of cyclosporine, azathioprine, rapamycin, and everolimus on the day of the fourth vaccine was also retrieved from the MDClone system.
TABLE 1. Demographic and clinical characteristics of RTRs, stratified by Omicron infection
Variable Total cohort (N = 447) Noninfected patients (N = 303) Infected patients (N = 144) P
RTR characteristics
Age, y, median (IQR) 61.5 (50.5–70.4) 63.2 (52.3–71.5) 57.4 (47.5–67.8) 0.0002**
Female sex, n (%) 313 (70) 218 (71.9) 95 (66) 0.19
Time from transplant to December 1, 2021, y, median (IQR) 4.6 (1.4–13.3) 5.6 (1.7–13.5) 3.5 (1.3–11.7) 0.12
ESRD cause, n (%)
ADPKD 70 (15.7) 54 (17.8) 16 (11.1) 0.05
Diabetic nephropathy 74 (16.5) 53 (17.5) 21 (14.6)
Glomerulonephritis 132 (29.5) 85 (28.1) 47 (32.6)
Nephrosclerosis 55 (12.3) 42 (13.9) 13 (9)
Other 74 (16.5) 41 (13.5) 33 (22.9)
Unknown 42 (9.4) 28 (9.2) 14 (9.7)
Dialysis pretransplant 293 (65.5) 193 (63.7) 100 (69.4) 0.44
Transplant number, n (%)
1 424 (94.8) 285 (94.1) 139 (96.5) 0.69
2 17 (3.8) 13 (4.3) 4 (2.8)
3 5 (1.1) 4 (1.3) 1 (0.7)
4 1 (0.2) 1 (0.3) 0 (0.0)
BNT162b2 mRNA vaccination and COVID-19 infection
COVID-19 infection in prior waves, n (%) 28 (6.3) 23 (7.6) 5 (3.5) 0.09
Received fourth vaccination,n (%) 256 (57.3) 185 (61.1) 71 (49.3) 0.02*
Days from fourth vaccination to Omicron infection,median (IQR) 33 (17.0–49.0) 33 (17.0–49.0)
N = 392 N = 270 N = 122
No. vaccines, n (%)
0 15 (3.8) 12 (4.4) 3 (2.5) 0.16
1 6 (1.5) 4 (1.5) 2 (1.6)
2 19 (4.8) 12 (4.4) 7 (5.7)
3 96 (24.5) 57 (21.1) 39 (32)
4 256 (65.3) 185 (68.5) 71 (58.2)
*P < 0.05; **P < 0.01.
ADPKD, autosomal dominant polycystic kidney disease; COVID-19, coronavirus disease 2019; ESRD, end-stage renal disease; IQR, interquartile range; RTR, renal transplant recipient.
TABLE 2. Multivariate logistic regression analysis for Omicron infection in RTRs
Effect Odds ratio (95% CI) P
Age, ≤60 vs >60 y 1.83 (1.21-2.77) 0.004**
Gender, M vs F 1.25 (0.81-1.94) 0.31
Fourth vaccine dose, yes vs no 0.63 (0.41-0.96) 0.03*
COVID-19 infection in prior waves, yes vs no 0.33 (0.12-0.92) 0.03*
*P < 0.05; **P < 0.01
CI, confidence interval; COVID-19, coronavirus disease 2019; F, female; M, male; RTR, renal transplant recipient.
Patients were instructed to report (using a specific questionnaire) any systemic (fever, fatigue, headache, myalgia, chills, nausea/vomiting, paresthesia) and local (pain, redness, or swelling at the injection site) reactions occurring within 30 d after fourth vaccine dose and were actively screened for any other systemic and local complaints.
Antibody Detection Assays
RBD IgG was measured using the SARS-CoV-2 IgG II Quant (6S60, Abbott) test. These commercial tests were performed according to the manufacturer’s instructions. For determining NAs, a SARS-CoV-2 pseudovirus (psSARS-2) neutralization assay was performed13 using a propagation-competent vesicular stomatitis virus spike (kindly provided by Gert Zimmer, University of Bern, Switzerland).
Antibody Detection Assays for Different SARS-CoV-2 Variants
To enable a comparison of the humoral responses to different SARS-CoV-2 variants after the fourth vaccine dose, microneutralization assays with wild-type virus, B.1.617.2 (Delta), and Omicron variant isolates were performed on serum samples obtained from 30 RTRs immediately before and 3 wk after the fourth booster dose, as previously described.14
Statistical Analysis
Descriptive statistics were expressed as frequencies and percentages for categorical data and as median values with interquartile range (IQR) for continuous variables. All continuous variables were assessed for normality by the Kolmogorov-Smirnov test and log-transformed as appropriate. Differences in baseline characteristics between the groups were tested using the chi-square or Fisher exact test for the categorical variables or a t test for the continuous variables. To compare the humoral responses before and after the fourth vaccine dose, a paired t test and McNemar test were used.
Multivariable logistic regression analysis was used to identify factors associated with the vaccine-induced antibody response. The variables used in the multivariate analysis were those with a P value of <0.15 in the univariate analysis and those of clinical and biological relevance. Results are presented as odds ratio (OR), 95% confidence intervals (CIs), and P values.
The humoral responses to different SARS-CoV-2 variants before and after the fourth BNT162b2 mRNA vaccine in 30 RTRs were compared by a mixed model for repeated measures adjusted for multiple comparisons by the Tukey test.
All data analyses were performed with the SAS 9.4 software (Cary, NC). Scatter plots of log-transformed IgG and NA were obtained using GraphPad Prism 5.0 (GraphPad Software, Inc, San Diego, CA). A P value of <0.05 was considered as the cutoff for statistical significance.
RESULTS
Cohort Characteristics of 447 RTRs
Cohort characteristics, including the cause of end-stage renal disease, transplant number, dialysis pretransplant, and number of COVID-19 vaccines, are presented in Table 1. The median age was 61.5 y (IQR, 50.5–70.4); 313 (70%) were females; and the median time from transplant to December 1, 2021, was 4.6 y (IQR, 1.4–13.3). Of our RTR cohort, 28 patients (6.3%) had PCR-confirmed COVID-19 infection in waves that preceded the Omicron wave, and 256 patients (57.3%) received the fourth BNT162b2 dose.
Univariate and Multivariable Comparison of Noninfected Versus Omicron-infected RTRs
During the 4-mo period from December 1, 2021, until the end of March 2022, 144 of 447 RTRs (32.2%) fell ill with the Omicron infection. Infected patients were younger, with a median age of 57.4 y (IQR, 47.5–67.8), than noninfected RTRs, with a median age of 63.2 y (IQR, 52.3–71.5; P = 0.0002). Of the infected RTRs, 71 (49.3%) had received the fourth vaccine dose versus 185 (61.1%) noninfected RTRs who had been 4-dose vaccinated (P = 0.02; Table 1).
In a multivariable logistic regression analysis for Omicron infection in RTRs adjusted for age, gender, status of fourth vaccination before the infection, and COVID-19 infection in prior waves (Table 2), age ≤60 y increased the likelihood of Omicron infection by 83% compared with age >60 y (OR 1.83; 95% CI, 1.21-2.77; P = 0.04). In addition, the fourth vaccination and COVID-19 disease in preceding waves reduced the odds for contracting the infection by 37% and 67%, respectively (OR 0.63; 95% CI, 0.41-0.96; P = 0.03 and OR 0.33; 95% CI, 0.12-0.92; P = 0.03, respectively).
Univariate Comparison of RTRs Infected With the Omicron Variant Who Received the Fourth Vaccine Dose Versus Infected RTRs Who Did Not Receive the Fourth Dose
Of 144 RTRs infected with the Omicron variant, 71 (49.3%) had received the fourth dose. Infected patients who had received the fourth vaccine were older (median age 60.2 y; IQR, 47.3–70.6) than those who had not (median age 55.3 y; IQR, 47.6–64.4; P = 0.049). Of the female RTRs, 54 (76.1%) received the fourth vaccine versus 41 (56.2%) who did not (P = 0.01). The rate of COVID-19 infection in prior waves was higher in those who did not receive the fourth vaccine versus those who did (Table 3).
TABLE 3. Demographic and clinical characteristics of RTRs infected with Omicron, stratified by preinfection fourth BNT162b2 vaccination status
Variable Total cohort (N = 144) Fourth vaccine administered (N = 71) No fourth vaccine (N = 73) P
RTR characteristics
Age, y, median (IQR) 57.4 (47.5–67.8) 60.2 (47.3–70.6) 55.3 (47.6–64.4) 0.049*
Female sex, n (%) 95 (66) 54 (76.1) 41 (56.2) 0.01*
Transplant to December 1, 2021, y,median (IQR) 3.5 (1.3–11.7) 3.8 (1.4–9.9) 3.2 (0.9–13.4) 0.49
ESRD cause, n (%)
ADPKD 16 (11.1) 9 (12.7) 7 (9.6) 0.91
Diabetic nephropathy 21 (14.6) 10 (14.1) 11 (15.1)
Glomerulonephritis 47 (32.6) 23 (32.4) 24 (32.9)
Nephrosclerosis 13 (9.03) 8 (11.3) 5 (6.8)
Other 33 (22.9) 15 (21.1) 18 (24.7)
Unknown 14 (9.7) 6 (8.5) 8 (11)
Dialysis pretransplant, n (%) 100 (69.4) 51 (71.8) 49 (67.1) 0.38
Transplant number, n (%)
1 139 (96.5) 69 (97.2) 70 (95.9) 0.37
2 4 (2.8) 1 (1.4) 3 (4.1)
3 1 (0.7) 1 (1.4) 0 (0)
COVID-19 infection in prior waves, n (%) 5 (3.5) 0 (0) 5 (6.8) 0.02*
*P < 0.05; **P < 0.01.
ADPKD, autosomal dominant polycystic kidney disease; ESRD, end-stage renal disease; IQR, interquartile range; RTR, renal transplant recipient.
TABLE 4. Omicron infection self-reported disease severity assessed by questionnaire results in RTRs, stratified by preinfection fourth BNT162b2 vaccination status
Variable, n (%) Total cohort (N = 137) Fourth vaccine administered (N = 68) No fourth vaccine (N = 69) P
Hospitalization 7 (5.1) 3 (4.4) 4 (5.7) 0.73
Fever 51 (37.2) 20 (29.4) 31 (44.9) 0.06
Chills 45 (32.8) 15 (22.1) 30 (43.5) 0.008**
Cough 80 (58.4) 34 (50) 46 (66.7) 0.048*
Sore throat 46 (33.6) 19 (27.9) 27 (39.1) 0.16
Headache 55 (40.1) 23 (33.8) 32 (46.4) 0.13
Fatigue/weakness 78 (56.9) 32 (47.1) 46 (66.7) 0.02*
Myalgia 57 (41.6) 23 (33.8) 34 (49.3) 0.06
Nausea/vomiting 12 (8.7) 2 (2.9) 10 (14.5) 0.02*
Abdominal pain/diarrhea 22 (16) 5 (7.4) 17 (24.6) 0.006**
Back pain 31 (22.6) 12 (17.6) 19 (27.5) 0.17
Loss of smell and taste 18 (13.1) 6 (8.8) 12 (18.8) 0.09
N = 134 N = 66 N = 68
Disease severity
Mild 86 (64.2) 52 (78.8) 37 (54.4) 0.008**
Moderate 21 (15.7) 8 (12.1) 13 (19.1)
Severe 24 (17.9) 6 (9.1) 18 (26.5)
Received treatment 31 (23.1) 19 (28.8) 12 (17.6) 0.11
*P < 0.05; **P < 0.01.
RTR, renal transplant recipient.
Univariate Comparison for Omicron Disease Severity Based on Questionnaire Results of RTR Infected With the Omicron Variant Who Received the Fourth Vaccine Dose Versus Infected RTRs Who Did Not
Of 137 RTRs infected with the Omicron variant who completed the questionnaire, 80 (58.4%) had a cough, 78 (56.9%) experienced fatigue or weakness, 57 (41.6%) had myalgia, 51 (37.2%) had fever, 46 (33.6%) complained of sore throat, and 45 (32.8%) had chills during the disease. RTRs who did not receive the fourth dose before the Omicron infection had higher rates of self-reported fever (44.9% not vaccinated versus 29.4% vaccinated, P = 0.06), chills (43.5% versus 22.1%; P = 0.008), cough (66.7% versus 50%; P = 0.048), fatigue or weakness (66.7% versus 47.1%; P = 0.02), myalgia (49.3% versus 33.8%; P = 0.06), nausea or vomiting (14.5% versus 2.9%; P = 0.02), and abdominal pain or diarrhea (24.6% versus 7.4%; P = 0.006) during the course of the disease. In addition, 18 (26.5%) of the nonvaccinated RTRs ranked their disease as severe compared with only 6 (9.1%) of those vaccinated before the infection, whereas 52 (78.8%) of vaccinated RTRs evaluated their disease as mild as opposed to only 37 (54.4%) of those not vaccinated (P = 0.008). A total of 31 RTRs (23.1%) received treatment with molnupiravir or nirmatrelvir/ritonavir during the disease, with no significant difference between vaccinated to nonvaccinated RTRs (Table 4).
Cohort Characteristics of a Subgroup of 74 RTRs Participating in a Prospective Study for Assessment of the Humoral Response to the Fourth BNT162b2 Vaccine Dose
Of 447 RTRs, 303 were not infected with the Omicron variant. In a subgroup of 74 of 303 noninfected RTRs, the humoral response before and after fourth vaccine was analyzed. For this subgroup, the median age was 60.2 y (IQR, 53.3–69.8); 24 (32.4%) were females; and median body mass index was 26.7 kg/m2 (IQR, 23.3–30.8). The median time from transplant was 3.1 y; 81.8% had received a living donor transplant; and 70.3% had undergone pretransplant dialysis. Past medical histories of hypertension (71.6%), diabetes (37.8%), ischemic heart disease (10.8%), and congestive heart failure (5.4%) were recorded for these patients (Table 5). Of these patients, 44.6% received the 3-drug immunosuppression regimen of tacrolimus-MPA-prednisone, whereas 14.9% were treated with the 2-dose protocol of tacrolimus and MPA (Table 6). Overall, 91.9% of these RTRs were treated with a calcineurin inhibitor (86.5% with tacrolimus and 5.4% with cyclosporine), 59.5% with MPA, and 71.6% with prednisone.
TABLE 5. Demographic, clinical, and biochemical characteristics of RTRs, stratified by antibody response.
Variable Total cohort (N = 74) Negative (N = 4) Positive (N = 70) P
RTR characteristics
Age, y, median (IQR) 60.2 (53.3–69.8) 70 (62.9–78.4) 59.8 (52.4–69.7) 0.06
Female sex, n (%) 24 (32.4) 1 (25) 23 (32.9) 0.74
Transplant to fourth vaccine, y, median (IQR) 3.1 (1.5–8.3) 6.5 (3.1–11.4) 3.1 (1.4–8.3) 0.35
Third to fourth vaccine date, d, median (IQR) 173 (172–174) 172 (146–175) 173 (172–174) 0.53
Fourth vaccine to antibody testing date, d,median (IQR) 21 (21–21) 21 (21–21.5) 21 (21–21) 0.25
ESRD cause, n (%)
ADPKD 9 (12.2) 0 (0) 9 (12.9) 0.72
DN 16 (21.6) 1 (25) 15 (21.4)
Glomerulonephritis 19 (25.7) 2 (50) 17 (24.3)
Nephrosclerosis 11 (14.9) 1 (25) 10 (14.3)
Other 12 (16.2) 0 (0) 12 (17.1)
Unknown 7 (9.5) 0 (0) 7 (10)
ESRD secondary to DN 16 (21.6) 1 (25) 15 (21.4) 0.87
Dialysis pretransplant 52 (70.3) 1 (25) 51 (72.8) 0.14
Transplant number, n (%)
1 69 (93.2) 4 (100) 65 (92.9) 0.86
2 3 (4.1) 0 (0) 3 (4.3)
3 2 (2.7) 0 (0) 2 (2.86)
Donor type, n (%)
Living 81 (81.8) 3 (75) 56 (80) 0.81
Deceased 16 (16.2) 1 (25) 14 (20)
Medical history
Hypertension 53 (71.6) 2 (50) 51 (72.9) 0.32
SBP 3-mo average, median (IQR) 130.2 (121–148.6) 159 (136.0–177.5) 129 (119–144.7) 0.047*
DBP 3-mo average, median (IQR) 75.8 (71.0–80.0) 77 (73.0–83.0) 75.5 (70.3–80.0) 0.57
Ischemic heart disease 8 (10.8) 0 (0) 8 (11.4) 0.47
Congestive heart failure 4 (5.4) 0 (0) 4 (5.7) 0.62
Diabetes 28 (37.8) 2 (50) 26 (37.1) 0.61
HbA1C 3-mo average, %, mean ± SD 6.32 ± 1.01 — 6.32 ± 1.01
Weight, kg,median (IQR) 80.3 (70–94) 67 (63.5–81.6) 80.5 (71–94.9) 0.47
BMI, kg/m2,median (IQR) 26.7 (23.3–30.8) 23.1 (19.3–30.1) 26.9 (23.6–31) 0.26
Average laboratory results 1 mo before antibody testing day, median (IQR)
White blood cells, K/μL 7.1 (6.0–8.6) 6.9 (4.9–7.3) 7.2 (6.0–8.7) 0.41
Lymphocyte absolute, K/μL 1.6 (1.3–2.0) 1.5 (1.2–1.8) 1.7 (1.3–2.1) 0.57
Neutrophils absolute, K/μL 4.5 (3.6–5.5) 4.5 (2.8–5.1) 4.6 (3.6–5.5) 0.61
Hemoglobin, g/dL 13.4 (12.2–14.2) 13.9 (10.9–14.4) 13.3 (12.2–14.2) 0.85
Platelets, K/μL 186 (150–228) 207 (78–224) 183 (150–230) 0.72
Creatinine, mg/dL 1.2 (1.0–1.4) 2 (1.9–2.3) 1.2 (1.0–1.3) 0.006**
eGFR (CKD-EPI)a 63.7 (50.8–78.7) 34.2 (26.3–36.8) 64.3 (51.0–80.7) 0.008**
Glucose, mg/dL 108 (98–142) 108 (96–192) 108.9 (98–142) 0.79
Albumin, g/dL 4.1 (3.9–4.3) 3.7 (3.6–3.8) 4.1 (3.9–4.3) 0.08
Globulins, g/dL 2.6 (2.5–2.8) 2.6 (2.2–3.0) 2.6 (2.5–2.8) 0.82
C-reactive protein, mg/L 4.4 (1.7–10.8) 2 (0.9–7.7) 4.4 (1.7–10.9) 0.39
*P < 0.05; ** P < 0.01.
a eGFR was calculated according to the following CKD-EPI formula:
eGFR = 14 * min (Scr/k, 1)α * max(Scr/k, 1) – 1.209 * 0.993Age * 1.018 * 1.159 (if black)
(where Scr is standardized serum creatinine; k = 0.7 if female, 0.9 if male; α = –0.329 if female, –0.411 if male; min = the minimum of Scr/k of 1; max = the maximum of Scr/k or 1).
ADPKD, autosomal dominant polycystic kidney disease; BMI, body mass index; CKD-EPI, Chronic Kidney Disease Epidemiology Collaboration; DBP, diastolic blood pressure; DN, diabetic nephropathy; eGFR, estimated glomerular filtration rate; ESRD, end-stage renal disease; HbA1C, hemoglobin A1C; IQR, interquartile range; RTR, renal transplant recipients; SBP, systolic blood pressure.
TABLE 6. RTR immunosuppression treatment on fourth vaccine day, stratified by antibody response
Immunosuppressive therapy Total cohort (N = 74) Negative (N = 4) Positive (N = 70) P
Tacrolimus, n (%) 67 (90.5) 4 (100) 63 (90) 0.51
Tacrolimus daily dose (mg) on fourth vaccine date, median, IQR 2.5 (1.5–3.5) 2.8 (2.3–3.0) 2.3 (1.5–3.5) 0.76
Tacrolimus daily dose (mg) per weight (kg) on fourth vaccine day, median (IQR) 0.03 (0.02–0.05) 0.04 (0.03–0.05) 0.03 (0.02–0.05) 0.44
Tacrolimus trough level, 1-mo average before fourth vaccine day, μg/L,median (IQR) 6.9 (5.30–8.35) 6.75 (3.30–7.10) 6.9 (5.30–8.50) 0.52
MPA, n (%) 50 (67.6) 2 (50) 48 (68.6) 0.44
MPA daily dose (mg) on fourth vaccine date, median (IQR) 360 (0–720) 180 (0–540) 360 (0–720) 0.35
MPA daily dose (mg) per weight (kg) on fourth vaccine date, median (IQR) 6 (0–8.37) 2.57 (0–8.2) 6.21 (0–8.37) 0.61
Prednisone, n (%) 59 (79.7) 4 (100) 55 (78.6) 0.29
Prednisone daily dose (mg) on fourth vaccine date, median (IQR) 5.0 (3.0–5.0) 5.0 (5.0–7.5) 5.0 (2.5–5.0) 0.06
Prednisone daily dose (mg) per weight (kg) on fourth vaccine date, median (IQR) 0.05 (0.04–0.07) 0.08 (0.06–0.12) 0.05 (0.04–0.07) 0.07
Immunosuppressive regimen, n (%)
Tacrolimus + MPA + prednisone 33 (44.6) 2 (50) 31 (44.3) 0.82
Tacrolimus + MPA 11 (14.9) 0 (0) 11 (15.7) 0.39
Tacrolimus + prednisone 20 (27) 2 (50) 18 (25.7) 0.28
Cyclosporine 4 (5.4) 0 (0) 4 (5.7) 0.62
Azathioprine 3 (4) 0 (0) 3 (4.3) 0.67
mTORi (everolimus or sirolimus 4 (5.4) 0 (0) 4 (5.7) 0.62
IQR, interquartile range; MPA, mycophenolic acid; mTORi, mammalian target of rapamycin inhibitor; RTR, renal transplant recipient.
The median time from the fourth vaccine dose to antibody testing was 21 d (IQR, 21–21). Seventy RTRs (94.6%) had NAs ≥16 (positive-response group), and only 4 RTRs (5.4%) exhibited NAs <16 (negative-response group).
Univariate Comparison of Patients With a Positive Versus a Negative Response to the Fourth Vaccine Dose in the Subgroup of 74 RTRs
RTRs who responded to the fourth dose were younger, with a median age of 59.8 y (IQR, 52.4–69.7), as opposed to 70 y (IQR, 62.9–78.4) in nonresponders (P = 0.06). Average systolic blood pressure in the 3 mo before the fourth vaccination was lower in the responders than in the nonresponders (P = 0.047). Renal allograft function was significantly higher in the positive-response versus the negative-response patients (median estimated glomerular filtration rate of 64.3 mL/min, IQR [51–80.7] and 34.2 mL/min, IQR [26.3–36.8], respectively; P = 0.008). For all other demographic, clinical, and laboratory variables, the differences between the positive- and negative-response patients were not significant (Table 5). Given the small number of patients who did not respond to the vaccine, inferences cannot be made about the clinical significance of these findings (although statistically significant differences were found).
Different immunosuppressive medications and regimens, including the triple regimen containing MPA and the double regimen of tacrolimus and prednisone, were similar for the patients with positive and negative antibody responses (Table 6). The differences in the humoral response between the positive and negative responders to the fourth vaccine dose are shown in Table 7.
TABLE 7. RBD IgG and NAs in RTRs before and after fourth vaccine, stratified by antibody response to the fourth dose
Variable Total cohort (N = 74) Negative (N = 4) Positive (N = 70) P
Baseline immune status on fourth vaccine day
Positive NAs on fourth vaccine day, n (%) 58 (78.4) 1 (25) 57 (81.4) 0.008**
IgG-RBD GMT on fourth vaccine day (95% CI) 38.34 (20.8-70.8) 1.77 (0.01-1029) 45.72 (25.33-82.53) 0.01*
NA GMT on fourth vaccine day, (95% CI) 66.44 (41-107.7) 6.73 (0.14-319.4) 75.73 (49.96-122.1) 0.02*
Response to fourth vaccine
IgG-RBD GMT post fourth vaccine (95% CI) 646.5 (360.6-1159) 39.53 (0.02-103310) 758.4 (440.9-1305) 0.32
NA GMT post fourth vaccine (95% CI) 950.4 (550.3-1641) 2.83 (0.94-8.52) 1325 (833.1-2108) <0.0001**
*P < 0.05; **P < 0.001.
CI, confidence intervals; GMT, geometric mean titer; IgG, immunoglobulin; NA, neutralizing antibody; RBD, receptor-binding domain; RTR, renal transplant recipient.
For the 5 RTRs excluded from the study because they were infected with COVID-19 before the post fourth vaccine serology assessment, no significant differences in the humoral response were found between those infected before antibody testing and the other 74 members of the subgroup (Table 8).
TABLE 8. Immune status before versus after the fourth vaccine dose in 5 RTRs who had COVID-19 infection before post fourth vaccine serology testing compared with the remainder of the cohort
RTR infected with COVID-19 before post fourth vaccine serology testing (N = 5) Remainder of cohort (N = 74) P
Before fourth vaccine
IgG-RBD GMT (95% CI) 24.11 (0.41-1428) 38.34 (20.8-70.8) 0.71
NA GMT (95% CI) 64 (4.84-846.3) 66.44 (41-107.7) 0.97
After fourth vaccine
IgG-RBD GMT (95% CI) 878.1 (5.09-152E3) 646.5 (360.6-1159) 0.8
NA GMT (95% CI) 1463 (35.85-59731) 950.4 (550.3-1641) 0.69
CI, confidence interval; COVID-19, coronavirus disease 2019; GMT, geometric mean titer; IgG, immunoglobulin G; NA, neutralizing antibody; RBD, receptor-binding domain; RTR, renal transplant recipient.
Response to the Third Vaccine Dose Versus the Fourth Dose of the BNT162b2 mRNA Vaccine in RTRs
The positive-response rate based on NA titers increased from 78.4% immediately before the fourth vaccine dose to 94.6% (70/74) 3 wk (median 21 d, IQR [21–21]) after the fourth dose. Geometric mean titers (GMT) for RBD IgG and for NA increased from 38.34 (95% CI, 20.8-70.8) and 66.44 (95% CI, 41-107.7) on the day of the fourth vaccine, respectively, to 646.5 (95% CI, 360.6-1159) and 950.4 (95% CI, 550.3-1641), respectively, after the vaccine (P < 0.001). Both the rate and the intensity of response to the fourth dose were significantly higher than the values before the fourth dose (Table 7).
For 56 of 74 patients, we had assessed NAs before and 3 wk (median 21 d, IQR [21–21]) after the third dose. In those 56 RTRs, a positive-response rate, based on NA titers, increased from 58.9% (31/56) before the third vaccine to 91.1% (51/56) after the third dose. Three months after the third vaccine, we observed a decrease in GMT for NAs, as shown in Figure 1. The intensity of the response to the fourth dose was significantly higher than that observed after the third vaccine (GMT for NAs post third dose of 573.5 [95% CI, 299.9-1097] versus 1119 [95% CI, 627.1-1997] after the fourth dose, P = 0.0024).
FIGURE 1. Neutralizing antibody levels before the third vaccine, and 3 wk and 3 mo after the third vaccine.
Of the 58 recipients with a positive humoral response before the fourth dose of the vaccine, 57 (98.3%) remained positive after the fourth dose, with a significant increase in GMTs for RBD IgG and NAs. Sixteen patients (21.6%) had a blunted antibody response before the fourth vaccine (GMTs for RBD IgG and NAs of 0.49 [95% CI, 0.24-0.97] and 1.68 [95% CI, 1.02-2.76], respectively); among those, 13 (81.3%) exhibited a positive antibody response after the fourth dose, with a significant increase in GMTs for RBD IgG and NAs (Figure 2).
FIGURE 2. Antibody response before and after the fourth vaccine dose in RTRs with positive vs negative humoral responses before the fourth vaccine dose. A, GM of RBD IgG antibody levels. B, Neutralizing antibody titers. GM, geometrical mean; IgG, immunoglobulin G; RBD, receptor-binding domain; RTR, renal transplant recipient.
AEs in the Subgroup of 74 RTRs
AEs were common (82.4% of the subgroup) after administration of the fourth dose of the BNT162b2 vaccine. Local and systemic AEs were reported in 75.7% and 37.8% of the cohort, respectively, with 56 (75.7%) recipients experiencing pain at the injection site as the most frequent local adverse event. Systemic AEs, mainly fatigue, were reported for 20 RTRs (27%), with all other systemic AEs being experienced only by the positive responders. Recipients with a positive humoral response after the fourth dose were more likely to experience local and systemic AEs than nonresponders, but because of the low number of nonresponders, a statistically significant difference could not be detected (Table 9). No episodes of rejection were observed, and renal allograft function remained stable at a mean follow-up of 60 d after the third vaccine dose. No allergic responses were documented.
TABLE 9. Local and systemic AEs in RTRs reported after the fourth dose of BNT162b2 vaccine, stratified by antibody response
AEs Total cohort (N = 74) Negative (N = 4) Positive (N = 70) P
Local AEs, n (%)
Pain at injection site 56 (75.7) 2 (50) 54 (77.1) 0.22
Swelling 7 (9.4) 0 (0) 7 (10) 0.5
Redness 7 (6.8) 0 (0) 5 (7.1) 0.58
Systemic AEs, n (%)
Fever 3 (4.1) 0 (0) 3 (4.3) 0.67
Fatigue 20 (27) 1 (25) 19 (27.1) 0.92
Headache 8 (10.8) 0 (0) 8 (11.4) 0.47
Myalgia 13 (17.6) 0 (0) 13 (18.6) 0.34
Chills 4 (5.4) 0 (0) 4 (5.7) 0.62
Nausea/vomiting 2 (2.7) 0 (0) 2 (2.9) 0.73
Paresthesia 0 (0) 0 (0) 0 (0)
Any local AE 56 (75.7) 2 (50) 54 (77.1) 0.22
Any systemic AE 28 (37.8) 1 (25) 27 (38.6) 0.59
Any AE 61 (82.4) 2 (50) 59 (84.3) 0.08
AE, adverse event; RTR, renal transplant recipient.
Humoral Response to Different SARS-CoV-2 Variants Before and After the Fourth BNT162b2 mRNA Vaccine in 30 RTRs
The humoral response to the Omicron variant before the fourth vaccine dose was significantly reduced compared with the response to the wild-type virus and the Delta variant (GMT 2.96; 95% CI, 1.71-5.14 versus GMT 14.59; 95% CI, 6.85-31.07 and GMT 14.93; 95% CI, 7-31.86, respectively; P < 0.001). The antibody titer to the Omicron variant (GMT 30.55; 95% CI, 12.7-73.53) 3 wk after the fourth vaccine dose was also significantly reduced compared with the increases for the wild-type virus (239; 95% CI, 94.96-601.8) and the Delta variant (233.6; 95% CI, 102.5-532.4; P < 0.001; Figure 3).
FIGURE 3. Neutralizing antibody response before and after the fourth vaccine dose to the WT virus and the Delta and Omicron variants of SARS-CoV-2 in RTRs. CI, confidence interval; GMT, geometric mean titer; RTR, renal transplant recipient; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2; WT, wild type.
DISCUSSION
At the height of the Omicron wave of the epidemic, 144 of 447 RTRs (32.3%) in our study had contracted the infection, with 71 of the 144 (49.3%) of those infected with Omicron having received the fourth vaccine dose before the infection. Yet, a comparison of self-reported disease severity between nonvaccinated RTRs and those vaccinated with the fourth BNT162b2 dose before the infection showed that having received a fourth dose was associated with markedly reduced self-reported disease severity, both as assessed by self-report of the presence or absence of symptoms listed on a questionnaire and also by subjective report of mild/moderate/severe disease.
Of the 74 RTRs tested for the humoral response after the fourth vaccine dose, 70 (94.6%) responded to the vaccine. The rate of response increased from 78.4% before the fourth vaccine dose to 94.6% postvaccine, with a significant increase in RBD IgG and NA titers. In fact, the intensity of NA response to the vaccine was significantly stronger after the fourth dose than that observed after the third dose in a subgroup of 56 RTR. Given the high rate of Omicron infection after the fourth vaccination dose, we analyzed the humoral response to 3 different SARS-CoV-2 variants before and after the fourth dose. The antibody response to the Omicron variant was markedly lower than the responses to the wild-type virus and to the Delta variant before and after the fourth vaccine.
To the best of our knowledge, this is the first study showing the effect of the fourth BNT162b2 vaccine on the rate and severity of Omicron infection in a large cohort of RTRs. In addition, we demonstrated a strong humoral response to the fourth vaccine in a subgroup of 74 RTRs. However, the antibody response before and after the fourth vaccine dose differed significantly when tested against different SARS-CoV-2 variants, showing unequivocally a reduced humoral response to the Omicron variant. Our findings thus provide an explanation for the high rate of breakthrough Omicron infection after the third and fourth BNT162b2 vaccine doses in RTRs. Nonetheless, we have clearly shown that the fourth vaccine dose was associated with significantly reduced self-reported severity of Omicron disease in our RTR population.
In line with our data, a small cohort of 18 solid organ transplant recipients developed high antibody titers after a fourth injection to negative and positive responders to 3 doses of the vaccine.15 In another study, 43% of 49 RTRs with a negative serology after the third injection seroconverted after the fourth mRNA vaccination, although the response remained weak and was probably not protective against COVID-19.16 In a different study, a fourth dose of mRNA vaccine given to 92 RTRs with low IgG titers 1 mo after the third dose elicited a satisfactory antibody response in 48% and 52% of those who received the BNT162b2 and the mRNA-1273 vaccines, respectively.17 Moreover, a fourth mRNA-1273 dose significantly increased the neutralizing response against the Delta variant, with the rate of NA rising from 16% to 66%.18
Taken together, the data indicate that a fourth vaccine dose is beneficial in mounting an antibody response in transplant recipients. It is thought that the higher antibody titers detected post–COVID-19 infection versus postvaccination19 can be attributed to the higher antigen dose exposure with infection versus vaccination. Previous publications have reported that stronger immune stimuli with additional or higher vaccine doses are required to generate better immunogenicity in immunocompromised patients.20,21 Similarly, it can be assumed that additional doses of COVID-19 vaccine in RTRs increase the antigen load exposure, thereby stimulating a stronger humoral response.
Interestingly, an age >60 versus <60 y was protective of Omicron infection in multivariable analysis. This is most probably related to older patients being more careful about social distancing and masking. Among 144 RTRs infected with the Omicron variant, those who received the fourth vaccine were older, indicating that older patients are also more watchful about getting vaccinated.
Our data indicate a substantially reduced NA response to the Omicron variant after the third and fourth BNT162b2 vaccine doses, with a concomitant high rate of Omicron infection in vaccinated RTRs. Nonetheless, we also demonstrated robust protection against severe Omicron disease provided by the fourth vaccine dose. Similarly, in a study estimating the rate of confirmed Omicron infection and severe disease in >1 million participants >60 y, a fourth dose provided added short-term protection against confirmed infections and severe illness compared with 3 vaccine doses given at least 4 mo earlier. Interestingly, protection from confirmed Omicron infection waned rapidly from a peak in the fourth week after vaccination, whereas protection against severe illness had not decreased by the sixth week after the fourth dose.11 In addition, BNT162b2 vaccine was shown to provide 70% protection against hospitalization with Omicron in South Africa.22
Because the multiple spike (S) protein mutations in the Omicron variant contribute to reduced vaccine-elicited antibody neutralization23-26 and reduced protection from infection—as reflected in the attenuated humoral response to the Omicron variant shown in our study—it seems likely that cellular immunity plays an important role in ameliorating the severity of COVID-19 caused by the Omicron variant. It has been shown that the BNT162b2 vaccine stimulated extensive crossreactive cellular immunity, manifested in specific CD8+ and CD4+ T-cell responses against SARS-CoV-2 variants, including the Omicron variant, with minimal escape at the T-cell level.26,27 Similarly, in a study of 70 participants, comprising individuals vaccinated with Ad26.CoV2.S or BNT162b2 and unvaccinated but convalescent COVID-19 individuals, 70% to 80% of the CD4+ and CD8+ T-cell response to the spike was maintained, and the magnitude of Omicron crossreactive T cells was similar to that for the Beta and Delta variants.28 In addition, greater crossreactivity of vaccine-induced cellular immune responses compared with humoral responses against the Alpha, Beta, and Gamma variants, consistent with T-cell responses targeting multiple regions in the spike protein and hence contributing to the preserved cellular immune responses to Omicron, has previously been shown.29 Taken together, the abovementioned data confirm robust CD4+ and CD8+ T-cell responses that largely crossreact with Omicron after vaccination and infection.
The retained T-cell immunity to the Omicron variant as shown in the abovementioned studies, despite a significant reduction of the Omicron-specific humoral response, thus may explain the marked protection against severe disease, whereas the infection rate with Omicron of 4-dose vaccinated RTRs was relatively high. It can be postulated that the number of NAs in vaccinated RTRs is insufficient to prevent the infection, but the cellular defense, specifically CD8+ cells capable of clearing viral infections, probably provides protection against severe disease.
Certain limitations should be taken into consideration in interpreting our results. The use of patients’ self-assessed severity of illness introduces an element of subjectivity, which is a limitation relative to the assessments that can be performed on inpatients under direct observation. It is not known whether vaccination status might influence how patients report their severity of illness. Nonetheless, given the acknowledged limitations of a survey study in this regard, we feel this information is still of value to clinicians. Omicron infection was confirmed by PCR to SARS-CoV-2 but not by a specific PCR to the Omicron variant. Nonetheless, the PCR testing was performed during a period in which the B.1.1.529 (Omicron) variant of SARS-CoV-2 was predominant. Longer follow-up is needed to evaluate the protection afforded by the fourth dose against confirmed infection and severe illness over time. Confounding factors may exist, such as personal and barrier measures differences between RTRs who received the fourth dose and those who did not. In addition, for severe illness, differences in the prevalence of coexisting medical conditions could potentially have affected the results. Finally, the study was not designed as an efficacy trial (there was no control group), but a significant correlation was demonstrated between NAs and protection from SARS-CoV-2. The implications of our findings for the subgroups of 74 and 30 RTRs are limited by the small number of patients and the short follow-up period after vaccination. Antibodies may wane over time, and the half-life of the neutralizing response cannot be predicted. Furthermore, cellular immunity was not assessed.
In conclusion, a fourth BNT162b2 vaccination should be encouraged for RTRs to prevent the undesirable consequences of severe Omicron illness in this vulnerable population.
Supplementary Material
The authors declare no funding or conflicts of interest.
T.H. contributed in conception and design, data acquisition, data interpretation, writing, and revising. A.B.-D., R.H., and P.B. contributed in data collection. E.M. and E.G. contributed in revising. L.O. contributed in data analysis. V.I., R.D., K.A., and N.A. contributed in data acquisition. O.B. and M.M. contributed in data interpretation. Y.L. contributed in conception and design and data interpretation. G.R. contributed in conception and design, data interpretation, and revising.
Supplemental Visual Abstract; http://links.lww.com/TP/C599.
Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com).
==== Refs
REFERENCES
1. Fall A Eldesouki RE Sachithanandham J . A quick displacement of the SARS-CoV-2 variant Delta with Omicron: unprecedented spike in COVID-19 cases associated with fewer admissions and comparable upper respiratory viral loads. medRxiv. 2022; 2022.01.26.22269927
2. Davies MA Kassanjee R Rousseau P . Outcomes of laboratory-confirmed SARS-CoV-2 infection in the Omicron-driven fourth wave compared with previous waves in the Western Cape Province, South Africa. Preprint. Posted January 12, 2022. medRxiv. doi:10.1101/2022.01.12.22269148
3. Rincon-Arevalo H Choi M Stefanski AL . Impaired humoral immunity to SARS-CoV-2 BNT162b2 vaccine in kidney transplant recipients and dialysis patients. Sci Immunol. 2021;6 :eabj1031.34131023
4. Boyarsky BJ Werbel WA Avery RK . Antibody response to 2-dose SARS-CoV-2 mRNA vaccine series in solid organ transplant recipients. JAMA. 2021;325 :2204–2206.33950155
5. Sattler A Schrezenmeier E Weber UA . Impaired humoral and cellular immunity after SARS-CoV-2 BNT162b2 (tozinameran) prime-boost vaccination in kidney transplant recipients. J Clin Invest. 2021;131 :150175.
6. Cucchiari D Egri N Bodro M . Cellular and humoral response after MRNA-1273 SARS-CoV-2 vaccine in kidney transplant recipients. Am J Transplant. 2021;21 :2727–2739.34036720
7. Reischig T Kacer M Vlas T . Insufficient response to mRNA SARS-CoV-2 vaccine and high incidence of severe COVID-19 in kidney transplant recipients during pandemic. Am J Transplant. 2022;22 :801–812.34860470
8. Hod T Ben-David A Olmer L . Humoral response of renal transplant recipients to the BNT162b2 SARS-CoV-2 mRNA vaccine using both RBD IgG and neutralizing antibodies. Transplantation. 2021;105 :e234–e243.34310101
9. Hall VG Ferreira VH Ku T . Randomized trial of a third dose of mRNA-1273 vaccine in transplant recipients. N Engl J Med. 2021;385 :1244–1246.34379917
10. Hod T Ben-David A Olmer L . BNT162b2 third booster dose significantly increases the humoral response assessed by both RBD IgG and neutralizing antibodies in renal transplant recipients. Transpl Int. 2022;35 :10239.35387393
11. Bar-On YM Goldberg Y Mandel M . Protection against COVID-19 by BNT162b2 booster across age groups. N Engl J Med. 2021;385 :2421–2430.34879188
12. Spitzer A Angel Y Marudi O . Association of a third dose of BNT162b2 Vaccine with incidence of SARS-CoV-2 infection among health care workers in Israel. JAMA. 2022;327 :341–349.35006256
13. Lustig Y Sapir E Regev-Yochay G . BNT162b2 COVID-19 vaccine and correlates of humoral immune responses and dynamics: a prospective, single-centre, longitudinal cohort study in health-care workers. Lancet Respir Med. 2021;9 :999–1009.34224675
14. Lustig Y Nemet I Kliker L . Neutralizing response against variants after SARS-CoV-2 infection and one dose of BNT162b2. N Engl J Med. 2021;384 :2453–2454.33826815
15. Alejo JL Mitchell J Chiang TP . Antibody response to a fourth dose of a SARS-CoV-2 vaccine in solid organ transplant recipients: a case series. Transplantation. 2021;105 :e280–e281.34428188
16. Masset C Benotmane I Dantal J . A fourth SARS-CoV-2 mRNA vaccine in strictly seronegative kidney transplant recipients. Kidney Int. 2022;101 :825–826.35167873
17. Caillard S Thaunat O Benotmane I . Antibody response to a fourth messenger RNA COVID-19 vaccine dose in kidney transplant recipients: a case series. Ann Intern Med. 2022;175 :455–456.35007148
18. Benotmane I Bruel T Planas D . A fourth dose of the mRNA-1273 SARS-CoV-2 vaccine improves serum neutralization against the Delta variant in kidney transplant recipients. Kidney Int. 2022;101 :1073–1076.35231463
19. Hall VJ Foulkes S Saei A ; SIREN Study Group. COVID-19 vaccine coverage in health-care workers in England and effectiveness of BNT162b2 mRNA vaccine against infection (SIREN): a prospective, multicentre, cohort study. Lancet. 2021;397 :1725–1735.33901423
20. Pratt PK Jr Nunes D Long MT . Improved antibody response to three additional hepatitis B vaccine doses following primary vaccination failure in patients with inflammatory bowel disease. Dig Dis Sci. 2019;64 :2031–2038.30945037
21. Natori Y Shiotsuka M Slomovic J . A double-blind, randomized trial of high-dose vs standard-dose influenza vaccine in adult solid-organ transplant recipients. Clin Infect Dis. 2018;66 :1698–1704.29253089
22. Collie S Champion J Moultrie H . Effectiveness of BNT162b2 vaccine against Omicron variant in South Africa. N Engl J Med. 2022;386 :494–496.34965358
23. Cameroni E Bowen JE Rosen LE . Broadly neutralizing antibodies overcome SARS-CoV-2 Omicron antigenic shift. Nature. 2022;602 :664–670.35016195
24. Cele S Jackson L Khoury DS ; NGS-SA; COMMIT-KZN Team. Omicron extensively but incompletely escapes Pfizer BNT162b2 neutralization. Nature. 2022;602 :654–656.35016196
25. Cao Y Wang J Jian F . Omicron escapes the majority of existing SARS-CoV-2 neutralizing antibodies. Nature. 2022;602 :657–663.35016194
26. GeurtsvanKessel CH Geers D Schmitz KS . Divergent SARS-CoV-2 Omicron-reactive T and B cell responses in COVID-19 vaccine recipients. Sci Immunol. 2022;7 :eabo2202.35113647
27. Liu J Chandrashekar A Sellers D . Vaccines elicit highly conserved cellular immunity to SARS-CoV-2 Omicron. Nature. 2022;603 :493–496.35102312
28. Keeton R Tincho MB Ngomti A . T cell responses to SARS-CoV-2 spike cross-recognize Omicron. Nature. 2022;603 :488–492.35102311
29. Alter G Yu J Liu J . Immunogenicity of Ad26.COV2.S vaccine against SARS-CoV-2 variants in humans. Nature. 2021;596 :268–272.34107529
| 36367927 | PMC9746231 | NO-CC CODE | 2022-12-15 23:21:55 | no | Transplantation. 2023 Jan 8; 107(1):192-203 | utf-8 | Transplantation | 2,022 | 10.1097/TP.0000000000004383 | oa_other |
==== Front
Transplantation
Transplantation
TP
Transplantation
0041-1337
1534-6080
Lippincott Williams & Wilkins Hagerstown, MD
36228269
00028
10.1097/TP.0000000000004411
3
Original Clinical Science—General
Real-world Evidence of COVID-19 Vaccines Effectiveness in Solid-organ Transplant Recipient Population in Colombia: A Study Nested in the Esperanza Cohort
Pinto-Álvarez Mariana MPH 1
Fernández-Niño Julián A. PhD 23
Arregocés-Castillo Leonardo DrPH 1
Rojas-Botero Maylen L. PhD 4
Palacios Andrés F. MSc 1
Galvis-Pedraza Maryory MSc 1
Ruiz-Gomez Fernando DrPH 5
1 Dirección de Medicamentos y Tecnologías de Salud, Ministerio de Salud y Protección Social, Bogotá DC, Colombia.
2 Department of International Health, Johns Hopkins Bloomberg School of Public Health, Baltimore, MD.
3 Departamento de Salud Pública, Universidad del Norte, Barranquilla, Colombia.
4 Facultad Nacional de Salud Pública, Universidad de Antioquia, Medellín, Colombia.
5 Ministerio de Salud y Protección Social, Bogotá DC, Colombia.
Correspondence: Julián A. Fernández-Niño, PhD, Department of International Health, Johns Hopkins Bloomberg School of Public Health, 615 N. Wolfe St, Room E8532, Baltimore, MD 21205. ([email protected]).
08 12 2022
1 2023
08 12 2022
107 1 216224
18 7 2022
23 8 2022
09 9 2022
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
2022
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
Background.
Solid-organ transplant recipients (SOTRs) have a higher risk of coronavirus disease 2019 (COVID-19) complications and death and a less powerful and lasting response to vaccines and to natural infection. In Colombia, this population was prioritized in the National Vaccination Plan against COVID-19 and received vaccines from different platforms. The aim of this study was to estimate the effectiveness of the complete vaccination schedule and of the vaccine booster for COVID-19 administered to SOTRs in Colombia.
Methods.
A nested-cohort was assembled within the population-based ESPERANZA cohort and included the subset of 16 y and older SOTRs (n = 6963); the follow-up period spanned March 11, 2021, to May 11, 2022. The vaccine effectiveness was estimated with Cox proportional-hazards models so that the overall effectiveness of the complete vaccination schedule, the vaccine booster, each used vaccine, and the homologous and heterologous schedules were estimated, adjusting by the main confounders.
Results.
The overall effectiveness of being fully vaccinated was 73.7% (95% confidence interval [CI], 68.9%-77.0%) to prevent COVID-19 infection, 83.7% (95% CI, 78.7%-87.5%) to prevent hospitalization, and 92.1% (95% CI, 88.8%-94.4%) to prevent death due to COVID-19. Similarly, the effectiveness of the vaccine booster was 76.7% (95% CI, 70.6%-81.5%), 86.9% (95% CI, 79.4%-91.6%), and 94.5% (95% CI, 89.8%-97.1%) to prevent confirmed COVID-19 infection, hospitalization, and death due to COVID-19, respectively. In both cases, there were no statistically significant differences across age groups.
Conclusions.
Findings from this work show a high protection of vaccination against infection, hospitalization, and death due to COVID-19 in SOTRs, which increases with the vaccine booster.
SDCT
==== Body
pmcINTRODUCTION
Infection by the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), which causes the coronavirus disease 2019 (COVID-19), was declared a pandemic on March 11, 2020, by the World Health Organization.1 The first confirmed infection case in the Colombian territory was reported on March 6, 2020, which prompted the declaration of the health emergency on March 12, 2020.2
When the Colombian government began immunization for COVID-19, it acquired 5 vaccine types: ChAdOx1 nCoV-19 (AstraZeneca), CoronaVac (Sinovac), Ad26.COV2.S (Janssen), mRNA-1273 (Moderna), and BNT162b2 (Pfizer-BioNTech). Given the global and national scarcity context, the National Vaccination Plan against COVID-19 assigned the first vaccines to people who were more likely to develop complications or die because of COVID-19.
The prioritization process of the National Vaccination Plan in Colombia was determined by an ethical framework and based on the existing best epidemiological evidence at that time.3 This defined the order of equitable access to biologicals based on the population’s risk, through 5 phases. Phase 1included adults over 80 y of age and health workers in COVID-19 areas; Phase 2 included people between 60 and 79 y old and other health workers; Phase 3 included people between 50 and 59 y old, patients with selected underlying diseases, including solid-organ transplant recipients (SOTRs) and other people who, because of their occupation, presented a higher risk of infection, complication, or death from COVID-19; Phase 4 included people between 40 and 49 y old, and people who were in places with risk of outbreaks; and Phase 5 included the population between 3 and 39 y old, which was not previously prioritized.3 It is worth noting that all 5 vaccines were used in the SOTR population because no specific vaccine platform was determined for them.4
SOTRs are more prone to a SARS-CoV-2 infection because of their immunosuppression and their lower likelihood to develop an effective immune response to vaccination5-11 because their immune response to natural infection is also less powerful and lasting, rendering them more vulnerable to reinfections.12 Additionally, chronic immunosuppression might reduce the infectious dose necessary to cause COVID-19 and hinder the immune control once the infection has been established, which increases the risk of severe infection and complications.13 Furthermore, it is hypothesized that SOTRs could shed higher viral loads for longer periods than healthy hosts, which could increase their chances to spread the infection to other people.12
Similarly, previous studies have found that both humoral and cellular immune responses to vaccines and natural infection are weaker in SOTRs, whereby boosters and additional doses are required5-7,14,15 to maintain the protection against COVID-19 infection and severe disease.14 Additionally, immunosuppression has also been identified to be caused by the use of certain substances, such as antimetabolites, calcineurin inhibitors, and monoclonal antibodies, which explains the insufficient immune responses to current COVID-19 vaccination schemes in SOTRs.5,7-9,16,17
Knowing the effectiveness of vaccination in real-life conditions will allow us to evaluate the impact of prioritization in countries where SOTRs were prioritized, to wholly estimate the impact of vaccination, and to adjust the vaccination schedules in this risk group. Therefore, the aim of this study was to estimate the effectiveness of the complete vaccination schedule and of the vaccine booster for COVID-19 administered to SOTRs in Colombia.
MATERIALS AND METHODS
Design and Population Study
A nested-cohort was assembled within the ESPERANZA cohort, which is a population-based cohort made up of all Colombian residents who were eligible to receive a COVID-19 vaccine and has methodology that has been described elsewhere.18 Our nested-cohort included all SOTRs aged 16 and older that were registered in Red data—the National Health Institute (NHI) database. The follow-up period went from March 11, 2021, when the first individuals completed their vaccination schedule, to May 11, 2022, which corresponds to the latest update on the national statistics.
Data Sources
All data were obtained from the Integrated Social Protection Information System (in Spanish, Sistema Integrado de Información de la Protección Social), which is the official health statistics data source in Colombia. Information from it included people who were cross-referenced in 8 Social Protection Information System data records by using an individual anonymized number that is encrypted and automatically generated by the information system to protect the person’s identity: (1) Red data include all SOTRs who reside in Colombia; (2) MIVACUNA contains sociodemographic data from vaccine candidates who were later vaccinated against COVID-19, according to the National Vaccination Plan against COVID-19; (3) PAIWEB registers people who have received any vaccine in Colombia and the basic vaccine information, such as dose, vaccine type, date, and vaccine location; (4) SEGCOVID contains information about confirmed COVID-19 cases; (5) SISMUESTRAS stores the results from polymerase chain reaction (PCR) and antigen tests conducted in Colombia; (6) Single Registry of Affiliates to the Social Protection System—Births and Deaths (Registro Único de Afiliados al Sistema de la Protección Social—Nacimientos y Defunciones, in Spanish) records death causes in Colombia; (7) the high-cost disease registry (Cuenta de Alto Costo, in Spanish) includes data about people with diseases that require a larger budget, that is, chronic kidney disease (CKD), high blood pressure (HBP), diabetes mellitus (DM), cancer, and HIV infection; and (8) unique affiliate database (Base de Datos Única de Afiliados, in Spanish) provides information regarding the affiliation regime to the health system. The listed databases are public, although they have restricted access and are currently available to the Ministry of Health and Social Protection.
Inclusion and Exclusion Criteria
This study first included male and female SOTRs aged 16 and older residing in Colombia, regardless of their vaccination status. Subsequently, individuals were excluded if they (1) had a history of confirmed COVID-19 infection, (2) had an incomplete vaccination schedule, or (3) reported inconsistencies in their vaccination records (ie, implausible vaccine dates or doses). Definitions of a complete vaccination schedule were those originally established by the manufacturer and adopted by the HSPM.19 Figure 1 shows the complete selection process.
FIGURE 1. Flowchart of the selection process of the analytic sample. The fully vaccinated group includes 2788 people who later received at least one booster dose. COVID-19, coronavirus disease 2019; SOTR, solid-organ transplant recipient; NHI, National Health Institute of Colombia.
Exposure Groups
Three groups were formed based on the subjects’ vaccination status: unvaccinated, fully vaccinated, and vaccinated with booster. Unvaccinated individuals were those who did not receive any vaccine during the study period. The definition of fully vaccinated people depended on the administered vaccine; hence, for AstraZeneca, Pfizer, Moderna, and Sinovac, 2 doses with a 28-d period (21 d for Pfizer) between doses was considered to be a complete schedule. It is worth noting that in the case of longer periods between doses because of any cause, people could complete their vaccination schedule without recommencing it, unless they had received a vaccine unavailable in Colombia. For the Janssen vaccine, a sole dose was deemed a complete vaccination scheme. Vaccinated with booster was defined as people who received at least 1 additional vaccine dose from the same or from a different platform; SOTRs were allowed to get a booster 1 mo after completing the vaccination schedule.20 Allocation to either the fully vaccinated or vaccinated with booster groups was done 15 d after completing the vaccination schedule or receiving the first booster, respectively. In Colombia, heterologous vaccination was used for both the booster, from September 15, 2021,20 and the initial vaccination schedule, from March 18, 2022.21
Outcomes
The study outcomes were (1) COVID-19 infection, defined as a COVID-19 diagnosis confirmed by PCR or antigen tests (these tests had to be validated by the NHI in Colombia) and registered in SISMUESTRAS; (2) hospitalization due to COVID-19, defined as having entered the general hospitalization service or the intensive care unit and having COVID-19 as one of the hospitalization causes at any moment of the hospital stay, as registered in SEGCOVID; and (3) confirmed death because of COVID-19, defined as having a confirmed COVID-19 diagnosis as the basic cause of death in the death certificate, as consulted in the Single Registry of Affiliates to the Social. Suspected deaths were not included in this study.
Covariates
Additional variables that have been deemed as relevant confounders in previous studies were also measured to include them in the analysis: age (y); sex (male versus female); affiliation regime to the health system (contributory versus subsidized); municipality of residence; comorbidities diagnosis (yes versus no), such as CKD, cancer, DM, HBP, and HIV infection; and the prevalent SARS-CoV-2 variant at the time of the COVID-19 infection (this information was taken from www.covariants.org).
Statistical Analysis
Categorical variables were described with absolute frequencies and proportions, whereas quantitative variables were described with central tendency (medians) and dispersion (range and interquartile range [IQR]) measures. Subjects’ characteristics were compared across exposure groups.
To estimate the overall vaccination effectiveness, a survival analysis was performed by using Cox proportional-hazards models to estimate the reduction in the risk of death, hospitalization, and infection in fully vaccinated individuals and in people vaccinated with booster. These models were adjusted for the confounders listed as covariates; the prevalent SARS-CoV-2 variant at the time of infection was adjusted for in the models to control the transmission risk. For those unvaccinated who did not develop any of the study outcomes, the infection risk given by a specific variant was randomly assigned proportional to its dominance during the study period. Additionally, all the time-to-event from the unvaccinated subjects during the study period was considered in the models.
Multiple types of right-censoring could occur, given by people who died by nonrelated COVID-19 causes, fully vaccinated individuals who received a booster or subjects who finished the follow-up period without developing any of the study outcomes. These censoring were considered while constructing the models. The statistical analysis was carried out by using R (4.2.0 version) and its survival (3.3.1 version) and ggplot2 (3.3.6 version) packages to perform the survival analysis and to create the graphs, respectively.
Ethics
This study used secondary data sources from public information systems. The research team did not have access to personal data from the participants at any moment and all used information was anonymized. Given that this study is classified as a research without risk according to the Colombian legislation,22 an approval from an Ethics Committee was not required.
RESULTS
The inclusion and exclusion criteria yielded a sample of 6963 SOTRs during the study period (March 11, 2021–May 11, 2022), from which 85% (n = 5925) were fully vaccinated (this figure includes 2072 individuals vaccinated with booster) and 15% (n = 1038) remained unvaccinated throughout the whole follow-up. Out of the 6963 SOTRs, 42.1% were female, and the median age was 52 y (IQR: 39–62; range: 16–97), whereas the median age of the unvaccinated group was 44 y (IQR: 33–56). Additionally, 76.7% of the participants belonged to the contributory health regime and 82.4% had at least 1 comorbidity, in which CKD (72.3%) and HBP (70.3%) were the most frequent diagnosis. Table 1 describes the main characteristics of the study individuals by exposure group.
TABLE 1. Solid-organ transplant recipients’ sociodemographic and clinical characteristics, ESPERANZA cohort
Variable Unvaccinated(n = 1038)n (%) Fully vaccinateda(n = 5925)n (%) Vaccine booster(n = 2072)n (%) Total(n = 6963)n (%)
Age (y)
Median (IQR) 44 (33–56) 53 (40–63) 56 (44–65) 52 (39–62)
Range 16–97 16–90 16–90 16–97
16–59 859 (82.8) 3987 (67.3) 1257 (60.7) 4846 (69.6)
60 and older 179 (17.2) 1938 (32.7) 815 (39.3) 2117 (30.4)
Sex
Female 460 (44.3) 2469 (41.7) 865 (41.7) 2929 (42.1)
Male 578 (55.7) 3456 (58.3) 1207 (58.3) 4034 (57.9)
Health system affiliation regime
Contributory 638 (61.5) 4704 (79.4) 1763 (85.1) 5342 (76.7)
Subsidized 400 (38.5) 1221 (20.6) 309 (14.9) 1621 (23.3)
Comorbidities
None 224 (21.6) 999 (16.9) 381 (18.4) 1223 (17.6)
≥1 comorbidity 814 (78.4) 4926 (83.1) 1691 (81.6) 5740 (82.4)
Cancer 21 (2.0) 183 (3.1) 77 (3.7) 204 (2.9)
Diabetes mellitus 123 (11.8) 1052 (17.8) 413 (19.9) 1175 (16.9)
Chronic kidney disease 710 (68.4) 4327 (73.0) 1453 (70.1) 5037 (72.3)
High blood pressure 700 (67.4) 4194 (70.8) 1413 (68.2) 4894 (70.3)
HIV infection 5 (0.5) 17 (0.3) 9 (0.4) 22 (0.3)
Prevalent SARS-CoV-2 variant at the time of infection
Delta 113 (10.9) 960 (16.2) 624 (30.1) 1.073 (15.4)
Delta/Mu 90 (8.7) 312 (5.3) 0 (0.0) 402 (5.8)
Mu 519 (50.0) 2581 (43.6) 0 (0.0) 3100 (44.5)
Omicron 316 (30.4) 2072 (35) 1448 (69.9) 2388 (34.3)
Initial vaccine schedule manufacturerb
AstraZeneca NA 671 (11.8) 214 (11.4) 671 (11.8)
Janssen NA 421 (7.4) 68 (3.6) 421 (7.4)
Moderna NA 234 (4.1) 44 (2.4) 234 (4.1)
Pfizer NA 2996 (52.7) 1056 (56.4) 2996 (52.7)
Sinovac NA 1362 (24) 489 (26.1) 1362 (24)
a This group includes people with full schedule, which is made up of 2 subgroups: 1) those who completed the schedule and did not receive any booster during the entire study period and 2) those who completed the schedule and had a booster but were included analytically in the period before receiving the booster (censored booster doses).
b Refers to the initial schedule’s manufacturer. The additional dose is not considered for the group that received a booster, because this information is described in detail in Table 2.
IQR, interquartile range; NA, not applicable; SARS-CoV-2, severe acute respiratory syndrome coronavirus 2.
In the fully vaccinated group, the most frequently used vaccines were Pfizer (52.7%), Sinovac (24.0%), AstraZeneca (11.8%), Janssen (7.4%), and Moderna (4.1%). Furthermore, most subjects received a homologous schedule (n = 3425) in this group, with mainly Pfizer being administered (n = 1720), followed by Sinovac (n = 785), AstraZeneca (n = 407), Jansen (n = 349), and Moderna (n = 164). Regarding the vaccinated with booster group who had a homologous schedule (ie, 3 doses from the same manufacturer), Pfizer (61.7%), Sinovac (20.2%), and AstraZeneca (11.4%) were the most used. In the case of heterologous schedule in those who had a booster, the main combinations were Pfizer/Moderna (29.1%), Pfizer/AstraZeneca (21.8%), and Sinovac/Pfizer (22.8%). The complete description of all combinations is shown in Table 2.
TABLE 2. Complete vaccination schedule and booster in solid-organ transplant recipients in Colombia, by manufacturer, ESPERANZA cohort
Vaccination schedule Homologous(n = 947)
n %
Pfizer 584 61.7
AstraZeneca (AZ) 108 11.4
Moderna 11 1.2
Sinovac 192 20.2
Janssen 52 5.5
Heterologous (n = 990)
n %
2 AZ + Pfizer 67 6.8
2 AZ + Moderna 37 3.7
2 AZ + Janssen 4 0.4
2 Pfizer + AZ 216 21.8
2 Pfizer + Moderna/Pfizer 289 29.1
2 Pfizer + Janssen 9 1.0
2 Moderna + AZ 5 0.5
2 Moderna + Pfizer/Moderna 27 2.7
2 Moderna + Janssen 1 0.1
2 Sinovac + AZ 73 7.4
2 Sinovac + Pfizer/Moderna 226 22.8
2 Sinovac + Janssen 7 0.7
Other combinations 29 3.0
Records that did not specify the manufacturer of the vaccination schedule were excluded from this table
As to the outcomes occurrence in the exposure groups, the risk of COVID-19 infection, hospitalization, and death due to COVID-19 was 26.4%, 9.6%, and 7.5%, respectively, in the unvaccinated group. On the other hand, figures were significantly lower in the fully vaccinated group, so that the risk of COVID-19 infection was 11.5%; risk of hospitalization, 2.7%; and risk of death due to COVID-19, 1.2%. Moreover, these risks were even lower in the subset that received a booster: 7.8% for COVID-19 infection; 2.7% for hospitalization; and 0.7% for death due to COVID-19. The listed differences between groups were statistically significant (P < 0.01). Table 3 lists the outcomes’ occurrence and their time-to-event in detail, according to the exposure group.
TABLE 3. Study outcomes in unvaccinated and fully vaccinated solid-organ transplant recipients in Colombia, ESPERANZA cohort
Outcome Unvaccinated(n = 1038)n (%) Fully vaccinateda(n = 5925)n (%) Vaccine booster(n = 2072)n (%) Total(n = 6963)n (%)
Confirmed COVID-19 infection 272 (26.4) 679 (11.5) 153 (7.8) 951 (13.7)
COVID-19 hospitalization 99 (9.6) 160 (2.7) 33 (1.7) 259 (3.7)
Death due to COVID-19 77 (7.5) 70 (1.2) 14 (0.7) 147 (2.1)
Time-to-outcome (d)b
Time-to-infection; median (IQR) 411 (255–411) 292 (222–309) 103 (62–147) 294 (223–316)
Time-to-hospitalization; median (IQR) 411 (411–411) 298 (264–313) 111 (71–153) 300 (268–338)
Time-to-death; median (IQR) 411 (411–411) 298 (267–315) 112 (72–153) 301 (272–344)
a This group includes people with full schedule, which is made up of 2 subgroups: 1) those who completed the schedule and did not receive any booster during the entire study period and 2) those who completed the schedule and had a booster but were included analytically in the period before receiving the booster (censored booster doses).
b The follow-up time of unvaccinated individuals was longer for all the assessed outcomes. Given that this group was made up of individuals who remained unvaccinated throughout the whole study period, they contributed a larger time-to-event than vaccinated people, whose time-to-event only counted after receiving the vaccine.
COVID-19, coronavirus disease 2019; IQR, interquartile range.
In relation to the survival analysis, a lower survival was found in those unvaccinated compared with fully vaccinated individuals. Furthermore, the difference in the survival time when comparing unvaccinated subjects to those who received a booster was not as large as that found when the comparison was made between the unvaccinated and the fully vaccinated groups; and this trend persisted over time. The unadjusted survival analysis curves are depicted in Figure 2.
FIGURE 2. Kaplan-Meier survival curves for COVID-19 infection, hospitalization, and death, according to vaccination status in solid-organ transplant recipients in Colombia. ESPERANZA cohort. COVID-19, coronavirus disease 2019.
On the other hand, the overall effectiveness of the complete vaccination schedule was 73.3% (95% CI, 68.9%-77.0%) to prevent COVID-19 infection; 83.7% (95% CI, 78.7%-87.5%) to prevent hospitalization; and 92.1% (95% CI, 88.8%-94.4%) to prevent death due to COVID-19. Table 4 shows the effectiveness estimates of being fully vaccinated according to age groups.
TABLE 4. COVID-19 complete schedule effectiveness for infection, hospitalization, and death in solid-organ transplant recipients in Colombia, by age group, ESPERANZA cohort
COVID-19 complete schedule effectiveness
Age group Infection Hospitalization Death
%(95% CI) P %(95% CI) P %(95% CI) P
All age groups 73.3 (68.9-77.0) <0.001 83.7 (78.7-87.5) <0.001 92.1 (88.8-94.4) <0.001
16–59 73.1 (67.8-77.5) <0.001 83.3 (76.6-88.1) <0.001 93.6 (89.5-96.1) <0.001
60 and older 72.6 (63.6-79.3) <0.001 82.5 (72.8-88.8) <0.001 91.0 (84.8-94.7) <0.001
Estimates adjusted for sex, affiliation regime, municipality of residence, presence of comorbidities (CKD, cancer, DM, HBP, and HIV infection), and prevalent variant at the time of infection. Unvaccinated individuals were the reference group in all models.
CI, confidence interval; CKD, chronic kidney disease; DM, diabetes mellitus; HBP, high blood pressure.
Likewise, the effectiveness of the vaccine booster to prevent the study outcomes was higher for death due to COVID-19 (94.5%; 95% CI, 89.8%-97.1%), followed by hospitalization (86.9%; 95% CI, 79.4%-91.6%) and COVID-19 infection (76.7%; 95% CI, 70.6%-81.5%). Similar results were obtained in both age groups (16–59 y and 60 y and older), with only a minimum difference between age groups by outcome. The complete estimates of the vaccine booster effectiveness according to age groups are presented in Table 5.
TABLE 5. COVID-19 booster effectiveness for infection, hospitalization, and death in solid-organ transplant recipients in Colombia, by age group, ESPERANZA cohort
COVID-19 booster effectiveness
Age group Infection Hospitalization Death
%(95% CI) P %(95% CI) P %(95% CI) P
All age groups 76.7 (70.6-81.5) <0.001 86.9 (79.4-91.6) <0.001 94.5 (89.8-97.1) <0.001
16–59 75.9 (67.1-82.3) <0.001 84.2 (69.6-91.8) <0.001 95.8 (86.2-98.8) <0.001
60 and older 70.2 (56.5-79.6) <0.001 86.4 (74.2-92.8) <0.001 91.1 (80.9-95.9) <0.001
Estimates adjusted for sex, affiliation regime, municipality of residence, presence of comorbidities (CKD, cancer, DM, HBP, and HIV infection), and prevalent variant at the time of infection. Unvaccinated individuals were the reference group in all models.
CI, confidence interval; CKD, chronic kidney disease; DM, diabetes mellitus; HBP, high blood pressure.
Finally, the effectiveness of all the complete vaccination and booster schedules are presented in Tables S1 and S2 (SDC, http://links.lww.com/TP/C615). A high effectiveness of the analyzed vaccines was observed with all the homologous and heterologous combinations analyzed. However, some estimates were very imprecise with wide confidence intervals given the small sample size for some of the studied groups. Finally, it is important to highlight that in all cases the effectiveness in preventing hospitalization and death due to COVID-19 was greater than in preventing the occurrence of confirmed infection.
DISCUSSION
This research found a high effectiveness of the complete vaccination schedule and the vaccination with booster in SOTRs, which represents a significant impact of vaccination in immunized SOTRs when compared with unvaccinated individuals. It is important to highlight that our findings are a measure of the infection and complications risk reduction within the SOTR population; hence, they cannot be directly extrapolated to other populations, nor are they comparable to effectiveness estimates from immunocompetent people, who are known to have a better response to vaccines.23 To make comparisons against other groups, impact measures (absolute estimates) would be required, which is beyond the scope of the present investigation.
The obtained results also that suggest the protection granted by immunization in SOTRs begins with a complete vaccination schedule, initially defined by the manufacturers, and that a booster could strengthen such protection in time. On the other hand, this research included one of the biggest SOTR samples evaluated to this day, which not only was assessed in real-life conditions but also allowed the estimation of the vaccination effectiveness by using national epidemiologic data, whereas most of the published research in the SOTR population regarding this topic are immunogenicity studies.8,14,15,24-26 Furthermore, this investigation also estimated the effectiveness of several heterologous schedules, including combinations of inactivated, adenovirus vector-based and mRNA-based vaccines and adjusted for prevalent SARS-CoV-2 variants in Colombia, evidencing a high effectiveness of the COVID-19 vaccines in the SOTR population.
The need to vaccinate immunosuppressed individuals was also underlined because our findings showed immunization significantly reduced mortality and morbidity relative to not receiving any vaccine in this group. This strengthens the public policy of vaccination aimed at preventing disease and the risk prioritization as a main component of public health, primary healthcare systems, and national vaccination programs. Accordingly, our findings also suggest the prioritization of SOTRs in Colombia might have significantly impacted the reduction of the morbimortality of this population subset.
Our work also allows recognition of vaccination as a potential cost-effective strategy in terms of the burden of disease caused by COVID-19 and years of potential life lost because of mortality or disability in a young population that has several comorbidities.27-29 There are also implications for clinical practice because receiving immunosuppressive drugs is a direct risk factor for COVID-19 death,30 although vaccination could be a preventive intervention for it. It also has to be considered that both the ability to prevent infection by activating the immune system and the risk of COVID-19 infection are related to the person’s immunosuppression status and to its competence to mount an immune response; therefore, the greater the immunodeficiency, the higher chance of an inadequate response to the biologic or the vaccine-induced immunization. Risk factors for an insufficient immune response comprise several individual aspects, such as age and receiving immunosuppressive therapy, which, in the case of SOTRs, needs to be considered along the underlying disease that caused the organ transplant in the first place (eg, kidney or liver failure). Thus, an adequate immune response cannot be assumed in all cases despite vaccination confers benefits to immunosuppressed individuals; therefore, the relevance of booster doses.31-34 Unfortunately, this study could not collect information about the transplanted organ or the therapy the participants were using, which impeded the analysis of their role in the effectiveness of vaccines.
Previous investigations have also found vaccination to be a good strategy in SOTRs to prevent COVID-19. For example, a research found a reduction in the incidence of symptomatic COVID-19 in vaccinated SOTRs (0.065 per 1000 person-days; 95% CI, 0.024-0.17) compared with unvaccinated subjects (0.34 per 1000 person-days; 95% CI, 0.26-0.44), which evidences a high effectiveness in this risk group.35 Other examples include an investigation carried out in Israel, where a cohort was immunized with mRNA-based vaccines and the effectiveness for symptomatic COVID-19 infection was found to be 71% (95% CI, 37%-87%) in immunosuppressed patients, whereas it was 94% (95% CI, 88%-97%) in general population36; and retrospective studies that have suggested a lower vaccination effectiveness to prevent COVID-19-related hospitalization in immunocompromised patients, as shown in a population with an immunosuppression prevalence of 44%.37
With regard to the booster dose, a study also found that only two-thirds of the included SOTRs generated anti-SARS-CoV-2 antibodies.38 This correlates to our findings, in which the survival time between the unvaccinated and vaccinated with boosted individuals was lower than when the comparison was made between the unvaccinated and the fully vaccinated subjects. Additionally, previous publications have also reported a poor response to COVID-19 vaccines in SOTRs, as indicated in a meta-analysis that estimated seroconversion was 16 times less likely to occur after vaccination in SOTRs.39
Results on this subject are heterogeneous and show that differences in control groups influence the conclusions. It is important to clarify that information related to the effectiveness of booster vaccines in SOTR was not found because most published studies assessed the vaccine-induced immune response, with a focus on the humoral response, but did not estimate its impact on the protection against SARS-CoV-2, which does not allow a direct comparison with our results.8,11,40
The limitations of this work comprise of high effectiveness estimates of the COVID-19 vaccines because the comparison was made against unvaccinated SOTRs instead of nontransplanted individuals. Moreover, these estimates might be affected by the lack of inclusion of certain covariates that could act as potential confounders, such as other diseases, the type of the administered immunosuppressive drugs, the educational level or the transplanted organ (although kidney transplants are the most frequent in Colombia). Lastly, vaccination effectiveness throughout time was not estimated, which is one of the main questions to be answered, considering the insufficient immune response seen in SOTRs. New studies are required that aim to not only respond this query but also address the impact of the hybrid immunity and compare the vaccines, effectiveness over time against other immunocompromised individuals and the general population.
In conclusion, our research evidences the relevant and coherent measures taken by the Colombian government when implementing the COVID-19 vaccination, which focused on prioritizing the most vulnerable groups, intervening the possible virus-related mortality causes, and decreasing the health inequities potentially caused by the COVID-19 syndemic. This study also serves as input to keep the recommendation to prioritize SOTR vaccination worldwide and to guarantee the timely access to booster doses.
Supplementary Material
This study was conducted using financial and economic resources provided by the Ministry of Health and Social Protection of Colombia (Ministerio de Salud y Protección Social).
M.P.A., J.F.N., L.A.C., and M.R.B. did design and statistical analysis conceptualization. J.F.N., L.A.C., and A.F.P. did data acquisition. M.P.A., J.F.N., L.A.C., M.R.B., A.F.P., and M.G.P. did data verification. M.P.A. and A.F.P. did dataset construction. M.P.A., J.F.N., and A.F.P. did statistical analysis. M.P.A., J.F.N., L.A.C., and M.R.B. drafted of the article. M.P.A., J.F.N., L.A.C., M.R.B., A.F.P., M.G.P., and F.R.G. did critical revision and discussion of the article.
J.F.N., L.A.C., and F.R.G. were members of the Colombian COVID-19 vaccine advisory committee. All other authors declare no competing interests.
This manuscript has been previously submitted to Social Science Research Network as a preprint; https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4166380.
Supplemental Visual Abstract; http://links.lww.com/TP/C616.
Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com).
==== Refs
REFERENCES
1. World Health Organization. WHO Target Product Profiles for COVID-19 Vaccines. 2022. Available at https://www.who.int/publications/m/item/who-target-product-profiles-for-covid-19-vaccines.
2. Ministerio de Salud y de la Protección Social. Resolución 385. Bogotá D.C., Colombia; 2020.
3. Departamento Administrativo De La Presidencia De La República (DAPRE). Decreto 109 de 2021: Por El Cual Se Adopta El Plan Nacional de Vacunación Contra El COVID-19 y Se Dictan Otras Disposiciones. Bogotá D.C., Colombia; 2021.
4. Ministerio de Salud y de la Protección Social, Departamento Nacional de Planeación, Ministerio de Hacienda y Crédito Público, Instituto de Evaluación Tecnológica en Salud. Plan Nacional de Vacunación Contra El COVID-19. Bogotá D.C., Colombia; 2021. Available at https://www.minsalud.gov.co/sites/rid/Lists/BibliotecaDigital/RIDE/VS/pnv-contra-covid-19.pdf.
5. Grupper A Rabinowich L Schwartz D . Reduced humoral response to mRNA SARS-CoV-2 BNT162b2 vaccine in kidney transplant recipients without prior exposure to the virus. Am J Transplant. 2021;21 :2719–2726.33866672
6. Itzhaki Ben Zadok O Shaul AA Ben-Avraham B . Immunogenicity of the BNT162b2 mRNA vaccine in heart transplant recipients—a prospective cohort study. Eur J Heart Fail. 2021;23 :1555–1559.33963635
7. Prendecki M Thomson T Clarke CL . Comparison of humoral and cellular responses in kidney transplant recipients receiving BNT162b2 and ChAdOx1 SARS-CoV-2 vaccines. medRxiv. [Preprint. July 14, 2021]. doi:10.1101/2021.07.09.21260192.
8. Kamar N Abravanel F Marion O . Three doses of an mrna covid-19 vaccine in solid-organ transplant recipients. N Engl J Med. 2021;385 :661–662.34161700
9. Cucchiari D Egri N Bodro M . Cellular and humoral response after MRNA-1273 SARS-CoV-2 vaccine in kidney transplant recipients. Am J Transplant. 2021;21 :2727–2739.34036720
10. Bruminhent J Boongird S Assanatham M . Anti–SARS-CoV-2 spike protein S1 receptor-binding domain antibody after vaccination with inactivated whole-virus SARS-CoV-2 in end-stage kidney disease patients: an initial report. 2021;100 :1136–1138.
11. Werbel WA Boyarsky BJ Ou MT . Safety and immunogenicity of a third dose of SARS-CoV-2 vaccine in solid organ transplant recipients: a case series. Ann Intern Med. 2021;174 :1330–1332.34125572
12. Limaye A Hardinger K . COVID-19: issues related to solid organ transplantation. 2022. Available at https://www.uptodate.com/contents/covid-19-issues-related-to-solid-organ-transplantation. Accessed July 16, 2022.
13. Jordan RE Adab P Cheng KK . Covid-19: risk factors for severe disease and death. BMJ. 2020;368 :m1198.32217618
14. Hall VG Ferreira VH Ku T . Randomized trial of a third dose of mRNA-1273 vaccine in transplant recipients. N Engl J Med. 2021;385 :1244–1246.34379917
15. Kumar D Ferreira VH Hall VG . Neutralization of SARS-CoV-2 variants in transplant recipients after two and three doses of mRNA-1273 vaccine: secondary analysis of a randomized trial. Ann Intern Med. 2022;175 :226–233.34807716
16. Holden IK Bistrup C Nilsson AC . Immunogenicity of SARS-CoV-2 mRNA vaccine in solid organ transplant recipients. J Intern Med. 2021;290 :1264–1267.34237179
17. Ramanathan M Murugesan K Yang LM . Cell-mediated and humoral immune response to 2-dose SARS-CoV2 mRNA vaccination in immunocompromised patient population. medRxiv. [Preprint. July 23, 2021]. doi:10.1101/2021.07.21.21260921.
18. Arregocés-Castillo L Fernández-Niño J Rojas-Botero M . Effectiveness of COVID-19 vaccines in older adults in Colombia: a retrospective, population-based study of the ESPERANZA cohort. Lancet Healthy Longev. 2022;3 :e242–e252.35340743
19. Ministerio de Salud y de la Protección Social. Resolución 1151. Colombia; 2021.
20. Ministerio de Salud y de la Protección Social. Resolución 1426. Bogotá D.C., Colombia; 2021.
21. Ministerio de Salud y de la Protección Social. Resolución 419. Bogotá D.C., Colombia; 2022.
22. Ministerio de Salud de la República de Colombia. Resolución 8430 de 1993. Bogotá D.C., Colombia; 1993.
23. Ssentongo P Ssentongo AE Voleti N . SARS-CoV-2 vaccine effectiveness against infection, symptomatic and severe COVID-19: a systematic review and meta-analysis. BMC Infect Dis. 2022;22 :439.35525973
24. Ali H Alterki A Sindhu S . Robust antibody levels in both diabetic and non-diabetic individuals after BNT162b2 mRNA COVID-19 vaccination. Front Immunol. 2021;12 :752233.34899701
25. Davidov Y Tsaraf K Cohen-Ezra O . Immunogenicity and adverse effects of the 2-dose BNT162b2 messenger RNA vaccine among liver transplantation recipients. Liver Transpl. 2022;28 :215–223.34767690
26. Jurdi A Al Gassen RB Borges TJ . Diminished antibody response against SARS-CoV-2 Omicron variant after third dose of mRNA vaccine in kidney transplant recipients. medRxiv. [Preprint. January 6, 2022]. doi:10.1101/2022.01.03.22268649.
27. Vaezi A Meysamie A . COVID-19 vaccines cost-effectiveness analysis: a scenario for Iran. Vaccines (Basel). 2021;10 :37.35062698
28. Pearson CAB Bozzani F Procter SR . COVID-19 vaccination in Sindh Province, Pakistan: a modelling study of health impact and cost-effectiveness. PLOS Med. 2021;18 :e1003815.34606520
29. Reddy KP Fitzmaurice KP Scott JA . Clinical outcomes and cost-effectiveness of COVID-19 vaccination in South Africa. Nat Commun. 2021;12 :6238.34716349
30. Kim AHJ Sparks JA . Immunosuppression and SARS-CoV-2 breakthrough infections. Lancet Rheumatol. 2022;4 :e379–e380.35527809
31. Gangappa S Kokko KE Carlson LM . Immune responsiveness and protective immunity after transplantation. Transpl Int. 2008;21 :293–303.18225995
32. Eckerle I Rosenberger KD Zwahlen M . Serologic vaccination response after solid organ transplantation: a systematic review. PLoS One. 2013;8 :e56974.23451126
33. Chong PP Avery RK . A Comprehensive review of immunization practices in solid organ transplant and hematopoietic stem cell transplant recipients. Clin Ther. 2017;39 :1581–1598.28751095
34. Kotton CN Hibberd PL . Immunizations in solid organ transplant candidates and recipients. 2022. Available at https://www.uptodate.com/contents/immunizations-in-solid-organ-transplant-candidates-and-recipients. Accessed July 16, 2022.
35. Aslam S Adler E Mekeel K . Clinical effectiveness of COVID-19 vaccination in solid organ transplant recipients. Transpl Infect Dis. 2021;23 :e13705.34324256
36. Chodick G Tene L Rotem RS . The effectiveness of the two-dose BNT162b2 vaccine: analysis of real-world data. Clin Infect Dis. 2022;74 :472–478.33999127
37. Tenforde MW Patel MM Ginde AA ; Influenza and Other Viruses in the Acutely Ill (IVY) Network. Effectiveness of severe acute respiratory syndrome coronavirus 2 messenger RNA vaccines for preventing coronavirus disease 2019 hospitalizations in the United States. Clin Infect Dis. 2022;74 :1515–1524.34358310
38. Kamar N Abravanel F Marion O . Anti-SARS-CoV-2 spike protein and neutralizing antibodies at 1 and 3 months after three doses of SARS-CoV-2 vaccine in a large cohort of solid organ transplant patients. Am J Transplant. 2022;22 :1467–1474.35000296
39. Lee ARYB Wong SY Chai LYA . Efficacy of Covid-19 vaccines in immunocompromised patients: systematic review and meta-analysis. BMJ. 2022;376 :e068632.35236664
40. Benotmane I Gautier G Perrin P . Antibody response after a third dose of the mRNA-1273 SARS-CoV-2 vaccine in kidney transplant recipients with minimal serologic response to 2 doses. JAMA. 2021;326 :1063–1065.34297036
| 36228269 | PMC9746232 | NO-CC CODE | 2022-12-15 23:21:55 | no | Transplantation. 2023 Jan 8; 107(1):216-224 | utf-8 | Transplantation | 2,022 | 10.1097/TP.0000000000004411 | oa_other |
==== Front
Transplantation
Transplantation
TP
Transplantation
0041-1337
1534-6080
Lippincott Williams & Wilkins Hagerstown, MD
36117251
00025
10.1097/TP.0000000000004340
3
Original Clinical Science—General
PASC in Solid Organ Transplant Recipients With Self-reported SARS-CoV-2 Infection
Alasfar Sami MD 1
Chiang Teresa Po-Yu MD, MPH 2
Snyder Andrew J. BS 3
Ou Michael T. BS 2
Boyarsky Brian J. MD, PhD 2
Abedon Aura T. BS 2
Alejo Jennifer L. MD 2
Cook Sydney BS 4
Cochran Willa CRNP 1
Brigham Emily MD 56
Parker Ann M. MD, PhD 1
Garonzik-Wang Jacqueline MD, PhD 7
Massie Allan B. PhD 8
Brennan Daniel C. MD 1
Vannorsdall Tracy PhD 9
Segev Dorry L. MD, PhD 8
Avery Robin K. MD 1
1 Department of Medicine, Johns Hopkins University School of Medicine, Baltimore, MD.
2 Department of Surgery, Johns Hopkins University School of Medicine, Baltimore, MD.
3 University of Washington School of Medicine, Seattle, WA.
4 Georgetown University School of Medicine, Washington, DC.
5 Department of Medicine, University of British Columbia, Vancouver, BC.
6 Vancouver Coastal Health Research Institute, Vancouver, BC.
7 Department of Surgery, University of Wisconsin School of Medicine and Public Health, Madison, WI.
8 Department of Surgery, New York University Grossman School of Medicine, New York, NY.
9 Department of Psychiatry and Behavioral Sciences, Johns Hopkins University School of Medicine, Baltimore, MD.
Correspondence: Sami Alasfar, MD, Department of Medicine, Johns Hopkins University School of Medicine, 610 N. Wolfe St, Baltimore, MD 20286. ([email protected]).
19 9 2022
1 2023
19 9 2022
107 1 181191
2 3 2022
5 7 2022
29 7 2022
Copyright © 2022 Wolters Kluwer Health, Inc. All rights reserved.
2022
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
Background.
Postacute sequelae of SARS-CoV-2 infection (PASC) is an increasingly recognized phenomenon and manifested by long-lasting cognitive, mental, and physical symptoms beyond the acute infection period. We aimed to estimate the frequency of PASC symptoms in solid organ transplant (SOT) recipients and compared their frequency between those with SARS-CoV-2 infection requiring hospitalization and those who did not require hospitalization.
Methods.
A survey consisting of 7 standardized questionnaires was administered to 111 SOT recipients with history of SARS-CoV-2 infection diagnosed >4 wk before survey administration.
Results.
Median (interquartile range) time from SARS-CoV-2 diagnosis was 167 d (138–221). Hospitalization for SARS-CoV-2 infection was reported in 33 (30%) participants. Symptoms after the COVID episode were perceived as following: significant trauma (53%), cognitive decline (50%), fatigue (41%), depression (36%), breathing problems (35%), anxiety (23%), dysgeusia (22%), dysosmia (21%), and pain (19%). Hospitalized patients had poorer median scores in cognition (Quick Dementia Rating System survey score: 2.0 versus 0.5, P = 0.02), quality of life (Health-related Quality of Life survey: 2.0 versus 1.0, P = 0.015), physical health (Global physical health scale: 10.0 versus 11.0, P = 0.005), respiratory status (Breathlessness, Cough and Sputum Scale: 1.0 versus 0.0, P = 0.035), and pain (Pain score: 3 versus 0 out of 10, P = 0.003). Among patients with infection >6 mo prior, some symptoms were still present as following: abnormal breathing (42%), cough (40%), dysosmia (29%), and dysgeusia (34%).
Conclusions.
SOT recipients reported a high frequency of PASC symptoms. Multidisciplinary approach is needed to care for these patients beyond the acute phase.
SDCT
==== Body
pmcINTRODUCTION
The SARS-CoV-2 is responsible for COVID-19. In the acute phase, SARS-CoV-2 affects multiple organ systems, including the respiratory, gastrointestinal, and neurologic systems.1 Early in the pandemic, people with mild to moderate COVID-19 were believed to have a short-term course of acute illness. However, emerging data have showed the persistence of symptoms in a subgroup of patients, which can manifest for several weeks to months.2-6 In fact, it is now accepted that the course of SARS-CoV-2 infection is composed of 3 phases: acute infection in the first 2 wk, postacute hyperinflammatory state between weeks 2 and 4, and postacute sequelae, which manifest 5 wk or later following the acute infection.7
Postacute sequelae of SARS-CoV-2 infection (PASC), that is, COVID-19 “Long Hauler,” is an increasingly recognized phenomenon and manifested by long-lasting respiratory symptoms, cardiac dysfunction, kidney dysfunction, distortion of smell and taste senses, and changes in mental, neurocognitive, and physical function resulting in decreased quality of life.5,8,9 There is no unified definition for PASC, but it has been suggested to include the persistence of acute COVID-19 symptoms or development of sequelae beyond 4 wks from the time of COVID-19 diagnosis.7,8 The estimated prevalence of PASC in the general population ranges between 5% and 80% with the highest percentage to date reported among people who were hospitalized for COVID-19.10
In parallel, solid organ transplantation (SOT) and immunosuppressive drugs are known to have significant neurotropic effects, which can lead to mental, neurocognitive, and physical disorders potentially exacerbating these long-term effects of COVID-19 infection.11-14 In addition, individuals with SOT are high risk for developing severe SARS-CoV-2 infection, and therefore might have worse long-term effects.15-19 Mental, neurocognitive, and physical manifestations of PASC seem likely to occur in SOT recipients with SARS-CoV-2 infection. However, data on their prevalence and impact on quality of life in SOT recipients are limited.
In this study, we aimed to estimate the frequency of cognitive, mental, and physical impairments in SOT recipients with history of SARS-CoV-2 infection in the short-term (1–6 mo) and long-term periods (>6 mo), as reported by study participants through a series of surveys. We also compared the frequency of these complications between those with SARS-CoV-2 infection requiring hospitalization and those who did not require hospitalization.
MATERIALS AND METHODS
Study Design
This is a cross-sectional survey study conducted among a total of 111 patients with history of SOT and a self-reported diagnosis of SARS-CoV-2 infection. The study was conducted over the period from June 2021 to August 2021. A Research Electronic Data Capture form was created to administer the survey.20 The study was approved by the Institutional Review Board at the Johns Hopkins School of Medicine. Participants were consented electronically.
Study Population
Participants were recruited from a larger prospective observational study of vaccine safety and efficacy outcomes among SOT recipients as previously reported.21,22 In the parent study, SOT recipients from multiple different US transplant centers self-enrolled online, and hence all information was self-reported by study participants, and the investigators did not have access to patient medical records from their transplant centers. English-speaking SOT recipients ≥18 y old who self-reported as testing positive for SARS-CoV-2 were eligible to participate in the current survey study. Participant demographics (age, sex, race, education, occupation, transplant type and date, date of SARS-CoV-2 infection diagnosis, medications) were obtained via self-report.
Patients were divided into 2 groups: those who reported hospital admission due to SARS-CoV-2 infection (n = 33 patients), and those who reported not being admitted to the hospital because of SARS-CoV-2 infection (n = 78). To assess persistence of symptoms, we also stratified our population according to the time when they were diagnosed with SARS-CoV-2 infection: those with remote infection who were diagnosed over 6 mo before the date of survey administration, (n = 35), and those with recent infection diagnosed within the past 6 mo (n = 76). We excluded those with SARS-CoV-2 infection within the last month.
Description of Survey
The following standardized questionnaires were used in the survey:
Quick Dementia Rating System, Patient Version
The Quick Dementia Rating System (QDRS) patient version is a 10-item questionnaire assessing subjective changes across 10 domains of cognition, mood, and daily functioning. Scores range from 0 to 30 with higher scores representing greater perceived impairment.23,24 Scores ≥1.5 and ≥6 reflect clinically meaningful mild and moderate functional decline, respectively. The QDRS survey questions can be found in Tables S1 and S2 (SDC, http://links.lww.com/TP/C552). The second response of each item was assigned a score of 0.5. The rest of the responses were assigned a score of 1.
Patient Health Questionnaire-9
The Patient Health Questionnaire-9 (PHQ-9) is a 9-item depression screening questionnaire designed for use in medical settings. Participants are asked to rate 9 depression symptoms on a 0 (not at all) to 3 (nearly every day) scale. Cut points have been established for mild (5 of 27), moderate (10 of 27), moderately severe (15 of 27), and severe symptoms (20 of 27).25 The PHQ-9 survey questions can be found in Tables S3 and S4 (SDC, http://links.lww.com/TP/C552).
The General Anxiety Disorder-7
The General Anxiety Disorder-7 (GAD-7) is a 7-item anxiety screening questionnaire. Participants are asked to rate how bothered they have been by 7 anxiety symptoms on a 0 (not at all) to 3 (nearly every day) scale. Cut points have been established for normal (≤4 of 21), mild (5–9 of 21), moderate (10–14 of 20), and severe (>15 of 21) symptoms.26 The GAD-7 survey questions can be found in Tables S5 and S6 (SDC, http://links.lww.com/TP/C552).
Impact of Events Scale-6
The Impact of Events Scale-6 (IES-6) is a 6-item posttraumatic stress screening questionnaire that has been validated for use in medical populations. Participants are asked to rate statements on a 0 (not at all) to 4 (quite a bit) scale.27,28 IES-6 score of 0–1.74 indicates minimal trauma and score of 1.75 and higher indicates significant trauma. The IES-6 survey questions can be found in Tables S7 and S8 (SDC, http://links.lww.com/TP/C552).
Health-related Quality of Life EuroQol-5D
This survey assesses health-related quality of life by surveying 5 dimensions: mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. Each dimension has 5 levels: no problems, slight problems, moderate problems, severe problems, and extreme problems.29,30 The EuroQol-5D survey questions can be found in Tables S9 and S10 (SDC, http://links.lww.com/TP/C552).
The PROMIS Global Physical Health Scale
The patient-reported outcomes measurement information system (PROMIS) Global Physical Health Scale (GHS) instrument consists of 10 global health items that represent 5 core PROMIS domains (physical function, pain, fatigue, emotional distress, social health). We used 4 questions that cover 3 out of the 5 domains: physical function, pain, and fatigue. Three of these are administered using 5-category response scales, and 1 item (rating of pain on average) uses a response scale of 0–10.31 The PROMIS GHS survey questions can be found in Tables S11 and S12 (SDC, http://links.lww.com/TP/C552).
Breathlessness, Cough, and Sputum Scale
Breathlessness, Cough, and Sputum Scale (BCSS) is a patient-reported outcome measure evaluating symptoms in patients with chronic obstructive pulmonary disease and has been used to assess respiratory symptoms in other pulmonary conditions.32 The BCSS is a daily diary that asks patients to rate the severity of the 3 symptoms, each on a 5-point scale; higher scores indicate more severe symptoms.33,34 The BCSS survey questions can be found in Table S13 (SDC, http://links.lww.com/TP/C552).
Nonstandardized COVID-19 Symptoms Questionnaire
We developed a 3-item questionnaire to assess severity of SARS-CoV-2 infection-specific symptoms: dysosmia, dysgeusia, and diarrhea. Each question is on 4-point scale. Respondents were asked about smell and taste symptoms on the day they took the survey, and about diarrhea in the past 2 wk before the survey. The survey questions can be found in Table S14 (SDC, http://links.lww.com/TP/C552).
Survey Administration
An electronic survey, hosted on Research Electronic Data Capture database hosted at Johns Hopkins, was created and consisted of patient demographics (age, sex, race, education, occupation, transplant type and date, date of COVID-19 diagnosis, medications), in addition to the surveys mentioned above.
Analysis
We compared the characteristics of patients between the groups using Fisher’s exact test for categorical variables and Wilcoxon rank-sum test for continuous variables. An α of 0.05 was used to determine statistical significance. Item-level analyses are exploratory in nature. All analyses were performed using Stata/SE 15.1.
The raw score of QDRS was transformed into 2 severity strata with scores ≥1.5 and ≥6 reflecting clinically meaningful mild and moderate functional decline, respectively. It was also stratified into 6 categories that include normal (<1.5), mild functional decline (1.5–5.9), moderate functional decline (6–12.4), moderately severe functional decline (12.5–17.4), and severe functional decline (>17.4). The raw score of PHQ-9 for severity of depression was transformed into minimal (0–4), mild (5–9), moderate (10–14), moderately severe (15–19), and severe depression (20–27). The raw score of GAD-7 for anxiety was transformed into no anxiety (0–4), mild (5–9), moderate (10–14), and severe anxiety (15–21). IES-6 was transformed into minimal trauma (0–1.74) and traumatic (1.75 and above). The extent of physical impairment in smell, taste, and bowel movements was evaluated separately using a 4-tier severity index. All the other scales (EQ-5D, PROMIS GHS, BCSS) were analyzed using raw scores. For the PROMIS survey, we used 4 questions that cover 3 out of the 5 domains of the survey. We then used univariable ordinal logistic regression to study factors (age, sex, race, type of organ transplant, mycophenolate mofetil, or mycophenolic acid use) potentially associated with the risk of having a higher level of severity in QDRS, PHQ-9, GAD-7, and the extent of physical impairment in taste, smell, and bowel movements. Univariable logistic regression examined factors associated with clinically significant trauma-related distress defined by IES-6. Univariable Poisson regression examined factors associated with higher raw score in EQ-5D, PROMIS GHS, and BCSS.
RESULTS
Population Characteristics
The surveys were distributed to 167 eligible SOT recipients, of whom 111 (66%) responded (Table 1). Median (interquartile range) age was 58 y (46, 65). Most patients were female (55%) and White (76.6%). Most patients were recipients of kidney transplants (60.4%) followed by liver (20.7%), heart (11.7%), lung (9%), and kidney/pancreas (7.2%) transplants. Around 30% (33 of 111) reported hospitalization for COVID-19 infection. Using available vaccination records by the time when this project was analyzed, 19 of the 111 participants reported having received at least 1 COVID-19 vaccination before the self-reported diagnosis of COVID-19 (n = 18 two doses, n = 1 one/first dose). An additional 63 participants reported having received vaccination (n = 1 three doses, n = 56 two doses, n = 6 one/first dose) after their COVID-19 diagnosis.
TABLE 1. Patient characteristics
Characteristic Value
Age, median (IQR) 57.1 (46.0, 64.8)
Sex
Male 33 (29.7%)
Female 61 (55.0%)
No answer 17 (15.3%)
Race
White 85 (76.6%)
Black or African American 4 (3.6%)
Asian 1 (0.9%)
Multiracial 3 (2.7%)
No answer 1 (0.9%)
Hispanic
Yes 4 (3.6%)
No 90 (81.1%)
No answer 17 (15.3%)
Organ Transplant Kidney 67 (60.4%)
Liver 23 (20.7%)
Kidney/Pancreas 8 (7.2%)
Heart 13 (11.7%)
Lung 10 (9.0%)
Median y from transplant (IQR) 6.0 (2.3, 11.6)
Median d from diagnosis of SARS-CoV-2 infection (IQR) 159 (126., 206.)
Hospitalization for SARS-CoV-2 infection
No 78 (70.3%)
Yes 33 (29.7%)
Hospitalization category
Non-ICU 25 (22.5%)
ICU 8 (7.2%)
Mechanical ventilation
Yes 2 (1.8%)
Highest grade in school (K-12)?
6th grade 1 (0.9%)
12th grade 110 (99.1%)
Highest level of education
High school (or GED) degree 2 (1.8%)
Some college 18 (16.2%)
Technical/certificate school 5 (4.5%)
Associates degree 14 (12.6%)
Bachelor’s degree 34 (30.6%)
Master’s degree 25 (22.5%)
Doctoral degree 11 (9.9%)
No answer 1 (0.9%)
Work
Retired 30 (27.0%)
Management 13 (11.7%)
Healthcare providers 15 (13.5%)
On disability 6 (5.4%)
Other occupations 41 (36.9%)
No answer 6 (5.4%)
Use of Mycophenolate or Azathioprine
No 37 (33.3%)
Yes 61 (55.0%)
No answer 13 (11.7%)
Use of corticosteroids for immunosuppression
No 33 (29.7%)
Yes 65 (58.6%)
No answer 13 (11.7%)
Use of calcineurin inhibitor
No 18 (16.2%)
Yes 80 (72.1%)
No answer 13 (11.7%)
IQR, interquartile range.
Survey Responses
Quick Dementia Rating System Survey
Results of the QDRS survey are shown in Figure 1, Tables 2–4. Around 50% of patients reported at least mild functional decline defined by a QDRS score of ≥1.5. This percentage was higher in the hospitalized group compared with the nonhospitalized group although not statistically significant (61% versus 45%, P = 0.15) (Table 4), and was not different between those with recent versus remote infection (50% versus 49%, P = 1.0). Hospitalized participants had worse scores in the components of memory, decision-making and problem-solving abilities, level of activities inside and outside the home, and mood changes (Tables S1 and S2, SDC, http://links.lww.com/TP/C552).
FIGURE 1. Distribution of QRDS category by time from infection and hospitalization status. QDRS, Quick Dementia Rating System.
Patient Health Questionnaire-9 Survey
Results of the PHQ-9 survey are shown in Figure 2, Tables 2–4. Around 36% of patients reported at least mild depression symptoms defined by a PHQ-9 score of 5 and above. This percentage was not different between the hospitalized and nonhospitalized groups (39% versus 35%, P = 0.50) (Table 4) and was not different between those with recent versus remote COVID-19 infection (37% versus 34%, P = 1.0). Hospitalized patients had worse scores in the components of sleeping difficulties, appetite changes, and self-esteem (Tables S3 and S4, SDC, http://links.lww.com/TP/C552).
FIGURE 2. Distribution of PHQ-9 category by time from infection and hospitalization status among respondents. PHQ-9, Patient Health Questionnaire-9.
General Anxiety Disorder-7 Survey
Results of the GAD-7 survey are shown in Figure 3, Tables 2–4. Around 23% of patients reported at least mild anxiety defined by a GAD-7 score of 5 and above. This percentage was not different between the hospitalized and nonhospitalized groups (27% versus 22%, P = 0.62), or between those with recent versus remote COVID-19 infection (21% versus 29%, P = 0.47). There were no differences in answers to the different components of the GAD-7 survey between the hospitalized and nonhospitalized groups (Tables S5 and S6, SDC, http://links.lww.com/TP/C552).
FIGURE 3. Distribution of GAD-7 category by time from infection and hospitalization status among respondents. GAD-7, Generalized Anxiety Disorder-7.
Impact of Events Scale-6 Survey
Results of the IES-6 survey are shown in Figure 4, Tables 2–4. Around 53% of patients reported significant trauma-related distress defined by an IES-6 score of 1.75 and above. This was not different between the hospitalized and nonhospitalized groups (58% versus 51%, P = 0.39), or between those with recent versus remote COVID-19 infection (57% versus 46%, P = 0.40). There were no differences in answers to the components of the IES-6 survey between the hospitalized and nonhospitalized groups (Tables S7 and S8, SDC, http://links.lww.com/TP/C552).
FIGURE 4. Distribution of IES-6 category by time from infection and hospitalization status among respondents. IES-6, Impact of Events Scale-6.
Health-related Quality of Life EuroQol-5D Survey
Results of the EuroQol-5D are shown in Figure 5, Tables 2–4. Hospitalized patients had worse scores in the components of pain and anxiety/depression. Mobility problems and pain were still common in the remote COVID-19 infection group, 40% and 55%, respectively (Tables S9 and S10, SDC, http://links.lww.com/TP/C552).
FIGURE 5. Sum scores for EuroQol-5D questionnaire grouped by time from infection and hospitalization status.
The PROMIS Global Health Instrument
Results of the PROMIS survey are shown in Figure 6, Tables 2–4. Among the components of the PROMIS survey, 83% reported mild fatigue, and 41% reported moderate fatigue. Hospitalized patients had worse perceived pain score on a 0–10 scale (3 versus 0, P = 0.003). Participants who reported hospitalization more frequently had perceived problems with daily activities such as walking, climbing stairs, carrying groceries, or moving a chair (66% versus 40%, exact P = 0.013) (Table S11 and S12, SDC, http://links.lww.com/TP/C552).
FIGURE 6. Sum scores for PROMIS Global Physical Health questionnaire grouped by time from infection and hospitalization status.
BCSS
Results of the BCSS are shown in Figure 7, Tables 2–4. Among components of the BCSS survey, 35% of all patients had perceived abnormal breathing and 31% reported some cough. Hospitalized patients reported more cough (45% versus 24%, exact P = 0.021) and up to 40% of patients with remote COVID-19 infection still reported some cough (Table S13, SDC, http://links.lww.com/TP/C552).
FIGURE 7. Distribution of BCSS category by time from infection and hospitalization status among respondents. BCSS, Breathlessness, Cough‚ and Sputum Scale.
COVID-19 Symptoms Questionnaire
Overall, 22% of patients reported dysosmia and up to 29% of patients with remote COVID-19 infection reported dysosmia (Table S14, SDC, http://links.lww.com/TP/C552). Around 22% of patients reported dysgeusia and up to 34% of patients with remote COVID-19 infections reported dysosmia. Around 32% reported at least mild diarrhea and that percentage remained elevated even in the remote infection group (32%). There was no difference in the percentage of people with abnormal smell sensation, taste sensation, or diarrhea between the hospitalized and nonhospitalized groups, or between the recent and remote COVID-19 infection groups (Table S14, SDC, http://links.lww.com/TP/C552).
Regression Analysis
There was no association between age, sex, race, type of organ transplant, mycophenolate use, and the results of the QDRS, PHQ-9, GAD-7, IES-6, EuroQol-5D, PROMIS, and BCSS surveys except that Mycophenolate use was associated with less perception of trauma (odds ratio 0.38; confidence intervals, 0.15–0.97) and older age was associated with worse BCSS score (Beta Coefficient 0.02; confidence intervals 0.00–0.04).
DISCUSSION
In this observational study of long-term sequelae of SARS-CoV-2 infection in SOT recipients, there was a high frequency of perceived cognitive, mental, and physical impairments in the period beyond the acute phase of SARS-CoV-2 infection. Some of these impairments were more common and profound in those who reported hospitalization because of SARS-CoV-2 infection, and some continued to be reported >6 mo after diagnosis. The high frequency of reported impairments in the group with remote infection (>6 mo) could have been related, at least in part, to the limited treatment options available early in the pandemic course. This may indicate decreased risk of PASC among patients whose infection fell later in the course of the pandemic, and who were thus able to benefit from treatments that were unavailable early in the pandemic.
The early recognition of these long-term consequences has led to the development of multidisciplinary clinics in some institutions.8,35,36 Literature on the long-term sequelae of SARS-CoV-2 infection in the general population is evolving. However, most studies in the general population understandably focused on people who were hospitalized for SARS-CoV-2 infection. Individuals with SOT constitute a unique group because of their preexisting comorbidities and burden of immunosuppression. Data on long-term sequelae of SARS-CoV-2 infection in SOT patients, whether hospitalized or not, are very limited.18
In this study, up to 50% of SOT recipients were found to have evidence of perceived cognitive, mood, and functional decline by the QDRS questionnaire. In particular, those who reported hospitalization for SARS-CoV-2 infection were more likely to report a decline in mood (60%), decision-making and problem-solving (54%), and memory (45%). With regard to mental disorders, there was a remarkably high frequency of reported symptoms of depression, anxiety, and significant trauma related to being diagnosed with SARS-CoV-2 infection. The frequency of reported depression and anxiety symptoms was higher than their reported prevalence in SOT recipients without SARS-CoV-2 infection in literature‚ which is 17.3%–25.5% for depression and 10%–21% for anxiety-related disorders.37,38 Although we could not ascertain definitively that these disorders developed after the diagnosis of SARS-CoV-2 infection, around 35% of respondents self-reported that these mood changes developed after the diagnosis of SARS-CoV-2 infection. Studies of mental disorders in the general population after the diagnosis of SARS-CoV-2 infection reported a prevalence of depression of 23% and anxiety of 15%–23% at 4–6 mo after hospitalization with SARS-CoV2 infection.5,36,39 We observed a similar frequency of anxiety in SOT patients (23%) and a much higher frequency of depression alone (36%) at the same time interval. In 1 study in the general population, 31% of patients reported posttraumatic stress disorder 1–2 mo after hospitalization for SARS-CoV2 infection, lower than what we observed in our study.40 We observed a much higher percentage of feelings of trauma in our population‚ close to 53%, although variation in definitions might account for some differences.
Studies of physical symptoms and quality of life in the general population after the diagnosis of SARS-CoV-2 infection reported a prevalence of 2%–19.6% for muscle and/or joint pain in patients who were followed for up of 3–6 mo.4,5 This is lower than the frequency of perceived pain of 66% that we observed in SOT patients who were hospitalized for SARS-CoV-2 infection. Fatigue was reported in 34.8%–64% in the general population during follow-up of up to 6 mo after hospitalization.4,5,9,40-43 We observed higher prevalence of fatigue in SOT patients who were hospitalized for SARS-CoV-2 infection. Around 81% of SOT patients who were hospitalized for SARS-CoV-2 infection reported at least mild fatigue and 45% reported at least moderate fatigue.
Although patients who had SARS-CoV-2 infection within the past 4 wk were excluded, perceived respiratory symptoms were still very common in our cohort. Around 35% of patients continued to report abnormal breathing and cough at the time of the survey. Studies of the general population who were hospitalized for SARS-CoV-2 infection reported dyspnea and cough in 11.1%–43% and 2.1%–16.7%, respectively, during a follow-up of 1–6 mo.4,9,42-44 We observed a higher frequency of dyspnea (48%) and cough (45%) in hospitalized SOT patients during the same time interval.
In addition, dysosmia, dysgeusia, and diarrhea were also very frequent in our cohort, even in those who had the infection >6 mo ago: 29%, 34%, and 32%, respectively. Dysosmia and dysgeusia were reported in 12%–22.7% in the general population 1–6 mo after the infection‚ which is close to the percentage that we observed (21%–22%) in the 1–6 mo period.4,9,41,42,44 This percentage becomes smaller 7%–11% in studies that assessed this parameter at 6 mo in the general population, but continued to be high in our cohort beyond 6 mo (29%–30%).5 Studies in the general population after the diagnosis of SARS-CoV-2 infection reported a prevalence of 10.5% for diarrhea 2–6 mo after hospitalization with SARS-CoV-2 infection.4,5,42 It is possible that the higher prevalence of diarrhea in our population is due to gastrointestinal side effects of immunosuppressive drugs.
Our study has some limitations. The parent study was a longitudinal assessment of SOT recipients from multiple centers who self-enrolled online; hence, no non-SOT patients were enrolled, and the study team did not have access to patients’ medical records from their various transplant centers. Because of the nature of the study design, we do not have baseline assessments before SARS-CoV-2 infection, and it is possible that some of these impairments predated the SARS-CoV-2 infection. However, there were several questions in the surveys that specifically asked whether symptoms developed after the SARS-CoV-2 infection. Because of the nature of the study and lack of contemporaneous control group, there could be confounding factors and bias such as recall and nonresponse bias that may have impacted the results. However, we were focused on describing potential PASC in this particular population of interest, which was also not yet well documented in the literature. In addition, participants’ recruitment into this study is based on self-report of COVID-19 disease. Nevertheless, it is worth noting that a significant proportion of people with suspected COVID-19 disease relied on home testing to confirm diagnosis during more recent waves of COVID-19. Patients from communities of color and of lower socioeconomic status, who have had a disproportionate impact from COVID-19, were underrepresented in this study.45-47 As a result, the true prevalence of PASC in this high-risk population may be even higher than our findings estimate.
In conclusion, our study highlights the major self-reported sequelae of SARS-CoV-2 infection after recovery from acute COVID-19 and showed a high frequency of these perceived complications in both hospitalized and nonhospitalized SOT recipients. These impairments appear to be more common than what has been reported in the general population. These results are highly concerning and should serve as a call to develop an evidence-based multidisciplinary team approach for caring for these patients, and perform further rigorously designed studies for better understanding of this phenomenon. A comprehensive understanding of patient care needs beyond the acute phase will help in the development of infrastructure for COVID-19 clinics for this vulnerable group that will be equipped to provide integrated multispecialty care in the outpatient setting.
TABLE 2. Median score of responses to surveys by SARS-CoV-2 infection hospitalization status
Test Construct/ability Raw score All median (IQR), N (111) NonhospitalizedMedian (IQR)n (78) HospitalizedMedian (IQR)n (33) P
QDRS Change in cognition, mood, and daily function /30 1 (0–2.5) 0.5 (0–2) 2 (0–4.5) 0.02
PHQ-9 Self-rated depression /27 3 (0–7) 2 (0–6) 4 (2–8) 0.14
GAD-7 Self-rated anxiety /21 1 (0–4) 1 (0–4) 2 (0–5) 0.28
IES-6 Trauma-related distress /24 2 (0–5) 2 (0–5) 2 (1–4) 0.50
EuroQol-5D Quality of life /25 1 (0–4) 1 (0–4) 2 (1–6) 0.015
PROMIS Global Physical Health Physical function /15 11 (9–12) 11 (10–13) 10 (8–11) 0.005
BCSS Cough and SOB /15 0 (0–2) 0 (0–1) 1 (0–3) 0.022
BCSS, Breathlessness, Cough‚ and Sputum Scale; GAD-7, Generalized Anxiety Disorder-7; IES-6, Impact of Events Scale-6; PHQ-9, Patient Health Questionnaire-9; IQR, interquartile range; QDRS, Quick Dementia Rating System.
TABLE 3. Median score of responses to surveys by time from SARS-CoV-2 infection
Test Construct/ability Raw score All median (IQR)n (111) <6 mo median (IQR)n (76) >6 mo median (IQR)N (35) P
QDRS Change in cognition, mood, and daily function /30 1 (0–2.5) 1.25 (0–2.5) 1 (0–3) 0.49
PHQ-9 Self-rated depression /27 3 (0–7) 3.5 (.5–6) 3 (0–8) 0.97
GAD-7 Self-rated anxiety /21 1 (0–4) 1 (0–4) 1 (0–5) 0.97
IES-6 Trauma-related distress /24 2 (0–5) 2 (0–5) 1 (0–5) 0.82
EuroQol-5D Quality of life /25 1 (0–4) 1 (0–4) 2 (0–5) 0.057
PROMIS Global Physical Health Physical Function /15 11 (9–12) 11 (10–12) 10 (8–13) 0.33
BCSS Cough and SOB /15 0 (0–2) 0 (0–1) 1 (0–3) 0.097
BCSS, Breathlessness, Cough‚ and Sputum Scale; GAD-7, Generalized Anxiety Disorder-7; IES-6, Impact of Events Scale-6; PHQ-9, Patient Health Questionnaire-9; IQR, interquartile range; QDRS, Quick Dementia Rating System.
TABLE 4. Classification of response to QDRS, PHQ-9, GAD-7, and IES-6 questionnaires by hospitalization status and time from SARS-CoV-2 infection
Result OverallN (111) Nonhospitalizedn (78) Hospitalizedn (33) P <6 mon (76) >6 mon (35) P
QDRS
0–1.4 Normal 56 (50.5%) 43 (55%) 13 (39%) 0.15 38 (50%) 18 (51%) 0.64
1.5–5.9 Mild decline 43 (38.7%) 29 (37%) 14 (42%) 31 (41%) 12 (34%)
6–12.4 Moderate decline 12 (10.8%) 6 (8%) 6 (18%) 7 (9%) 5 (14%)
12.5–17.4 Moderately severe decline 0 0 0 0 0
>17.4 Severe decline 0 0 0 0 0
Abnormal 55 (50%) 35 (45%) 20 (61%) 0.15 38 (50%) 17 (49%) 1.00
PHQ-9
0–4 Minimal depression 63 (56.8%) 47 (60%) 16 (48%) 0.68 44 (58%) 19 (54%) 0.84
5–9 Mild depression 25(22.5%) 18 (23%) 7 (21%) 18 (24%) 7 (20%)
10–14 Moderate depression 9 (8.1%) 5 (6%) 4 (12%) 5 (7%) 4 (11%)
15–19 Moderately severe 5 (4.5%) 3 (4%) 2 (6%) 4 (5%) 1 (3%)
20–27 Severe 1 (0.9%) 1 (1%) 0 (0%) 1 (1%) 0 (0%)
No answer 8 (7.2%) 4 (5%) 4 (12%) 4 (5%) 4 (11%)
≥5 40 (36%) 27 (35%) 13 (39%) 0.50 28 (37%) 12 (34%) 1.00
GAD-7
0–4 No anxiety 80 (72.1%) 58 (74%) 22 (67%) 0.80 56 (74%) 24 (69%) 0.66
5–9 Mild anxiety 20 (18.0%) 13 (17%) 7 (21%) 13 (17%) 7 (20%)
10–14 Mod anxiety 4 (3.6%) 3 (4%) 1 (3%) 2 (3%) 2 (6%)
>15 Severe anxiety 2 (1.8%) 1 (1%) 1 (3%) 1 (1%) 1 (3%)
No answer 5 (4.5%) 3 (4%) 2 (6%) 4 (5%) 1 (3%)
≥5 26 (23%) 17 (22%) 9 (27%) 0.62 16 (21%) 10 (29%) 0.47
IES-6
<1.75 No significant trauma 48 (43.2%) 37 (47%) 11 (33%) 0.39 31 (41%) 17 (49%) 0.40
≥1.75 Significant trauma 59 (53.2%) 40 (51%) 19 (58%) 43 (57%) 16 (46%)
No answer 4 (3.6%) 1 (1%) 3 (9%) 2 (3%) 2 (6%)
GAD-7, Generalized Anxiety Disorder-7; IES-6, Impact of Events Scale-6; PHQ-9, Patient Health Questionnaire-9; QDRS, Quick Dementia Rating System.
Supplementary Material
This work was supported by the Ben-Dov and Trokhan Patterson families; grants F32DK124941 (Dr Boyarsky), T32DK007713 (Dr. Alejo), K01DK101677 (Dr Massie), and K23DK115908 (Dr Garonzik-Wang) from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK); grant K24AI144954 (Dr Segev) from the National Institute of Allergy and Infectious Diseases (NIAID)
S.A. received research/grant support from Shire and CareDx. A.M.P. received consulting from Aidar Health and speaking for Vizient. D.L.S. received consulting and speaking honoraria from Sanofi, Novartis, CLS Behring, Jazz Pharmaceuticals, Veloxis, Mallinckrodt, Thermo Fisher Scientific, Astra Zeneca, Regeneron. R.K.A. received research/grant support from Aicuris, Astellas, Chimerix, Merck, Oxford Immunotec, Qiagen, Regeneron, and Takeda/Shire.
S.A., B.J.B., J.G.-W., T.V., D.L.S., and R.K.A. participated in research design. S.A., T.P.-Y.C., A.J.S, A.T.A., J.L.A., S.C., W.C., E.B., A.M.P., J.G.-W., A.B.M., D.C.B., T.V., D.L.S., and R.K.A. participated in the writing of the article. S.A., T.P.-Y.C., A.J.S, M.T.O., B.J.B, A.T.A., and J.L.A. participated in the performance of the research. S.A., T.P.C., and A.B.M participated in data analysis.
Supplemental digital content (SDC) is available for this article. Direct URL citations appear in the printed text, and links to the digital files are provided in the HTML text of this article on the journal’s Web site (www.transplantjournal.com).
==== Refs
REFERENCES
1. Stokes EK Zambrano LD Anderson KN . Coronavirus disease 2019 case surveillance—United States, January 22-May 30, 2020. MMWR Morb Mortal Wkly Rep. 2020;69 :759–765.32555134
2. Infectious Diseases Society of America. Post COVID conditions (e.g. long COVID). Available at https://www.idsociety.org/covid-19-real-time-learning-network/disease-manifestations--complications/post-covid-syndrome. Accessed November 9, 2021.
3. Augustin M Schommers P Stecher M . Post-COVID syndrome in non-hospitalised patients with COVID-19: a longitudinal prospective cohort study. Lancet Reg Health Eur. 2021;6 :100122.34027514
4. Moreno-Pérez O Merino E Leon-Ramirez JM ; COVID19-ALC research group. Post-acute COVID-19 syndrome. Incidence and risk factors: a Mediterranean cohort study. J Infect. 2021;82 :378–383.33450302
5. Huang C Huang L Wang Y . 6-month consequences of COVID-19 in patients discharged from hospital: a cohort study. Lancet. 2021;397 :220–232.33428867
6. Malik J Zaidi SMJ Iqbal R . Post-acute COVID-19 syndrome and its prolonged effects: an updated systematic review. medRxiv. 2021:2021.05.09.21256911.
7. Datta SD Talwar A Lee JT . A proposed framework and timeline of the spectrum of disease due to SARS-CoV-2 infection: illness beyond acute infection and public health implications. JAMA. 2020;324 :2251–2252.33206133
8. Nalbandian A Sehgal K Gupta A . Post-acute COVID-19 syndrome. Nat Med. 2021;27 :601–615.33753937
9. Carfì A Bernabei R Landi F ; Gemelli Against COVID-19 Post-Acute Care Study Group. Persistent symptoms in patients after acute COVID-19. JAMA. 2020;324 :603–605.32644129
10. Yomogida K Zhu S Rubino F . Post-acute sequelae of SARS-CoV-2 infection among adults aged ≥18 years—long beach, California, April 1-December 10, 2020. MMWR Morb Mortal Wkly Rep. 2021;70 :1274–1277.34529639
11. Faravelli I Velardo D Podestà MA . Immunosuppression-related neurological disorders in kidney transplantation. J Nephrol. 2021;34 :539–555.33481222
12. Arnold R Pussell BA Pianta TJ . Association between calcineurin inhibitor treatment and peripheral nerve dysfunction in renal transplant recipients. Am J Transplant. 2013;13 :2426–2432.23841745
13. Sharma P Chen V Fung CM . COVID-19 outcomes among solid organ transplant recipients: a case-control study. Transplantation. 2021;105 :128–137.32890139
14. Avery RK . COVID-19 therapeutics for solid organ transplant recipients; 6 months into the pandemic: where are we now? Transplantation. 2021;105 :56–60.33141805
15. Alasfar S Avery RK . The impact of COVID-19 on kidney transplantation. Nat Rev Nephrol. 2020;16 :568–569.32788747
16. Pereira MR Mohan S Cohen DJ . COVID-19 in solid organ transplant recipients: Initial report from the US epicenter. Am J Transplant. 2020;20 :1800–1808.32330343
17. Trapani S Masiero L Puoti F ; Italian Network of Regional Transplant Coordinating Centers Collaborating Group; Italian Surveillance System of COVID-19, Italian Society for Organ Transplantation (SITO), The Italian Board of Experts in Liver Transplantation (I-BELT) Study Group, Italian Association for the Study of the Liver (AISF), Italian Society of Nephrology (SIN), SIN-SITO Study Group. Incidence and outcome of SARS-CoV-2 infection on solid organ transplantation recipients: a nationwide population-based study. Am J Transplant. 2021;21 :2509–2521.33278850
18. Duivenvoorden R Vart P Noordzij M ; ERACODA Collaborators. Clinical, functional, and mental health outcomes in kidney transplant recipients 3 months after a diagnosis of COVID-19. Transplantation. 2022;106 :1012–1023.35320154
19. Azzi Y Bartash R Scalea J . COVID-19 and solid organ transplantation: a review article. Transplantation. 2021;105 :37–55.33148977
20. Harvey LA . REDCap: web-based software for all types of data storage and collection. Spinal Cord. 2018;56 :625.29977003
21. Boyarsky BJ Werbel WA Avery RK . Immunogenicity of a single dose of SARS-CoV-2 messenger RNA vaccine in solid organ transplant recipients. JAMA. 2021;325 :1784–1786.33720292
22. Boyarsky BJ Werbel WA Avery RK . Antibody response to 2-Dose SARS-CoV-2 mRNA vaccine series in solid organ transplant recipients. JAMA. 2021;325 :2204–2206.33950155
23. Galvin JE . The Quick Dementia Rating System (QRDS): a rapid dementia staging tool. Alzheimers Dement (Amst). 2015;1 :249–259.26140284
24. Berman SE Koscik RL Clark LR . Use of the Quick Dementia Rating System (QDRS) as an initial screening measure in a longitudinal cohort at risk for Alzheimer’s disease. J Alzheimers Dis Rep. 2017;1 :9–13.28819654
25. Kroenke K Spitzer RL Williams JB . The PHQ-9: validity of a brief depression severity measure. J Gen Intern Med. 2001;16 :606–613.11556941
26. Spitzer RL Kroenke K Williams JB . A brief measure for assessing generalized anxiety disorder: the GAD-7. Arch Intern Med. 2006;166 :1092–1097.16717171
27. Hosey MM Leoutsakos JS Li X . Screening for posttraumatic stress disorder in ARDS survivors: validation of the Impact of Event Scale-6 (IES-6). Crit Care. 2019;23 :276.31391069
28. Thoresen S Tambs K Hussain A . Brief measure of posttraumatic stress reactions: impact of Event Scale-6. Soc Psychiatry Psychiatr Epidemiol. 2010;45 :405–412.19479171
29. Herdman M Gudex C Lloyd A . Development and preliminary testing of the new five-level version of EQ-5D (EQ-5D-5L). Qual Life Res. 2011;20 :1727–1736.21479777
30. Rabin R de Charro F . EQ-5D: a measure of health status from the EuroQol Group. Ann Med. 2001;33 :337–343.11491192
31. Hays RD Bjorner JB Revicki DA . Development of physical and mental health summary scores from the Patient-Reported Outcomes Measurement Information System (PROMIS) global items. Qual Life Res. 2009;18 :873–880.19543809
32. Leidy NK Schmier JK Jones MK . Evaluating symptoms in chronic obstructive pulmonary disease: validation of the Breathlessness, Cough and Sputum Scale. Respir Med. 2003;97 Suppl A :S59–S70.12564612
33. McCarroll ML Pohle-Krauza RJ Volsko TA . Use of the breathlessness, cough, and sputum scale (BCSS(©)) in pulmonary rehabilitation. Open Respir Med J. 2013;7 :1–5.23494521
34. Leidy NK Rennard SI Schmier J . The breathlessness, cough, and sputum scale: the development of empirically based guidelines for interpretation. Chest. 2003;124 :2182–2191.14665499
35. Brigham E O’Toole J Kim SY . The Johns Hopkins Post-Acute COVID-19 Team (PACT): a multidisciplinary, collaborative, ambulatory framework supporting COVID-19 survivors. Am J Med. 2021;134 :462–467.e1.33444589
36. Vanichkachorn G Newcomb R Cowl CT . Post-COVID-19 syndrome (Long Haul Syndrome): description of a multidisciplinary clinic at mayo clinic and characteristics of the initial patient cohort. Mayo Clin Proc. 2021;96 :1782–1791.34218857
37. Dew MA Kormos RL DiMartini AF . Prevalence and risk of depression and anxiety-related disorders during the first three years after heart transplantation. Psychosomatics. 2001;42 :300–313.11496019
38. Dew MA Roth LH Schulberg HC . Prevalence and predictors of depression and anxiety-related disorders during the year after heart transplantation. Gen Hosp Psychiatry. 1996;18 :48S–61S.8937923
39. Vannorsdall TD Brigham E Fawzy A . Cognitive dysfunction, psychiatric distress, and functional decline after COVID-19. J Acad Consult Liaison Psychiatry. 2022;63 :133–143.34793996
40. Halpin SJ McIvor C Whyatt G . Postdischarge symptoms and rehabilitation needs in survivors of COVID-19 infection: a cross-sectional evaluation. J Med Virol. 2021;93 :1013–1022.32729939
41. Carvalho-Schneider C Laurent E Lemaignen A . Follow-up of adults with noncritical COVID-19 two months after symptom onset. Clin Microbiol Infect. 2021;27 :258–263.33031948
42. Arnold DT Hamilton FW Milne A . Patient outcomes after hospitalisation with COVID-19 and implications for follow-up: results from a prospective UK cohort. Thorax. 2021;76 :399–401.33273026
43. Garrigues E Janvier P Kherabi Y . Post-discharge persistent symptoms and health-related quality of life after hospitalization for COVID-19. J Infect. 2020;81 :e4–e6.32853602
44. Chopra V Flanders SA O’Malley M . Sixty-day outcomes among patients hospitalized with COVID-19. Ann Intern Med. 2021;174 :576–578.33175566
45. Kim SJ Bostwick W . Social vulnerability and racial inequality in COVID-19 deaths in Chicago. Health Educ Behav. 2020;47 :509–513.32436405
46. Webb Hooper M Nápoles AM Pérez-Stable EJ . COVID-19 and racial/ethnic disparities. JAMA. 2020;323 :2466–2467.32391864
47. Magesh S John D Li WT . Disparities in COVID-19 outcomes by race, ethnicity, and socioeconomic status: a systematic-review and meta-analysis. JAMA Netw Open. 2021;4 :e2134147.34762110
| 36117251 | PMC9746234 | NO-CC CODE | 2022-12-15 23:21:55 | no | Transplantation. 2023 Jan 19; 107(1):181-191 | utf-8 | Transplantation | 2,022 | 10.1097/TP.0000000000004340 | oa_other |
==== Front
Am J Cardiol
Am J Cardiol
The American Journal of Cardiology
0002-9149
1879-1913
Excerpta Medica
S0002-9149(21)00797-9
10.1016/S0002-9149(21)00797-9
Article
Contents
20 9 2021
15 10 2021
20 9 2021
157 xxii
2019
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmc
| 0 | PMC9746243 | NO-CC CODE | 2022-12-15 23:21:56 | no | Am J Cardiol. 2021 Oct 15; 157:x-xii | utf-8 | Am J Cardiol | 2,021 | 10.1016/S0002-9149(21)00797-9 | oa_other |
==== Front
Nurs Adm Q
Nurs Adm Q
NURAQ
Nursing Administration Quarterly
0363-9568
1550-5103
Wolters Kluwer Health, Inc.
36469374
10.1097/NAQ.0000000000000561
nuraq4701p55
3
Original Articles
Nursing The Future 2.0
Reimagining New Graduate Transition in the COVID-19 Era (Part II)
Duchscher Judy PhD, MN, BScN, RN [email protected]
Corneau Kathryn BScN, RN [email protected]
Thompson Rivers University School of Nursing, Kamloops, British Columbia, Canada.
Correspondence: Judy Duchscher, PhD, MN, BScN, RN, Thompson Rivers University School of Nursing, Nursing and Population Health Bldg, Room 255, 805 TRU Way, Kamloops, BC V2C 0C8, Canada ([email protected]).
1 2023
01 12 2022
01 12 2022
47 1 Innovative Healthcare Models 5563
© 2023 Wolters Kluwer Health, Inc. All rights reserved.
2023
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
For newly graduated nurses (NGNs), the characteristically challenging and dynamic period of transition from student to professional practitioner is being further strained by global crises and the uncertainty and insecurity they motivate, health care systems and institutional restructuring, and extreme workload burdens. A novel approach to aiding the transition of NGNs is detailed in this article, culminating in the offering of an inclusive framework of potential strategies aimed at supporting NGNs and those who lead, manage, and educate them. This approach outlines strategies of support deliverable by both centralized and local means and acknowledging contemporary needs such as workload burdens and generationally-sensitive employee needs. Nursing The Future is a platform that uniquely situates an evidence-based, grassroots-driven response to the needs of NGNs, while encouraging collaborative partnering of health care institutions with governmental, professional, and regional advanced education bodies. This is the second article in a 2-part series that builds on the historical and developmental intents of Nursing The Future as an organization and outlines how evidence-informed, creative, and affordable grassroots-driven supports may be offered to NGNs for the purpose of sustaining and advancing our future nurse professionals.
new graduate nurse
non medical
nursing theory
professional role transition
residency
transition to practice
==== Body
pmcNURSING THE FUTURE 2.0
In March of 2020, the global landscape of health care was drastically, and it would appear enduringly, altered. With the change brought about by a global pandemic came a dramatic surge in intensity, uncertainty, and instability within both the culture and context of nursing practice. It seemed that more than ever, the sustainability of quality health care in Canada was dependent upon the collective ability of nurses to optimize their work environment and our ability to empower the newest professionals to deliver a high standard of care to individuals at the most vulnerable moments of their lives. Due largely to the health care demands resulting from COVID-19, and arguably the toll taken on senior practicing nurses that played out in unprecedented rates of attrition, the need for newly graduated nurses (NGNs) escalated in the years following the pandemic. These demands have served as a catalyst for the current global health human resource crisis.
In responding to this crisis, there is little doubt that countries will need to continue to work with relative urgency to ensure the optimal preparation, workplace integration, and retention of the newest members of the nursing profession. To actualize this agenda, they will need concrete measures to support and protect novice practitioners who consider their disrupted education and the stressors in the contemporary nursing workplace to be impeding their practice and professional development. The collective decisions made regarding the nursing workforce over the coming years may be the tipping point in our essential task of building (or some might argue “rebuilding”) a highly functioning and resilient workforce.
In the United States, it has been reported that the mitigation of a one percentage change in RN employee turnover saves a hospital US $262 300 yearly.1 The anticipated influx of new professionals needed to replace an aging and exhausted nursing workforce, and the now unprecedented job stress that is being experienced by new recruits in hospital settings, should alarm both education and health service sectors. Unless steps are taken to recognize, understand, and provide support for the resolution of the issues inherent in the traumatic and stressful socialization of NGNs to professional practice, we are in danger of not being able to replenish and rejuvenate this critical component of our health care workforce during this unprecedented care crisis and beyond. Nursing The Future 2.0 (NTF) evolved out of this urgent call for support in balancing the stark realities of practice presented here with the ongoing advancement of theoretical, ethical, and practice foundations of nursing as a discipline. By supporting NGNs in their social integration as well as their practical and theoretical application, and by uniting them with the senior nurses upon whom the health care system has traditionally relied, we can reasonably expect to optimize their decision-making, increase the quality of their clinical judgments, and foster a quality workplace culture. It is equally easonable to assume that these actions will favorably influence the NGNs work satisfaction and ultimately enhance the clinical outcomes of the patients they care for. The creation, implementation, and evaluation of transition-specific programs, policies, and practices that seek to recruit, retain, and rejuvenate our health care workforce in Canada depend on linkages between research, policy, and a current and accurate profile of the issues facing practicing NGNs in this country.
Canada is recognized as a leader of innovative global health care initiatives. In collaboration with partners in postsecondary education, the health care service industry, professional nursing organizations and associations, unions, and ministries, provincial communities are charged with understanding nursing needs as they work to replenish and revitalize their nursing workforce. Understanding the challenges faced by NGNs as they make the transition to professional practice, and supporting programs that can optimize a collaborative educational, governmental, and industry-centered response to these issues, will assist provinces to recruit and retain energized and motivated young professional nurses in Canada. In support of this mandate, the Canadian Nurses Association's document, titled Toward 2020: Vision for Nursing, emphasized the need for the nursing profession to take control of its future, not react to it.2 The vision and mandate of the NTF support platform epitomize Michael Villeneuve's assertion that “if nurses aren't building their future, someone else surely will.” Further to this, in a position statement released March 2020, the Canadian Association of Schools of Nursing clearly stated that efforts to optimize entry to practice for NGNs while supporting the delivery of health care services during this challenging time were a priority. The current crisis places NGNs entering practice for the first time in an unprecedentedly unstable position.3 The Canadian health care system is experiencing tremendous tension and, in places, it is on the verge of being entirely overwhelmed. As a result of evolving world crises (ie, political, health, climate, and economic systems strain), our existing nurses are being excessively burdened not only by stressed workplaces but equally also by the daily witnessing of trauma in both their professional and personal lives.4–7 It is reasonable to assume that these stressors will tax those nurses as they seek to support NGNs entering the practice context.
An urgent collective approach to supporting NGNs in the workplace that includes the health care sector, nursing education, government, and the entire professional nursing community is required. The creation, implementation, and evaluation of a national transition support program are the gateway to the capacity of our provinces and territories to recruit, retain, and rejuvenate their health care workforces. Actualizing this mandate requires that we form deep and broad linkages between research, policy, and strategy that represent a current and accurate profile of the issues facing practicing NGNs in this country. This current tragedy has the potential to be a catalyst for positive change in our health care community.
Commencing in July 2021, Duchscher and a team of nurses from across Canada developed the current version of NTF (www.nursingthefuture.ca) as a Web-based platform that was intended to serve as: a support network for NGNs moving into professional practice for the first time; a knowledge-generating and initiative-sharing platform about NGN professional role transition (PRT) for health care educators and employers; and a source of information that can assist in our understanding of practice patterns and workplace issues being faced by NGNs and their senior nursing colleagues. Creating this network occurred in close consultation with NGNs and health care practice partners across the country (eg, health care regions, educational institutions, ministries of health, professional organizations, and unions) with the intent to harness the creativity, ingenuity, and commitment to excellence that exist throughout our national nursing community. The comprehensive array of strategies served to assist NGNs with their social integration into the workplace, supported the practical and theoretical application of their knowledge in new clinical situations, united them with their senior nursing mentors, and offered leadership opportunities for nurses across this country. The outcome of an improvement in workplace culture and the fostering of a collaborative ethos amongst our nursing family is the optimization of decision-making in practice, an increase in the quality of the clinical judgments that are the cornerstone of excellent care, and a workplace that welcomes and unifies its practitioners. While long-term funding for the ongoing efforts of NTF has not yet been secured, the proposed actions of this support platform have demonstrated a significant influence on both NGNs and their senior colleagues. This said, the intent of this article is to share the platform strategies so that all institutions and health regions can enact similar programs with the potential to optimize professional integration of all nurses, increase work satisfaction, unite the nursing profession, and positively impact the health outcomes of our citizens.
The foundational elements of this initiative are as follows:
Support for NGNS in all sectors and practice contexts include, but are not limited to, hospitals, long-term care settings, and rural and remote practice settings: NTF is built upon an inclusive, bilingual Web-based platform and framed by contemporary evidence related to the PRT of the NGN;
Multiple strategies target the dissemination of knowledge related to PRT and the support of mentoring relationships between NGNs and their senior practice partners in acute care, long-term care, community, home care, and public health contexts;
Virtual and face-to-face chats offer connections, problem-solving, and professional collaborations with experienced practitioners and more senior NGNs;
“Normalization” of the transition experience is facilitated through information sessions on the Stages of Transition and Transition Shock;
Video- and text-based information on coping highlights both the personal and professional (patient/client/family/community) health and social struggles being experienced in our world secondary to the suffering and isolation of individuals and families in our communities; and
The acknowledgment of historical colonial trespasses and crimes committed against Indigenous communities, as well as current inequity experienced by those in particularly rural settings, has informed resource decisions, consultations, content, and goals across NTF. Strength-based approaches to supports for Indigenous NGNs returning to, and supporting, their home communities are facilitated by forging strong linkages between local and regional Elders and leaders, Indigenous professional groups (eg, Canadian Indigenous Nurses Association), and Indigenous NGNs.
Building and strengthening of the larger community of nursing includes licensed/registered practical, registered, and registered psychiatric nurses: Strategies and initiatives within the NTF platform consider all areas of practice (urban/acute care/community/public health/rural/remote); and
NTF includes all nursing scopes and represents this broad demographic on the Nursing the Future Project Team, embracing all scopes of practice for inclusion on local, regional, and strategy-specific consultation teams;
Health care sector employers, nursing educational institutions, ministries of health and ministries of advanced education, skills, and training are key partners in any strategy that addresses preparatory education and workforce integration and retention of the health care workforce: Leaders at various levels (local, regional, provincial, territorial, and national) were invited to participate in consultation sessions over the course of the development of NTF and are included in ongoing strategy and initiative evolution; and
Ongoing and continuous feedback loops are infused into communication coming into and going out of NTF (eg, NTF Newsletter Monthly Distribution to NTF Members/Provincial and Regional Employment Boards/Student and NGN Chat Platforms/Educational Podcasts on “Preparing for Your Transition to Professional Practice”).
A strong nursing and health care workforce requires ongoing attention to the development and nurturing of professional competence, advanced practice, and leadership potential in NGNs: Various avenues for facilitating an interface of NGNs with their senior nursing leaders are integrated into NTF strategies and approaches (eg, interviews with more senior NGNs, senior nursing partners and nursing leaders, frameworks for provinces and regions to establish their own New Graduate Transition Facilitation Networks, and a Virtual Conference on New Graduate Transition); and
Working alongside, and learning from experienced clinical, academic, and health care leaders, NGNs are encouraged and facilitated to expand their professional development to include advanced skills preparation and education.
Driven by emerging data, current workplace needs, and previous iterations of NTF, Duchscher developed strategies aimed at meeting the diverse needs of NGNs (Table).
Table. Strategies for Successful Transition of Newly Graduated Nurses
Strategy Description
Transition Theory Duchscher's Transition Shock model and Stages of Transition theory are detailed on these pages—numerous documents offer students, NGNs, and their supporters insight into, and frameworks for, application and implementation of the most current knowledge of transition from student to professional nurse. Non-nursing transition theorists are also featured.
The Greenhouse Established as both a virtual strategy and a face-to-face strategy, The Greenhouse offers multiple avenues of leadership capacity building, including activities for: (1) NGNs (<1 y of practice) who are not necessarily ready for formal leadership positions but who are looking to increase their skills related to relationship building, clinical confidence, professional networking, nursing advocacy, or crisis communication; (2) recently graduated nurses (RGNs) (1-2 y of practice) who are exploring ways to enhance their knowledge-sharing capacity through various approaches to scholarship that include discovery (research), evidence integration and application, or teaching; considering clinical leadership positions like taking on charge positions; or considering clinical advancement opportunities such as patient care coordination; (3) graduates interested in advancing their professional knowledge in areas such as specialty practice, nursing education, graduate studies, clinical research, and/or formal nursing leadership roles.
The Greenhouse incorporates 3 platforms: (1) Building Capacity: The Road to Leadership fosters grassroots leadership conversations between NGNs, RGNs and senior nursing leaders; (2) Leaning In consists of a series of interviews and Master Classes from nursing leaders. The goal of these sessions is to advance research, innovation, and grassroots practice initiatives through knowledge translation and dissemination, professional mentorship, and leadership skill building; (3) The Incubator consists of face-to-face workshops addressing and evolving the leadership needs of NGNs and RGNs. The Incubator was created to occur as a PRE-CONFERENCE Workshop at the Workplace Integration for New Nurses (WINN)/NTF Annual Conference.
Question Period Hundreds of questions are answered related to preparing for and then making the initial entry into professional practice. Both grassroots- and research-inspired perspectives target students, NGNs, and NGN supporters.
The Interview Nurses share their reflections on and stories about practicing nursing today in a video format—what nursing means, what makes nursing practice unique, what challenges they overcame and how. Health care topics important to NGNs are the focus.
Speaker Series Leaders in diverse fields share their perspectives on the issues and challenges of nursing and health care today. A lens on how the issues impact NGN practice and how new professionals might influence the issues is offered.
Book Club A monthly archived selection of books on a host of provocative, contemporary nursing and health care issues is featured, including 1:1 interviews with renowned authors.
The Podcast Audio interviews with a multitude of guests on all subjects impacting nursing students and NGNs in practice are featured here. With topics ranging from coping with the stress of transition and remaining resilient in a health care system under pressure to understanding the stages of PRT, recognizing and dealing with moral distress and compassion fatigue, combating bullying in the workplace, or learning how to structure communication with physicians, The Podcast is diverse and comprehensive.
Nursing Excellence Awards Five awards celebrate NGNs, frontline nurses, educators, and nurse leaders positively impacting the nursing profession and the health care system. The Strength in the Storm Award is conferred upon a new graduate, frontline nurse, senior nurse, educator or nurse leader who has consistently overcome disruption, chaos and turmoil during a global pandemic affecting both personal and professional aspects of their lives. Called upon to make sacrifice after sacrifice, this award recognizes the contribution, leadership and steady hand of nurses in Canada, who are stabilizing forces in their families, communities and profession. The Teaching Excellence Award recognizes the substantial contributions of committed, engaging, and innovative educators in nursing education across the country. This award is granted to an educator or team of educators who make a broad range of contributions to the nursing education community including, but not limited to, the creation of new courses and programs, devising and implementing innovative strategies for instruction, conducting research on teaching and learning issues and other important work that occurs inside and outside the classroom. The New Graduate Excellence Award is offered to a new graduate (in professional nursing practice < 1 year) who displays humility and trustworthiness, elevating the level of dialogue in conversations about nursing, leadership and change. This individual has displayed evidence of courage and compassion, volunteers and contributes to community and professional initiatives. They have displayed exemplar growth in leadership over time and have the potential to positively impacts the nursing profession. The Torch Award acknowledges leadership in new graduate issues, and is awarded to a nurse leader, researcher, educator or frontline practitioner who recognizes the complexity of the transition experience, promotes initiatives aimed at supporting the healthy integration of new grads in professional practice settings, researches new graduate issues with the goal of advancing the knowledge base in this area of scholarship, evaluates new graduate initiatives with the goal of improving them for the future, implements supportive networks for the benefit of the new graduate population. Finally, the Odyssey Award goes to an individual who has played an essential role in the development of a new graduate nurse(s) in the workplace. It recognizes the importance of experienced colleagues in facilitating a healthy professional role transition for new nurses. The award is preferential to applicants who consistently demonstrate NTF's values in the workplace.
CONNECT Synchronous virtual sessions attended in the hundreds feature educational content on contemporary issues facing students and NGNs in practice—discussion and debate are encouraged, and the issues discussed are immediately impactive to NGNs transitioning into practice (ie, NCLEX preparation, dealing with horizontal aggression from colleagues, time management).
Blog Some people like to listen, some are visual, while others like to read. These short bursts of text introduce NGNs and students to featured scholars from around the world who are experts on nursing and health care.
Newsletter This quarterly publication features editorials on a variety of topics (regular columns such as “‘Let's Talk Transition” plus a variety of other guest editorials that speak to contemporary subjects—ie, Indigenous Nurses and the TRC). Each newsletter features a senior nurse (i.e., “Mentor Profile”) and an NGN (“TrailBlazer”) from across Canada, drawing attention to the importance of mentorship and profiling the outstanding work being done by the newest members of our profession.
International Conference on Workplace Integration for New Nurses (WINN) This annual conference hosts an international audience for the purpose of sharing research, innovation, and capacity-building initiatives specific to the education, orientation, transition, integration, and stabilization of NGNs in the workforce. Integrated into the annual conference platform are specific workshops that discuss topics exclusive to students and NGNs entering practice for the first time. A New Graduate Planning Committee consisting of NTF leaders, senior students and NGNs determines both content and format for the NGN-specific knowledge dissemination.
On Call NGN Crisis Line for general advice and counsel.
Regional/National/International Resources Web pages with regional resources that include employment opportunities and information about NGN orientation and transition support programs in partnership and regarding health authorities across Canada. NTF also recognizes the unique needs of internationally educated nurses in Canada who constitute approximately 8.4% of the employed nursing workforce.8
Welcome to Our World This multipronged platform was developed in recognition of the vital nature of mentorship between novice and experienced nurses. Initiatives for this strategy were intended to emphasize support, connection, and collaboration, with the primary goal of embracing our newest professional members, encouraging professional development in and amongst all nurses, and bridging the gap between school and the workplace from social, developmental, and professional perspectives.
Storm Chasers One of the primary subelements of the Welcome to Our World strategy to support NGNs is a platform for ‘boots on the ground’ (unit/institutionally based) support within the workplace. Working closely with health regions across the country, recent graduates (1-3 y of practice experience) will be selected by their workplaces to undergo a specialized certification program through NTF to advance their knowledge, understanding, and application of transition theory, conflict resolution, and crisis management skills. These Storm Chasers will then be called upon to respond to NGNs struggling with transition. Planned as a self-sustaining initiative, the NGNs supported by Storm Chasers will later be offered opportunities to ‘pay it forward’ by becoming Storm Chasers themselves.
Shock Absorbers The second of the primary subelements in the “Welcome to Our World” support strategy consists of senior nurses (>3 y of experience) offering mentorship to NGNs with the intent to support them through both the precarious ‘transition shock’ period of their transition (2 wk-2 mo postorientation) and mentoring them through their initial transition year. Working closely with health regions across the country, graduate nurses with >3 y of practice experience will be selected by their workplaces to undergo a specialized certification program through NTF to advance their knowledge, understanding, and application of transition theory with a dual focus on both transition shock and the stages of PRT for the NGN. These Shock Absorbers will be paired with NGNs upon hire and will meet regularly with them over the course of their initial transition year to identify and work through the challenges posed by their transition. Unlike the crisis response approach of the Storm Chasers, Shock Absorbers are there to counter the day-to-day stressors of PRT in the NGN. The intent is to reduce the need for Storm Chaser interventions to urgent and emergent issues only and to support NGNs in their long-term professional advancement planning.
Social Media Platforms currently include Twitter/Facebook/Instagram/Snapchat/LinkedIn and provide opportunities to post PRT information and support ideas for NGNs—decisions on media focus are driven by NGN consultation and empirics from or responses on social media.
Abbreviations: NGN, newly graduated nurse; NTF, Nursing The Future; PRT, professional role transition; RGN, recently graduated nurse; TRC, Truth and Reconciliation Commission of Canada.
CONCLUSION
Challenges to delivering timely and appropriate health care are being experienced worldwide. Although these challenges did not begin with the COVID-19 pandemic, this global health care crisis clearly revealed the fractures and fissures in our system. While transitioning from our precarious and reactive intrapandemic state to one of a new post-COVID-19 reality, identified gaps in our support of practicing nurses must be intentionally and strategically addressed. The trend of highly knowledgeable senior professionals seeking out alternate means of employment to regain their sense of well-being post-COVID-19 will continue to devastate our health human resources and limit our capacity for novice-expert knowledge and professional practice transfer. The reality of workplace trauma, role strain, and moral distress for nurses has been captured in unprecedented levels of burnout and subsequent workplace exiting. While expanding the capacity to optimize individual coping mechanisms when working in the face of human tragedy simply makes sense, the expectation that individual practitioner resilience is or should be the primary response to wide-ranging and pervasive systems issues is dangerously myopic; the challenges we face are multidimensional and require a like response. Despite the existence of supportive transition programs, what remains constant is the time needed for NGNs to engage in the crucial transition period as they develop their professional confidence and capacity. Engaging in an analysis of the dichotomies present in institutional and professional concepts and standards of care is more likely to offer us a diverse and proliferative framework for resolution.
For the current and future betterment of all professional nurses, we argue that an understanding and application of existing and emerging transition theories, consideration of NGN practice and job satisfaction evidence, and evolving insights into the tumultuous context of the contemporary workplace cements the urgent need for formal and standardized PRT programs. NTF has demonstrated how support for the NGN can be approached with fiscal responsibility, cyclical growth, evolution of the new nurses' professional capabilities, and a critical unifying of new and experienced nursing knowledge and skill for optimal outcomes. We suggest that partnerships between government, institutional, professional, regulatory, and other representative groups are required to foster meaningful and engaged dialogue that begins and ends with the evidence on NGN transition and entry to practice challenges. While we witness escalating rates of attrition internationally through the loss of new and seasoned nurses, it is imperative that we focus as fervently on the retention of our nurses as we do on their recruitment.9–11 NGNs require transition supports to endure current professional challenges, and it can be argued that these can be provided through an evidence-driven, sustainable, and cost-effective means; NTF works.
The authors extend their gratitude to the Canadian Nurses Foundation COVID-19 Fund for Nurses and its donors that supported this work. The authors give a grateful acknowledgment to the Canadian Nurses Association for its administrative support throughout this grant period.
The authors declare no conflict of interest.
==== Refs
REFERENCES
1. NSi Nursing Solutions, Inc. 2022 NSI national health care retention & RN staffing report. www.nsinursingsolutions.com. Accessed September 13, 2022.
2. Villeneuve M Macdonald J . Toward 2020: visions for nursing. http://www.cna-aiic.ca. Published 2006. Accessed September 13, 2022.
3. Canadian Association of Schools of Nursing. Nursing education during the COVID-19 pandemic. https://www.casn.ca/wp-content/uploads/2020/03/COVID-19-POSITION-STATEMENT.pdf. Published 2020. Accessed September 20, 2022.
4. Smith SM Buckner M Jessee MA Robbins V Horst T Ivory CH . Impact of COVID-19 on new graduate nurses' transition to practice: loss or gain? Nurse Educ. 2021;46 (4 ):209–214. doi:10.1097/NNE.0000000000001042.33988534
5. Lake ET Narva AM Holland S Hospital nurses' moral distress and mental health during COVID-19. J Adv Nurs. 2022;78 (3 ):799–809. doi:10.1111/jan.15013.34402538
6. Jackson D Bradbury-Jones C Baptiste D Life in the pandemic: some reflections on nursing in the context of COVID-19 [Editorial]. J Clin Nurs. 2020;29 (13/14 ):2041–2043. doi:10.1111/JOCN.15257.32281185
7. Hu D Kong Y Li W Frontline nurses' burnout, anxiety, depression, and fear statuses and their associated factors during the COVID-19 outbreak in Wuhan, China: a large-scale cross-sectional study. EClinicalMedicine. 2020;24 :100424. doi:10.1016/J.ECLINM.2020.100424.32766539
8. Organisation for Economic Co-operation and Development (OECD). Health Care Resources: Nurses. Paris, France: OECD; 2021.
9. Buerhaus PI Staiger DO Auerbach DI Yates MC Donelan K . Nurse employment during the first fifteen months of the COVID-19 pandemic. Health Aff (Millwood). 2022;41 (1 ):79–85. doi:10.1377/hlthaff.2021.01289.34982625
10. Stevenson RL Maclaren J Vaulkhard K . Commentary: the nursing workforce: who will be left to answer the call? Nurs Leadersh. 2021;34 (4 ):31–35. doi:10.12927/CJNL.2021.26692.
11. Auerbach DI Buerhaus PI Donelan K Staiger DO . A worrisome drop in the number of young nurses. Health Affairs Forefront. April 13, 2022. https://www.healthaffairs.org/do/10.1377/forefront.20220412.311784. Accessed May 30, 2022.
| 36469374 | PMC9746247 | NO-CC CODE | 2022-12-15 23:21:56 | no | Nurs Adm Q. 2023 Jan 1; 47(1):55-63 | utf-8 | Nurs Adm Q | 2,022 | 10.1097/NAQ.0000000000000561 | oa_other |
==== Front
Nurs Adm Q
Nurs Adm Q
NURAQ
Nursing Administration Quarterly
0363-9568
1550-5103
Wolters Kluwer Health, Inc.
36469375
10.1097/NAQ.0000000000000564
nuraq4701p72
3
Original Articles
The Creation of a Novel Undergraduate Nursing Employee/Student Hybrid Role in the COVID-19 Response
An Alberta Experience
Shajani Zahra EdD, RN [email protected]
Laing Catherine M. PhD, RN [email protected]
Robinson Fadumo MN, RN [email protected]
Yun Lira PhD [email protected]
Patterson J. David MN, RN [email protected]
Rieder Linda MN, RN [email protected]
University of Calgary, Calgary, Alberta, Canada (Drs Shajani and Laing, and Mr Patterson); and Alberta Health Services, Edmonton, Alberta, Canada (Mss Robinson and Rieder and Dr Yun).
Correspondence: Zahra Shajani, EdD, RN, Faculty of Nursing, University of Calgary, 2500 University Dr, NW Calgary, AB T2N 1N4, Canada ([email protected]).
1 2023
01 12 2022
01 12 2022
47 1 Innovative Healthcare Models 7283
© 2023 Wolters Kluwer Health, Inc. All rights reserved.
2023
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
The COVID-19 pandemic impacted nursing education and health care systems alike. Increases in staff absenteeism along with increased hospitalizations have strained health systems across the globe. Postsecondary institutions (PSIs) were required to remove students from clinical placements, thus delaying nursing students' ability to complete their programs, and in turn, contributing to the nursing workforce challenges. Health care organizations and PSIs had to collaborate innovatively to support the health care response to the pandemic while continuing to educate and graduate students to expand the nursing workforce. In Alberta, the collaboration between the health system and PSIs led to the creation of an undergraduate nursing employee/student hybrid (UNE/Hybrid) role. This role was not only a response to the nursing workforce challenges created by the pandemic, but it provided nursing students with positive learning clinical placements ensuring that they completed their program in a timely manner. This role was designed to assist with the fourth wave of the pandemic (omicron variant), which was expected to be the most severe wave in terms of hospitalizations and increased staff absences. The UNE/Hybrid role allowed nursing students to complete the required learning for their final preceptorships and/or complete leadership placements in a paid role while being integrated into the unit culture and becoming part of the team. The initiative's results, including its successes, challenges, and lessons, are discussed.
COVID-19
education
nursing
nursing students
nursing workforce
==== Body
pmcTHE CORONAVIRUS DISEASE-2019 (COVID-19) pandemic caused unprecedented strain in all industries on a global scale, and this has been more evident in health care.1 As the pandemic persisted and hospitalizations increased, nurses bore the brunt of this strain. Very quickly, it became clear to health leaders that a severe nursing shortage loomed, and a drastic response was warranted. Concurrently for postsecondary institutions (PSIs), the ability to graduate nursing students was jeopardized as many of them either missed clinical hours or clinical placements were cancelled. Nurse absenteeism, burnout, and attrition, coupled with a disruption to the regular number of graduating nurses ready to enter the workforce, became the foremost problem for health care systems.2–4
Nursing workforce gaps were not introduced by the pandemic, but rather were exacerbated by it. Historically, a key problem in the nursing workforce is forecasting demand and the lack of shared accountability to address the workforce challenges among many stakeholders involved.4 The lack of shared accountability for forecasting of the needs of the nursing workforce and planning to address workforce challenges is a Canada-wide gap.4 Bourgeault4(p13) noted a lack of coordinating superstructure over the entire Canadian health care system and its stakeholders, which creates challenges in having basic information on trends in the nursing workforce over time (ie, “the who, what, and where of the nursing workforce”). In addition to the local and national absence of predictive workforce plans that could remedy the increased workload, the pandemic added complexity of nursing workforce challenges as a result of forecasting limitations and the unique physical and mental occupational hazards of the profession, causing a great increase in absenteeism and thus increased workloads.
Strategies to increase workforce capacity such as international recruitment, educational seat expansions, financial incentives, and relocation bonuses have been utilized; however, these traditional approaches have had mixed effectiveness.3 Stevenson et al3 argued for new approaches such as establishing supportive new graduate initiatives and collaboration among stakeholders accountable for the overall nursing workforce (ie, PSIs, government, employers, and regulators) as the key to addressing the nursing shortage in the postpandemic era.
ADDRESSING NURSING WORKFORCE CHALLENGES IN ALBERTA
In Alberta, the collaboration between 2 provincial health care organizations (Alberta Health Services and Covenant Health) and 8 PSIs in Alberta led to the creation of an undergraduate nursing employee/student hybrid (UNE/Hybrid) role as a response to the nursing workforce challenges created by the pandemic. The undergraduate nurse employee (UNE)/Hybrid initiative was one of the key strategies used in Alberta to respond to the omicron wave of the pandemic. Conducted from January to April 2022,5,6 this initiative was designed to increase the health care workforce while ensuring process change focused on positive clinical placements and an opportunity to remain in clinical practice in order for nursing students to complete their program requirements and graduate on schedule.
The purpose of this article is to illustrate the collaboration between a major health care organization in western Canada and 8 PSIs in developing a UNE/Hybrid initiative and how this was evaluated by the students who took the UNE/Hybrid role in response to the nursing workforce challenges caused by the COVID-19 pandemic.
THE STUDENT AS AN EMPLOYEE
Students have traditionally wished for a “great placement” for their final focus that will allow them to consolidate the learning received as an undergraduate nursing student while, hopefully, will provide a place of employment once the final practicum is completed. As future employees, students learn unit culture, patient populations, and the daily business of their placement site. PSI placement teams and clinical coordinators are mindful to remind students that their final focus placements are not job interviews, but rather an introduction to working as a registered nurse, with the same schedules and shifts of their preceptors. This traditional cognitive apprenticeship role has best been described as focusing on “learning through guided experience on cognitive and metacognitive skills and processes.”7 The independence from faculty, and growing relationship with the “expert” practitioner, allows the student to gain additional confidence and skills prior to graduation and entry into a workforce that may, or may not, be able to hire them. Partnering with the major health care organizations within the province is this unique staffing role benefits not only the student, who will be able to move into a staff role with greater ease, but the health care organizations as well. The organizations gain unit-trained and freshly mentored staff while the students gain an employer upon graduation.
A UNE/HYBRID ROLE
The UNE role is an unregulated role that provides direct patient care under the supervision of a registered nurse or a registered psychiatric nurse. The existing UNE role requires prior competency to perform nursing activities with patients. A key difference in the UNE/Hybrid role is the RN preceptor coaches and mentors in addition to providing overall supervision of the student. This allows the student to perform clinical skills if they have previously received theoretical/laboratory instruction. For example, a student who has completed theory and laboratory practice to initiate a vascular access device can gain skill competency in the UNE/Hybrid role. The participating students were those who were required to complete 300 to 388 hours of their final consolidated preceptorship (from 8 PSIs). The students were given a choice to move their clinical placement to care areas in most need due to the pandemic (eg, medical, surgical, or emergency departments) if they were not already placed in those areas with their original practicum placement. Students were paired with at least one preceptor (or mentor) who provided direct supervision and opportunities to consolidate their learning in theory, laboratory, and clinical. The relationship between the preceptor and the student is a key component in the process of socializing nursing students into the profession.8
Assignment of care was determined by the nature and level of risk of the nursing activity, the established competency of the UNE/Hybrid student, the employer policy, and the practice setting. UNE/Hybrid students only performed skills for which the theoretical knowledge was covered in the nursing program, although the unit may have additional learning materials for students to complete activities. UNE/Hybrid students were provided with opportunities to perform skills in clinical settings with direct or indirect supervision and coaching by a registered nurse who is authorized and competent to perform and consents to supervise. The student could process orders if this was included in the nursing program; however, they were prohibited from taking verbal or telephone medication orders and providing supervision to other employees (such as health care aides or other students).
INITIATIVE ROLLOUT AND SUPPORT THE UNE/HYBRID ROLE
Two fundamental principles guided this unique collaboration between the health care organization and 8 PSIs: triad wrap-around support and student employment status. The triad support consisted of the preceptor (or leadership mentor for the leadership student), unit manager/clinical nurse educator (CNE), and the faculty advisor from the PSI assigned to the student. Nursing managers, educators, and faculty representatives (course leads) received information and resources to understand the unique needs of these students and the importance of regular touch points, appropriate assignments, and communication for addressing concerns and challenges. Preceptors had access to all support resources developed including a role description, frequently asked questions, role comparisons (UNE-UNE/Hybrid student), and email access to the provincial team supporting this initiative for specific questions. Significant effort was made to ensure student learning needs were not jeopardized in this unique arrangement, which was taking place in a stressed system. With respect to employment status, 487 nursing students were hired into the 2 participating health care organizations. Employing the students, even though they were spending most (if not all) of their paid hours learning, was important for two reasons: first, it was necessary to have the students stay on to support patient care if the pandemic worsened and resulted in the discontinuation of clinical placements, as was the case in earlier waves of the pandemic; second, employing the students into a central staffing pool but placing them into specific units created an opportunity to integrate them into the team, thus paving the way for postgraduation recruitment. These potential recruitment and retention strategies would only be effective on the premise of students being placed in welcoming and supportive environments so that they could be easily integrated into the team, which would increase the chances of students remaining with their home units and applying to graduate and registered nurse positions.
The UNE/Hybrid initiative was evaluated using a process improvement lens with the aim of examining whether the initiative met the goal of ensuring students completed their clinical placements successfully and had positive experiences in the UNE/Hybrid role. It is also hoped that the findings from this evaluation will inform the delivery of similar initiatives as viable recruitment and retention strategies to address future nursing workforce challenges while protecting the learning process of nursing students.
METHODS
Student survey
Two online surveys were developed and distributed to all 487 students from 8 participating PSIs to evaluate the initiative. The surveys were completed at the midpoint (end of February and early March 2022) and the end of the initiative (end of April). The midpoint survey included questions related to students' experiences with their preceptor(s), unit managers, CNEs, and unit staff, as well as students' perceived confidence and competence in providing care for patients, and satisfaction with the UNE/Hybrid role. The endpoint survey evaluated topics similar to the midpoint survey, with additional questions to assess individuals' perceived confidence and preparedness in their ability to provide safe patient care as graduate nurses. Responses to questions were on a 4-point Likert scale ranging from strongly disagree (1) to strongly agree (4). Both the mid- and endpoint surveys included open-ended questions for students to provide additional information about their experiences, what they liked about the UNE/Hybrid role, and suggestions for improvement.
RESULTS
Of 487 UNE/Hybrid students invited to participate in the midpoint survey, 225 responded, with 186 surveys fully completed and 39 surveys partially completed. For the endpoint survey, of 108 responses, 106 were completed and 2 were partially completed. Results from both surveys are shown in Tables 1 to 4.
Table 1. Frequencies and Percentages of Responses From Midpoint Student Survey (n = 201)
n (%)
I meet regularly with my preceptor
Strongly disagree 5 (2.5)
Disagree 7 (3.5)
Agree 41 (20.4)
Strongly agree 148 (73.6)
My preceptor provides supervision when needed
Strongly disagree 2 (1.0)
Disagree 4 (2.0)
Agree 43 (21.4)
Strongly agree 152 (75.6)
My preceptor encourages me to ask questions and engage in self-reflection
Strongly disagree 6 (3.0)
Disagree 9 (4.5)
Agree 46 (22.9)
Strongly agree 140 (69.7)
My preceptor appears to understand the final preceptor stage of nursing education and tailors expectations accordingly
Strongly disagree 8 (4.0)
Disagree 8 (4.0)
Agree 50 (24.9)
Strongly agree 135 (69.7)
My preceptor and I are given adequate opportunity to focus on my specific learning needs
Strongly disagree 9 (4.5)
Disagree 22 (10.9)
Agree 52 (25.9)
Strongly agree 118 (58.7)
My preceptor provides ongoing constructive feedback on my performance
Strongly disagree 7 (3.5)
Disagree 13 (6.5)
Agree 63 (31.3)
Strongly agree 118 (58.7)
I have developed a trusting relationship with my preceptor
Strongly disagree 6 (3.0)
Disagree 10 (5.0)
Agree 43 (21.4)
Strongly agree 142 (70.6)
I have regular check-ins with my manager/CNE
Strongly disagree 22 (10.9)
Disagree 68 (33.8)
Agree 78 (38.8)
Strongly agree 32 (15.9)
Missing 1 (0.5)
My manager/CNE encourages me to ask questions to support my role
Strongly disagree 12 (6.0)
Disagree 44 (21.9)
Agree 91 (45.3)
Strongly agree 53 (26.4)
Missing 1 (0.5)
My manager/CNE creates an environment where I could be easily immersed in the unit
Strongly disagree 6 (3.0)
Disagree 19 (9.5)
Agree 111 (55.2)
Strongly agree 64 (31.8)
Missing 1 (0.5)
The team on my unit was aware of my unique UNE/Hybrid role
Strongly disagree 16 (8.0)
Disagree 44 (21.9)
Agree 87 (43.3)
Strongly agree 53 (26.4)
Missing 1 (0.5)
I felt very welcome as a new staff member
Strongly disagree 4 (2.0)
Disagree 17 (8.5)
Agree 101 (50.2)
Strongly agree 78 (38.8)
Missing 1 (0.5)
There was no horizontal violence (eg, bullying and conflict between groups) on my unit
Strongly disagree 10 (5.0)
Disagree 19 (9.5)
Agree 81 (40.3)
Strongly agree 90 (44.8)
Missing 1 (0.5)
I am becoming confident in my ability to provide safe patient care on this unit
Strongly disagree 1 (0.5)
Disagree 4 (2.0)
Agree 90 (44.8)
Strongly agree 102 (50.7)
Missing 4 (2.0)
I am clear about my role as a UNE/Hybrid and the roles of those in other professions
Strongly disagree 9 (4.5)
Disagree 23 (11.4)
Agree 95 (47.3)
Strongly agree 70 (34.8)
Missing 4 (2.0)
I feel control over decisions related to my patient's care
Strongly disagree 2 (1.0)
Disagree 12 (6.0)
Agree 122 (60.7)
Strongly agree 61 (30.3)
Missing 4 (2.0)
My experience as a UNE/Hybrid role is based on my learning needs
Strongly disagree 6 (3.0)
Disagree 25 (12.4)
Agree 92 (45.8)
Strongly agree 74 (36.8)
Missing 4 (2.0)
I have an overall positive experience working as a UNE/Hybrid role on the current unit
Strongly disagree 3 (1.5)
Disagree 10 (5.0)
Agree 86 (42.8)
Strongly agree 98 (48.8)
Missing 4 (2.0)
If a friend asked for your recommendation on whether or not to apply for a similar UNE/Hybrid role, would you recommend they apply?
Definitely not recommend 2 (1.0)
Probably not recommend 12 (6.0)
Probably recommend 61 (30.3)
Definitely recommend 111 (55.2)
Missing 15 (7.5)
If there is an opportunity to work in a graduate nurse position at the completion of your practicum/coursework at the same place of a UNE/Hybrid role, would you take the job?
Definitely not take the job 7 (3.5)
Probably not take the job 14 (7.0)
Probably take the job 50 (24.9)
Definitely take the job 115 (57.2)
Missing 15 (7.5)
Abbreviations: CNE, clinical nurse educator; UNE/Hybrid, undergraduate nursing employee/student hybrid.
Table 2. Descriptive Statistic (Means and Standard Deviations) Results From Midpoint Student Survey
Mean SD Range (Min-Max)
Students' experience with preceptor(s)
I meet regularly with my preceptor 3.65 0.67 3 (1-4)
My preceptor provides supervision when needed 3.72 0.55 3 (1-4)
My preceptor encourages me to ask questions and engage in self-reflection 3.59 0.72 3 (1-4)
My preceptor appears to understand the final preceptor stage of nursing education and tailors expectations accordingly 3.55 0.75 3 (1-4)
My preceptor and I are given adequate opportunity to focus on my specific learning needs 3.39 0.85 3 (1-4)
My preceptor provides ongoing constructive feedback on my performance 3.45 0.77 3 (1-4)
Students' experience with unit manager/CNE/unit staff
I have developed a trusting relationship with my preceptor 3.60 0.72 3 (1-4)
I have regular check-ins with my manager/CNE 2.60 0.89 3 (1-4)
My manager/CNE encourages me to ask questions to support my role 2.93 0.85 3 (1-4)
My manager/CNE creates an environment where I could be easily immersed in the unit 3.17 0.71 3 (1-4)
The team on my unit was aware of my unique UNE/Hybrid role 2.89 0.89 3 (1-4)
I felt very welcome as a new staff member 3.27 0.70 3 (1-4)
There was no horizontal violence (eg, bullying and conflict between groups) on my unit 3.26 0.83 3 (1-4)
Students' perceived confidence, competence, and overall satisfaction
I am becoming confident in my ability to provide safe patient care on this unit 3.49 0.57 3 (1-4)
I am clear about my role as a UNE/Hybrid and the roles of those in other professions 3.15 0.80 3 (1-4)
I feel control over decisions related to my patient's care 3.23 0.60 3 (1-4)
My experience as a UNE/Hybrid role is based on my learning needs 3.19 0.77 3 (1-4)
I have an overall positive experience working as a UNE/Hybrid role on the current unit 3.42 0.66 3 (1-4)
If a friend asked for your recommendation on whether or not to apply for a similar UNE/Hybrid role, would you recommend they apply? 3.33 0.67 3 (1-4)
If there is an opportunity to work in a graduate nurse position at the completion of your practicum/coursework at the same place of a UNE/Hybrid role, would you take the job? 3.29 0.79 3 (1-4)
Abbreviations: CNE, clinical nurse educator; UNE/Hybrid, undergraduate nursing employee/student hybrid.
Table 3. Frequencies and Percentages of Responses From Endpoint Student Survey (n = 107)
n (%)
My preceptor(s)/mentor(s) supported me in my learning and growth
Strongly disagree 5 (4.7)
Disagree 4 (3.7)
Agree 25 (23.4)
Strongly agree 72 (67.3)
Missing 1 (0.9)
The unit manager on my unit supported me in my learning and growth
Strongly disagree 7 (6.5)
Disagree 6 (5.6)
Agree 41 (38.3)
Strongly agree 52 (48.6)
Missing 1 (0.9)
The CNE(s) on my unit supported me in my learning and growth
Strongly disagree 4 (3.7)
Disagree 8 (7.5)
Agree 41 (38.3)
Strongly agree 52 (48.6)
Missing 2 (1.9)
The staff on my unit were welcoming and supportive in my learning and growth
Strongly disagree 5 (4.7)
Disagree 5 (4.7)
Agree 30 (28.0)
Strongly agree 66 (61.7)
Missing 1 (0.9)
I feel confident in my ability to provide safe patient care as a graduate nurse
Strongly disagree 4 (3.7)
Disagree 2 (1.9)
Agree 49 (45.8)
Strongly agree 52 (48.6)
My experience as a UNE/Hybrid prepared me for my role as a graduate nurse
Strongly disagree 4 (3.7)
Disagree 5 (4.7)
Agree 45 (42.1)
Strongly agree 52 (48.6)
Missing 1 (0.9)
I had an overall positive experience working as a UNE/Hybrid role on the current unit
Strongly disagree 5 (4.7)
Disagree 5 (4.7)
Agree 30 (28.0)
Strongly agree 66 (61.7)
Missing 1 (0.9)
If a friend asked for your recommendation on whether or not to apply for a similar UNE/Hybrid role, would you recommend they apply?
Definitely not recommend 2 (1.9)
Probably not recommend 3 (2.8)
Probably recommend 34 (31.8)
Definitely recommend 68 (63.6)
Abbreviations: CNE, clinical nurse educator; UNE/Hybrid, undergraduate nursing employee/student hybrid.
Table 4. Descriptive Statistic (Means and Standard Deviations) Results From Endpoint Student Survey
Mean SD Range (Min-Max)
Students' experience with preceptor/unit manager/CNE/unit staff
My preceptor(s)/mentor(s) supported me in my learning and growth 3.55 0.78 3 (1-4)
The unit manager on my unit supported me in my learning and growth 3.30 0.85 3 (1-4)
The CNE(s) on my unit supported me in my learning and growth 3.34 0.78 3 (1-4)
The staff on my unit were welcoming and supportive in my learning and growth 3.48 0.80 3 (1-4)
Students' perceived confidence, preparedness, and overall satisfaction
I feel confident in my ability to provide safe patient care as a graduate nurse 3.39 0.71 3 (1-4)
My experience as a UNE/Hybrid prepared me for my role as a graduate nurse 3.37 0.75 3 (1-4)
I had an overall positive experience working as a UNE/Hybrid role on the current unit 3.48 0.80 3 (1-4)
If a friend asked for your recommendation on whether or not to apply for a similar UNE/Hybrid role, would you recommend they apply? 2.64 0.54 3 (1-4)
Abbreviations: CNE, clinical nurse educator; UNE/Hybrid, undergraduate nursing employee/student hybrid.
Midpoint survey
Student responses from the survey indicated that overall they received great support from, and had positive relationships with their preceptors. Most students responded that their preceptors provided supervision when needed (97%) and gave ongoing constructive feedback on their performance (90%). Students' experience of working with their unit managers, CNEs, and the unit staff was also mostly positive such that the manager and CNE encouraged them to ask questions (72%) and created an environment where students could be easily immersed in the unit (82%). Most students felt welcomed as new staff members (89%) and became confident in their ability to provide safe patient care on the unit (95%). Almost 92% reported they had an overall positive experience working in the UNE/Hybrid role on their current unit. Open-ended comments by students indicated that taking a UNE/Hybrid role provided them with an opportunity for professional growth and confidence while fostering new relationships with the staff as employee/team members. They were able to gain an understanding of unit culture and routines and had opportunities to practice independently and autonomously. Suggestions for improvement included: a more clearly defined scope of the UNE/Hybrid role; clarification of workload expectations; and education for staff members about the UNE/Hybrid role (eg, paid preceptorship). Tables 1 and 2 outline the results from the midpoint survey.
Endpoint survey
Most of the students who completed the endpoint survey indicated that they had support in their learning and growth from their preceptors (91%), unit managers (87%), CNEs (87%), and unit staff (90%). Students agreed that they felt confident in their ability to provide safe patient care as (soon-to-be) graduate nurses (94%), and their experience in the UNE/Hybrid role prepared them for the graduate nurse role (91%). Students reported that they had an overall positive experience working as a UNE/Hybrid on their current unit (90%) and would recommend applying for a similar UNE/Hybrid role to a friend (95%). Comments from students in terms of their experiences indicated that those who reported a “great experience” with their preceptors received constructive criticism and feedback, and challenges for professional growth and confidence. Negative comments pertained to less experienced preceptors who did not provide a robust learning environment, and preceptors with unrealistically high expectations, both of which contributed to student anxiety, and are not unique to the UNE/Hybrid role. Several positive comments from students included their unit managers and CNEs as great resources for their learning, and extra support provided to them in terms of the scope of practice as a student within the unit. A few students were not able to interact much with their managers and/or CNEs as they were not readily available on the unit (eg, small rural sites). Most UNE/Hybrid individuals stated that unit staff members were mostly kind and approachable and provided a sense of belonging to students as team members. Students also reported witnessing significant staff burnout and fatigue on their units due to the demands of the pandemic and staffing shortages. Tables 3 and 4 outline the results from the endpoint survey.
PRACTICE IMPLICATIONS
The aim of this initiative was to think “outside the box” with a collaborative initiative between the health system and PSIs to support the nursing workforce during the surge of the COVID-19 pandemic. This initiative created a work-integrated experience for undergraduate nursing students to be employed in the newly created UNE/Hybrid role. Crisis-driven innovations are reported in the literature1,9 and highlight the importance of innovativeness in health care and nursing education—areas not typically associated with innovation. This initiative was ambitious and required shared responsibility from both the health system and PSIs. It required ongoing communication to work together to support student learning and acclimatize them into a complex and challenging nursing workforce environment.
The UNE/Hybrid role may be thought of as a disruptive innovation in health care, defined as an innovation causing radical change, often resulting in a new way of doing things, and having rippling effects throughout the system.11 The UNE/Hybrid initiative not only helped stabilize the nursing workforce, but also allowed nursing students to be part of the solution in responding to the COVID-19 pandemic: senior nursing students stepped in and were accountable for their nursing practice in a safe environment, with support from their preceptors, CNEs, unit managers, and faculty advisors. From an academic institutional perspective, it enabled students to transition from undergraduate to graduate nurses without delay and ensured students had a positive clinical placement where both institutions have shared responsibilities for positive student learning environment. The collaboration between the health care organization and the PSIs enabled students to utilize their clinical placement/employment setting as a platform for skill development, confidence, and positive nursing experiences.
This initiative has emphasized the importance of considering how work-integrated experiences can be leveraged to support the development of nursing students as well as health care systems during challenging times. The scale of the initiative, and the overall positive experience for participants, was only possible due to intentional relationship building, shared responsibility, open communication, and commitment to the collaboration process.
The COVID-19 pandemic brought challenges for both health systems and PSIs. It took a crisis to bring together stakeholders and shift mindsets from traditional approaches to more flexible and innovative ways of thinking with respect to nursing education and workforce planning/stabilization. In this case, undergraduate nursing students needed to be in a practicum that developed their nursing skills and allowed them to meet the entry-to-practice requirements of the nursing profession as they transitioned to newly graduated nurses. The collaboration between the health system and PSIs was an innovative way to address the nursing shortage, stabilize the workforce, and allow senior nursing students the opportunity to both finish their undergraduate education and contribute to the COVID-19 pandemic response. Student feedback via mid- and endpoint surveys confirmed that their development, knowledge, autonomy, and confidence were fostered and supported, and led to a successful transition into the nursing profession.
This role impacted student learning, and it provided an opportunity for leadership experience to develop students' confidence and competence as they prepare to transition into their role as the registered nurse. The clinical placement provided positive learning interactions that collaborated working with registered nurses, and other allied health professionals to develop their own relational capacity, decision-making abilities, and thinking about becoming leaders and innovators in their own practice. For instance, in the placement where they needed to work in interprofessional teams, they were able to consolidate their learning and put it into practice. This experience allowed students to understand their role within the health care systems and change some of the current practices and demonstrate thinking outside of norms.
When health care systems and PSIs collaborate on innovative solutions to crises such as the recent pandemic, everyone benefits. With this kind of collaboration, there is an increased likelihood for learning opportunities, the development of innovative solutions to complex problems, and approaches to both education and health care delivery in mutually beneficial and sustainable ways. With ongoing budget cuts, nursing workforce shortages, and overall strain on the health care system, a coordinated effort from all stakeholders is not only needed: it is critical.
LESSONS LEARNED: A PATH FORWARD
For both the health care system and nursing education, the pandemic proved a unique opportunity for systemic creativity. The UNE/Hybrid role was a successful and innovative project that, like all disruptive innovations, came with several lessons learned: Future initiatives must consider the role and the engagement of unit staff and preceptors early in the process. Preceptor buy-in is especially important in evolving the preceptor model of learning.
The role and the expectation of the clinical managers also require particular attention. The combination of increased workload for clinical managers and lack of understanding of the UNE/Hybrid role may have negatively impacted students, particularly during the early phases of this initiative. While these issues were resolved over the course of the project, some were left with negative feelings about the UNE/Hybrid role, thus demonstrating the need for increased communication and education.
Some attention must be given to the long-term effects of this initiative; for example, whether these students were hired into RN roles in the area and how they were supported and retained.
Limitations: There was lack of previous data to compare with, due to the shared patient assignment, whereas the traditional UNE role was different.
The implementation had urgency and time limitations.
There is limited research on student employment within the nursing work-force.
There is currently a gap in the literature on early student integration into the nursing workforce, and also on strategies and options for continuing with nursing clinical placements during times of health care system crisis. Supervised internship, dedicated educational units, and incentives to precept are potential areas to explore. Both practice and theoretical discussion are needed within the nursing profession regarding postpandemic clinical education. An evidence-based risk-benefit analysis from both student experience and system needs perspectives is also needed. Limitations and barriers to students integrated into clinical settings in times of health care system crisis also indicate the need for the development of an educational conceptual framework. These limitations could have a profound impact on nursing clinical education and equally on the nursing workforce. Therefore, it is critical to evaluate and projects such as this and to provide evidence of their effectiveness as we continue to face the challenges of workforce shortages within the health care system.
FUTURE RESEARCH
There are limited studies related to initiatives such as the UNE/Hybrid role in Canada. It will be critical for future research to focus not only on the long-term efficacy of this role and the impact on the health system/nursing workforce, but also on whether the continued use of the UNE/Hybrid role in nonpandemic times is a useful workforce strategy. In addition, a better understating of the role nursing leaders play, the tools and the skills they need to create clinical positive learning environments, and a culture that promotes coaching and mentorship as an essential nursing skill will be needed. With an aging population, the retirement of baby boomers, and burnout from the pandemic, the nursing profession will continue to be challenged in significant ways. Without investment in this area (both capital and social), PSIs will face issues recruiting enough preceptors10; having clinical placements to continue, especially senior placements; and health care systems will continue to be challenged with a potentially catastrophic nursing shortage.
CONCLUSION
The COVID-19 pandemic created a perfect storm for the nursing profession. The workforce will need a multilevel transformation to evolve and become prepared for the next global pandemic. Nurses are the bedrock of the health care system, and their experience is a warning about the vulnerability of the system. With a depleted health care workforce, an aging population, and an increased demand for health care services, Canada will need a transformative solution to maintain its public health care system. Innovative solutions such as the UNE/Hybrid role are important to consider as responses to this problem. Communication, collaboration, and creativity were key to the success of this UNE/Hybrid role as an innovative response to the unprecedented demands of the pandemic.
Source of Funding: Alberta Health Services Funding.
The authors declare no conflicts of interest.
==== Refs
REFERENCES
1. Fernandez AA Shaw GP . Academic leadership in a time of crisis: the coronavirus and COVID-19. J Leadersh Stud. 2020;14 (1 ):39–45. doi:10.1002/jls.21684.
2. Murphy GT MacKenzie A MacQuarrie C Sampalli T Rigby J . Evidence to care: learning from a case study of health workforce planning and COVID-19 response in Nova Scotia. Can J Nurs Leadersh. 2021;34 (4 ):19–30. doi:10.12927/cjnl.2021.26693.
3. Stevenson RL Maclaren J Vaulkhard K . Commentary—the nursing workforce: who will be left to answer the call? Nurs Leadersh (Tor Ont). 2021;34 (4 ):31–35. doi:10.12927/cjnl.2021.26692.
4. Bourgeault IL . Vulnerabilities in the nursing workforce in Canada: the anatomy and physiology of nursing workforce challenges and potential solutions for better planning, policy and management. Nurs Leadersh (Tor Ont). 2021;34 (4 ):11–18. doi:10.12927/cjnl.2021.26694.35039114
5. Snowdon W . Student nurses joining Alberta hospitals are eager to ease Omicron pressure. CBC News. https://www.cbc.ca/news/canada/edmonton/alberta-student-nurses-ahs-omicron-covid-1.6329550. Published January 28, 2022. Accessed June 9, 2022.
6. Short D . Nursing students joining COVID-19 front line to be paid: AHS. Calgary Herald. https://calgaryherald.com/news/local-news/nursing-students-will-be-paid-have-opportunity-at-full-time-employment-as-they-join-the-covid-19-frontline-ahs. Published January 23, 2022. Accessed June 1, 2022.
7. Collins A Brown JS Newman SE . Cognitive apprenticeship: teaching the crafts of reading, writing, and mathematics. In: Resnick LB , ed. Knowing, Learning, and Instruction: Essays in Honor of Robert Glaser. Hillsdale, NJ:Lawrence Erlbaum Associates;1989::453–494.
8. Rosli NA Choo TY Idris DR . Impact of preceptorship models for undergraduate nursing students and its implementation: systematic review. Int J Nurs Educ. 2022;14 (1 ):111. doi:10.37506/ijone.v14i1.17764.
9. Weberg D Chan GK Dickow M . Disrupting nursing education in light of COVID-19. Online J Issues Nurs. 2021;26 (1 ). doi:10.3912/OJIN.Vol26No01Man04.
10. Boamah SA Callen M Cruz E . Nursing faculty shortage in Canada: a scoping review of contributing factors. Nurs Outlook. 2021;69 (4 ):574–588. doi:10.1016/j.outlook.2021.01.018.33707118
11. Hogg AS . You are what you treat: Transforming the health care business model so companies—and people—thrive. Christensen Institute. https://www.Christenseninstitute.org/publications/doh-business-models/ Published May 2022. Accessed June 23, 2022.
| 36469375 | PMC9746248 | NO-CC CODE | 2022-12-15 23:21:56 | no | Nurs Adm Q. 2023 Jan 1; 47(1):72-83 | utf-8 | Nurs Adm Q | 2,022 | 10.1097/NAQ.0000000000000564 | oa_other |
==== Front
Nurs Adm Q
Nurs Adm Q
NURAQ
Nursing Administration Quarterly
0363-9568
1550-5103
Wolters Kluwer Health, Inc.
36469372
10.1097/NAQ.0000000000000566
nuraq4701p31
3
Original Articles
Reducing Staff Turnover and Clinician Burnout With a Structured Support Group During the COVID-19 Pandemic
Drexler Diane DNP, MBA, RN, FACHE [email protected]
Cornell Diane MSN, RN, CCRN-K [email protected]
Cherrie Carrie BSN, RN, CCRN-K [email protected]
Consolo Christina MSN, RN, CCRN [email protected]
Doonan Ronda L. PsyD [email protected]
Community Memorial Hospital, Ventura, California.
Correspondence: Diane Drexler, DNP, MBA, RN, FACHE, Community Memorial Hospital, 147 N. Brent St, Ventura, CA 93003 ([email protected]).
1 2023
01 12 2022
01 12 2022
47 1 Innovative Healthcare Models 3140
© 2023 Wolters Kluwer Health, Inc. All rights reserved.
2023
This article is made available via the PMC Open Access Subset for unrestricted re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the COVID-19 pandemic or until permissions are revoked in writing. Upon expiration of these permissions, PMC is granted a perpetual license to make this article available via PMC and Europe PMC, consistent with existing copyright protections.
Retention and burnout have always been a challenge for nurse leaders, but the pandemic brought these concerns to a whole new level. And now the Great Resignation is affecting health care. So how can nurse leaders at hospitals and health care systems create a supportive environment for staff during a public health emergency? Structured support groups are a viable option for emphasizing self-care and wellness. We explain why we decided to form a structured support group for our intensive care unit nurses and illustrate the results from our clinical research team. In addition, we share feedback we received from participating nurses and offer advice on forming a structured support group in acute care settings. This strategy resulted in a change in the participant's behaviors after attending the structured emotional support group. This finding aligns with the literature, which supports strategies to protect nurses' mental well-being and to take preventive measures in critical situations. Using this as a foundation, a structured emotional support group can change nurse engagement and involvement in their process and practice, during times of crisis. Many other benefits could be realized from this strategy such as improved nursing practice and processes, improved nurse satisfaction, and improved recruitment and retention.
clinician burnout
COVID-19 pandemic
reduce staff turnover
structured emotional support
==== Body
pmcTHE STRUCTURED support group was conducted within a 28-bed critical care unit (CCU) in a community hospital in a suburban area of southern California. The facility is a teaching hospital within a health care system that has a main hospital, a critical access hospital, and multiple ambulatory clinics. Prior to the COVID-19 pandemic, the unit reported minimal turnover, stable staffing, high morale, strong communication, and cohesiveness among frontline staff, department leadership, and organization administration.
During the COVID-19 pandemic, CCU nurses experienced unprecedented, sustained, intense stress related to increased patient acuity, increased patient volumes coupled with staffing shortages, and increased demands on nursing when patient families were not allowed to visit. Pandemic-related stressors included the need to learn new technology as the hospital introduced extracorporeal membrane oxygenation (ECMO) for the first time. The severity of illness related to COVID-19 drove an increase in patients on continuous renal replacement therapy, requirements to prone, and other less familiar therapies that required staff training and learning in a rapid manner. In addition, the higher patient census, the higher acuity, and the increase in unfamiliar therapies and medications led to substantially increased workloads. The unit saw high volumes of futile care, death, and prolonged dying processes that were not experienced or anticipated prepandemic. The nurses were responsible for delivering tragic news to patients' family members on a repeated basis.
THE ISSUE
Observations by leadership during the pandemic included the unit experiencing increased numbers of sick calls, escalations in leaves of absence, and increased registered nurse (RN) turnover, in particular, on night shift. In addition, staff self-reported feelings of frustration and impatience, which sometimes led to lateral violence, misunderstandings, and mistreatment of coworkers. Staff members reported sleep disturbances, feelings of wanting to avoid work, requesting assignments away from COVID-19 patients, wanting to decrease work commitments such as committee involvements, clinical advancement projects, precepting new employees, and unwillingness to work extra shifts or precept new employees. Leadership observed staff disinterest in learning new skills and avoidance of volunteering for educational opportunities. There were increased signs of emotional exhaustion including avoiding talking to patient family members when there were no positive patient changes to report, crying before, during, and after their shifts, and sadness that they could not help each other more. Nursing reported that they felt “abandoned” as ancillary services such as palliative care, social services, pharmacy, dietary, information services, and rehabilitation services tended to stay away from the “hot zones.” Hiring travelers proved to be challenging for existing staff. Leadership noticed resentment and a perception that travelers were making much more money than they were, yet they had to provide training for the temporary staff. There was a perception that the travelers were not as vested or engaged in the unit culture. Sometimes travelers were treated as members of the core staff “family” only to see them leave as contracts expired, resulting in additional feelings of loss.
Organization leadership sought to support the CCU staff in several ways and provided regular hands-on assistance in proning. There was increased staff rounding on all shifts. The Quality Department and Social Services collaborated to provide critical incident debriefings 24 hours per day. Department leadership created streamlined documentation standards and treatment protocols without diminishing standards of patient care. Department leadership created a Relaxation/Wellness Room with a massage chair, essential oils, coloring pages, yoga mats, and other tools for self-care. There was restructured physician staffing to provide additional coverage and support on the night shift with an assigned resident in the department. There was increased focus on providing additional equipment and supplies when needed. The organization continued to have smaller-scale celebrations and gifts for Nurses' Day and Hospital Week. The nurses were supported in creating small “compassion gifts” (coffee mugs, candy, tea, snacks) to give when they saw a coworker struggling. Leadership maintained unit routines such as staff meetings using a Microsoft Teams format.
Despite these interventions, leadership continued to observe signs of fatigue, sadness, and moral distress. These observations led to an attempt to address the nurses' concerns and feelings, by developing a structured support group that was facilitated by a clinical psychologist. All staff members were invited to attend the Teams (WebEx) meeting. Feedback following the initial meeting led to the design of a weekly group meeting that is still in process as of the writing of this study. The group met weekly and was attended by many CCU staff members. Staff members were assured that no one from management would be at the support group sessions to encourage open, honest communication and a safe environment to vent. Nursing staff were encouraged to attend whenever they wanted, as many times as they wanted, and with whatever level of participation they felt comfortable with. During one session, the group's facilitator sought permission from the group participants to invite Administration representatives (COO, CNO, and VP of Quality) to the next group meeting to participate by listening only. Direct CCU Management did not attend the support group in any capacity.
There was anticipation that a support group such as this would provide immediate emotional support to staff and also enable the CCU staff to develop strategies to improve resiliency and emotional well-being. As the pandemic continued longer than most anticipated, the leadership wanted to determine the effectiveness of the structured support group in reducing the feelings of emotional distress and work-related stressors among nurses in the CCU.
Hospital leadership wanted to evaluate the structured support group and its role in reducing the feelings of moral distress among nurses in the CCU. Leaders wanted to assure that a support group such as this would enable the CCU staff to develop strategies to improve resiliency and emotional well-being.
To evaluate the structured peer support group on nurse resilience, emotional well-being, retention, job satisfaction, and suggestions for improvement, hospital leadership instituted a cross-sectional survey in April 2022 to collect data regarding the group's experience.
THE EVALUATION
Our clinical research team evaluated the effectiveness of the structured support group. With approval from the institutional review board, we invited our critical care nursing staff to participate in a survey in April 2022. Nurses were asked to evaluate the support group on resilience, emotional well-being, retention, and job satisfaction and suggest improvements.
The aim of this cross-sectional evaluation was to garner honest answers about the stress of working through the pandemic and whether or not the participant found value in emotional support services. The support group was implemented in hopes of providing the CCU staff a safe place to develop strategies to improve resiliency and emotional well-being.
The objectives were to: Determine the prevalence of emotional well-being and job satisfaction among CCU nursing staff prior to the pandemic and at the time of the survey;
Determine the prevalence of emotional distress among CCU nursing staff who were exposed to structured emotional support during the COVID-19 pandemic;
Determine if offering structured emotional support for future unprecedented clinical experiences is a valuable tool for staff; and
Determine what improvements and/or changes should be made in emotional support offerings to better serve those seeking support.
METHODS
E-mails containing the consent and survey were sent to all critical care nursing staff members regardless of whether or not they participated in the structured support group. Participants were given 3 weeks to complete the anonymous survey. Reminder e-mails were sent 1 and 2 weeks after the initial invitation to participate. Data were analyzed after 1 week of study completion.
Eligibility criteria
To be included in the evaluation, the following criteria were required: critical care RN or nurse tech working at the organization in Ventura at the time of the study, at least 18 years old, and agreed to the information outlined in the electronic consent form. Since the structured emotional support group was only offered to CCU RNs or nurse techs, CCU physicians and other CCU staff were excluded from participating in the study.
Study population
The survey was offered to all 80 nurses in the CCU at an organization in Ventura in April 2022. A 50% (40 participants) response rate was anticipated; however, after 3 weeks only 24 individuals submitted the study survey, which yielded a 30% response rate.
Data collection
Data were collected by means of an anonymous electronic survey administered by SurveyMonkey. Critical care nursing staff were invited to participate in the study via flyers, verbal invitation during daily rounding/huddles and staff meetings, and via a personal e-mail invitation sent to their hospital e-mail address. The e-mail invitation included a SurveyMonkey link, which contained an electronic acknowledgment (consent) to participate in the study. Participants agreed to be a part of the study once they reviewed the consent and continued to the survey questions. Participants were given 3 weeks to complete the survey. Reminder e-mails were sent 1 and 2 weeks after the initial invitation to participate.
All survey questions, with the exception of 2 open-response questions, were multiple-choice to ensure the anonymity of the study participants. In addition, certain demographic questions such as race, ethnicity, educational level, and income level were not included in the survey to avoid potential identification of study participants.
Analysis methods
Frequency methods were applied to provide summary statistics for covariates (Table 1). Bivariate models (Fisher's exact test) were also conducted to determine if there was a significant relation between any use of psychological or emotional support and select covariates (Table 2) and between the use of organization-offered CCU structured support and select covariates (Table 3). Stata 17 (College Station, Texas) was used for all statistical analyses.
Table 1. CCU RN or Nurse Tech Survey Participants (N = 24)
Demographics n (%)
Age,a y
≤24 0 (0.0)
25-34 10 (41.7)
35-44 6 (25.0)
45-54 3 (12.5)
55-64 4 (16.7)
Gender identitya
Male 3 (12.5)
Female 20 (83.3)
Other 0 (0.0)
Prefer not to answer 0 (0.0)
Number of years as an RN or nurse techa
0-5 4 (16.7)
6-10 10 (41.7)
11-15 2 (8.3)
16-25 3 (12.5)
25+ 4 (16.7)
Number of years as a CCU RN or nurse techa
0-5 10 (41.7)
6-10 6 (25.0)
11-15 1 (4.2)
16-25 4 (16.7)
25+ 2 (8.3)
Number of years as at CMH Venturaa
0-5 8 (33.3)
6-10 8 (33.3)
11-15 2 (8.3)
16-25 4 (16.7)
25+ 1 (4.2)
Emotional well-being prior to the pandemic
Excellent 4 (16.7)
Good 16 (66.7)
Fair 3 (12.5)
Poor 1 (4.2)
Very poor 0 (0.0)
Prefer not to answer 0 (0.0)
Emotional well-being at the time of survey vs prior to the pandemic
Better 2 (8.3)
Somewhat better 3 (12.5)
Undecided 5 (20.8)
Somewhat worse 8 (33.3)
Worse 6 (25.0)
Satisfaction with work prior to the pandemic
Completely satisfied 2 (8.3)
Very satisfied 11 (45.8)
Moderately satisfied 9 (37.5)
Slightly satisfied 0 (0.0)
Not at all satisfied 0 (0.0)
Prefer not to answer 2 (8.3)
Satisfaction with work at the time of survey
Completely satisfied 0 (0.0)
Very satisfied 4 (16.7)
Moderately satisfied 9 (37.5)
Slightly satisfied 7 (29.2)
Not at all satisfied 3 (12.5)
Prefer not to answer 1 (4.2)
There is value in group support or therapy
Strongly agree 9 (37.5)
Agree 9 (37.5)
Undecided 5 (20.8)
Disagree 1 (4.2)
Strongly disagree 0 (0.0)
Utilized psychological or emotional support during the pandemic
Yes 19 (79.2)
No 5 (20.8)
Utilized the following methods of emotional support during the pandemicb
Individual therapy (community-based) 11 (45.8)
Individual therapy (CMH-sponsored/EAP) 2 (8.3)
Group therapy 3 (12.5)
Group support (community-based) 3 (12.5)
Group support (CMH-based) 16 (66.7)
Spiritual support 7 (29.2)
Informal medical peer support 8 (33.3)
Informal family/friend support 11 (45.8)
Abbreviations: CCU, critical care unit; CMH, Community Memorial Hospital; EAP, Employee Assistance Program; RN, registered nurse.
aMissing data for 1 participant.
bResponses are not mutually exclusive.
Table 2. Prevalence of the Utilization of Psychological or Emotional Support During the Pandemic (N = 24)a
Demographics No Use of Emotional Support (0) (N = 5), n (%) Use of Emotional Support (1) (N = 19), n (%) P
Age,b y .74
≤24 0 (0.0) 0 (0.0)
25-34 3 (60.0) 7 (36.8)
35-44 1 (20.0) 5 (26.0)
45-54 1 (20.0) 2 (10.5)
55-64 0 (0.0) 4 (21.1)
Gender identityb .54
Male 1 (20.0) 2 (10.5)
Female 4 (80.0) 16 (84.2)
Other 0 (0.0) 0 (0.0)
Prefer not to answer 0 (0.0) 0 (0.0)
Number of years as an RN or nurse techb .00
0-5 4 (80.0) 0 (0.0)
6-10 0 (0.0) 10 (52.6)
11-15 0 (0.0) 2 (10.5)
16-25 1 (20.0) 2 (10.5)
25+ 0 (0.0) 4 (21.1)
Number of years as a CCU RN or nurse techb .39
0-5 4 (80.0) 6 (31.6)
6-10 0 (0.0) 6 (31.6)
11-15 0 (0.0) 1 (5.3)
16-25 1 (20.0) 3 (15.8)
25+ 0 (0.0) 2 (10.5)
Number of years as at CMH Venturab .14
0-5 4 (80.0) 4 (21.1)
6-10 0 (0.0) 8 (42.1)
11-15 0 (0.0) 2 (10.5)
16-25 1 (20.0) 3 (15.8)
25+ 0 (0.0) 1 (5.3)
Emotional well-being prior to the pandemic .21
Excellent 0 (0.0) 4 (21.1)
Good 3 (60.0) 13 (68.4)
Fair 2 (40.0) 1 (5.3)
Poor 0 (0.0) 1 (5.3)
Very poor 0 (0.0) 0 (0.0)
Prefer not to answer
Emotional well-being at the time of survey vs prior to the pandemic .51
Better 0 (0.0) 2 (10.5)
Somewhat better 1 (20.0) 2 (10.5)
Undecided 2 (40.0) 3 (15.8)
Somewhat worse 2 (40.0) 6 (31.6)
Worse 0 (0.0) 6 (31.6)
Satisfaction with work prior to the pandemic .11
Completely satisfied 0 (0.0) 2 (10.5)
Very satisfied 2 (40.0) 9 (47.4)
Moderately satisfied 1 (20.0) 8 (42.1)
Slightly satisfied 0 (0.0) 0 (0.0)
Not at all satisfied 0 (0.0) 0 (0.0)
Prefer not to answer 2 (40.0) 0 (0.0)
Satisfaction with work at the time of survey .25
Completely satisfied 0 (0.0) 0 (0.0)
Very satisfied 0 (0.0) 4 (21.1)
Moderately satisfied 3 (60.0) 6 (31.6)
Slightly satisfied 1 (20.0) 6 (31.6)
Not at all satisfied 0 (0.0) 3 (15.8)
Prefer not to answer 1 (20.0) 0 (0.0)
There is value in group support or therapy .22
Strongly agree 0 (0.0) 9 (47.4)
Agree 3 (60.0) 6 (31.6)
Undecided 2 (40.0) 3 (15.8)
Disagree 0 (0.0) 1 (5.3)
Strongly disagree 0 (0.0) 0 (0.0)
Abbreviations: CCU, critical care unit; CMH, Community Memorial Hospital; RN, registered nurse.
a(0) and (1) indicate number of participants in that category.
bMissing data for 1 participant.
Table 3. Prevalence of Use of Organization-Offered Structured CCU Support Group (N = 19)a
Demographics No Use of Structured Group (0) (N = 1), n (%) Use of Structured Group (1) (N = 18), n (%) P
Age,b y 1.00
≤24 0 (0.0) 0 (0.0)
25-34 1 (100.0) 6 (33.3)
35-44 0 (0.0) 5 (27.8)
45-54 0 (0.0) 2 (11.1)
55-64 0 (0.0) 4 (22.2)
Gender identityb 1.00
Male 0 (0.0) 2 (11.1)
Female 1 (100.0) 15 (83.3)
Other 0 (0.0) 0 (0.0)
Prefer not to answer 0 (0.0) 0 (0.0)
Number of years as an RN or nurse techb 1.00
0-5 0 (0.0) 0 (0.0)
6-10 1 (100.0) 9 (50.0)
11-15 0 (0.0) 2 (11.1)
16-25 0 (0.0) 2 (11.1)
25+ 0 (0.0) 4 (22.2)
Number of years as a CCU RN or nurse techb 1.00
0-5 1 (100.0) 5 (27.8)
6-10 0 (0.0) 6 (33.3)
11-15 0 (0.0) 1 (5.6)
16-25 0 (0.0) 3 (16.7)
25+ 0 (0.0) 2 (11.1)
Number of years as at CMH Venturab .56
0-5 1 (100.0) 3 (16.7)
6-10 0 (0.0) 8 (44.4)
11-15 0 (0.0) 2 (11.1)
16-25 0 (0.0) 3 (16.7)
25+ 0 (0.0) 1 (5.6)
Emotional well-being prior to the pandemic .11
Excellent 0 (0.0) 4 (22.2)
Good 0 (0.0) 13 (72.2)
Fair 1 (100.0) 0 (0.0)
Poor 0 (0.0) 1 (5.6)
Very poor 0 (0.0) 0 (0.0)
Prefer not to answer
Emotional well-being at the time of survey vs prior to the pandemic 1.00
Better 0 (0.0) 2 (11.1)
Somewhat better 0 (0.0) 2 (11.1)
Undecided 0 (0.0) 3 (16.7)
Somewhat worse 0 (0.0) 6 (33.3)
Worse 1 (100.0) 5 (27.8)
Satisfaction with work prior to the pandemic 1.00
Completely satisfied 0 (0.0) 2 (11.1)
Very satisfied 1 (100.0) 8 (44.4)
Moderately satisfied 0 (0.0) 8 (44.4)
Slightly satisfied 0 (0.0) 0 (0.0)
Not at all satisfied 0 (0.0) 0 (0.0)
Prefer not to answer 0 (0.0) 0 (0.0)
Satisfaction with work at the time of survey 1.00
Completely satisfied 0 (0.0) 0 (0.0)
Very satisfied 0 (0.0) 4 (22.2)
Moderately satisfied 1 (100.0) 5 (27.8)
Slightly satisfied 0 (0.0) 6 (33.3)
Not at all satisfied 0 (0.0) 3 (16.7)
Prefer not to answer 0 (0.0) 0 (0.0)
There is value in group support or therapy 1.00
Strongly agree 1 (100.0) 8 (44.4)
Agree 0 (0.0) 6 (33.3)
Undecided 0 (0.0) 3 (16.7)
Disagree 0 (0.0) 1 (5.6)
Strongly disagree 0 (0.0) 0 (0.0)
Abbreviations: CCU, critical care unit; CMH, Community Memorial Hospital; RN, registered nurse.
a(0) and (1) indicate number of participants in that category.
bMissing data for 1 participant.
RESULTS
Table 1 provides descriptive statistics for age, gender identity, number of years as an RN or nurse tech, number of years as a CCU RN or nurse tech, number of years at the Community Memorial Hospital in Ventura, emotional well-being prior to the pandemic, emotional well-being at the time of survey versus prior to the pandemic, satisfaction with work prior to the pandemic, satisfaction with work at the time of survey, if there is value in group support or therapy, the utilization of psychological or emotional support during the pandemic, and the types of emotional support methods utilized during the pandemic. Of the 24 participants, approximately 41.7% (n = 10) were between the ages of 25 and 34 years and 83.3% (n = 20) were female. In terms of years of service, 41.7% (n = 10) of participants indicated having worked as an RN or nurse tech for 6 to 10 years, 41.7% (n = 10) of participants have worked as a CCU RN or nurse tech for 0 to 5 years, and approximately 66.6% (n = 16) of the participants indicated having worked at the organization in Ventura between 0 and 10 years. When reporting their emotional well-being prior to the pandemic, 83.3% (20) of participants indicated an emotional well-being of “good” or “excellent” in comparison with 16.7% (n = 4) of participants who indicated an emotional well-being of “fair” or “poor.” When asked about their emotional well-being at the time of the survey versus prior to the pandemic, 20.8% (n = 5) of participants indicated “somewhat better” or “better,” another 20.8% (n = 5) of participants were undecided, and 58.3% (n = 14) of participants indicated “somewhat worse” or “worse.”
Fisher's exact test was used to determine if there was a significant association between utilization of psychological or emotional support and select covariates from Table 1 (Table 2) and to determine if there was a significant association between the use of organization-offered CCU support group and select covariates from Table 1 (Table 3). The bivariate model outlined in Table 2 indicated that there is a statistical significance between the utilization of psychological or emotional support and the number of years as an RN or nurse tech (P < .001). However, the bivariate model outlined in Table 3 yielded no statistically significant results.
Participants who attended the organization-offered support group were able to provide additional feedback in the form or 2 free response questions that asked (1) what they found most helpful about the support group and (2) if they had any suggestions for improvements moving forward. Overwhelmingly, participants shared that they found it helpful to share their experience with their peers because it made them realize that they are not alone. They also shared that they found it helpful to have a clinical psychologist facilitate the group's discussion to help them process their difficult experience. In terms of improvements, participants shared that they would like for leadership to acknowledge their difficult experience, for the group meetings to continue meeting regularly, for weekly discussion topics, and to emphasize confidentiality among the group.
Nursing leader's observations after starting the support group included the following: Increased willingness to participate in precepting new employees;
Increased engagement in learning activities—2 employees even traveled out of state for ECMO training;
Return of focus to unit-based governance councils and projects to improve the department;
Less lateral violence;
Increased evidence of empathy;
Greater ability of staff to identify and intervene when another staff member is distressed;
Recognition that other clinicians (physicians, pharmacists, therapists, etc) also experienced moral distress and needed emotional support;
Elimination of requests to “opt out” of being assigned COVID-19 patients;
Increased vocalization of staff that they appreciated the support group;
Staff were more emotionally prepared for the COVID-19 surge in January 2022;
Emotions switched from sadness to anger when patients/family/visitors refused to get vaccinated;
Staff perceived value in support group being available to them; and
Leadership did presentation for staff to remember the joy in their work and provided tools to find joy.
DISCUSSION
The purpose of this evaluation was to determine if a structured emotional support group reduced distress in critical care nurses during the COVID-19 pandemic. The sample size for this study was very small; however, this evaluation proposes suggestions on how to improve this type of program in the future (eg, giving participants a pre- and postintervention group participation survey to evaluate them at baseline and postintervention, randomized study with an intervention and control arm, etc). The participants provided great feedback that we can implement in future events. This evaluation results suggest that the structured emotional support group had the desired impact of improved staff morale and job satisfaction.
CONCLUSIONS
This evaluation found a statistically significant difference in the participant's behaviors before and after attending the structured emotional support group. This finding aligns with the literature, which supports strategies to protect nurses' mental well-being1,2 and to take preventive measures in critical situations.3–6 Using this as a foundation, a structured emotional support group can change nurse engagement and involvement in their process and practice, during times of crisis. Many other benefits could be realized from this model, such as improved nursing practice and processes, reduced clinician burnout, improved nurse satisfaction, and improved recruitment and retention.
Conflicts of Interest: None to declare.
==== Refs
REFERENCES
1. Green AA Kinchen EV . The effects of mindfulness meditation on stress and burnout in nurses. J Holist Nurs. 2021;39 (4 ):356–368. doi:10.1177/08980101211015818.33998935
2. Pollock A Campbell P Cheyne J . Interventions to support the resilience and mental health of frontline health and social care professionals during and after a disease outbreak, epidemic or pandemic: a mixed methods systematic review. Cochrane Database Syst Rev. 2020;11 (11 ):CD013779. doi:10.1002/14651858.CD013779.33150970
3. Li M Shu Q Huang H Bo W Wang L Wu H . Associations of occupational stress, workplace violence, and organizational support on chronic fatigue syndrome among nurses. J Adv Nurs. 2020;76 (5 ):1151–1161. doi:10.1111/jan.14312.32011012
4. Rose S Hartnett J Pillai S . Healthcare worker's emotions, perceived stressors and coping mechanisms during the COVID-19 pandemic. PLoS One. 2021;16 (7 ):e0254252. doi:10.1371/journal.pone.0254252.34242361
5. Sierakowska M Doroszkiewicz H . Stress coping strategies used by nurses during the COVID-19 pandemic. PeerJ. 2022;10 :e13288. doi:10.7717/peerj.13288.35529493
6. Yu F Raphael D Mackay L Smith M King A . Personal and work-related factors associated with nurse resilience: a systematic review. Int J Nurs Stud. 2019;93 :129–140. doi:10.1016/j.ijnurstu.2019.02.014.30925279
| 36469372 | PMC9746250 | NO-CC CODE | 2022-12-15 23:21:56 | no | Nurs Adm Q. 2023 Jan 1; 47(1):31-40 | utf-8 | Nurs Adm Q | 2,022 | 10.1097/NAQ.0000000000000566 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Elsevier Inc.
S0002-9343(21)00321-1
10.1016/j.amjmed.2021.04.027
Advancing High Value Health Care
Ensuring Primary Care Diagnostic Quality in the Era of Telemedicine
Willis Joel Steven DO, PA, MA, MPhil a⁎
Tyler Carl Jr MD, MSc b
Schiff Gordon D. MD c
Schreiner Katherine BA d
a Assistant Professor, Division of Family Medicine, Associate Medical Director, GW Immediate Primary Care, George Washington University, Washington, DC
b Professor of Family Medicine and Community Health, Cleveland Clinic Lerner College of Medicine, Case Western Reserve University School of Medicine, Cleveland, Ohio
c Associate Professor of Medicine, Harvard Medical School, Brigham and Women's Hospital, Boston, Mass
d Medical Student, George Washington School of Medicine and Health Sciences, Washington, DC
⁎ Requests for reprints should be addressed to Joel Steven Willis, DO, PA, MA, MPhil, Assistant Professor, Division of Family Medicine, Associate Medical Director, GW Immediate Primary Care, George Washington University, 2120 L Street NW, Ste 450, Washington, DC, 20037.
27 5 2021
9 2021
27 5 2021
134 9 11011103
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcTelemedicine applications have been used for decades, most commonly in specific areas of medicine (eg, dermatology, pathology, radiology) and in specific contexts (eg, rural or other under-resourced areas).1 The coronavirus disease 2019 (COVID-19) pandemic has markedly accelerated the expansion of the telemedicine platform, especially video visits, for primary care clinicians and practices relatively unfamiliar with its use. Although telemedicine usage may be receding from its peak pandemic rollout, clinicians will likely continue to employ these video visits for patient care. Within this context of new and widespread ongoing use, critical evaluation of telemedicine's benefits and harms related to diagnosis is necessary and linked to efforts to minimize potential diagnostic errors and to establish best clinical practices.
Potential for Diagnostic Error
Existing research regarding diagnostic accuracy in telemedicine has focused primarily on tele-dermatology or specific clinical contexts (eg, stroke identification).2, 3, 4 There has been far less examination of diagnostic accuracy in the broader range types and severity of diagnoses that typically present to primary care clinicians. Despite this, primary care clinicians have been thrust almost overnight, onto a fundamentally different platform for diagnosis and management of clinical questions: Does the patient appear well, mildly or moderately ill, or in need of emergency medical services? What tests, referrals, or procedures are needed or appropriate lacking a traditional physical examination?
There is an urgent need to evaluate the extent to which our current technological resources fall short of providing the truest clinical picture to inform these diagnostic decisions. This is not to say that face-to-face health care encounters unfailingly result in accurate diagnoses; indeed, there is robust literature to the contrary.5 However, in concordance with broader ongoing efforts to improve on diagnostic accuracy in medicine, we must recognize the specific challenges posed by telemedicine.6 , 7
Diagnostic Advantages and Disadvantages
A 5-component structural model of primary care telemedicine encounters provides a framework to examine the elements impacting telemedicine diagnostic quality and error (Table ).Table Factors Influencing Diagnostic Accuracy in Primary Care Telemedicine
TableDomain Specifics
Patient Technology availability, physical environment, social environment (privacy, distractions, embarrassment), health literacy, technical proficiency (experience with communication technologies), additional support from family members/others
Physician Office versus home environment, staff support, clinical encounter workflow, EMR skills, health information access, clinical diagnostic experience, telemedicine training and proficiency, engagement versus burnout
Telemedicine/EMR Platform Audio-video quality, bandwidth, connectivity time, embedded smart technologies, design for user ease and telemedicine integration and workflow, multiple other general features that support diagnosis8
Clinical Contexts Visit agenda, patient-physician rapport, diagnostic problem urgency, complexity, and clinician familiarity, pre- and postvisit time, logistics, and support
Health System Regulations, reimbursement, health system supports, scheduling, and productivity requirements (patients per hour), patient continuity, continuous learning, and improvement
EMR = electronic medical record.
The first component is the patient, whose video visit occurs in a specific context and environment, influenced by family and other social networks. In the secure environment of the home, patients may recount their illness and symptoms with greater detail, fostering more accurate diagnosis. Additionally, family member contributions to a patient's history may improve diagnosis, occurring more easily in the virtual platform (eg, “Honey, did you tell them about the sore on your foot?”). Likewise, examination of a patient's environment may provide diagnostic clues not visible in routine clinic visits (eg, spotting fall risks such as throw rugs in the home of an elderly patient). More challenging are situations in which the lack of privacy in the home may hinder discussion about sensitive topics such as anxiety, depression, or symptoms suggestive of sexually transmitted infections. Riskier still would be discussion of imminent threats to personal safety like domestic abuse.
The second component of telemedicine is the physician, who may be working less closely with the office-based health care team. In practice-based telemedicine, the introduction of video visits into outpatient clinics has changed workflow and team dynamics, creating environments where physicians may be working solo or with reduced support, altering routine diagnostic cues and data. Physicians in an outpatient practice may still have access to data such as prior patient records, but they are disadvantaged by loss of other data such as vital signs and clinical observations gathered by other health care team members. The current lack of standardization in telemedicine workflows and in comprehensive health care team training in the virtual setting may create wide variability in information-gathering and even in deciding which clinical complaints are appropriate for telemedicine encounters.
Third, the telemedicine platform itself introduces potential for diagnostic error. On the one hand, video visits may enhance diagnostic timeliness by removing access barriers and shortening delays in initial diagnostic evaluation. Indeed, the convenience and expanded access to care through telemedicine has been associated with high patient satisfaction.9 However, the sudden and substantial demand for telemedicine during the pandemic has led to use of readily available consumer-grade videoconferencing platforms (Zoom, Facetime, and Skype), devoid of any “smart” technology that might enhance clinician observation or diagnostic accuracy. Such platforms were originally designed to connect individuals socially or in office meeting conditions—not for Health Insurance Portability and Accountability Act (HIPAA)-compliant, high-resolution clinical observation, on which a patient's medical encounters and diagnosis may depend.10 Additionally, the diagnostic utility of video visits depends on both clear audio and image quality. Either feature may be impaired by an unstable Internet connection, limited capacity of a patient's device, or disrupting features of a patient's environment (poor lighting, background noise), all of which may contribute to diagnostic error. Further, a clinician's inability to expedite connectivity/technology issues required for video teleconferencing may consume 5-10 minutes of a 15-minute appointment, obviously compromising valuable encounter and diagnostic assessment time.
Fourth, the context of the video visit itself can significantly influence the quality of care provided. Certain types of clinical encounters may be better suited for the telemedicine platform. For example, telemedicine visits regarding follow-up for patients’ chronic conditions such as diabetes or high blood pressure with their established primary care physician may be conducted remotely with less chance of clinical error compared with a visit involving an unknown patient-clinician dyad with acute symptoms (ie, shortness of breath or chest pain). Explicit evidence-based criteria regarding which types of encounters are appropriate for primary care telemedicine visits have not been established. Challenges in building rapport, especially for a new patient encounter, may also introduce situations where a patient's level of comfort may lead to less disclosure of information and hinder an accurate diagnostic assessment.
Fifth, the system through which telemedicine currently operates in the United States was built and disseminated quickly to meet emergent needs. Regulations hindering ease of telemedicine use were initially relaxed.11 However, the ways by which legal and financial supports for telemedicine evolve moving forward will shape potential for diagnostic error reduction in the future. For example, a downgraded payment structure for video visits may directly affect a clinician's time and resources to accurately diagnose a patient's condition. Similarly, health care system support and training of end users to wield this new medium have the potential to lessen—or exacerbate—possible threats to diagnostic accuracy and patient safety.
Setting a Research Agenda
The preceding model provides a framework to examine factors that may influence diagnostic error in telemedicine. Examining these components could guide explorations of the multidimensional processes relevant to telemedicine practice, ranging from psychosocial (eg, patient reluctant to disclose symptoms within earshot of others) to health professional education (eg, deficits in physician training and telemedicine diagnosis) to improving the telemedicine platform (eg, “smart information technology”) to contextual factors (eg, establishing appropriate triage algorithms) to systemic issues (eg, adequate time allotted per encounter and resources devoted to infrastructure development and training by end-users).
Traditional strategies for decreasing diagnostic error must be reconceptualized, redesigned, and tailored to the specific capacities and limitations of video visits. Creating, implementing, and evaluating these strategies will require cross-disciplinary insights and collaboration. Applying concepts and methods from social psychology and communication science will be needed to provide insights for optimizing communication through virtual media. Even something as seemingly basic as teaching patients to adjust the camera on their device may be decisive for enhanced clues needed for accurate diagnosis (ie, showing a dermal lesion or jugular venous distention clearly).
Improvements in video platform technology may help avert diagnostic errors due to poor visual or auditory clarity but are likely marginal compared to broader nontechnical issues. Borrowing from industrial and systems engineering principles, innovative larger system workflow redesign that redirect specific clinical tasks or scenarios to or away from exclusive reliance on telemedicine platforms (eg, leveraging home visiting nurses or physical therapists or shorten follow-up intervals to better operationalize “the test of time”). There is limited research to date measuring similarities and differences in diagnostic accuracy for specific, common conditions via virtual platforms compared with face-to-face encounters.12 However, at the very least, the inability to form a corporeal connection and perform certain physical examination assessments over video platforms will likely continue to pose a major limitation for reliable diagnosis and introduce opportunities for medical error and mismanagement of both acute and chronic conditions.
Embracing the Future
Although the conveniences of telemedicine use for both primary care patients and clinicians are many, benefits must be carefully weighted with limitations, particularly for medical diagnosis. Telemedicine encounters do not happen in a vacuum. Pre-encounter information solicitation as well as more frequent and systematic postvideo encounter follow-up can surely enhance diagnostic thoroughness and safety. The framework of management reasoning—defined as the process of making decisions about patient management, bundling choices about treatment, follow-up visits, future testing, and allocation of limited resources—may aid reframing telemedicine diagnosis and the clinical reasoning behind it.13 Ultimately a directed focus on understanding the extent and sources of potential diagnostic error in telemedicine coupled with mitigation of its modifiable factors will allow an expanded capacity to provide safe, patient-centered primary care services.
Funding: None.
Conflicts of Interest: None.
Authorship: All authors contributed to data gathering and writing/editing this manuscript.
==== Refs
References
1 The University of New Mexico. Project Echo. Available at: https://hsc.unm.edu/echo/. Accessed February 19, 2021.
2 Solenski NJ Telestroke Neuroimaging Clin N Am 28 4 2018 551 563 10.1016/j.nic.2018.06.012 30322592
3 Trettel A Eissing L Augustin M Telemedicine in dermatology: findings and experiences worldwide - a systematic literature review J Eur Acad Dermatol Venereol 32 2 2018 215 224 10.1111/jdv.14341 28516492
4 Bashshur RL Krupinski EA Weinstein RS Dunn MR Bashshur N. The empirical foundations of telepathology: evidence of feasibility and intermediate effects Telemed J E Health 23 3 2017 155 191 10.1089/tmj.2016.0278 28170313
5 Panesar SS deSilva D Carson-Stevens A How safe is primary care? A systematic review BMJ Qual Saf 25 7 2016 544 553 10.1136/bmjqs-2015-004178
6 National Academies of Sciences, Engineering, and Medicine Improving Diagnosis in Health Care 2015 The National Academies Press Washington, DC
7 Coalition to Improve Diagnosis. Society to Improve Diagnosis in Medicine. Available at: https://www.improvediagnosis.org/coalition/. Accessed January 23, 2021.
8 El-Kareh R Hasan O Schiff GD. Use of health information technology to reduce diagnostic errors BMJ Qual Saf 22 2013 ii40 ii51 10.1136/bmjqs-2013-001884
9 Ramaswamy A Yu M Drangsholt S Patient satisfaction with telemedicine during the COVID-19 pandemic: retrospective cohort study J Med Internet Res 22 9 2020 e20786 10.2196/20786 32810841
10 Contreras CM Metzger GA Beane JD Dedhia PH Ejaz A Pawlik TM. Telemedicine: patient-provider clinical engagement during the COVID-19 pandemic and beyond J Gastrointest Surg 24 7 2020 1692 1697 10.1007/s11605-020-04623-5 32385614
11 Rockwell KL Gilroy AS. Incorporating telemedicine as part of COVID-19 outbreak response systems Am J Manag Care 26 4 2020 147 148 10.37765/ajmc.2020.42784 32270980
12 Akhtar M Van Heukelom PG Ahmed A Telemedicine physical examination utilizing a consumer device demonstrates poor concordance with in-person physical examination in emergency department patients with sore throat: a prospective blinded study Telemed J E Health 24 10 2018 790 796 10.1089/tmj.2017.0240 29470127
13 Cook D Sherbino J Durning S. Management reasoning beyond the diagnosis JAMA 319 22 2018 2267 2268 doi: 10.1001.2018.4385 29800012
| 34051151 | PMC9746257 | NO-CC CODE | 2022-12-15 23:21:56 | no | Am J Med. 2021 Sep 27; 134(9):1101-1103 | utf-8 | Am J Med | 2,021 | 10.1016/j.amjmed.2021.04.027 | oa_other |
==== Front
Am J Med
Am J Med
The American Journal of Medicine
0002-9343
1555-7162
Elsevier Inc.
S0002-9343(21)00309-0
10.1016/j.amjmed.2021.04.020
Commentary
Rediscovering Meaning and Purpose: An Approach to Burnout in the Time of COVID-19 and Beyond
Southwick Steven MD abc⁎
Wisneski Leonard MD bde
Starck Patricia RNPhD bf
a Yale University School of Medicine, New Haven, Conn
b Viktor Frankl Institute of Logotherapy, Abilene, Texas
c Icahn School of Medicine at Mount Sinai, New York, NY
d University of Colorado School of Medicine, Aurora
e George Washington School of Medicine, Washington, DC
f University of Texas Health Science Center, Houston
⁎ Requests for reprints should be addressed to Steven Southwick, MD, Department of Psychiatry, Yale University School of Medicine, 27 Castle Rock, Branford, CT 06405.
11 5 2021
9 2021
11 5 2021
134 9 10651067
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcRediscovering Meaning and Purpose
Medicine is in the midst of a crisis, resulting in high rates of burnout. Over the past few decades there has been an erosion of meaning, largely caused by the de-humanizing commodification of medicine with its inattention to humanistic values, where the focus is on rapid throughput and profit, and where practitioners are forced to spend much of their time engaged in tasks and activities unrelated to what they find most meaningingful. This surge in burnout has been magnified by the ongoing Coronavirus disease 2019 (COVID-19) pandemic.
We believe the current level of staff burnout is directly related to loss of meaning and purpose in daily work. In this viewpoint, we propose that medical institutions and health care providers embrace the meaning-centered teaching originally articulated by Viktor Frankl, MD, PhD and expanded by scholars around the globe and from various professional fields.
Frankl (1905-1997), a Viennese neurologist, psychiatrist, and philosopher, proposed that the primary motivation of human behavior was not pleasure (Freud) or power (Adler), but the desire to seek meaning and purpose in one's life. He called his style of clinical practice logotherapy (meaning-centered therapy) and published numerous ground-breaking books, including Man's Search for Meaning 1 and The Doctor and the Soul.2
How Important Are Meaning and Purpose in Medical Practice?
In a study of 465 academic physicians,3 those who spent <20% of their time at work engaged in the activity they found most personally meaningful had significantly higher rates of burnout (53.8%) compared with those who spent >20% of their work time (29.9%) engaged in such activities. Further, the less time engaged in the most meaningful activity was the strongest predictor of burnout, followed by working more hours per week, younger age, and working as a generalist rather than subspecialist.
In a series of publications based on a survey of 2000 full-time members of the American Academy of Family Physicians,4 the factor most strongly associated with physician happiness was career purpose, which included career satisfaction, joy in work, spiritual purpose, and meaning in patient relationships. A sense of calling and having long-term relationships with patients were strongly associated with high life meaning. In most cases, extrinsic motivators, such as annual income or other work-related characteristics, were not related to well-being, life satisfaction, life meaning, or career commitment.
That meaning, purpose, and calling are so strongly linked to physician well-being is consistent with a large body of research linking meaning and purpose to mental, physical, and spiritual health. For example, in the Health and Retirement Study of 6985 subjects who were followed over a 4-year period, stronger purpose in life was associated with decreased mortality.5 Meaning and purpose is also strongly associated with resilience,6 something that is essential for all medical practitioners.
Recently, a number of institutions have recommended that a greater sense of meaning and purpose be re-infused into the practice of medicine. A Press Ganey white paper7 on burnout and resilience in medicine emphasized the need to “more reliably find meaning, pleasure and respect” in work; and the Accreditation Council for Graduate Medical Education8 recommended that trainees have more direct contact with patients as a way to enhance meaning. The call to focus on meaning and purpose is even more pressing during the COVID-19 pandemic. In a 2020 survery of over 2300 physicians, the Physicians Foundation9 found that 30% of respondents reported feelings of hopelessness or having no purpose as a result of pandemic-related changes in their practice.
What Do Doctors Find Meaningful About Their Work?
In a qualitatitive study conducted at annual meetings of the American College of Physicians and the Society of General Internal Medicine,10 3 major themes emerged: 1) experiencing a change in perspective about human nature, themselves, their roles, illness, or patient care after involvement in a profound or emotional event; 2) connecting with patients in moments of intimacy; and 3) feeling that they had made a difference in patients’ lives. The author's were “struck” that nearly all participants wrote about nontechnical humanistic interactions with patients rather than diagnostic and therapeutic successes.
While the COVID-19 pandemic has resulted in a renewed sense of purpose for some health care workers, for others it has meant reduced time with patients in moments of intimacy. Further, some have had to wrestle with moral and ethical dilemmas including having to make life-and-death decisions based on limited resources and watching patients suffer and sometimes die alone. These issues can contribute to guilt, sadness, and burnout.
How Can Franklian Principles Enhance Meaning at Work for Medical Professionals?
Frankl advocated finding meaning through: 1) creative acts or what we give to the world; 2) experiences that we take from the world; and 3) attitude.1 , 2 , 11 Creative acts include tasks to be accomplished, one's career, making a difference in someone's life, and alleviating suffering. Experiential values include witnessing courage in the face of tragedy and experiencing deeply emotional encounters with suffering patients. Attitudinal values means choosing an attitude toward an unalterable fate such as chronic illness or imminent death, what Frankl called the last of the human freedoms.
How Does Frankl's Philosophical Approach Differ from Other Teachings?
Compared with other philosophical and psychological teachings, logo-philosophy sees life in terms of solutions rather than problems, focuses on goals rather than obstacles, emphasizes discovering rather than uncovering, and is holistic rather than reductionistic. It is future oriented, focuses on personal strengths, emphasizes individual values, and places responsibility for change on patients, practitioners, and institutions.12 With systems, Franklian interventions can help organizations build frameworks, structures, and supports to maintain a sense of meaning in the work environment.
Conclusion
The practice of medicine has changed substantially over the past few decades, with an accompanying loss of meaning and purpose. As noted by Christine Sinsky of the American Medical Association, “At the highest level we are disconnected from our purpose and have lost touch with the things that give joy and meaning to our work.”13 The cost is enormous. We must not forget the wisdom of of William Osler, who believed that the medical profession was not a business, but rather a calling.
Based on a large body of psychological and organizational research, we believe that helping employees, trainees, and medical institutions rediscover meaning and purpose in their work will improve morale and motivation, reduce burnout and depression, lessen staff turnover, and improve health care outcomes. While numerous health care experts and organizations have identified the need to rediscover and foster meaning in medicine, practical approaches have been relatively elusive. The basic principles and teachings of Frankl's logotherapy have time-tested answers for discovering meaning and purpose in life.
Health care, with its mission of service to patients, families, and communities, provides a unique opportunity for all health care employees to experience high levels of meaning in their work. It is up to medical leaders to create an environment with policies, procedures, and training opportunities that help all employees understand and connect with the organization's ultimate mission of service.
Frankl believed that every crisis presents an opportunity for growth. We propose that teaching the principles put forth by Viktor Frankl to health care students, medical caregivers, hospital administrators, and patients would help to bring professional satisfaction and meaning back to medicine in the time of COVID-19 and beyond.
Acknowledgment
The authors acknowledge Robert Barnes, PhD and Bernadette Lowthert, MBA for their valuable contributions.
Funding: None.
Conflicts of Interest: SS receives royalties for the book Resilience: The Science of Mastering Life's Greatest Challenges. Cambridge University Press, 2018.
Authorship: All authors have participated in the preparation of the manuscript.
==== Refs
References
1 Frankl VE Man's Search for Meaning: An Introduction to Logotherapy 1985 Pocket Books New York
2 Frankl VE “The Doctor and the Soul” From Psychotherapy to Logotherapy 1977 Vintage Books New York
3 Shanafelt TD West CP Sloan JA Career fit and burout among academic faculty Arch Intern Med 169 10 2009 990 995 19468093
4 Tak HJ Curlin FA Yoon JD Association of intrinsic motivating factors and markers of physician well-being: a national physician survey Gen Intern Med 32 7 2017 739 746
5 Alimujiang A Wiensch A Boss J Association between life purpose and mortality among US adults older that 50 years JAMA Netw Open 2 5 2019 e194270
6 Southwick SM Charney DS Resilience: The Science of Mastering Life's Greatest Challenges 2018 Cambridge University Press Cambridge, England
7 Press Ganey 2018 White Paper, Burnout and Resilience. A Framework for Data Analysis and a Positive Path Forward 2018 Available at: https://www.pressganey.com/blog/burnout-and-resilience-a-framework-for-data-analysis. Accessed May 16, 2021
8 Hipp HM Rialon KL Nevel K Kothari AN Jardine LDA ”Back to bedside": residents' and fellows' perspectives on finding meaning in work J Grad Med Educ 9 2 2017 269 273 28439376
9 The Physicians Foundation 2020 Survey of America's Physicians, Part 2 of 3: COVID-19's Impact on Physician Wellbeing The Physicians Foundation 2020
10 Horowitz CR Suchman AL Branch WT Frankel RM What Do Doctors Find Meaningful About Their Work? Ann Intern Med 2003 772 777 12729445
11 Shantall T The Life-changing Impact of Viktor Frankl's Logotherapy 2020 Springer Nature Switzerland AG Cham, Switzerland
12 Starck PL. Theory of Meaning. In: Smith MJ, Liehr PR, eds. Middle Range Theory for Nursing. New York: Springer Publishing; 2014:87–108.
13 Wright AA Katz IT Beyond Burnout - Redesigning Care to Restore Meaning and Sanity for Physicians N Engl J Med 2018 309 311 29365301
| 33989605 | PMC9746258 | NO-CC CODE | 2022-12-15 23:21:56 | no | Am J Med. 2021 Sep 11; 134(9):1065-1067 | utf-8 | Am J Med | 2,021 | 10.1016/j.amjmed.2021.04.020 | oa_other |
==== Front
Am J Surg
Am J Surg
American Journal of Surgery
0002-9610
1879-1883
Elsevier Inc.
S0002-9610(21)00067-2
10.1016/j.amjsurg.2021.01.040
My Thoughts / My Surgical Practice
The development of a virtual pilot for the American Board of Surgery Certifying examination
Jones Andrew T. ∗
Barry Carol L.
Ibáñez Beatriz
LaPlante Michelle
Buyske Jo
American Board of Surgery, United States
∗ Corresponding author.
3 2 2021
4 2021
3 2 2021
221 4 764767
19 1 2021
29 1 2021
30 1 2021
© 2021 Elsevier Inc. All rights reserved.
2021
Elsevier Inc.
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
==== Body
pmcThe COVID-19 pandemic has had substantial effects on the ability of testing organizations to deliver assessments. Organizations like the College Board, for example, have cancelled delivery of the SATs.1 Many medical certification organizations have also cancelled or delayed exams and are still in the process of developing contingency plans on how to safely test candidates during the pandemic.2 , 3 In the field of surgery, the pandemic has disrupted the normal processes of assessment and subsequent board certification. Historically, the final step in the certification process for general surgeons is to take and pass the oral certifying examination (CE), traditionally administered in person several times a year at a hotel in Philadelphia. Under normal circumstances, around 370 candidates, examiners, and administrators would convene in Philadelphia for each exam, with candidates moving from room to room and being examined by pairs of examiners over 90 minutes.
As the pandemic began to spread across the country, the ABS determined that having surgeons travel long distances and gather in one location would directly endanger the surgeons who were involved, while also greatly increasing the risk of the spread of the virus to the broader public. The ABS acted quickly to cancel its scheduled April in-person oral examination. As the pandemic evolved, the ABS made sequential real-time decisions to cancel the scheduled May in-person oral examination for Vascular Surgery as well as the June examination for general surgery. It became apparent that the likelihood of large gatherings reliably occurring anytime in the near future was quite low. The ABS made the decision to pilot virtual administrations of the CE.
Principles
The ABS began the development of the virtual CE pilot by espousing a set of principles. The first principle was that the ABS would not require candidates, examiners, or staff to travel, as that might put them at risk. Individuals were only expected to test in a location that would minimize both the risk of their exposure to COVID-19 as well as the risk of them spreading the virus. This meant that the ABS would allow candidates to test from home or in an office setting, as long as it was a quiet environment without the presence of other individuals.
The next principle was that the ABS wanted to maximize the integrity of the exam delivery. The ABS serves the public by only certifying individuals who have met its defined standard for general surgery or surgical specialties. Having someone take the test in one’s place, sharing exam content to aid others, or otherwise cheating before, during, or after the exam threaten the integrity of these certification decisions and therefore the public trust in the value of certification. It is critical that the individual who completes the exam is who they purport to be, that they are not able to gain assistance from someone else while testing, and that the risk of obtaining exam material for future unauthorized distribution is minimized. In the traditional hotel, in-person setting, the ABS has more control over the environment and can better ensure the integrity of the exam and the certification decisions made thereafter. In the virtual setting, these factors become harder to control, and the ABS therefore needed to take as many precautions as possible to help mitigate the risks of a security breach.
The final major principle was that the ABS needed to have some tolerance for risk, understanding that not all exams would be successfully administered. The ABS realized that by switching to a virtual format there would be technical issues that were outside the control of the organization (e.g., power failures) that could interrupt the exam. A “definitions of success” document was generated and included successful delivery of 90% of the scheduled exams as a metric. This was based on the oft quoted but unscientific assertion of a 15% industry wide failure rate for virtual exams.
Design structure
With these principles in mind, the ABS began to design the first virtual pilot administration of the exam. The first decision was to design the overall structure of the exam. The ABS determined that the virtual administration should be highly similar to the in-person format, with candidates being examined by three pairs of examiners within a similar period of time (the in-person format normally has appointments that last 1.5 hours). The next decision was to determine how many candidates to examine for the first virtual administration. Since this was the first administration of its kind, the ABS decided to start with a small group of 18 candidates, consistent with the surgical saw “small holes small problems big holes big problems”. One candidate did not test and the actual number of candidates was 17. Table 1 shows the planned schedule for the delivery on the first day of the pilot.Table 1 Day 1 schedule of exam deliveries (All times eastern).
Table 1Day Session Time Examiner Group 1 Examiner Group 2
Team 1a Team 1b Team 1c Team 2a Team 2b Team 2c
Candidate Check-In (7:30 a.m.)
1 1 (8AM) 8:00–8:30 a.m. Cand1 Cand2 Cand3 Cand4 Cand5 Cand6
10-Minute Transition/Break (or testing time for late start)
8:40–9:10 a.m. Cand3 Cand1 Cand2 Cand6 Cand4 Cand5
10-Minute Transition/Break (or testing time for late start)
9:20–9:50 a.m. Cand2 Cand3 Cand1 Cand5 Cand6 Cand4
Candidate Check-In (10:30 a.m.)
1 2 (11AM) 11:00–11:30 a.m. Cand7 Cand8 Cand9 Cand10 Cand11 Cand12
10-Minute Transition/Break (or testing time for late start)
11:40–12:10 a.m. Cand9 Cand7 Cand8 Cand12 Cand10 Cand11
10-Minute Transition/Break (or testing time for late start)
12:20–12:50 p.m. Cand8 Cand9 Cand7 Cand11 Cand12 Cand10
Debrief with Examiners (1:30 p.m.)
Abbreviations: Cand = Candidate.
This schedule differed from in-person administration in that candidates needed to arrive 30 minutes prior to the start of their exam for an individual check-in process with ABS staff and in that there were 10-min breaks for transitioning between pairs of examiners. Normally, candidates would check in as a large group at the hotel before starting their exam process. Furthermore, the usual situation was that candidates would have 2 min to transition between hotel rooms. The extra 10 minutes of break time allowed for appointments to run long in case there were any interruptions to the virtual delivery during the exam. This helped to minimize the risk of small technology issues leading to an invalid exam result. Another key difference was that the ABS minimized the number of sessions in a day. Normally, the in-person exam would have four 1.5-h exam blocks per day. In the virtual pilot format, the ABS reduced this to scale down the scope of the pilot delivery.
Delivery platform/process
The ABS currently uses the Google product G-Suite for its email and other IT infrastructure needs. Given the short time frame between the cancellation of the in-person exams and the need to deliver a virtual pilot, the ABS decided to use a tool that would be familiar to the staff for conducting the virtual administrations (i.e., Google Meet). This required either setting up examiner-centered appointments (as the traditional in-person exams) or candidate-centered appointments, where either the candidate or the examiner would have to move in and out of appointments to adhere to the schedule. The ABS decided to make the appointments candidate-centered (one appointment per candidate) and have the examiners rotate between the candidate virtual rooms. This process allowed the ABS to have a proctor that continuously monitored the candidate during the exam to provide support and minimize any security risks. Fig. 1 illustrates the inversion of the normal in-person flow on exam day to the flow for the virtual exam.Fig. 1 Comparison of in-person versus online candidate/examiner flow.
Fig. 1
In addition to using Google Meet for virtual appointments, the ABS had to develop a secure method for sharing exam content with examiners. At an in-person exam, examiners have paper copies of exam books that are tracked and returned to the ABS at the end of each administration. To ensure content integrity in a virtual format, the ABS shared electronic versions of the exams with examiners in a way that would prevent printing and downloading, and that would ensure that the content was no longer accessible after the exam. The format of the content was re-arranged so that it would display appropriately in an electronic format.
When exams are administered in the in-person format, scores are documented on paper and scanned for reporting. These scores also go through a rigorous quality control process to ensure that they are accurately recorded. This process does not translate well to an online delivery and needed to be modified. The ABS therefore developed an online score collection form which also had quality control measures in place for accurate score collection.
In sum, each of these pieces (the virtual meeting room, the exam content delivery, and the scoring process) were each critical components. They were independent of each other and needed to be integrated into an intuitive process for examiners and candidates to follow on exam day. As all of these steps were new to the examiners, candidates, and proctors, it was important that the ABS oriented them to how the technology would function on exam day. Prior to the exam, the ABS conducted systems checks for all candidates and examiners to ensure that they could access and use Google Meet on exam day. All examiners participated in training sessions prior to the administration that explained each part of the process and allowed the examiners to see how the technology would function. Finally, all proctors were also trained on their roles for ensuring the integrity of the exam and helping to troubleshoot any technology issues.
As noted above, the ABS was prepared for some degree of technical failure during the exam administration. However, the ABS also took multiple measures to mitigate the impact of technical issues on exam delivery. For example, there were backup proctors who were available in the event that the primary proctor had a connectivity issue. Furthermore, the ABS had backup examiners/observers who were available in the event that one of the primary examiners had a technical failure. These backup examiners could step in and deliver questions to the candidate if needed. Additionally, ABS staff utilized a group messaging chat to communicate internally and texts to communicate rapidly with examiners. IT staff was also live monitoring connections to try to pre-emptively identify examiners or candidates who might be likely to lose their connections. Finally, all sessions were recorded so that if a primary examiner had a technical failure, they could go back and review anything they missed and still provide scores for a candidate. These recordings were destroyed shortly after exam delivery.
Security measures
Given that ensuring the integrity of the exam was of utmost importance, the ABS took numerous precautions for security. First, as noted previously, each candidate was assigned a proctor for the duration of their exam. This proctor was responsible for the candidate check-in process, monitoring the candidate for the duration of the exam, and for helping to troubleshoot any technical issues that arose. During the check-in process, the proctor verified that the candidate had a state-issued photo identification that matched the individual on camera and was the individual who was scheduled to be examined at that time. The proctor had the candidate complete additional security measures, which included taking a room scan with a camera, emptying their pockets, and observing the candidate turning off their mobile phone. Prior to starting the exam, the candidate shared their desktop view with the proctor and opened up their task manager so that any additional programs besides the Google Meet virtual exam room were closed. Additionally, one of the security measures that Google Meet allowed was that anyone outside of the organization (e.g., candidates, examiners) would have to be admitted into the appointment by the proctor. This minimized the risk of someone hacking into or interrupting the exam.
Outcomes
Prior to the administration of the exam, the ABS defined several criteria for a successful pilot. These included but were not limited to all candidates, proctors, and examiners arriving at the correct virtual appointments, candidates completing their exam (with an acknowledgment that not all candidates would necessarily be able to), Google Meet functioning well with few connectivity drops, scores being accurately recorded, and no known security breaches occurring during the exam. Minor connectivity issues did occur (e.g., brief audio or video glitches) and one examiner did lose connectivity for several minutes. When this occurred, the backup examiner delivered the case and the examiner who lost connectivity reviewed a recording of the session later in the day for scoring purposes. The ABS was able to meet all criteria, and every candidate completed their exam. During an in-person administration, candidates are allowed to request to invalidate their CE results immediately after testing if they feel that any incidents occurred (e.g., someone knocked on the hotel room door) which caused a distraction or resulted in an unfair exam. The ABS applied this same rule to the virtual delivery; none of the candidates felt that any issues occurred which would invalidate the results of their exam. Fig. 2, Fig. 3 show some of the results from the candidate post-exam survey. As the results show, the vast majority of candidates thought that the technology functioned well and were satisfied overall with their exam experience. Cost savings to candidates were not measurable but were significant by avoiding the costs of flights, hotels, transportation, meals, childcare, and valuable time away from practice. Costs to the ABS were substantially similar to those of in-person exams, with expenses shifted from airfare and hotel rooms to having to pay for a substantial increase in staff resources to deliver the exams, technology platforms, and proctors. Last, cost savings and some revenue preservation for examiners can be assumed, primarily due to the saved time of travel. For in-person exams examiners travelled the day before the exams as well as immediately after the end of the exam, with a minimum of 4 days away from work and home.Fig. 2 Candidate perceptions of the technology during the exam.
Fig. 2
Fig. 3 Candidate overall satisfaction with delivery of the exam.
Fig. 3
Lessons learned
After the first smaller pilot with 17 candidates, the ABS repeated the process with a larger group of 54 candidates. Both of these pilots were successful, and all candidates completed their exams with minimal issues. These pilots were the first examples of remote administrations of oral exams from a member board of the American Board of Medical Specialties (ABMS). While the ABS has viewed these pilots as largely successful, it also had some lessons learned that informed plans to scale up to a virtual solution to examine over 2000 candidates between October, 2020 and June, 2021. For example, while each major component of the solution was functional (Google Meet, exam content delivery, and score collection), it is clear that it was not optimal. Examiners had to switch between programs to complete each part of the process, for example.
A better solution would integrate all, or at least some, of these components into a more streamlined application. Additionally, logistics of an all-volunteer surgeon examiner pool mean that the ABS is not able to have a third backup examiner/observer for the scale of exams that need to be delivered. Subsequent exam deliveries have two examiners. If one of the examiners has a technology issue, the other can continue to deliver the exam. The examiner with the connectivity issue is able to review the recording of the exam after the administration and provide scores for the candidate. Moreover, the ABS depended upon staff across the entire organization to act as proctors for this administration. The necessary and routine work of the ABS was still required, which, given lean ABS staffing, meant that staff proctoring is not scalable for subsequent large-scale deliveries. The ABS therefore recruited and trained outside proctors to help deliver subsequent exams. Moreover, the ABS shifted the schedule of the delivery to allow for better scheduling for West Coast examiners by starting later in the day. Finally, virtual oral exam results will be analyzed to evaluate the comparability of the scores from the virtual administrations to the scores from the in-person exams to ensure that candidates are still meeting the equivalent standard expected of a board-certified surgeon.
Looking to the future
The switch to a virtual delivery model may enable innovations above and beyond the current changes that may help to maximize objectivity in scoring. For example, the ABS could potentially implement image blurring to mask the identity of the candidates, which may have potential ramifications for bias. Additionally, voice baffling may be an additional layer of de-identification that can minimize potential issues with examiner bias. The ABS may also be able to add more raters to the scoring process by having one set of examiners deliver the exam and a completely different set score the exams using recordings or transcripts of the exams, so that they are completely blinded to the examiners and candidates. Furthermore, some combination of transcripts and natural language processing may help to further enhance the scoring of the exam. Examiner training for new examiners should become easier, with novices having the ability to participate in live training without having to travel to an event. All of these concepts would be significant changes to the exam and would require evaluation before being implemented in an actual exam setting. However, they have the potential to enhance the assessment of surgical judgment and improve fairness in making decisions for board certification.
==== Refs
References
1 Anderson N. College Board cancels June SAT tests and floats an ‘unlikely’ scenario: College admission exams at home The Washington Post website www.washingtonpost.com/education/2020/04/15/college-board-cancels-june-sat-tests-floats-an-unlikely-scenario-college-admission-exams-home/ Published April 15, 2020. Accessed July 17, 2020
2 Coronavirus updates (COVID-19) American board of internal medicine website www.abim.org/media-center/Coronavirus-Updates.aspx
3 Stempniak M. American Board of Radiology canceling exams in response to coronavirus outbreak Radiology Business website www.radiologybusiness.com/topics/leadership/american-board-radiology-canceling-exams-response-coronavirus-outbreak Published March 6, 2020. Accessed July 17, 2020
| 33563463 | PMC9746259 | NO-CC CODE | 2022-12-15 23:21:56 | no | Am J Surg. 2021 Apr 3; 221(4):764-767 | utf-8 | Am J Surg | 2,021 | 10.1016/j.amjsurg.2021.01.040 | oa_other |
==== Front
Psychother Forum
Psychotherapie Forum
0943-1950
1613-7604
Springer Vienna Vienna
218
10.1007/s00729-022-00218-4
Editorial
Psychotherapie und Gesellschaft/en
Schigl Brigitte
Lerch Leonore [email protected]
psychotherapie forum, Österreichischer Bundesverband für Psychotherapie (ÖBVP), Löwengasse 3/3/4, 1030 Wien, Österreich
13 12 2022
12
23 11 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Austria, part of Springer Nature 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
==== Body
pmcPsychotherapie findet nicht in einem neutralen, werte- und herrschaftsfreien Raum statt, sondern bildet gesellschaftliche Strukturen und Prozesse ab. Sie ist Teil gesellschaftlichen Handelns. Die Flucht- und Migrationsbewegungen als auch die Corona-Pandemie – sowie die Politiken im Umgang mit diesen Phänomenen – verstärkten in den letzten Jahren das Auseinanderdriften von gesellschaftlichen Bewegungen bzw. die Polarisierung und Radikalisierung von Gesellschaften.
Als wir dieses Heft planten, war die Covid Krise auf einem Höhepunkt, der Krieg in der Ukraine noch nicht vorhersehbar, die Energiepreise niedrig, die Inflation im erwarteten Rahmen. Die derzeitige Kumulation von Krisen und die immer deutlichere Infragestellung des bisherigen Lebensstils, vor allem in westlichen Gesellschaften – Stichwort Global Warming –, hat inzwischen große Teile der Gesellschaften erreicht; die prekärsten und vulnerabelsten Gruppen merken dies wie immer am stärksten.
Und Psychotherapie? Genügt sie sich darin, eine Reparaturmaßnahme für Burnout Betroffene zu sein? Oder ihren individuumzentrierten Blick auf die Effektstärken in der Verringerung klinischer Symptome einzelner Patient:innen zu richten und auszublenden, dass diese nach ihrer Entlassung aus dem stationären Setting, in dem die Psychotherapie untersucht wurde, in ein möglicherweise krank machendes Umfeld zurückkommen?
Mit diesem Heft möchten wir die psychotherapeutische Praxis als eine gesellschaftliche Praxis beleuchten: (Wie) Beeinflussen Gesellschaften die Psychotherapie, unsere Rolle als Psychotherapeut*innen sowie die psychotherapeutische Praxis? (Wie) Beeinflusst Psychotherapie die Gesellschaft/en? Dies war auch die Fragestellung in unserem Call for Papers für diese Doppelausgabe.
Es trafen sehr unterschiedliche und spannende Arbeiten ein. Auffallend war für uns als Herausgeberinnen, dass zahlreiche Autor:innen die maximale Artikellänge von 28.000 Zeichen überschritten. Es war schwierig für viele, ihre Texte auf das entsprechende Format auszurichten. Das Thema sprengt offenbar die Grenzen und soviel muss erklärt, soviel will gesagt werden …
Wir freuen uns, Ihnen eine breite Palette ganz unterschiedlicher Blickwinkel auf das Thema nahebringen zu können.
Den Auftakt bilden der Artikel von Johanna Muckenhuber: Die Gesellschaft und die Couch. Auseinandersetzungen mit individuellen und sozialen Bedingungen für eine emanzipatorische Psychotherapie, der Beitrag von Silke Birgitta Gahleitner, Karsten Giertz, Cornelia Caspari, Peter Caspari, Heiner Keupp: Der Preis der Psychotherapie – Argumente für eine Wiederbelebung der sozialen Perspektive im psychotherapeutischen Denken und Handeln sowie der Artikel von Martin Luger & Elisabeth Fehrmann: Psychotherapie als Gesellschaftspraxis. Integrativ-therapeutische und systemische Beiträge zu einer kontextsensitiven Psychotherapie, die sich alle mit gesellschaftlichen und sozialen Bedingungen von Psychotherapie auseinandersetzen.
Die Beiträge von Angelika Grubner: Mutter und Psychotherapie und von Sascha Schipflinger: Drohender Kontrollverlust: Überlegungen zu autoritären Anpassungsprozessen im gesellschaftlichen, klinischen und institutionellen Kontext beleuchten die Etablierung der Psychotherapie aus einer feministischen Perspektive sowie Fragestellungen zu Macht und Kontrolle.
Insbesondere mit der psychotherapeutischen Versorgung beschäftigen sich Henriette Löffler-Stastka & Gabriele Rieß im Artikel: VersorgungsNOT – Psychotherapie als zentrale, aber marginalisierte Versorgungsleistung im Gesundheitssystem sowie Brigitte Fiala-Baumann, Helga Ploner, Dominik Witzmann und Andrea Jesser im Beitrag: Säuglings‑, Kinder- und Jugendlichen- (SKJ) Psychotherapien während der Covid-19 Pandemie: Ergebnisse einer Studie unter psychodynamischen Psychotherapeut*innen in Österreich.
Zwei Autor:innen setzen sich mit der Psychotherapie von marginalisierten Gruppen auseinander, nämlich Sabine Tiefenthaler in ihrem Artikel: Intersektionale Diskriminierung: Erfahrungen und Perspektiven in der Psychotherapie mit Frauen mit Fluchtbiografien sowie Olga Kostoula in ihrem Beitrag: Psychotherapie im Migrationskontext.
Die Autor:innen Ursula Grillmeier-Rehder, Karoline Hochreiter, Ida-Maria Kisler, Christian Korunka und Brigitte Schigl versuchen mit dem Beitrag: Gemeinsame Perspektiven Humanistischer Psychotherapien in Covid-Zeiten am Beispiel eines Blicks auf die Covid Krise eine gemeinsame Perspektive der unterschiedlichen humanistischen Verfahren herauszuarbeiten.
Abschließend diskutiert Elisabeth Pölleritzer das Buch von Luise Reddemann: Die Welt als unsicherer Ort – Psychotherapeutisches Handeln in Krisenzeiten. Nina Belinda Ute Stephan stellt die Publikation von Eben Louw & Katja Schwabe: Rassismussensible Beratung und Therapie von geflüchteten Menschen vor. Die Rezension von Anita Scheuermann bietet einen Einblick in des Buch von Bernd Rieken, Reinhold Popp, Paolo Raile (Hrsg.): Eco-Anxiety – Zukunftsangst und Klimawandel. Interdisziplinäre Zugänge.
Wir wünschen Ihnen mit diesen sehr unterschiedlichen Beiträgen eine anregende Lektüre!
Brigitte Schigl & Leonore Lerch
Verantwortliche Editorinnen in Chief der Doppelausgabe „Psychotherapie und Gesellschaft/en“
Interessenkonflikt
B. Schigl und L. Lerch geben an, dass kein Interessenkonflikt besteht.
Hinweis des Verlags
Der Verlag bleibt in Hinblick auf geografische Zuordnungen und Gebietsbezeichnungen in veröffentlichten Karten und Institutsadressen neutral.
| 0 | PMC9746552 | NO-CC CODE | 2022-12-15 00:03:54 | no | Psychother Forum. 2022 Dec 13;:1-2 | utf-8 | null | null | null | oa_other |
==== Front
J Immigr Minor Health
J Immigr Minor Health
Journal of Immigrant and Minority Health
1557-1912
1557-1920
Springer US New York
1428
10.1007/s10903-022-01428-3
Original Paper
Predictors of Alcohol Use Among Latinx Men in South Florida: Machismo as a Correlate of Alcohol Use Frequency and Quantity
http://orcid.org/0000-0002-4799-9910
Rojas Patria [email protected]
125
Wang Weize [email protected]
23
Sanchez Mariana [email protected]
12
Ravelo Gira [email protected]
25
Ángel Cano Miguel [email protected]
4
Galvez Gemma [email protected]
1
Li Tan [email protected]
3
C. Penn Alvonee [email protected]
1
F. Colon-Burgos Jose [email protected]
2
De La Rosa Mario [email protected]
25
Behar-Zusman Victoria [email protected]
5
1 grid.65456.34 0000 0001 2110 1845 Department of Health Promotion and Disease Prevention, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
2 grid.65456.34 0000 0001 2110 1845 CRUSADA, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
3 grid.65456.34 0000 0001 2110 1845 Department of Biostatistics, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
4 grid.65456.34 0000 0001 2110 1845 Department of Epidemiology, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
5 grid.26790.3a 0000 0004 1936 8606 CLaRO, University of Miami; School of Nursing and Health Studies, University of Miami, Coral Gables, FL 33124 USA
13 12 2022
17
8 11 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Previous studies have found Latinx cultural values to be positively associated with healthy behaviors. This study aims to examine socioeconomic and cultural correlates of alcohol use among Latinx adult men living in Miami-Dade County, Florida. The study sample included 122 Latinx adult men (mean age = 44, SD = 10), predominantly of South and Central American origin. Data was collected using REDCap. Interviews included the Timeline Follow-Back scale for alcohol use. Results indicate that Caribbean participants were significantly less likely to report drinking in the past 90 days (aOR = 0.08, p = 0.042) compared to their Venezuelan counterparts. Higher machismo scores were associated with low drinking frequency (aRR = 0.67, p = 0.043), while no significant associations were found between machismo and other drinking outcomes. Drinking quantity and frequency are significantly associated with higher income and authorized immigration status in the US among Latinx men in South Florida. Higher machismo scores were associated with low drinking frequency.
Keywords
Latinx men
Alcohol
Drinking frequency
Machismo
Caballerismo
Latino men
South Florida
http://dx.doi.org/10.13039/100000002 National Institutes of Health National Institutes of Health
==== Body
pmcBackground
The etiology of alcohol use among adult Latinx men has been primarily studied among Mexican American men who were born in the United States (U.S.) and elsewhere [1–3]. Castañeda [3] documented that cultural beliefs and values, such as gender norms (e.g., machismo and caballerismo), have been linked to alcohol use behaviors among Latinx men of Mexican descent. Machismo is a traditional Latinx male gender norm that encompasses hyper-masculinity, aggression, and dominance, and it has been associated with alcohol use risk factors [4]. However, within a more broadened definition, machismo includes positive aspects such as family protection, responsibility, and hard work [5]. Conversely, caballerismo focuses on emotional connectedness, familial cohesion, and social responsibility, and it has been associated as a protective health factor for outcomes such as self-esteem [4, 5]. Yet, the influence of machismo and caballerismo on the alcohol use behaviors of men of South and Central American or Cuban descent living in the U.S. has not been widely studied. The present study investigates alcohol use among a diverse sample of Latinx men living in Miami-Dade County (MDC).
Several papers published between 2018 and 2021 examined the relationship between masculinity and substance misuse; results from one study indicated that the perception and endorsement of machismo norms was associated with alcohol misuse [6]. Moreover, a literature review that examined traditional gender roles among Latinxs men highlighted that, across studies, those who closely aligned with traditional machismo were likely to have a strong association with alcohol use [7]. Young men are more likely to endorse machismo [4] along with choosing to drink more because drinking is perceived as normal masculine behavior; but older Latinx men may refrain from hazardous drinking in order to maintain dignity and family responsibilities [5]. Moreover, alcohol misuse has been associated with a lower level of empathy and lack of pro-social behaviors [8]. Despite receiving alcohol detox treatments, men with problematic alcohol use report lower pro-social behaviors than men who do not use alcohol [8]. Lower empathy and moral compass were consistently low even after men who drank a moderate amount of alcohol were compared with a control group [9]. More recent studies using a multidimensional definition of machismo have included positive social behavioral aspects of machismo such as care and protection of the family, respect, dignity and hard work. In those studies machismo has not been significantly associated with alcohol use [5, 10, 11]. However, Perrotte and Zamboanga [7] discussed caballerismo and its relation to increased well-being and increased conflict resolution, which have been found to be associated with less alcohol consumption [12]. Using the theory of planned behavior, the present study examined the association between gender norms and alcohol use among a diverse group of South Florida Latinx men.
Theoretical Framework
The Theory of Planned Behavior (TPB) has been utilized to predict human behavior particularly frequency of alcohol use. TPB components were not tested in this study, instead gender norms were predicted to be associated with drinking frequency. Thus, guided by TPB, the present study aimed to examine correlates of alcohol use frequency and quantity among Latinx adult men. Notably, in addition to their geographic distinctions, urban communities in MDC are predominantly composed of Latinx individuals of South American and Caribbean origin, whereas Latinx men in semirural areas of MDC are largely of Central American and Mexican origin. We hypothesized that machismo would be associated with higher alcohol use frequency and quantity for all participants.
Methods
The present study is a secondary analysis of a National Institutes of Health (NIH) funded community based clinical study that investigated the effectiveness of an HIV prevention program targeting Latinx fathers and sons in MDC. Participants were recruited using conventional community outreach activities such as placing printed fliers in community organizations, participating in community meetings, social media advertisements, and word-of-mouth. Using only baseline data collected from the fathers, the study sample included 122 male participants aged 18–66 (Mean = 44, SD = 10). Most participants were born in Venezuela (25%), and the second largest group were from other Mexico (22%) (See Table 1). The eligibility criteria for fathers included: (a) being 18 years or older, (b) being the father, or father figure, of an adolescent between the ages of 11–17, (c) living or working in MDC, (d) self-identifying as Latinx, (e) and consenting to participate in one of two randomly assigned groups (i.e., intervention or control groups). Data were collected in Spanish using REDCap survey software, and phone interviews were facilitated by bilingual and bicultural trained interviewers. Measures that were not already available in Spanish were translated into Spanish via translation/back translation methods using a well-established translation protocol with institutional review board approval. Interviewers received a two-day training on the protocol administration and received Collaborative Institutional Training Initiative (CITI) human subjects research certifications. 92% (92%) of the interviews were completed via phone due to the Severe Acute Respiratory Syndrome Corona Virus-2019 (SARS COVID-19) pandemic. This study was approved by the institutional review board of a large private university in Miami, Florida.
Table 1 Participants’ baseline characteristics (N=122)
Variable Mean (SD.) Median Range
Age in years 43.5 (9.8) 44.0 [18, 66]
Depressive symptoms 4.4 (3.6) 4.0 [0, 20]
Anxiety 2.9 (3.2) 2.0 [0, 16]
Years living in the U.S. 14.4 (11.6) 15.0 [1, 47]
Machismo 2.9 (1) 2.9 [1, 5.9]
Caballerismo 6 (0.7) 6.1 [3.3, 7]
Drinking frequency 7.7 (13.7) 3.5 [0, 90]
Drinking quantity 2.6 (2.8) 2.1 [0, 14]
Variable n %
Reported drinking
Yes 77 63.1
No 45 36.9
Binge drinking
Yes 44 36.1
No 78 63.9
Living area
Semi-rural 69 56.6
Urban 53 43.4
Country of origin
Caribbean 11 9.0
Central America 16 13.1
Mexico 27 22.1
Other South American Countries 22 18.0
U.S. 15 12.3
Venezuela 31 25.4
Household income in the last month
0-$999 19 15.8
$1000-$1999 41 34.2
$2000 or more 60 50.0
Education
Less than high school 33 27.1
High school or GED 24 19.7
Some college 29 23.8
College/university degree 36 29.5
Marital status
In a domestic relationship 14 11.5
Married 94 77.1
Single or separated 14 11.5
Employment status
Employed 99 81.2
Unemployed 23 18.9
Immigration status
Authorized 97 82.2
Unauthorized 21 17.8
Note.SD.=Standard deviation
Table 2 Estimates for drinking outcomes among adult Latin x men living in urban and semirural areas of Miami-Dade County, Florida
Predictors Drinking Status Drinking Frequency Drinking Quantity Binge Drinking
aORa [95% CI] aIRRb [95% CI] Estimatec [95% CI] aORa [95% CI]
Age in years 1.03 [0.96, 1.1] 1.02 [0.97, 1.07] 0.99 [0.97, 1.02] 0.99 [0.94, 1.05]
Years living in the U.S. 0.96 [0.91, 1.02] 0.97 [0.94, 1.01] 0.99 [0.96, 1.01] 0.99 [0.94, 1.05]
Machismo -- 0.67 [0.46, 0.99]* -- --
Household income
$1000-$1999 vs. 0-$999 0.06 [0.01, 0.37]** 0.31 [0.11, 0.89]* 0.49 [0.27, 0.9]* 0.49 [0.11, 2.12]
$2000 or more vs. 0-$999 0.24 [0.04, 1.35] 0.68 [0.24, 1.94] 0.81 [0.45, 1.45] 0.79 [0.18, 3.55]
Marital status
In a domestic relationship vs. Married 3.26 [0.51, 21.04] 1.89 [0.55, 6.51] 3.42 [1.85, 6.31]*** 6.18 [1.19, 32.01]
Single or separated vs. Married 0.55 [0.06, 4.77] 0.44 [0.12, 1.55] 1.77 [0.88, 3.58] 2.3 [0.34, 15.45]
Country of origin
Caribbean vs. Venezuela 0.08 [0.01, 0.91]* 0.54 [0.15, 1.96] 0.36 [0.14, 0.92]* 0.7 [0.09, 5.16]
Central America vs. Venezuela 0.23 [0.03, 2.02] 0.73 [0.21, 2.54] 1.32 [0.69, 2.52] 1.1 [0.18, 6.91]
Mexico vs. Venezuela 0.51 [0.04, 6.81] 2.51 [0.59, 10.67] 0.84 [0.38, 1.87] 0.86 [0.09, 7.96]
Other South American countries vs. Venezuela 0.79 [0.09, 6.92] 1.11 [0.36, 3.44] 0.96 [0.5, 1.85] 1.32 [0.25, 6.93]
U.S. vs. Venezuela 0.16 [0, 6.13] 0.41 [0.04, 4.43] 0.55 [0.03, 9.44] 0.55 [0.01, 27.75]
Employment status: Employed vs. Unemployed 6.57 [1.43, 30.07]* 2.8 [1.04, 7.55]* 1.85 [0.96, 3.58] 5.71 [1.19, 27.33]*
Education
High school or GED vs. Less than high school 0.58 [0.12, 2.68] 0.43 [0.15, 1.26] 1.09 [0.59, 2] 0.59 [0.14, 2.57]
Some college vs. Less than high school 0.21 [0.04, 1.19] 0.14 [0.05, 0.43]*** 0.66 [0.32, 1.35] 0.34 [0.07, 1.67]
College/university degree vs. Less than high school 0.27 [0.04, 1.89] 0.13 [0.04, 0.37]*** 0.97 [0.47, 2.02] 1.29 [0.24, 6.8]
Immigration status: Authorized vs. Unauthorized 2.97 [0.55, 16.03] 3.16 [1.12, 8.9]* 3.2 [1.32, 7.8]* 1.9 [0.39, 9.36]
Living area: Urban vs. Semi-rural 3.62 [0.7, 18.74] 2.47 [0.95, 6.37] 1.12 [0.57, 2.17] 1.04 [0.25, 4.32]
Depressive symptoms 1.15 [0.98, 1.34] -- 1.02 [0.96, 1.08] --
a. Estimates are adjusted odds ratios (aORs) from multiple logistic regression
b. Estimates are adjusted incident risk ratios (aIRRs) from negative binomial regression
c. Estimates are exponential correlation coefficients from log-normal regression
*p<0.05; **p<0.01; ***p<0.001
Measures
Demographic Variables
The following demographics were included: age, number of years living in the U.S., country of origin, educational status, marital status, immigration status, the level of urbanization of living area (urban vs. semirural), employment, and household income. Participants were asked to report their educational status, choosing from five categories ranging from “less than high school” to “college/graduate professional studies,” which we then recoded into three categories. Similarly, participants were asked to report their immigration status from 11 categories, ranging from “citizen” to “temporarily protected asylum seeker.” These categories were then recoded into a dichotomous variable of (1) authorized status or (2) unauthorized status. For the present study, marital status categories were recoded into (1) in a domestic relationship, (2) being married, (3) being single or separated, because most participants were married or had a romantic partner. Age was collected and analysed as a continuous variable, but household income in past month was recoded into three categories (0-$999; $1,000-$1,999; $2,000 or more).
Gender Norms
Traditional gender norms were measured using the Machismo and Caballerismo Scale, a widely recognized and validated 20-item bi-dimensional scale that measures machismo and caballerismo on two separate subscales [4]. Sample items for the machismo subscale include: ‘it’s important not to be the weakest man in a group’ and ‘real men never let their guard down.’ Items on the caballerismo subscale include ‘men should be affectionate to their children’ and ‘family is more important than the individual’ [7]. Items were measured using a 7-point Likert-type scale ranging from 1 (not at all) to 7 (very much so). Subscale scores were measured by calculating mean values, with higher values indicating greater adherence to traditional gender norms [4]. Good reliability was reported with a Cronbach’s alpha of 0.80 overall. Cronbach’s alphas in the present study were 0.85 in both the machismo and caballerismo subscales.
Alcohol use was measured using the Timeline Follow-Back method [13]. Participants reported their alcohol use frequency and quantity in the last 90 days. To assist participants in remembering their drinking behaviors, special dates such as holidays were used as anchor points. Initially, participants indicated their drinking behavior during the last 30 days; they then subsequently indicated their drinking behavior during the 60 days preceding the initial 30-day period. Quantity of alcohol was measured using standard serving sizes of drinks, such as a 12 oz beer bottle or can, or a 1 oz. hard liquor shot. Participants reported the number of drinks consumed per day. The reliability of this scale has been validated with similar populations. A dichotomous variable of alcohol consumption (yes/no) was created based on reported drinking in the past 90 days. Alcohol use frequency was the total number of days alcohol was consumed, and alcohol use quantity was the average number of standard drinks consumed in the past 90 days. Binge drinking was recoded into a dichotomous variable (yes/no) that indicated whether a participant had five or more drinks on any occasion.
Mental Health
Depressive symptoms were measured using the Patient Health Questionnaire-9 (PHQ-9), which is an adaptation of the Patient Health Questionnaire (PHQ-59 items). The tool has shown criteria validity [14]. The questionnaire has shown good diagnostic validity with comparable sensitivity and specificity for major depression symptoms among adults [15]. We used the Generalized Anxiety Disorder scale (GAD-7), which is a 7-item measure (4-point Likert-type scale) that assesses anxiety among participants in the past 2 weeks [16]. The Cronbach’s alpha for the PHQ-9 and GAD-7 scales was 0.79 and 0.85 respectively in this study.
Statistical Analysis
Descriptive statistics of participant demographic and alcohol use characteristics are presented in Table 1. The sample mean, standard deviation, median, minimum, and maximum were used to describe the continuous variables. Counts and percentages were used to describe categorical variables. We used logistic regression for alcohol consumption and binge drinking in the past 90 days. We performed negative binomial regression to predict drinking frequency due to its right-skewed count measure. We applied generalized linear model with a log-link function for predicting drinking quantity. To examine the association between the independent variables (machismo, caballerismo, and mental health) and the drinking outcomes, bivariate regression models for each drinking outcome were performed. Independent variables with a p-value at 0.1 or less were included in the multiple regression models, controlling for demographics, including participants’ age, years living in the U.S., household income, marital status, country of origin, employment status, education, immigration status, and living area. The Statistical Analysis System (SAS) 9.4 was used for all data analyses [17]. A statistical significance level of 0.05 was used to reject the null.
Results
Sociodemographic Characteristics
The present analysis included data from 122 Latinx men with a mean age of 44 years (SD = 10). Most men were authorized immigrants (82%), employed (81%), and married (77%) at the time of assessment. Only 12% of the participants were U.S.-born, while other participants immigrated from South America (43%), Mexico (22%), Central America (13%), and the Caribbean (9%). More than half of the participants (57%) were from semirural areas. Half of the participants (50%) had a household income of $2,000 or more in the last month. The mean number of years living in the U.S. was 14 (SD = 12).
Alcohol Use
63% (n = 77) of the participants reported alcohol consumption in the past 90 days. Machismo, caballerismo, or anxiety were not significantly associated with alcohol consumption in the past 90 days, according to results from the bivariate analysis. Thus, the scale scores were excluded from the multiple logistic regression for predicting drinking in the past 90 days. Results show that, compared to participants with a household income of 0-$999 in the past month, those with the household income in the $1,000-$1,999 range were 94% less likely to report drinking in the past 90 days (aOR = 0.06, p = 0.002), controlling for other variables in the model. However, no significant difference in alcohol consumption was found between people with a household income of $2,000 or above and 0-$999 (p > 0.05). Participants who were from the Caribbean were significantly less likely to report drinking in the past 90 days (aORs = 0.08, p = 0.042) compared to participants from Venezuela. Employed participants were more likely to report drinking (aOR = 6.57, p = 0.015) than unemployed participants in the past 90 days.
Drinking Frequency
The median drinking frequency in the past 90 days was 3.5 days. Caballerismo, depressive symptoms, and anxiety were excluded from the final negative binomial regression model because of the non-significant association with drinking frequency from the bivariate analysis. Results from the multiple regression model suggest that machismo was significantly negatively associated with drinking frequency, after controlling for other covariates in the model. A higher machismo score was associated with lower drinking frequency (aRR = 0.67, p = 0.043). Having a household income of $1000-$2000 was significantly associated with lower drinking frequency, compared to the household income of 0-$999 (aRR = 0.31, p = 0.030). Employed participants had more drinking frequency in the past 90 days (aRR = 2.8, p = 0.042) than unemployed participants. Compared to the participants with an educational level of less than high school, those with some college (aRR = 0.14, p < 0.001) or a college/university degree (aRR = 0.13, p < 0.001) reported drinking less frequently. Authorized immigrants were likely to have a higher number of drinking days than those who were unauthorized (aRR = 3.16, p = 0.030).
Drinking Quantity
Among participants, the mean number of drinks per day in which alcohol was consumed in the past 90 days was 2.6 (SD = 2.8). Results from the bivariate analysis show no significant correlation between drinking quantity and machismo, caballerismo, or anxiety. Hence, these variables were excluded from the final regression model. Results show that having a household income of $1000-$1999 was associated with decreased drinking quantity (β = 0.49, p = 0.020), compared to those with household income of 0-$999. Compared to married men, those who were in a domestic relationship reported higher drinking quantities (β = 3.42, p < 0.001). Compared to participants whose country of origin was Venezuela, Caribbean origin was associated with a significantly lower drinking quantity (β = 0.36, p = 0.032). Authorized immigrants reported greater drinking quantity (β = 3.16, p = 0.010). Age, years in the U.S., education, employment, participants’ living area, or depressive symptoms were not significantly associated with drinking quantity (p > 0.05).
Binge Drinking
About 36% of the participants reported binge drinking in the past 90 days ( > = 5 drinks per day in a single sitting). Machismo, caballerismo, depressive symptoms, or anxiety had no statistically significant association with binge drinking in the past 90 days and were excluded from the final regression model. Results from multiple logistic regression indicate employed participants were more likely to binge drink in the past 90 days (aOR = 5.71, p = 0.029). Results show that binge drinking was not significantly associated with age, years in the U.S., household income, marital status, country of origin, education, immigration status, or participants’ living area (p > 0.05).
Discussion
The present study highlights a negative correlation of machismo and drinking frequency, as men with higher scores of machismo reported drinking less frequently. Contrary to the literature on machismo that has characterized men with higher levels of machismo with higher levels of risk taking and substance abuse behavior [11, 18], the current study findings align with a wider perspective found in more recent ones [4, 19]. According to the more recent perspective of masculinity norms, there is a multidimensional aspect of machismo that encompasses attributes such as chivalry, respect, and honor [4, 12]. Studies on alcohol misuse also support the findings that there are higher levels of empathy and higher sense of morality among non-alcohol misusers [8, 9]. This result aligns with more positive aspects of machismo that have been associated with caballerismo [7] and that need to be further examined in future studies.
Similar to other studies [6], authorized immigrants were more likely to report drinking higher quantities and more frequently. There may be several factors associated with their frequency such as having a job, higher income, social stability, and less immigration stress (e.g., fear of deportation).
The present study represents a small example of the heterogeneity among Latinx communities. Our analysis demonstrates that future interventions may want to target low education and documented Latinx individuals. Alcohol frequency and quantity was more prevalent among households reporting less income; however, participants who were employed and lived in urban areas were more likely to report using alcohol in the last 90 days. However, the alcohol use prevalence identified in the present analysis needs to be examined further in future longitudinal studies that show potential differences among different Latinx subgroups.
Most participants from urban areas were of South American origin, primarily Venezuela, and they were more likely to report alcohol use than participants of Central America and the Caribbean region. These findings support a previous study that investigated the alcohol use trajectory of Latinx men prior to and after immigration to the United States [20]. De La Rosa et al. [20] which revealed a decrease in alcohol use frequency for men during their pre- to post-immigration period. Future studies should focus on expanding the sample and including more representation from South America and the Caribbean.
In general, a significant percentage of participants (36%) reported binge drinking in the past 90 days, although the mean number of drinks per day was approximately 3. Although higher drinking quantities were found among men who reported being in a domestic relationship rather than legally married, most participants in the study were married; hence, we cannot draw substantial clinical implications or significant conclusions. Our results indicate the need for other mixed methods research to examine this relationship further.
Limitations
Results from this study should be viewed with the known limitations of a small non-randomized self-reported retrospective data. Additionally, the Latinx population is a heterogeneous group, and the current community-based study sample was mostly well-educated, employed South and Central American immigrants; therefore, there is an underrepresentation of other large U.S. Latinx groups. Because this was a cross-sectional study, temporal relationships could not be established. In addition, it was not possible to examine time variance, given this study’s design, since this was a secondary data analysis. However, future studies might address the mixed findings regarding machismo using other methodological approaches.
Conclusion
The results from our analysis suggest that among adult Latinx men in MDC, alcohol use quantity and frequency are significantly associated with higher income and authorized immigration status. Contrary to our hypothesis, machismo was significantly associated to lower frequency of alcohol use but not quantity or binge drinking. 36% of participants reported binge drinking. Although this is a small cross-sectional community-based study, this is a modest but significant contribution to the knowledge gap in the literature on South and Central American immigrants, particularly Venezuelan immigrants, in Florida. Future studies that include biomarkers, larger samples, and control groups would advance our findings and clarify the association among machismo, alcohol use frequency and quantity. Although previous studies have examined the negative consequences of alcohol use frequency and quantity among Latinos living in the U.S., the present study present an important subgroup distinction and posits the need to continue research on the associations of gender norms and alcohol misuse among Latinx subgroups.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Caetano R Ramisetty-Mikler S Rodriguez LA The Hispanic Americans Baseline Alcohol Survey (HABLAS): rates and predictors of DUI across Hispanic national groups Accid Anal Prev 2008 40 2 733 41 10.1016/j.aap.2007.09.010 18329428
2. Caetano R Ramisetty-Mikler S Rodriguez LA The Hispanic Americans Baseline Alcohol Survey (HABLAS): the association between birthplace, acculturation and alcohol abuse and dependence across Hispanic national groups Drug Alcohol Depend 2009 99 1–3 215 21 10.1016/j.drugalcdep.2008.08.011 18945554
3. Castañeda SF Garcia ML Lopez-Gurrola M Stoutenberg M Emory K Daviglus ML Kaplan R Giachello AL Molina KM Perreira KM Youngblood ME Alcohol use, acculturation and socioeconomic status among Hispanic/Latino men and women: The Hispanic Community Health Study/Study of Latinos PLoS ONE 2019 14 4 e0214906 10.1371/journal.pone.0214906 30947280
4. Arciniega GM Anderson TC Tovar-Blank ZG Tracey TJ Toward a fuller conception of Machismo: Development of a traditional Machismo and Caballerismo Scale J Couns Psychol 2008 55 1 19 10.1037/0022-0167.55.1.19
5. Daniel-Ulloa J Sun C Rhodes SD The intersection between masculinity and health among rural immigrant Latino men Int J Men’s Health 2017 16 1 84 29422781
6. Balagopal G, Davidson S, Gill S, Barengo N, De La Rosa M, Sanchez M. The impact of cultural stress and gender norms on alcohol use severity among Latino immigrant men. Ethnicity & Health. 2021; Feb 9:1–9.
7. Perrotte JK, Zamboanga BL. Traditional gender roles and alcohol use among Latinas/os: A review of the literature. Journal of Ethnicity in Substance Abuse. 2019; Mar 25.
8. Martinotti G, Nicola MD, Tedeschi D, Cundari S, Janiri L. Empathy ability is impaired in alcohol-dependent patients. American Journal on Addictions. 2009 Jan 1; 18(2):157–61.
9. Francis KB Gummerum M Ganis G Howard IS Terbeck S Alcohol, empathy, and morality: acute effects of alcohol consumption on affective empathy and moral decision-making Psychopharmacology 2019 236 12 3477 96 10.1007/s00213-019-05314-z 31289885
10. Perrotte JK Zamboanga BL Kearns N Linking alcohol-specific masculine norms and drinking behavior among Latino men Psychol Men Masculinities 2020 21 3 490 10.1037/men0000252
11. Mirandé A. Hombres y machos: Masculinity and Latino culture. Routledge; 2018. Feb 12.
12. Corbin WR Farmer NM Nolen-Hoekesma S Relations among stress, coping strategies, coping motives, alcohol consumption and related problems: A mediated moderation model Addict Behav 2013 38 4 1912 9 10.1016/j.addbeh.2012.12.005 23380486
13. Sobell LC, Buchan G, Cleland P, Sobell MB, Fedoroff I, Leo GI: The reliability of the Timeline Followback method applied to drug, cigarette, and cannabis use. In 30th Annual Meeting of the Association for Advancement of Behavior Therapy, New York, NY. 1996.
14. Kroenke K Spitzer RL Williams JB The PHQ-9: validity of a brief depression severity measure J Gen Intern Med 2001 16 9 606 13 10.1046/j.1525-1497.2001.016009606.x 11556941
15. Spitzer RL Williams JB Kroenke KJ Test review: patient health questionnaire–9 (PHQ-9) Rehabilitation Couns Bull 2014 57 4 246 8 10.1177/0034355213515305
16. Spitzer RL Kroenke K Williams JB Löwe B A brief measure for assessing generalized anxiety disorder: the GAD-7 Arch Intern Med 2006 166 10 1092 7 10.1001/archinte.166.10.1092 16717171
17. SAS Institute Inc SAS/ACCESS® 9.4 Interface to ADABAS: Reference 2013 Cary SAS Institute Inc
18. Neff JA Prihoda TJ Hoppe SK ” Machismo,“ self-esteem, education, and high maximum drinking among anglo, black and Mexican-American male drinkers J Stud Alcohol 1991 52 5 458 63 10.15288/jsa.1991.52.458 1943101
19. Falicov CJ Changing constructions of machismo for Latino men in therapy: “The devil never sleeps.” Fam Process 2010 49 3 309 29 10.1111/j.1545-5300.2010.01325.x 20831763
20. De La Rosa M Sanchez M Wang W Cano MA Rojas P Alcohol use trajectories of adult Latinx immigrants during their first decade in the United States Addict Behav 2020 106 106352 10.1016/j.addbeh.2020.106352 32145498
| 36512291 | PMC9746553 | NO-CC CODE | 2022-12-15 23:21:56 | no | J Immigr Minor Health. 2022 Dec 13;:1-7 | utf-8 | J Immigr Minor Health | 2,022 | 10.1007/s10903-022-01428-3 | oa_other |
==== Front
J Immigr Minor Health
J Immigr Minor Health
Journal of Immigrant and Minority Health
1557-1912
1557-1920
Springer US New York
1428
10.1007/s10903-022-01428-3
Original Paper
Predictors of Alcohol Use Among Latinx Men in South Florida: Machismo as a Correlate of Alcohol Use Frequency and Quantity
http://orcid.org/0000-0002-4799-9910
Rojas Patria [email protected]
125
Wang Weize [email protected]
23
Sanchez Mariana [email protected]
12
Ravelo Gira [email protected]
25
Ángel Cano Miguel [email protected]
4
Galvez Gemma [email protected]
1
Li Tan [email protected]
3
C. Penn Alvonee [email protected]
1
F. Colon-Burgos Jose [email protected]
2
De La Rosa Mario [email protected]
25
Behar-Zusman Victoria [email protected]
5
1 grid.65456.34 0000 0001 2110 1845 Department of Health Promotion and Disease Prevention, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
2 grid.65456.34 0000 0001 2110 1845 CRUSADA, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
3 grid.65456.34 0000 0001 2110 1845 Department of Biostatistics, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
4 grid.65456.34 0000 0001 2110 1845 Department of Epidemiology, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
5 grid.26790.3a 0000 0004 1936 8606 CLaRO, University of Miami; School of Nursing and Health Studies, University of Miami, Coral Gables, FL 33124 USA
13 12 2022
17
8 11 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Previous studies have found Latinx cultural values to be positively associated with healthy behaviors. This study aims to examine socioeconomic and cultural correlates of alcohol use among Latinx adult men living in Miami-Dade County, Florida. The study sample included 122 Latinx adult men (mean age = 44, SD = 10), predominantly of South and Central American origin. Data was collected using REDCap. Interviews included the Timeline Follow-Back scale for alcohol use. Results indicate that Caribbean participants were significantly less likely to report drinking in the past 90 days (aOR = 0.08, p = 0.042) compared to their Venezuelan counterparts. Higher machismo scores were associated with low drinking frequency (aRR = 0.67, p = 0.043), while no significant associations were found between machismo and other drinking outcomes. Drinking quantity and frequency are significantly associated with higher income and authorized immigration status in the US among Latinx men in South Florida. Higher machismo scores were associated with low drinking frequency.
Keywords
Latinx men
Alcohol
Drinking frequency
Machismo
Caballerismo
Latino men
South Florida
http://dx.doi.org/10.13039/100000002 National Institutes of Health National Institutes of Health
==== Body
pmcBackground
The etiology of alcohol use among adult Latinx men has been primarily studied among Mexican American men who were born in the United States (U.S.) and elsewhere [1–3]. Castañeda [3] documented that cultural beliefs and values, such as gender norms (e.g., machismo and caballerismo), have been linked to alcohol use behaviors among Latinx men of Mexican descent. Machismo is a traditional Latinx male gender norm that encompasses hyper-masculinity, aggression, and dominance, and it has been associated with alcohol use risk factors [4]. However, within a more broadened definition, machismo includes positive aspects such as family protection, responsibility, and hard work [5]. Conversely, caballerismo focuses on emotional connectedness, familial cohesion, and social responsibility, and it has been associated as a protective health factor for outcomes such as self-esteem [4, 5]. Yet, the influence of machismo and caballerismo on the alcohol use behaviors of men of South and Central American or Cuban descent living in the U.S. has not been widely studied. The present study investigates alcohol use among a diverse sample of Latinx men living in Miami-Dade County (MDC).
Several papers published between 2018 and 2021 examined the relationship between masculinity and substance misuse; results from one study indicated that the perception and endorsement of machismo norms was associated with alcohol misuse [6]. Moreover, a literature review that examined traditional gender roles among Latinxs men highlighted that, across studies, those who closely aligned with traditional machismo were likely to have a strong association with alcohol use [7]. Young men are more likely to endorse machismo [4] along with choosing to drink more because drinking is perceived as normal masculine behavior; but older Latinx men may refrain from hazardous drinking in order to maintain dignity and family responsibilities [5]. Moreover, alcohol misuse has been associated with a lower level of empathy and lack of pro-social behaviors [8]. Despite receiving alcohol detox treatments, men with problematic alcohol use report lower pro-social behaviors than men who do not use alcohol [8]. Lower empathy and moral compass were consistently low even after men who drank a moderate amount of alcohol were compared with a control group [9]. More recent studies using a multidimensional definition of machismo have included positive social behavioral aspects of machismo such as care and protection of the family, respect, dignity and hard work. In those studies machismo has not been significantly associated with alcohol use [5, 10, 11]. However, Perrotte and Zamboanga [7] discussed caballerismo and its relation to increased well-being and increased conflict resolution, which have been found to be associated with less alcohol consumption [12]. Using the theory of planned behavior, the present study examined the association between gender norms and alcohol use among a diverse group of South Florida Latinx men.
Theoretical Framework
The Theory of Planned Behavior (TPB) has been utilized to predict human behavior particularly frequency of alcohol use. TPB components were not tested in this study, instead gender norms were predicted to be associated with drinking frequency. Thus, guided by TPB, the present study aimed to examine correlates of alcohol use frequency and quantity among Latinx adult men. Notably, in addition to their geographic distinctions, urban communities in MDC are predominantly composed of Latinx individuals of South American and Caribbean origin, whereas Latinx men in semirural areas of MDC are largely of Central American and Mexican origin. We hypothesized that machismo would be associated with higher alcohol use frequency and quantity for all participants.
Methods
The present study is a secondary analysis of a National Institutes of Health (NIH) funded community based clinical study that investigated the effectiveness of an HIV prevention program targeting Latinx fathers and sons in MDC. Participants were recruited using conventional community outreach activities such as placing printed fliers in community organizations, participating in community meetings, social media advertisements, and word-of-mouth. Using only baseline data collected from the fathers, the study sample included 122 male participants aged 18–66 (Mean = 44, SD = 10). Most participants were born in Venezuela (25%), and the second largest group were from other Mexico (22%) (See Table 1). The eligibility criteria for fathers included: (a) being 18 years or older, (b) being the father, or father figure, of an adolescent between the ages of 11–17, (c) living or working in MDC, (d) self-identifying as Latinx, (e) and consenting to participate in one of two randomly assigned groups (i.e., intervention or control groups). Data were collected in Spanish using REDCap survey software, and phone interviews were facilitated by bilingual and bicultural trained interviewers. Measures that were not already available in Spanish were translated into Spanish via translation/back translation methods using a well-established translation protocol with institutional review board approval. Interviewers received a two-day training on the protocol administration and received Collaborative Institutional Training Initiative (CITI) human subjects research certifications. 92% (92%) of the interviews were completed via phone due to the Severe Acute Respiratory Syndrome Corona Virus-2019 (SARS COVID-19) pandemic. This study was approved by the institutional review board of a large private university in Miami, Florida.
Table 1 Participants’ baseline characteristics (N=122)
Variable Mean (SD.) Median Range
Age in years 43.5 (9.8) 44.0 [18, 66]
Depressive symptoms 4.4 (3.6) 4.0 [0, 20]
Anxiety 2.9 (3.2) 2.0 [0, 16]
Years living in the U.S. 14.4 (11.6) 15.0 [1, 47]
Machismo 2.9 (1) 2.9 [1, 5.9]
Caballerismo 6 (0.7) 6.1 [3.3, 7]
Drinking frequency 7.7 (13.7) 3.5 [0, 90]
Drinking quantity 2.6 (2.8) 2.1 [0, 14]
Variable n %
Reported drinking
Yes 77 63.1
No 45 36.9
Binge drinking
Yes 44 36.1
No 78 63.9
Living area
Semi-rural 69 56.6
Urban 53 43.4
Country of origin
Caribbean 11 9.0
Central America 16 13.1
Mexico 27 22.1
Other South American Countries 22 18.0
U.S. 15 12.3
Venezuela 31 25.4
Household income in the last month
0-$999 19 15.8
$1000-$1999 41 34.2
$2000 or more 60 50.0
Education
Less than high school 33 27.1
High school or GED 24 19.7
Some college 29 23.8
College/university degree 36 29.5
Marital status
In a domestic relationship 14 11.5
Married 94 77.1
Single or separated 14 11.5
Employment status
Employed 99 81.2
Unemployed 23 18.9
Immigration status
Authorized 97 82.2
Unauthorized 21 17.8
Note.SD.=Standard deviation
Table 2 Estimates for drinking outcomes among adult Latin x men living in urban and semirural areas of Miami-Dade County, Florida
Predictors Drinking Status Drinking Frequency Drinking Quantity Binge Drinking
aORa [95% CI] aIRRb [95% CI] Estimatec [95% CI] aORa [95% CI]
Age in years 1.03 [0.96, 1.1] 1.02 [0.97, 1.07] 0.99 [0.97, 1.02] 0.99 [0.94, 1.05]
Years living in the U.S. 0.96 [0.91, 1.02] 0.97 [0.94, 1.01] 0.99 [0.96, 1.01] 0.99 [0.94, 1.05]
Machismo -- 0.67 [0.46, 0.99]* -- --
Household income
$1000-$1999 vs. 0-$999 0.06 [0.01, 0.37]** 0.31 [0.11, 0.89]* 0.49 [0.27, 0.9]* 0.49 [0.11, 2.12]
$2000 or more vs. 0-$999 0.24 [0.04, 1.35] 0.68 [0.24, 1.94] 0.81 [0.45, 1.45] 0.79 [0.18, 3.55]
Marital status
In a domestic relationship vs. Married 3.26 [0.51, 21.04] 1.89 [0.55, 6.51] 3.42 [1.85, 6.31]*** 6.18 [1.19, 32.01]
Single or separated vs. Married 0.55 [0.06, 4.77] 0.44 [0.12, 1.55] 1.77 [0.88, 3.58] 2.3 [0.34, 15.45]
Country of origin
Caribbean vs. Venezuela 0.08 [0.01, 0.91]* 0.54 [0.15, 1.96] 0.36 [0.14, 0.92]* 0.7 [0.09, 5.16]
Central America vs. Venezuela 0.23 [0.03, 2.02] 0.73 [0.21, 2.54] 1.32 [0.69, 2.52] 1.1 [0.18, 6.91]
Mexico vs. Venezuela 0.51 [0.04, 6.81] 2.51 [0.59, 10.67] 0.84 [0.38, 1.87] 0.86 [0.09, 7.96]
Other South American countries vs. Venezuela 0.79 [0.09, 6.92] 1.11 [0.36, 3.44] 0.96 [0.5, 1.85] 1.32 [0.25, 6.93]
U.S. vs. Venezuela 0.16 [0, 6.13] 0.41 [0.04, 4.43] 0.55 [0.03, 9.44] 0.55 [0.01, 27.75]
Employment status: Employed vs. Unemployed 6.57 [1.43, 30.07]* 2.8 [1.04, 7.55]* 1.85 [0.96, 3.58] 5.71 [1.19, 27.33]*
Education
High school or GED vs. Less than high school 0.58 [0.12, 2.68] 0.43 [0.15, 1.26] 1.09 [0.59, 2] 0.59 [0.14, 2.57]
Some college vs. Less than high school 0.21 [0.04, 1.19] 0.14 [0.05, 0.43]*** 0.66 [0.32, 1.35] 0.34 [0.07, 1.67]
College/university degree vs. Less than high school 0.27 [0.04, 1.89] 0.13 [0.04, 0.37]*** 0.97 [0.47, 2.02] 1.29 [0.24, 6.8]
Immigration status: Authorized vs. Unauthorized 2.97 [0.55, 16.03] 3.16 [1.12, 8.9]* 3.2 [1.32, 7.8]* 1.9 [0.39, 9.36]
Living area: Urban vs. Semi-rural 3.62 [0.7, 18.74] 2.47 [0.95, 6.37] 1.12 [0.57, 2.17] 1.04 [0.25, 4.32]
Depressive symptoms 1.15 [0.98, 1.34] -- 1.02 [0.96, 1.08] --
a. Estimates are adjusted odds ratios (aORs) from multiple logistic regression
b. Estimates are adjusted incident risk ratios (aIRRs) from negative binomial regression
c. Estimates are exponential correlation coefficients from log-normal regression
*p<0.05; **p<0.01; ***p<0.001
Measures
Demographic Variables
The following demographics were included: age, number of years living in the U.S., country of origin, educational status, marital status, immigration status, the level of urbanization of living area (urban vs. semirural), employment, and household income. Participants were asked to report their educational status, choosing from five categories ranging from “less than high school” to “college/graduate professional studies,” which we then recoded into three categories. Similarly, participants were asked to report their immigration status from 11 categories, ranging from “citizen” to “temporarily protected asylum seeker.” These categories were then recoded into a dichotomous variable of (1) authorized status or (2) unauthorized status. For the present study, marital status categories were recoded into (1) in a domestic relationship, (2) being married, (3) being single or separated, because most participants were married or had a romantic partner. Age was collected and analysed as a continuous variable, but household income in past month was recoded into three categories (0-$999; $1,000-$1,999; $2,000 or more).
Gender Norms
Traditional gender norms were measured using the Machismo and Caballerismo Scale, a widely recognized and validated 20-item bi-dimensional scale that measures machismo and caballerismo on two separate subscales [4]. Sample items for the machismo subscale include: ‘it’s important not to be the weakest man in a group’ and ‘real men never let their guard down.’ Items on the caballerismo subscale include ‘men should be affectionate to their children’ and ‘family is more important than the individual’ [7]. Items were measured using a 7-point Likert-type scale ranging from 1 (not at all) to 7 (very much so). Subscale scores were measured by calculating mean values, with higher values indicating greater adherence to traditional gender norms [4]. Good reliability was reported with a Cronbach’s alpha of 0.80 overall. Cronbach’s alphas in the present study were 0.85 in both the machismo and caballerismo subscales.
Alcohol use was measured using the Timeline Follow-Back method [13]. Participants reported their alcohol use frequency and quantity in the last 90 days. To assist participants in remembering their drinking behaviors, special dates such as holidays were used as anchor points. Initially, participants indicated their drinking behavior during the last 30 days; they then subsequently indicated their drinking behavior during the 60 days preceding the initial 30-day period. Quantity of alcohol was measured using standard serving sizes of drinks, such as a 12 oz beer bottle or can, or a 1 oz. hard liquor shot. Participants reported the number of drinks consumed per day. The reliability of this scale has been validated with similar populations. A dichotomous variable of alcohol consumption (yes/no) was created based on reported drinking in the past 90 days. Alcohol use frequency was the total number of days alcohol was consumed, and alcohol use quantity was the average number of standard drinks consumed in the past 90 days. Binge drinking was recoded into a dichotomous variable (yes/no) that indicated whether a participant had five or more drinks on any occasion.
Mental Health
Depressive symptoms were measured using the Patient Health Questionnaire-9 (PHQ-9), which is an adaptation of the Patient Health Questionnaire (PHQ-59 items). The tool has shown criteria validity [14]. The questionnaire has shown good diagnostic validity with comparable sensitivity and specificity for major depression symptoms among adults [15]. We used the Generalized Anxiety Disorder scale (GAD-7), which is a 7-item measure (4-point Likert-type scale) that assesses anxiety among participants in the past 2 weeks [16]. The Cronbach’s alpha for the PHQ-9 and GAD-7 scales was 0.79 and 0.85 respectively in this study.
Statistical Analysis
Descriptive statistics of participant demographic and alcohol use characteristics are presented in Table 1. The sample mean, standard deviation, median, minimum, and maximum were used to describe the continuous variables. Counts and percentages were used to describe categorical variables. We used logistic regression for alcohol consumption and binge drinking in the past 90 days. We performed negative binomial regression to predict drinking frequency due to its right-skewed count measure. We applied generalized linear model with a log-link function for predicting drinking quantity. To examine the association between the independent variables (machismo, caballerismo, and mental health) and the drinking outcomes, bivariate regression models for each drinking outcome were performed. Independent variables with a p-value at 0.1 or less were included in the multiple regression models, controlling for demographics, including participants’ age, years living in the U.S., household income, marital status, country of origin, employment status, education, immigration status, and living area. The Statistical Analysis System (SAS) 9.4 was used for all data analyses [17]. A statistical significance level of 0.05 was used to reject the null.
Results
Sociodemographic Characteristics
The present analysis included data from 122 Latinx men with a mean age of 44 years (SD = 10). Most men were authorized immigrants (82%), employed (81%), and married (77%) at the time of assessment. Only 12% of the participants were U.S.-born, while other participants immigrated from South America (43%), Mexico (22%), Central America (13%), and the Caribbean (9%). More than half of the participants (57%) were from semirural areas. Half of the participants (50%) had a household income of $2,000 or more in the last month. The mean number of years living in the U.S. was 14 (SD = 12).
Alcohol Use
63% (n = 77) of the participants reported alcohol consumption in the past 90 days. Machismo, caballerismo, or anxiety were not significantly associated with alcohol consumption in the past 90 days, according to results from the bivariate analysis. Thus, the scale scores were excluded from the multiple logistic regression for predicting drinking in the past 90 days. Results show that, compared to participants with a household income of 0-$999 in the past month, those with the household income in the $1,000-$1,999 range were 94% less likely to report drinking in the past 90 days (aOR = 0.06, p = 0.002), controlling for other variables in the model. However, no significant difference in alcohol consumption was found between people with a household income of $2,000 or above and 0-$999 (p > 0.05). Participants who were from the Caribbean were significantly less likely to report drinking in the past 90 days (aORs = 0.08, p = 0.042) compared to participants from Venezuela. Employed participants were more likely to report drinking (aOR = 6.57, p = 0.015) than unemployed participants in the past 90 days.
Drinking Frequency
The median drinking frequency in the past 90 days was 3.5 days. Caballerismo, depressive symptoms, and anxiety were excluded from the final negative binomial regression model because of the non-significant association with drinking frequency from the bivariate analysis. Results from the multiple regression model suggest that machismo was significantly negatively associated with drinking frequency, after controlling for other covariates in the model. A higher machismo score was associated with lower drinking frequency (aRR = 0.67, p = 0.043). Having a household income of $1000-$2000 was significantly associated with lower drinking frequency, compared to the household income of 0-$999 (aRR = 0.31, p = 0.030). Employed participants had more drinking frequency in the past 90 days (aRR = 2.8, p = 0.042) than unemployed participants. Compared to the participants with an educational level of less than high school, those with some college (aRR = 0.14, p < 0.001) or a college/university degree (aRR = 0.13, p < 0.001) reported drinking less frequently. Authorized immigrants were likely to have a higher number of drinking days than those who were unauthorized (aRR = 3.16, p = 0.030).
Drinking Quantity
Among participants, the mean number of drinks per day in which alcohol was consumed in the past 90 days was 2.6 (SD = 2.8). Results from the bivariate analysis show no significant correlation between drinking quantity and machismo, caballerismo, or anxiety. Hence, these variables were excluded from the final regression model. Results show that having a household income of $1000-$1999 was associated with decreased drinking quantity (β = 0.49, p = 0.020), compared to those with household income of 0-$999. Compared to married men, those who were in a domestic relationship reported higher drinking quantities (β = 3.42, p < 0.001). Compared to participants whose country of origin was Venezuela, Caribbean origin was associated with a significantly lower drinking quantity (β = 0.36, p = 0.032). Authorized immigrants reported greater drinking quantity (β = 3.16, p = 0.010). Age, years in the U.S., education, employment, participants’ living area, or depressive symptoms were not significantly associated with drinking quantity (p > 0.05).
Binge Drinking
About 36% of the participants reported binge drinking in the past 90 days ( > = 5 drinks per day in a single sitting). Machismo, caballerismo, depressive symptoms, or anxiety had no statistically significant association with binge drinking in the past 90 days and were excluded from the final regression model. Results from multiple logistic regression indicate employed participants were more likely to binge drink in the past 90 days (aOR = 5.71, p = 0.029). Results show that binge drinking was not significantly associated with age, years in the U.S., household income, marital status, country of origin, education, immigration status, or participants’ living area (p > 0.05).
Discussion
The present study highlights a negative correlation of machismo and drinking frequency, as men with higher scores of machismo reported drinking less frequently. Contrary to the literature on machismo that has characterized men with higher levels of machismo with higher levels of risk taking and substance abuse behavior [11, 18], the current study findings align with a wider perspective found in more recent ones [4, 19]. According to the more recent perspective of masculinity norms, there is a multidimensional aspect of machismo that encompasses attributes such as chivalry, respect, and honor [4, 12]. Studies on alcohol misuse also support the findings that there are higher levels of empathy and higher sense of morality among non-alcohol misusers [8, 9]. This result aligns with more positive aspects of machismo that have been associated with caballerismo [7] and that need to be further examined in future studies.
Similar to other studies [6], authorized immigrants were more likely to report drinking higher quantities and more frequently. There may be several factors associated with their frequency such as having a job, higher income, social stability, and less immigration stress (e.g., fear of deportation).
The present study represents a small example of the heterogeneity among Latinx communities. Our analysis demonstrates that future interventions may want to target low education and documented Latinx individuals. Alcohol frequency and quantity was more prevalent among households reporting less income; however, participants who were employed and lived in urban areas were more likely to report using alcohol in the last 90 days. However, the alcohol use prevalence identified in the present analysis needs to be examined further in future longitudinal studies that show potential differences among different Latinx subgroups.
Most participants from urban areas were of South American origin, primarily Venezuela, and they were more likely to report alcohol use than participants of Central America and the Caribbean region. These findings support a previous study that investigated the alcohol use trajectory of Latinx men prior to and after immigration to the United States [20]. De La Rosa et al. [20] which revealed a decrease in alcohol use frequency for men during their pre- to post-immigration period. Future studies should focus on expanding the sample and including more representation from South America and the Caribbean.
In general, a significant percentage of participants (36%) reported binge drinking in the past 90 days, although the mean number of drinks per day was approximately 3. Although higher drinking quantities were found among men who reported being in a domestic relationship rather than legally married, most participants in the study were married; hence, we cannot draw substantial clinical implications or significant conclusions. Our results indicate the need for other mixed methods research to examine this relationship further.
Limitations
Results from this study should be viewed with the known limitations of a small non-randomized self-reported retrospective data. Additionally, the Latinx population is a heterogeneous group, and the current community-based study sample was mostly well-educated, employed South and Central American immigrants; therefore, there is an underrepresentation of other large U.S. Latinx groups. Because this was a cross-sectional study, temporal relationships could not be established. In addition, it was not possible to examine time variance, given this study’s design, since this was a secondary data analysis. However, future studies might address the mixed findings regarding machismo using other methodological approaches.
Conclusion
The results from our analysis suggest that among adult Latinx men in MDC, alcohol use quantity and frequency are significantly associated with higher income and authorized immigration status. Contrary to our hypothesis, machismo was significantly associated to lower frequency of alcohol use but not quantity or binge drinking. 36% of participants reported binge drinking. Although this is a small cross-sectional community-based study, this is a modest but significant contribution to the knowledge gap in the literature on South and Central American immigrants, particularly Venezuelan immigrants, in Florida. Future studies that include biomarkers, larger samples, and control groups would advance our findings and clarify the association among machismo, alcohol use frequency and quantity. Although previous studies have examined the negative consequences of alcohol use frequency and quantity among Latinos living in the U.S., the present study present an important subgroup distinction and posits the need to continue research on the associations of gender norms and alcohol misuse among Latinx subgroups.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Caetano R Ramisetty-Mikler S Rodriguez LA The Hispanic Americans Baseline Alcohol Survey (HABLAS): rates and predictors of DUI across Hispanic national groups Accid Anal Prev 2008 40 2 733 41 10.1016/j.aap.2007.09.010 18329428
2. Caetano R Ramisetty-Mikler S Rodriguez LA The Hispanic Americans Baseline Alcohol Survey (HABLAS): the association between birthplace, acculturation and alcohol abuse and dependence across Hispanic national groups Drug Alcohol Depend 2009 99 1–3 215 21 10.1016/j.drugalcdep.2008.08.011 18945554
3. Castañeda SF Garcia ML Lopez-Gurrola M Stoutenberg M Emory K Daviglus ML Kaplan R Giachello AL Molina KM Perreira KM Youngblood ME Alcohol use, acculturation and socioeconomic status among Hispanic/Latino men and women: The Hispanic Community Health Study/Study of Latinos PLoS ONE 2019 14 4 e0214906 10.1371/journal.pone.0214906 30947280
4. Arciniega GM Anderson TC Tovar-Blank ZG Tracey TJ Toward a fuller conception of Machismo: Development of a traditional Machismo and Caballerismo Scale J Couns Psychol 2008 55 1 19 10.1037/0022-0167.55.1.19
5. Daniel-Ulloa J Sun C Rhodes SD The intersection between masculinity and health among rural immigrant Latino men Int J Men’s Health 2017 16 1 84 29422781
6. Balagopal G, Davidson S, Gill S, Barengo N, De La Rosa M, Sanchez M. The impact of cultural stress and gender norms on alcohol use severity among Latino immigrant men. Ethnicity & Health. 2021; Feb 9:1–9.
7. Perrotte JK, Zamboanga BL. Traditional gender roles and alcohol use among Latinas/os: A review of the literature. Journal of Ethnicity in Substance Abuse. 2019; Mar 25.
8. Martinotti G, Nicola MD, Tedeschi D, Cundari S, Janiri L. Empathy ability is impaired in alcohol-dependent patients. American Journal on Addictions. 2009 Jan 1; 18(2):157–61.
9. Francis KB Gummerum M Ganis G Howard IS Terbeck S Alcohol, empathy, and morality: acute effects of alcohol consumption on affective empathy and moral decision-making Psychopharmacology 2019 236 12 3477 96 10.1007/s00213-019-05314-z 31289885
10. Perrotte JK Zamboanga BL Kearns N Linking alcohol-specific masculine norms and drinking behavior among Latino men Psychol Men Masculinities 2020 21 3 490 10.1037/men0000252
11. Mirandé A. Hombres y machos: Masculinity and Latino culture. Routledge; 2018. Feb 12.
12. Corbin WR Farmer NM Nolen-Hoekesma S Relations among stress, coping strategies, coping motives, alcohol consumption and related problems: A mediated moderation model Addict Behav 2013 38 4 1912 9 10.1016/j.addbeh.2012.12.005 23380486
13. Sobell LC, Buchan G, Cleland P, Sobell MB, Fedoroff I, Leo GI: The reliability of the Timeline Followback method applied to drug, cigarette, and cannabis use. In 30th Annual Meeting of the Association for Advancement of Behavior Therapy, New York, NY. 1996.
14. Kroenke K Spitzer RL Williams JB The PHQ-9: validity of a brief depression severity measure J Gen Intern Med 2001 16 9 606 13 10.1046/j.1525-1497.2001.016009606.x 11556941
15. Spitzer RL Williams JB Kroenke KJ Test review: patient health questionnaire–9 (PHQ-9) Rehabilitation Couns Bull 2014 57 4 246 8 10.1177/0034355213515305
16. Spitzer RL Kroenke K Williams JB Löwe B A brief measure for assessing generalized anxiety disorder: the GAD-7 Arch Intern Med 2006 166 10 1092 7 10.1001/archinte.166.10.1092 16717171
17. SAS Institute Inc SAS/ACCESS® 9.4 Interface to ADABAS: Reference 2013 Cary SAS Institute Inc
18. Neff JA Prihoda TJ Hoppe SK ” Machismo,“ self-esteem, education, and high maximum drinking among anglo, black and Mexican-American male drinkers J Stud Alcohol 1991 52 5 458 63 10.15288/jsa.1991.52.458 1943101
19. Falicov CJ Changing constructions of machismo for Latino men in therapy: “The devil never sleeps.” Fam Process 2010 49 3 309 29 10.1111/j.1545-5300.2010.01325.x 20831763
20. De La Rosa M Sanchez M Wang W Cano MA Rojas P Alcohol use trajectories of adult Latinx immigrants during their first decade in the United States Addict Behav 2020 106 106352 10.1016/j.addbeh.2020.106352 32145498
| 36513769 | PMC9746554 | NO-CC CODE | 2022-12-15 23:21:56 | no | Nat Rev Microbiol. 2022 Dec 13;:1 | latin-1 | Nat Rev Microbiol | 2,022 | 10.1038/s41579-022-00844-4 | oa_other |
==== Front
J Immigr Minor Health
J Immigr Minor Health
Journal of Immigrant and Minority Health
1557-1912
1557-1920
Springer US New York
1428
10.1007/s10903-022-01428-3
Original Paper
Predictors of Alcohol Use Among Latinx Men in South Florida: Machismo as a Correlate of Alcohol Use Frequency and Quantity
http://orcid.org/0000-0002-4799-9910
Rojas Patria [email protected]
125
Wang Weize [email protected]
23
Sanchez Mariana [email protected]
12
Ravelo Gira [email protected]
25
Ángel Cano Miguel [email protected]
4
Galvez Gemma [email protected]
1
Li Tan [email protected]
3
C. Penn Alvonee [email protected]
1
F. Colon-Burgos Jose [email protected]
2
De La Rosa Mario [email protected]
25
Behar-Zusman Victoria [email protected]
5
1 grid.65456.34 0000 0001 2110 1845 Department of Health Promotion and Disease Prevention, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
2 grid.65456.34 0000 0001 2110 1845 CRUSADA, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
3 grid.65456.34 0000 0001 2110 1845 Department of Biostatistics, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
4 grid.65456.34 0000 0001 2110 1845 Department of Epidemiology, Robert Stempel College of Public Health and Social Work, Florida International University, 11200 SW 8th Street, Miami, FL 33199 USA
5 grid.26790.3a 0000 0004 1936 8606 CLaRO, University of Miami; School of Nursing and Health Studies, University of Miami, Coral Gables, FL 33124 USA
13 12 2022
17
8 11 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Previous studies have found Latinx cultural values to be positively associated with healthy behaviors. This study aims to examine socioeconomic and cultural correlates of alcohol use among Latinx adult men living in Miami-Dade County, Florida. The study sample included 122 Latinx adult men (mean age = 44, SD = 10), predominantly of South and Central American origin. Data was collected using REDCap. Interviews included the Timeline Follow-Back scale for alcohol use. Results indicate that Caribbean participants were significantly less likely to report drinking in the past 90 days (aOR = 0.08, p = 0.042) compared to their Venezuelan counterparts. Higher machismo scores were associated with low drinking frequency (aRR = 0.67, p = 0.043), while no significant associations were found between machismo and other drinking outcomes. Drinking quantity and frequency are significantly associated with higher income and authorized immigration status in the US among Latinx men in South Florida. Higher machismo scores were associated with low drinking frequency.
Keywords
Latinx men
Alcohol
Drinking frequency
Machismo
Caballerismo
Latino men
South Florida
http://dx.doi.org/10.13039/100000002 National Institutes of Health National Institutes of Health
==== Body
pmcBackground
The etiology of alcohol use among adult Latinx men has been primarily studied among Mexican American men who were born in the United States (U.S.) and elsewhere [1–3]. Castañeda [3] documented that cultural beliefs and values, such as gender norms (e.g., machismo and caballerismo), have been linked to alcohol use behaviors among Latinx men of Mexican descent. Machismo is a traditional Latinx male gender norm that encompasses hyper-masculinity, aggression, and dominance, and it has been associated with alcohol use risk factors [4]. However, within a more broadened definition, machismo includes positive aspects such as family protection, responsibility, and hard work [5]. Conversely, caballerismo focuses on emotional connectedness, familial cohesion, and social responsibility, and it has been associated as a protective health factor for outcomes such as self-esteem [4, 5]. Yet, the influence of machismo and caballerismo on the alcohol use behaviors of men of South and Central American or Cuban descent living in the U.S. has not been widely studied. The present study investigates alcohol use among a diverse sample of Latinx men living in Miami-Dade County (MDC).
Several papers published between 2018 and 2021 examined the relationship between masculinity and substance misuse; results from one study indicated that the perception and endorsement of machismo norms was associated with alcohol misuse [6]. Moreover, a literature review that examined traditional gender roles among Latinxs men highlighted that, across studies, those who closely aligned with traditional machismo were likely to have a strong association with alcohol use [7]. Young men are more likely to endorse machismo [4] along with choosing to drink more because drinking is perceived as normal masculine behavior; but older Latinx men may refrain from hazardous drinking in order to maintain dignity and family responsibilities [5]. Moreover, alcohol misuse has been associated with a lower level of empathy and lack of pro-social behaviors [8]. Despite receiving alcohol detox treatments, men with problematic alcohol use report lower pro-social behaviors than men who do not use alcohol [8]. Lower empathy and moral compass were consistently low even after men who drank a moderate amount of alcohol were compared with a control group [9]. More recent studies using a multidimensional definition of machismo have included positive social behavioral aspects of machismo such as care and protection of the family, respect, dignity and hard work. In those studies machismo has not been significantly associated with alcohol use [5, 10, 11]. However, Perrotte and Zamboanga [7] discussed caballerismo and its relation to increased well-being and increased conflict resolution, which have been found to be associated with less alcohol consumption [12]. Using the theory of planned behavior, the present study examined the association between gender norms and alcohol use among a diverse group of South Florida Latinx men.
Theoretical Framework
The Theory of Planned Behavior (TPB) has been utilized to predict human behavior particularly frequency of alcohol use. TPB components were not tested in this study, instead gender norms were predicted to be associated with drinking frequency. Thus, guided by TPB, the present study aimed to examine correlates of alcohol use frequency and quantity among Latinx adult men. Notably, in addition to their geographic distinctions, urban communities in MDC are predominantly composed of Latinx individuals of South American and Caribbean origin, whereas Latinx men in semirural areas of MDC are largely of Central American and Mexican origin. We hypothesized that machismo would be associated with higher alcohol use frequency and quantity for all participants.
Methods
The present study is a secondary analysis of a National Institutes of Health (NIH) funded community based clinical study that investigated the effectiveness of an HIV prevention program targeting Latinx fathers and sons in MDC. Participants were recruited using conventional community outreach activities such as placing printed fliers in community organizations, participating in community meetings, social media advertisements, and word-of-mouth. Using only baseline data collected from the fathers, the study sample included 122 male participants aged 18–66 (Mean = 44, SD = 10). Most participants were born in Venezuela (25%), and the second largest group were from other Mexico (22%) (See Table 1). The eligibility criteria for fathers included: (a) being 18 years or older, (b) being the father, or father figure, of an adolescent between the ages of 11–17, (c) living or working in MDC, (d) self-identifying as Latinx, (e) and consenting to participate in one of two randomly assigned groups (i.e., intervention or control groups). Data were collected in Spanish using REDCap survey software, and phone interviews were facilitated by bilingual and bicultural trained interviewers. Measures that were not already available in Spanish were translated into Spanish via translation/back translation methods using a well-established translation protocol with institutional review board approval. Interviewers received a two-day training on the protocol administration and received Collaborative Institutional Training Initiative (CITI) human subjects research certifications. 92% (92%) of the interviews were completed via phone due to the Severe Acute Respiratory Syndrome Corona Virus-2019 (SARS COVID-19) pandemic. This study was approved by the institutional review board of a large private university in Miami, Florida.
Table 1 Participants’ baseline characteristics (N=122)
Variable Mean (SD.) Median Range
Age in years 43.5 (9.8) 44.0 [18, 66]
Depressive symptoms 4.4 (3.6) 4.0 [0, 20]
Anxiety 2.9 (3.2) 2.0 [0, 16]
Years living in the U.S. 14.4 (11.6) 15.0 [1, 47]
Machismo 2.9 (1) 2.9 [1, 5.9]
Caballerismo 6 (0.7) 6.1 [3.3, 7]
Drinking frequency 7.7 (13.7) 3.5 [0, 90]
Drinking quantity 2.6 (2.8) 2.1 [0, 14]
Variable n %
Reported drinking
Yes 77 63.1
No 45 36.9
Binge drinking
Yes 44 36.1
No 78 63.9
Living area
Semi-rural 69 56.6
Urban 53 43.4
Country of origin
Caribbean 11 9.0
Central America 16 13.1
Mexico 27 22.1
Other South American Countries 22 18.0
U.S. 15 12.3
Venezuela 31 25.4
Household income in the last month
0-$999 19 15.8
$1000-$1999 41 34.2
$2000 or more 60 50.0
Education
Less than high school 33 27.1
High school or GED 24 19.7
Some college 29 23.8
College/university degree 36 29.5
Marital status
In a domestic relationship 14 11.5
Married 94 77.1
Single or separated 14 11.5
Employment status
Employed 99 81.2
Unemployed 23 18.9
Immigration status
Authorized 97 82.2
Unauthorized 21 17.8
Note.SD.=Standard deviation
Table 2 Estimates for drinking outcomes among adult Latin x men living in urban and semirural areas of Miami-Dade County, Florida
Predictors Drinking Status Drinking Frequency Drinking Quantity Binge Drinking
aORa [95% CI] aIRRb [95% CI] Estimatec [95% CI] aORa [95% CI]
Age in years 1.03 [0.96, 1.1] 1.02 [0.97, 1.07] 0.99 [0.97, 1.02] 0.99 [0.94, 1.05]
Years living in the U.S. 0.96 [0.91, 1.02] 0.97 [0.94, 1.01] 0.99 [0.96, 1.01] 0.99 [0.94, 1.05]
Machismo -- 0.67 [0.46, 0.99]* -- --
Household income
$1000-$1999 vs. 0-$999 0.06 [0.01, 0.37]** 0.31 [0.11, 0.89]* 0.49 [0.27, 0.9]* 0.49 [0.11, 2.12]
$2000 or more vs. 0-$999 0.24 [0.04, 1.35] 0.68 [0.24, 1.94] 0.81 [0.45, 1.45] 0.79 [0.18, 3.55]
Marital status
In a domestic relationship vs. Married 3.26 [0.51, 21.04] 1.89 [0.55, 6.51] 3.42 [1.85, 6.31]*** 6.18 [1.19, 32.01]
Single or separated vs. Married 0.55 [0.06, 4.77] 0.44 [0.12, 1.55] 1.77 [0.88, 3.58] 2.3 [0.34, 15.45]
Country of origin
Caribbean vs. Venezuela 0.08 [0.01, 0.91]* 0.54 [0.15, 1.96] 0.36 [0.14, 0.92]* 0.7 [0.09, 5.16]
Central America vs. Venezuela 0.23 [0.03, 2.02] 0.73 [0.21, 2.54] 1.32 [0.69, 2.52] 1.1 [0.18, 6.91]
Mexico vs. Venezuela 0.51 [0.04, 6.81] 2.51 [0.59, 10.67] 0.84 [0.38, 1.87] 0.86 [0.09, 7.96]
Other South American countries vs. Venezuela 0.79 [0.09, 6.92] 1.11 [0.36, 3.44] 0.96 [0.5, 1.85] 1.32 [0.25, 6.93]
U.S. vs. Venezuela 0.16 [0, 6.13] 0.41 [0.04, 4.43] 0.55 [0.03, 9.44] 0.55 [0.01, 27.75]
Employment status: Employed vs. Unemployed 6.57 [1.43, 30.07]* 2.8 [1.04, 7.55]* 1.85 [0.96, 3.58] 5.71 [1.19, 27.33]*
Education
High school or GED vs. Less than high school 0.58 [0.12, 2.68] 0.43 [0.15, 1.26] 1.09 [0.59, 2] 0.59 [0.14, 2.57]
Some college vs. Less than high school 0.21 [0.04, 1.19] 0.14 [0.05, 0.43]*** 0.66 [0.32, 1.35] 0.34 [0.07, 1.67]
College/university degree vs. Less than high school 0.27 [0.04, 1.89] 0.13 [0.04, 0.37]*** 0.97 [0.47, 2.02] 1.29 [0.24, 6.8]
Immigration status: Authorized vs. Unauthorized 2.97 [0.55, 16.03] 3.16 [1.12, 8.9]* 3.2 [1.32, 7.8]* 1.9 [0.39, 9.36]
Living area: Urban vs. Semi-rural 3.62 [0.7, 18.74] 2.47 [0.95, 6.37] 1.12 [0.57, 2.17] 1.04 [0.25, 4.32]
Depressive symptoms 1.15 [0.98, 1.34] -- 1.02 [0.96, 1.08] --
a. Estimates are adjusted odds ratios (aORs) from multiple logistic regression
b. Estimates are adjusted incident risk ratios (aIRRs) from negative binomial regression
c. Estimates are exponential correlation coefficients from log-normal regression
*p<0.05; **p<0.01; ***p<0.001
Measures
Demographic Variables
The following demographics were included: age, number of years living in the U.S., country of origin, educational status, marital status, immigration status, the level of urbanization of living area (urban vs. semirural), employment, and household income. Participants were asked to report their educational status, choosing from five categories ranging from “less than high school” to “college/graduate professional studies,” which we then recoded into three categories. Similarly, participants were asked to report their immigration status from 11 categories, ranging from “citizen” to “temporarily protected asylum seeker.” These categories were then recoded into a dichotomous variable of (1) authorized status or (2) unauthorized status. For the present study, marital status categories were recoded into (1) in a domestic relationship, (2) being married, (3) being single or separated, because most participants were married or had a romantic partner. Age was collected and analysed as a continuous variable, but household income in past month was recoded into three categories (0-$999; $1,000-$1,999; $2,000 or more).
Gender Norms
Traditional gender norms were measured using the Machismo and Caballerismo Scale, a widely recognized and validated 20-item bi-dimensional scale that measures machismo and caballerismo on two separate subscales [4]. Sample items for the machismo subscale include: ‘it’s important not to be the weakest man in a group’ and ‘real men never let their guard down.’ Items on the caballerismo subscale include ‘men should be affectionate to their children’ and ‘family is more important than the individual’ [7]. Items were measured using a 7-point Likert-type scale ranging from 1 (not at all) to 7 (very much so). Subscale scores were measured by calculating mean values, with higher values indicating greater adherence to traditional gender norms [4]. Good reliability was reported with a Cronbach’s alpha of 0.80 overall. Cronbach’s alphas in the present study were 0.85 in both the machismo and caballerismo subscales.
Alcohol use was measured using the Timeline Follow-Back method [13]. Participants reported their alcohol use frequency and quantity in the last 90 days. To assist participants in remembering their drinking behaviors, special dates such as holidays were used as anchor points. Initially, participants indicated their drinking behavior during the last 30 days; they then subsequently indicated their drinking behavior during the 60 days preceding the initial 30-day period. Quantity of alcohol was measured using standard serving sizes of drinks, such as a 12 oz beer bottle or can, or a 1 oz. hard liquor shot. Participants reported the number of drinks consumed per day. The reliability of this scale has been validated with similar populations. A dichotomous variable of alcohol consumption (yes/no) was created based on reported drinking in the past 90 days. Alcohol use frequency was the total number of days alcohol was consumed, and alcohol use quantity was the average number of standard drinks consumed in the past 90 days. Binge drinking was recoded into a dichotomous variable (yes/no) that indicated whether a participant had five or more drinks on any occasion.
Mental Health
Depressive symptoms were measured using the Patient Health Questionnaire-9 (PHQ-9), which is an adaptation of the Patient Health Questionnaire (PHQ-59 items). The tool has shown criteria validity [14]. The questionnaire has shown good diagnostic validity with comparable sensitivity and specificity for major depression symptoms among adults [15]. We used the Generalized Anxiety Disorder scale (GAD-7), which is a 7-item measure (4-point Likert-type scale) that assesses anxiety among participants in the past 2 weeks [16]. The Cronbach’s alpha for the PHQ-9 and GAD-7 scales was 0.79 and 0.85 respectively in this study.
Statistical Analysis
Descriptive statistics of participant demographic and alcohol use characteristics are presented in Table 1. The sample mean, standard deviation, median, minimum, and maximum were used to describe the continuous variables. Counts and percentages were used to describe categorical variables. We used logistic regression for alcohol consumption and binge drinking in the past 90 days. We performed negative binomial regression to predict drinking frequency due to its right-skewed count measure. We applied generalized linear model with a log-link function for predicting drinking quantity. To examine the association between the independent variables (machismo, caballerismo, and mental health) and the drinking outcomes, bivariate regression models for each drinking outcome were performed. Independent variables with a p-value at 0.1 or less were included in the multiple regression models, controlling for demographics, including participants’ age, years living in the U.S., household income, marital status, country of origin, employment status, education, immigration status, and living area. The Statistical Analysis System (SAS) 9.4 was used for all data analyses [17]. A statistical significance level of 0.05 was used to reject the null.
Results
Sociodemographic Characteristics
The present analysis included data from 122 Latinx men with a mean age of 44 years (SD = 10). Most men were authorized immigrants (82%), employed (81%), and married (77%) at the time of assessment. Only 12% of the participants were U.S.-born, while other participants immigrated from South America (43%), Mexico (22%), Central America (13%), and the Caribbean (9%). More than half of the participants (57%) were from semirural areas. Half of the participants (50%) had a household income of $2,000 or more in the last month. The mean number of years living in the U.S. was 14 (SD = 12).
Alcohol Use
63% (n = 77) of the participants reported alcohol consumption in the past 90 days. Machismo, caballerismo, or anxiety were not significantly associated with alcohol consumption in the past 90 days, according to results from the bivariate analysis. Thus, the scale scores were excluded from the multiple logistic regression for predicting drinking in the past 90 days. Results show that, compared to participants with a household income of 0-$999 in the past month, those with the household income in the $1,000-$1,999 range were 94% less likely to report drinking in the past 90 days (aOR = 0.06, p = 0.002), controlling for other variables in the model. However, no significant difference in alcohol consumption was found between people with a household income of $2,000 or above and 0-$999 (p > 0.05). Participants who were from the Caribbean were significantly less likely to report drinking in the past 90 days (aORs = 0.08, p = 0.042) compared to participants from Venezuela. Employed participants were more likely to report drinking (aOR = 6.57, p = 0.015) than unemployed participants in the past 90 days.
Drinking Frequency
The median drinking frequency in the past 90 days was 3.5 days. Caballerismo, depressive symptoms, and anxiety were excluded from the final negative binomial regression model because of the non-significant association with drinking frequency from the bivariate analysis. Results from the multiple regression model suggest that machismo was significantly negatively associated with drinking frequency, after controlling for other covariates in the model. A higher machismo score was associated with lower drinking frequency (aRR = 0.67, p = 0.043). Having a household income of $1000-$2000 was significantly associated with lower drinking frequency, compared to the household income of 0-$999 (aRR = 0.31, p = 0.030). Employed participants had more drinking frequency in the past 90 days (aRR = 2.8, p = 0.042) than unemployed participants. Compared to the participants with an educational level of less than high school, those with some college (aRR = 0.14, p < 0.001) or a college/university degree (aRR = 0.13, p < 0.001) reported drinking less frequently. Authorized immigrants were likely to have a higher number of drinking days than those who were unauthorized (aRR = 3.16, p = 0.030).
Drinking Quantity
Among participants, the mean number of drinks per day in which alcohol was consumed in the past 90 days was 2.6 (SD = 2.8). Results from the bivariate analysis show no significant correlation between drinking quantity and machismo, caballerismo, or anxiety. Hence, these variables were excluded from the final regression model. Results show that having a household income of $1000-$1999 was associated with decreased drinking quantity (β = 0.49, p = 0.020), compared to those with household income of 0-$999. Compared to married men, those who were in a domestic relationship reported higher drinking quantities (β = 3.42, p < 0.001). Compared to participants whose country of origin was Venezuela, Caribbean origin was associated with a significantly lower drinking quantity (β = 0.36, p = 0.032). Authorized immigrants reported greater drinking quantity (β = 3.16, p = 0.010). Age, years in the U.S., education, employment, participants’ living area, or depressive symptoms were not significantly associated with drinking quantity (p > 0.05).
Binge Drinking
About 36% of the participants reported binge drinking in the past 90 days ( > = 5 drinks per day in a single sitting). Machismo, caballerismo, depressive symptoms, or anxiety had no statistically significant association with binge drinking in the past 90 days and were excluded from the final regression model. Results from multiple logistic regression indicate employed participants were more likely to binge drink in the past 90 days (aOR = 5.71, p = 0.029). Results show that binge drinking was not significantly associated with age, years in the U.S., household income, marital status, country of origin, education, immigration status, or participants’ living area (p > 0.05).
Discussion
The present study highlights a negative correlation of machismo and drinking frequency, as men with higher scores of machismo reported drinking less frequently. Contrary to the literature on machismo that has characterized men with higher levels of machismo with higher levels of risk taking and substance abuse behavior [11, 18], the current study findings align with a wider perspective found in more recent ones [4, 19]. According to the more recent perspective of masculinity norms, there is a multidimensional aspect of machismo that encompasses attributes such as chivalry, respect, and honor [4, 12]. Studies on alcohol misuse also support the findings that there are higher levels of empathy and higher sense of morality among non-alcohol misusers [8, 9]. This result aligns with more positive aspects of machismo that have been associated with caballerismo [7] and that need to be further examined in future studies.
Similar to other studies [6], authorized immigrants were more likely to report drinking higher quantities and more frequently. There may be several factors associated with their frequency such as having a job, higher income, social stability, and less immigration stress (e.g., fear of deportation).
The present study represents a small example of the heterogeneity among Latinx communities. Our analysis demonstrates that future interventions may want to target low education and documented Latinx individuals. Alcohol frequency and quantity was more prevalent among households reporting less income; however, participants who were employed and lived in urban areas were more likely to report using alcohol in the last 90 days. However, the alcohol use prevalence identified in the present analysis needs to be examined further in future longitudinal studies that show potential differences among different Latinx subgroups.
Most participants from urban areas were of South American origin, primarily Venezuela, and they were more likely to report alcohol use than participants of Central America and the Caribbean region. These findings support a previous study that investigated the alcohol use trajectory of Latinx men prior to and after immigration to the United States [20]. De La Rosa et al. [20] which revealed a decrease in alcohol use frequency for men during their pre- to post-immigration period. Future studies should focus on expanding the sample and including more representation from South America and the Caribbean.
In general, a significant percentage of participants (36%) reported binge drinking in the past 90 days, although the mean number of drinks per day was approximately 3. Although higher drinking quantities were found among men who reported being in a domestic relationship rather than legally married, most participants in the study were married; hence, we cannot draw substantial clinical implications or significant conclusions. Our results indicate the need for other mixed methods research to examine this relationship further.
Limitations
Results from this study should be viewed with the known limitations of a small non-randomized self-reported retrospective data. Additionally, the Latinx population is a heterogeneous group, and the current community-based study sample was mostly well-educated, employed South and Central American immigrants; therefore, there is an underrepresentation of other large U.S. Latinx groups. Because this was a cross-sectional study, temporal relationships could not be established. In addition, it was not possible to examine time variance, given this study’s design, since this was a secondary data analysis. However, future studies might address the mixed findings regarding machismo using other methodological approaches.
Conclusion
The results from our analysis suggest that among adult Latinx men in MDC, alcohol use quantity and frequency are significantly associated with higher income and authorized immigration status. Contrary to our hypothesis, machismo was significantly associated to lower frequency of alcohol use but not quantity or binge drinking. 36% of participants reported binge drinking. Although this is a small cross-sectional community-based study, this is a modest but significant contribution to the knowledge gap in the literature on South and Central American immigrants, particularly Venezuelan immigrants, in Florida. Future studies that include biomarkers, larger samples, and control groups would advance our findings and clarify the association among machismo, alcohol use frequency and quantity. Although previous studies have examined the negative consequences of alcohol use frequency and quantity among Latinos living in the U.S., the present study present an important subgroup distinction and posits the need to continue research on the associations of gender norms and alcohol misuse among Latinx subgroups.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Caetano R Ramisetty-Mikler S Rodriguez LA The Hispanic Americans Baseline Alcohol Survey (HABLAS): rates and predictors of DUI across Hispanic national groups Accid Anal Prev 2008 40 2 733 41 10.1016/j.aap.2007.09.010 18329428
2. Caetano R Ramisetty-Mikler S Rodriguez LA The Hispanic Americans Baseline Alcohol Survey (HABLAS): the association between birthplace, acculturation and alcohol abuse and dependence across Hispanic national groups Drug Alcohol Depend 2009 99 1–3 215 21 10.1016/j.drugalcdep.2008.08.011 18945554
3. Castañeda SF Garcia ML Lopez-Gurrola M Stoutenberg M Emory K Daviglus ML Kaplan R Giachello AL Molina KM Perreira KM Youngblood ME Alcohol use, acculturation and socioeconomic status among Hispanic/Latino men and women: The Hispanic Community Health Study/Study of Latinos PLoS ONE 2019 14 4 e0214906 10.1371/journal.pone.0214906 30947280
4. Arciniega GM Anderson TC Tovar-Blank ZG Tracey TJ Toward a fuller conception of Machismo: Development of a traditional Machismo and Caballerismo Scale J Couns Psychol 2008 55 1 19 10.1037/0022-0167.55.1.19
5. Daniel-Ulloa J Sun C Rhodes SD The intersection between masculinity and health among rural immigrant Latino men Int J Men’s Health 2017 16 1 84 29422781
6. Balagopal G, Davidson S, Gill S, Barengo N, De La Rosa M, Sanchez M. The impact of cultural stress and gender norms on alcohol use severity among Latino immigrant men. Ethnicity & Health. 2021; Feb 9:1–9.
7. Perrotte JK, Zamboanga BL. Traditional gender roles and alcohol use among Latinas/os: A review of the literature. Journal of Ethnicity in Substance Abuse. 2019; Mar 25.
8. Martinotti G, Nicola MD, Tedeschi D, Cundari S, Janiri L. Empathy ability is impaired in alcohol-dependent patients. American Journal on Addictions. 2009 Jan 1; 18(2):157–61.
9. Francis KB Gummerum M Ganis G Howard IS Terbeck S Alcohol, empathy, and morality: acute effects of alcohol consumption on affective empathy and moral decision-making Psychopharmacology 2019 236 12 3477 96 10.1007/s00213-019-05314-z 31289885
10. Perrotte JK Zamboanga BL Kearns N Linking alcohol-specific masculine norms and drinking behavior among Latino men Psychol Men Masculinities 2020 21 3 490 10.1037/men0000252
11. Mirandé A. Hombres y machos: Masculinity and Latino culture. Routledge; 2018. Feb 12.
12. Corbin WR Farmer NM Nolen-Hoekesma S Relations among stress, coping strategies, coping motives, alcohol consumption and related problems: A mediated moderation model Addict Behav 2013 38 4 1912 9 10.1016/j.addbeh.2012.12.005 23380486
13. Sobell LC, Buchan G, Cleland P, Sobell MB, Fedoroff I, Leo GI: The reliability of the Timeline Followback method applied to drug, cigarette, and cannabis use. In 30th Annual Meeting of the Association for Advancement of Behavior Therapy, New York, NY. 1996.
14. Kroenke K Spitzer RL Williams JB The PHQ-9: validity of a brief depression severity measure J Gen Intern Med 2001 16 9 606 13 10.1046/j.1525-1497.2001.016009606.x 11556941
15. Spitzer RL Williams JB Kroenke KJ Test review: patient health questionnaire–9 (PHQ-9) Rehabilitation Couns Bull 2014 57 4 246 8 10.1177/0034355213515305
16. Spitzer RL Kroenke K Williams JB Löwe B A brief measure for assessing generalized anxiety disorder: the GAD-7 Arch Intern Med 2006 166 10 1092 7 10.1001/archinte.166.10.1092 16717171
17. SAS Institute Inc SAS/ACCESS® 9.4 Interface to ADABAS: Reference 2013 Cary SAS Institute Inc
18. Neff JA Prihoda TJ Hoppe SK ” Machismo,“ self-esteem, education, and high maximum drinking among anglo, black and Mexican-American male drinkers J Stud Alcohol 1991 52 5 458 63 10.15288/jsa.1991.52.458 1943101
19. Falicov CJ Changing constructions of machismo for Latino men in therapy: “The devil never sleeps.” Fam Process 2010 49 3 309 29 10.1111/j.1545-5300.2010.01325.x 20831763
20. De La Rosa M Sanchez M Wang W Cano MA Rojas P Alcohol use trajectories of adult Latinx immigrants during their first decade in the United States Addict Behav 2020 106 106352 10.1016/j.addbeh.2020.106352 32145498
| 0 | PMC9746555 | NO-CC CODE | 2022-12-15 23:21:56 | no | Padiatr Padol. 2022 Dec 13; 57(6):313 | latin-1 | Padiatr Padol | 2,022 | 10.1007/s00608-022-01040-6 | oa_other |
==== Front
Z Bild Forsch
Z Bild Forsch
Zeitschrift Fur Bildungsforschung
2190-6890
2190-6904
Springer Fachmedien Wiesbaden Wiesbaden
374
10.1007/s35834-022-00374-2
Aktuelle Bücher aus der ÖFEB
Aktuelle Buchveröffentlichungen von Mitgliedern der ÖFEB
13 12 2022
12
© Springer Fachmedien Wiesbaden GmbH, ein Teil von Springer Nature 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
==== Body
pmc Harald Burgsteiner, Georg Krammer (Hrsg.)
Impacts of COVID-19 Pandemic’s Distance Learning on Students and Teachers in Schools and in Higher Education
International Perspectives
Leykam, 2022, 525 Seiten, broschiert, 65 €, ISBN: 978-3-7011-0496‑3; 10.56560/isbn.978-3-7011-0496-3 (open access).
The worldwide imposed lockdowns forced schools and universities to digitise conventional teaching in a very short time and to convert teaching and learning formats partially or completely to Distance Learning. The changes in everyday teaching brought by Distance Learning were felt worldwide. With 22 double blind peerreviewed articles of researchers reporting on 17 different countries, the editors of this book want to shed light on the effects of Distance Learning in different regions of the world. This will allow for a value-free comparison of how the COVID-19 pandemic has been addressed in education in different parts of the world and what impacts it has had, is having or may have in the future.
Verlagstext
Martin Lehner
Mini-Aufgaben
Denkprozesse anregen und Neugier wecken
utb, 2022, 157 Seiten, kartoniert, 20 €, ISBN: 9783825258894; PDF: 19,99 €, eISBN: 9783838558899, 10.36198/9783838558899.
Die kognitive Aktivierung von Lernenden ist heute ein zentrales Element lernwirksamen Unterrichts, stellt für Lehrende aber nicht selten eine Herausforderung dar. Wie können Schülerinnen und Schüler oder auch Studierende dazu angeregt werden, sich mit den Lerninhalten aktiv auseinanderzusetzen?
Der Einsatz von sogenannten Mini-Aufgaben ist dafür besonders gut geeignet. Im Unterschied zu anderen Aufgabenformaten zielen Mini-Aufgaben darauf ab, in kurzer Zeit möglichst alle Lernenden anzusprechen, sie inhaltlich herauszufordern und damit auch kognitiv zu aktivieren. Dadurch eignen sich Mini-Aufgaben auch für die Förderung des Verstehens beim Lernen.
In diesem kompakten Buch zeigt Didaktik-Autor Martin Lehner leicht verständlich und anhand zahlreicher Beispiele, was es für richtig gute Mini-Aufgaben braucht und wie sie sinnvoll in den Unterricht eingebettet werden können.
Verlagstext
| 0 | PMC9746558 | NO-CC CODE | 2022-12-15 23:21:56 | no | Z Bild Forsch. 2022 Dec 13;:1-2 | utf-8 | Z Bild Forsch | 2,022 | 10.1007/s35834-022-00374-2 | oa_other |
==== Front
J Racial Ethn Health Disparities
J Racial Ethn Health Disparities
Journal of Racial and Ethnic Health Disparities
2197-3792
2196-8837
Springer International Publishing Cham
1480
10.1007/s40615-022-01480-7
Article
Barriers to Type 2 Diabetes Mellitus Management for Older Hmong Patients with Minimal English Language Skills: Accounts from Caregivers, Case Managers, and Clinicians
http://orcid.org/0000-0001-8008-343X
Park Linda [email protected]
1
Vang Addison 2
Yang Brittany 2
Quanbeck Andrew 1
1 grid.14003.36 0000 0001 2167 3675 Department of Family Medicine and Community Health, University of Wisconsin-Madison, Madison, WI USA
2 grid.14003.36 0000 0001 2167 3675 University of Wisconsin-Madison, Madison, WI USA
13 12 2022
18
28 7 2022
22 11 2022
23 11 2022
© W. Montague Cobb-NMA Health Institute 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Type 2 diabetes mellitus prevalence rates for Hmong Americans in Wisconsin are more than double that of non-Hispanic Whites. The Hmong’s history, lifestyle (dietary and behavioral patterns), and reliance on traditional medicine contribute to their increased risk of diabetes. This qualitative study aimed to better understand the barriers challenging older Hmong patients’ ability to manage diabetes. Asian Americans have long been overlooked in health-related research, but recent disaggregated data of specific ethnic groups reveal significant health inequities. Among the different ethnic groups, there is a significant lack of research on the Hmong Americans. Three participant groups (Hmong American family caregivers, Hmong American case managers, and clinicians from different racial backgrounds who provide care for Hmong patients) were recruited from the community and interviewed to understand the barriers experienced by older Hmong patients with minimal English language skills in managing their diabetes. Directed content analysis of the data resulted in three major themes: adherence to culture, health inequity, and managing diabetes. Subthemes included Hmong herbs and shamans, lack of trust in Western medicine, the significance of rice, language barriers, lack of cultural sensitivity, health literacy, monitoring glucose, medicine compliance, and nutrition. Minimal English language skills and low literacy rates (health and education) contribute to their strong adherence to cultural practices which challenges Western medicine, creating difficulty for older Hmong patients to manage their diabetes. Recognizing cultural differences and barriers will enable healthcare providers to improve and cater the treatment options, bridging the gap between older Hmong patients and Western medicine.
Keywords
Type 2 diabetes mellitus
Hmong
Health inequities
Language barrier
Health literacy
==== Body
pmcBackground and Significance
Recent data collected (2011–2012) from the National Health and Nutrition Examination Survey (NHANES) revealed that over 50% of Asian Americans living with type 2 diabetes are undiagnosed. In addition, the age-standardized total prevalence of type 2 diabetes (both diagnosed and undiagnosed) is 21%, which is comparable to total rates among Black Americans and Hispanic Americans, and nearly double the total rate among non-Hispanic Whites [1]. Diabetes is a growing public health crisis with a total economic cost of diagnosed diabetes of $327 billion in 2019—a 26% increase from $245 billion in 2012 [2]. Healthcare costs are 2.3 times greater for Americans with diabetes compared to those without diabetes. Type 2 diabetes is the seventh most common cause of death in the USA, but it is the fifth most common cause of death for Asian Americans [3].
Asian Americans have long been overlooked in health-related research compared to other racial and ethnic groups [1]. As an aggregate, they are the fastest growing racial group with an expected growth of 213% by 2050 (compared to the 49% growth rate of the USA) [4]. Disaggregating Asian American ethnic groups illuminates considerable health inequities, including diabetes prevalence rates [5–7]. Limited research has shown that Hmong Americans experience worse health outcomes compared to other Asian Americans. While one systematic review revealed that the Hmong American prevalence rate for “diagnosed” diabetes is 11.3% compared to 6% non-Hispanic White [8], a study in Wisconsin revealed the Hmong American diabetes prevalence rate is 19.1% compared to 7.8% for non-Hispanic White [9]. Some studies found the risk for diabetes is over three times higher for Hmong Americans compared to non-Hispanic White when adjusted for age, sex, BMI, and insurance [9, 10]. More important to note is that developing type 2 diabetes seems to be a local phenomenon for Hmong Americans, who are 20 times more likely to develop diabetes than Hmong living in Thailand [10].
The older Hmong in this study were resettled in the USA as part of a refugee program between the years 1975 and 2004. When the USA pulled out of Vietnam in 1975, the Hmong fled Laos to Thailand to escape genocide after aiding the US CIA in a “Secret War” during the Vietnam War. After more than 45 years, the 2019 US Census estimated that the Hmong American population was 327,000 [11]. The Hmong Americans’ transition from their traditional and agrarian lifestyle to a westernized US lifestyle resulted in a decrease in physical activity and a new diet consisting of high carbohydrates, high-fats, and low fiber [12, 13]. This has increased their risk for diseases such as diabetes, hypertension, heart disease, and obesity. This phenomenon, termed the “New World Syndrome,” also affected other populations who experienced a similar transition from traditional to westernized lifestyles (e.g., Native Americans, and Hispanic Americans) [10, 14, 15].
Very few studies examine the health inequities existing in Asian American communities [16], resulting in a lack of national studies that include Hmong Americans [17]. With the recency of Hmong migration to the USA (especially compared to East and South Asians), there is a dearth of literature on this population, including the lack of attention on the cultural and language barriers of the Hmong population living in the USA [12, 18–21] and the lack of cultural understanding and sensitivity to providing care for Hmong American patients [22–24]. These barriers significantly contribute to the high prevalence rate of diabetes in the Hmong American population.
This pilot study explored the needs and challenges of diabetes management for older Hmong American patients with minimal English language skills through the experiences of their bilingual adult children caregivers, bilingual Hmong case managers, and healthcare providers. This study aims to provide a better understanding of the health needs of and bring cultural awareness of the Hmong American community to the forefront. Recognizing cultural differences and barriers will enable healthcare providers to improve treatment options for these older patients, thus bridging the gap between older Hmong patients and Western medicine.
Methods
This qualitative study explored the experiences of three groups involved in the care of helping older Hmong patients with minimal English language skills manage their type 2 diabetes: family (caregivers), community (case managers), and health care (providers). We used purposive sampling to recruit participants from clinics and community settings [25]. We also used a snowball sampling technique by asking participants to refer other potential participants fitting our inclusion criteria [25].
This study was approved by the Health Sciences Institutional Review Board at the University of Wisconsin-Madison with a waiver of signed consent. All participants were emailed an informed consent to review, along with the confirmation time for an interview. We received verbal consent and answered any questions regarding the study at the beginning of the interview. Two Hmong undergraduate students completed their human subjects training and were HIPAA certified to assist in this pilot study.
Eligibility criteria included the following: (1) being bilingual in Hmong and English, (2) having a parent/patient with diabetes and with minimal English language skills, and (3) having previously cared for/worked with or are currently caring for or working with Hmong elders or patients (Hmong elder is a term of deference used within the family system and community). The final sample size was reached once theme saturation to the responses became redundant and attempts to uncover new themes failed to reveal novel data for each group [26]. Participants were compensated $50 in appreciation for their time.
Data Collection
Semi-structured interviews were conducted between December 2020 and May 2021. In following COVID-19 restrictions and guidelines, interviews were conducted over a secure Zoom link connected to the university. All interviews were audio-recorded and conducted in English. Interview times ranged from 30 to 90 min. Table 1 describes the categories of questions in the semi-structured interview guide. The interview guide was adapted slightly for each group of participants.Table 1 Categories of questions asked in semi-structured interview
Categories
Influence of culture on diabetes management
Challenges of diabetes management within a cultural context
Experiences and challenges of helping older Hmong patients with minimal English language skills manage their diabetes
Resources utilized during diabetes management
General healthcare experience
Data Analysis
All audio recordings of interviews were transcribed verbatim by a HIPAA-approved transcription services on campus. Transcripts were independently coded by study team members, who then met to review codes and categories. Discrepancies were discussed until a group consensus was reached.
Data was analyzed using directed content analysis [27, 28]. Categories identified during the interviews became the preliminary codebook. Each study team member independently coded each interview using the preliminary codebook. Discrepancies were resolved through group consensus. Categories and codes were refined and agreed upon by study team members using an iterative process to create a final codebook, which was used to code all transcripts. The codes then informed emerging themes. Each study member searched for patterns or themes across all interviews later defined and named for the master codebook.
Results
Participant Characteristics
We interviewed four caregivers, five community case managers, and four healthcare providers (N = 13). One participant declined to participate in the study. Ten of the 13 participants were Hmong (79%), 38% of the participants were female, and 62% were male. Table 2 reflects participant demographics.Table 2 Participant characteristics
N = 13
Characteristic Label N %
Gender Male
Female
5
8
38.5
61.5
Identification Caregiver*
Community case managers*
Healthcare provider**
4
5
4
30.8
38.5
30.8
Race and ethnicity Hmong American
Asian American
White American
10
1
2
76.9
7.7
15.4
*Caregivers and community case managers were all Hmong American participants
**Healthcare providers were comprised of one Hmong American, one Asian American, and two White American participants
We identified three major themes across all three groups of participants: adherence to culture, health inequity, and managing diabetes. The major themes were further divided into sub themes and elaborated below. Table 3 outlines the major themes and subthemes, including additional illustrative quotes.Table 3 Major themes, subthemes, and context of themes
Major themes Subthemes Context Additional illustrative quotes
Adherence to culture Hmong herbs and shaman Herbs are commonly used by the Hmong to treat ailments. The Hmong seek the aid of a shaman, which is one of their traditional practices. And sometimes, they like to use their herbal medication too, but some of those herbal medications [are not] guaranteed. I know many of them are good, but some of [them] are [not] effective so you have to take a lot.
Lack of trust in Western medicine The older Hmong patients are hesitant to seek care from hospitals, which leads them to continue seeking traditional medical help. And obviously trust is a big issue in the Hmong community, and they would tend to trust their family members more than, more than anybody else if they’re, you know. And that’s, that’s a big plus if you can educate the family and the patient trusts the family to give them the medications then that would be better overall…
Significance of rice White rice is a staple food in the diet of the Hmong people. They don’t think it’s the diet, they don’t think it’s because of the rice, they don’t think it’s, that they don’t, for like a balanced meal, they don’t believe in a balanced meal you know, they just eat rice with water, that’s health enough, that’s low calorie enough for them, that it shouldn’t’ spike up their blood sugar.
Health inequity Language barrier The older Hmong patients experience a language barrier to English, resulting in health inequities. Because they really, most of my patients I would say are illiterate and not just in English but in Hmong as well. So, it’s difficult to, you know to treat them because of the compliance issues. And they don’t really understand, a lot of the English words that we would try to, would try to educate them. Overall, yeah it’s a compliance issue.
Cultural insensitivity Lack of cultural sensitivity from Western treatment continues to burden the older Hmong patients, leading to health inequities from lack of compliance to treatment options. So, a common problem that occurs in diabetes is diet and a common mistake that is made is referring to a culturally non-informed dietician who basically says stop eating so much rice.
Health literacy Many older Hmong patients (with traditional health views and a lack of exposure to the practice of modern medicine) have lower levels of health literacy. I think they just focus on, okay, diabetes is blood sugar, so the main, the main word is sugar, so they think minimizing just any sweets and sugar is good for them. And so I think we do lack the education.
Managing T2DM Monitoring glucose The lack of confidence, physical touch of the lancet (a source of pain), and embarrassment makes glucose monitoring an unpleasant, albeit necessary, experience for the Hmong patients. I think a lot of the time they don’t understand why they have to do it like every, like that much, you know. And so, and it’s all the poking too, you know. And it is painful for some of them, so.
Medicine compliance The older Hmong patients often stop taking their medication upon experiencing symptoms associated with the medication or when they feel “better,” resulting in incomplete medication compliance. Medication compliance resumes when the older Hmong patients’ diabetes worsens, or they experience severe symptoms. In the older generation, their experience and their belief set is that when you feel bad, you take a medicine and it’s a good medicine if it makes you feel better fairly promptly. When you feel good, you don’t take a medicine, because you don’t need it and taking medicines when you feel good will harm your body, not help it.
Nutrition The older Hmong patients do not know foods to eat or not to eat. Their understanding of diet differs from the Western healthy dietary guidelines. They don’t think it’s the diet, they don’t think it’s because of the rice, they don’t think it’s, that they don’t, for like a balanced meal, they don’t believe in a balanced meal you know, they just eat rice with water, that’s health enough, that’s low calorie enough for them, that it shouldn’t’ spike up their blood sugar. So I think just the food, they don’t believe that it’s the food that, that’s why there’s blood sugar.
Adherence to Culture
For the older Hmong patients, language barriers limit their ability to acculturate and integrate into American society, including the western health care system [29–31]. Difficulty integrating also means that the older Hmong are more likely to maintain traditional cultural practices. Participants addressed this adherence to culture in three traditional practices and beliefs: herbs and shamans, a lack of trust in western medicine, and rice.
Hmong Herbs and Shaman
Caregiver participants shared ways in which the Hmong culture was involved in their parent’s diabetes management. They talked about how their parents utilized Hmong herbs and consulted with a shaman. Participants spoke about the importance of using traditional Hmong herbs and in their parent’s diabetes management. One participant stated:The herbals they have that they used for many years that’s passed on from their parents to them. So, it’s the things that they are not only familiar with but trust that it has helped them with all the symptoms. So- and I say try to help, because to them, they think that that’s helpful, but in some ways it may not...
Lack of Trust in Western Medicine
Many participants shared that their elderly patients are often hesitant to seek care from a hospital regarding their diabetes. Participants expressed that hospital visits are intimidating for the older Hmong patients, who question the effectiveness of the prescribed medication. However, each of the participant groups stated a different aspect of the older Hmong patient’s lack of trust in Western medicine. Case managers often escort their clients to clinic visits and many expressed similar concerns as this one case manager stated:I feel like they’re still a little bit hesitant about the western medication, which can be, you know, scary at times for them.
The caregivers also expressed that trust is a major issue in the Hmong community.And obviously trust is a big issue in the Hmong community, and they would tend to trust their family members more than, more than anybody else if they’re, you know. And that’s, that’s a big plus if you can educate the family and the patient trusts the family to give them the medications then that would be better overall...
Even providers mentioned the trust factor for their Hmong patients, particularly around medications when symptoms are minimal, but side effects are significant. This participant used this example and also added that even with suggestions on how to manage the diarrhea or if medications were switched afterwards, the element of trust is already lost:Say I really don’t notice much my blood sugar affecting me, but I agreed because my family pressures me to try the medicine and the metformin gives me diarrhea every day, I don’t trust your medicine. They cause me more problems than they do benefit.
Significance of Rice
Many Hmong elders prefer to eat traditional Hmong food on a daily basis, of which white rice is a staple and considered part of the balanced meal. A traditional Hmong diet consists of white rice along with a side dish. During the interview, the majority of the participants mentioned the impact of white rice on the Hmong culture and diet, making it difficult to substitute. One participant stated:If you tell a Hmong elder, don’t eat rice. It’s like what, you trying to kill me?
Health Inequity
While multiple factors contribute to the health inequities experienced by the older Hmong patients, three major topics that participants continuously raised were language barriers, cultural insensitivity, and health literacy.
Language Barriers
All participants talked about how the older Hmong patients faced challenges due to their minimal English language skills. Language barriers were exceptionally prominent between the elderly patients and the healthcare providers during hospital visits. Participants also spoke about how the older patients’ lack of English proficiency created challenges in managing their diabetes. One caregiver participant stated:One of the challenge is that my parents, they do not know English. And they do, they have no idea about what the doctor talk about, they have no idea sometime, you explain to them, they do not understand.
Cultural Insensitivity
Participants expressed the persistent lack of cultural awareness in the healthcare field towards Hmong and other minority patients. Hmong participants stated that some concepts, such as the “Asian” diet, are often not accounted for during the treatment plan of the patients. Comments such as this provided by one caregiver participant were mentioned by other Hmong participants:I met with the diabetic educator with my mom, it was not helpful at all, because the foods were completely different that the Hmong people normally consume on a daily basis.
Health Literacy
When asked about challenges to help Hmong elders manage their diabetes, the caregivers and case managers spoke about their parents’/clients’ lack of understanding of diabetes and chronic diseases. One caregiver stated:I think engrained in pretty much the Hmong culture is you take the medication and all of a sudden your disease is gone, so the concept of chronic disease was very difficult for them to understand that you have to take this medication daily to prevent the progression of your disease.
Managing Diabetes
Diabetes is a complex chronic disease that is already difficult to convey especially when symptoms are not always visible. Participants shared three areas most challenging for Hmong elders to manage their diabetes: monitoring their glucose, medicine compliance, and nutrition.
Monitoring Glucose
Caregivers expressed that their elderly parents found that pricking of the finger to monitor their blood glucose to be an unpleasant experience resulting in resistance to checking their glucose. Despite the unpleasantry, the caregivers found that routinely checking the blood glucose made managing their parent’s diabetes easier. One participant stated:Even though you feel perfectly fine, you actually may possibly have diabetes. Having the blood sugar monitor was something that was really helpful in helping with the diagnosis and when she started medications, she could check her blood sugar and see that it would come down, the symptoms that she was having of diabetes that she didn’t know were symptoms of diabetes were getting better.
Medicine Compliance
Community case managers and providers addressed challenges and barriers to managing the diabetes of their Hmong patients and clients. They spoke about the signs of resistance against treatment plans. In addition, there was mention of the inconsistency of the patients taking their medications as prescribed by their healthcare provider.Yes, we have clients that- usually it’s right at the beginning when they are diagnosed- they don’t believe it because they’re not that sick yet. They’ll take it maybe like for a month or two and then they start feeling better and they stop taking it and then they start getting sick again and then it shows that they’re numbers are going up and they’ll take it again.
Nutrition
All Hmong participants expressed that the elderly patients lack an understanding of the foods that they should and should not be consuming as part of their diabetes management.They just don’t know what they’re supposed to eat. So sometimes because they don’t know what they’re supposed to eat, then my clients just starve to hopefully see their A1C drop.
According to the participants, the Hmong elders’ low health literacy around diabetes management, particularly around nutrition, often conflicted with a traditional Hmong diet.
Discussion
Older Hmong patients face ongoing issues in properly managing their diabetes. Results from this pilot study suggest that adherence to Hmong culture and ongoing health inequity both contribute significantly towards this issue. Hmong participants expressed that the Hmong elders still heavily rely on the use of traditional practices such as Hmong herbs to treat their diabetes. Continued adherence to their cultural practices regarding health after decades of living in the US signals a lack of trust in western medicine. Many of the health inequities discussed by the participants resulted from language and cultural barriers, cultural insensitivity from the health care system, and patient health literacy.
The patients’ cultural adherence influences their trust and understanding of Western medicine and contributes to their heavy reliance on shamans and Hmong herbs. The lack of trust and understanding of Western medicine increased patients’ hesitancy in complying with prescribed treatment plans by their doctor aligns with other published studies [17, 32–34]. Participants’ descriptions of how the Hmong elders adhere to cultural traditional health practices and diet reflects their inability to manage diabetes is also confirmed in other studies that found older patients were more likely to treat themselves using herbal medicine before seeking care from a doctor [12, 17, 33, 35].
We found that language barriers contributed to their limited health literacy. Health literacy is a crucial aspect of healthcare and significantly impacts the treatment plans of all patients, but especially for patients experiencing language barriers with their providers and the health care system. Similar to our findings, Khuu et al. (2018) found a lack of understanding of health information and a low health literacy rate among Hmong American patients. A scoping review of Asian Americans and diabetes reported that the Hmong had the lowest health literacy regarding diabetes leading to misunderstanding, misconceptions, confusion, and lack of knowledge around this chronic disease [17, 35].
All participants addressed the issue about the lack of nutritional education and understanding of the appropriate foods to help manage diabetes. The older Hmong patients literally understand diabetes to mean sugar in the blood and do not understand why rice needs to be eliminated from their diet. This often results in some patients taking extreme measures to reduce their glucose by starving themselves. Culhane-Pera et al. (2007) found similar results in how the Hmong approach their diet. Our interviews also revealed that cultural insensitivity from the health care system significantly impacted patient response to their treatment plans and heavily contributed to the challenges for older Hmong patients. Non-Hmong nutritionists and some providers consistently advised the older Hmong patients to eliminate rice from their diet without understanding that they did not follow what is considered a “typical” American diet. Hmong participants stated that the Hmong elders preferred to prioritize rice over other nutrients. Franzen and Smith (2009) also found that the importance of rice to the Hmong diet makes it difficult for the Hmong to alter their dietary behavior. Hmong elders’ reluctance to alter their diet is consistent with other published studies [34].
Our findings suggest that a major contributing factor towards the quality of care that the Hmong receive is cultural insensitivity. Western healthcare can take steps to acknowledge the cultural sensitivity of the Hmong population such as how culture influences diet and understanding of health and disease. The Hmong are one of many Asian ethnic groups who experience continued cultural insensitivity from western health care providers regarding their diet and culture [22].
There were several limitations present in this pilot study. First, utilizing snowball sampling makes it difficult to generalize the data and it may not be representative of the whole Hmong American population or the Asian American community. Also, the Hmong community in Wisconsin may differ in some respects from the Hmong communities in California and Minnesota where the larger population increases the availability of more concordant care. Second, although we reached data saturation, given the smaller sample size for each interview group—caregiver, community case managers, healthcare provider—makes it difficult to generalize from the data and may also not be representative of the experiences of each group. Future research could expand the sample population to include the older Hmong patients and provide for a more comprehensive study.
Conclusion
This pilot study aimed to provide a better understanding of the health needs of and bring cultural awareness of the Hmong American population to the forefront. The lack of literature on the health of Hmong Americans also makes this study of greater importance. Hmong Americans are an at-risk group for developing type 2 diabetes mellitus. However, the health inequities, including those arising from language and cultural barriers between the patient and provider, significantly impacts the older Hmong patients’ ability to manage their diabetes and the care they receive from their healthcare provider. Oftentimes, their medical decisions are heavily dependent on their cultural practices as they have maintained a strong adherence to their culture based on our findings. Our findings also reveal that the barriers posed by discordant care and language barriers pose a major challenge for older Hmong patients. Recognizing these cultural differences and barriers will enable healthcare providers to improve and cater to the treatment of these older patients, thus bridging the gap between older Hmong patients and Western medicine and reducing the current health inequities. Our findings suggest that these issues are not cut and dry, but rather, complex.
Funding
This work was supported by the UW Institute for Clinical and Translational Research startup funds: PRIV-20170808.
Declarations
Ethical Approval
This study was approved by the University of Wisconsin-Madison Health Sciences Institutional Review Board and was considered a minimal risk study.
Competing Interests
The authors declare no competing interests.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Menke A Casagrande S Geiss L Cowie CC Prevalence of and trends in diabetes in the United States, 1988–2012 JAMA 2015 314 10 1021 1029 10.1001/jama.2015.10029 26348752
2. [ADA] ADA. The Cost of Diabetes 2021 [Available from: https://www.diabetes.org/about-us/statistics/cost-diabetes.
3. [CDC] CfDCaP. Diabetes and Asian Americans 2016 [Available from: https://www.cdc.gov/diabetes/library/spotlights/diabetes-asian-americans.html.
4. Budimen A, Ruiz NG. Asian Americans are the fastest growing racial or ethnic group n the U.S. Washington, D.C.: Pew Research Center; 2021. https://www.pewresearch.org/fact-tank/2021/04/09/asian-americans-are-the-fastest-growing-racial-or-ethnicgroup-in-the-u-s/.
5. Hastings KG Jose PO Kapphahan KI Frank ATH Goldstein BA Thompson CA Leading causes of death among Asian American subgroups (2003–2011) PLOS ONE 2015 10 4 e0124341 10.1371/journal.pone.0124341 25915940
6. Miller BA Chu KC Hankey BF Ries LAG Cancer incidence and mortality patterns among specific Asian and Pacific Islander populations in the U.S Cancer Causes Control 2008 19 227 56 10.1007/s10552-007-9088-3 18066673
7. Stewart SL Dang J Chen MS Diabetes prevalence and risk factors in four Asian American communities J Community Health 2016 41 6 1264 1273 10.1007/s10900-016-0214-6 27339064
8. Lor M. Systematic review: Health promotion and disease prevention among Hmong adults in the United States. J Racial Ethn health inequities. 2018;5(3). 10.1007/s40615-017-0410-9.
9. Thao KK Arndt B Tandias A Hanrahan L The prevalence of type 2 diabetes mellitus in a Wisconsin Hmong patient population WMJ 2015 114 5 190 195 26726339
10. Her C Mundt M Risk prevalence for type 2 diabetes mellitus in adult Hmong in Wisconsin: a pilot study WMJ 2005 104 5 70 77
11. Budimen A. Hmong in the U.S: Hmong population in the U.S. 2000–2019. Pew Research Center. https://www.pewresearch.org/social-trends/fact-sheet/asian-americans-hmong-in-the-u-s/#hmong-populationin-the-u-s-2000-2019.
12. Culhane-Pera KA Her C Her B "We are out of balance here": a Hmong cultural model of diabetes J Immigr Minor Health 2007 9 3 179 190 10.1007/s10903-006-9029-3 17245657
13. Franzen L Smith C Acculturation and environmental change impacts dietary habits among adult Hmong Appetite 2009 52 1 173 183 10.1016/j.appet.2008.09.012 18848592
14. Abate N Chandalia M The impact of ethnicity on type 2 diabetes J Diabetes Complicat 2003 17 1 39 58 10.1016/s1056-8727(02)00190-3
15. McCarty DJ Glucose intolerance in Wisconsin's Hmong population Wis Med J 2005 104 5 13 14
16. Gee GC Ro A Shariff-Marco S Chae D Racial discrimination and health among Asian Americans: evidence, assessment, and directions for future research Epidemiol Rev 2009 31 1 130 151 10.1093/epirev/mxp009 19805401
17. Mitchell-Brown F Newman S Diabetes mellitus and the Hmong: a scoping review of the literature Californian J Health Promot 2015 13 3 55 65 10.32398/cjhp.v13i3.1835
18. Perez MA, Thao C. Understanding barriers to prevention of ntshav qab zib / nsthaav qaab zib : A Hmong perspective. Hmong Stud J. 2009;10:1–23.
19. Yang A Xiong D Vang E Pharris MD Hmong American women living with diabetes J Nurs Scholarsh 2009 41 2 139 148 10.1111/j.1547-5069.2009.01265.x 19538698
20. Lor M Xiong P Schwei R Bowers B Jacobs EA Limited English proficient Hmong- and Spanish-speaking patients' perceptions of the quality of interpreter services Int J Nurs Stud 2016 54 75 83 10.1016/j.ijnurstu.2015.03.019 25865517
21. Bengiamin M, Chang X, Capitman JA. Understanding traditional Hmong health and prenatal care beliefs, practices, utilization and needs. Central Valley Health Policy Institute; 2011. https://chhs.fresnostate.edu/cvhpi/documents/hmong-report.pdf.
22. Uba L Cultural barriers to health care for Southeast Asian refugees Public Health Rep 1992 107 544 548 1410235
23. Bruce KH Schwei RJ Park LS Jacobs EA Barriers and facilitators to preventive cancer screening in Limited English Proficient (LEP) patients: physicians' perspectives Commun Med 2014 11 3 235 247 10.1558/cam.v11i3.24051 27499725
24. Frazen L Smith C Acculturation and environmental change impacts dietary habits among adult Hmong Appetite 2009 52 1 173 183 10.1016/j.appet.2008.09.012 18848592
25. Creswell JW, Plano Clark VL. Designing and conducting mixed methods research. Second edition. ed. Los Angeles: SAGE Publications; 2011. xxvi, 457 pages p.
26. Saunders B Sim J Kingstone T Baker S Waterfield J Bartlam B Burroughs H Jinks C Saturation in qualitative research: exploring its conceptualization and operationalization Qual Quant: Int J Methodol 2018 52 1893 1907 10.1007/s11135-017-0574-8
27. Graneheim UH Lundman B Qualitative content analysis in nursing research: concepts, procedures, and measure to achieve tustworthiness Nurse Educ Today 2004 24 105 112 10.1016/j.nedt.2003.10.001 14769454
28. Hsieh H-F Shannon SE Three approaches to qualitative content analysis Qual Health Res 2005 15 1277 1288 10.1177/1049732305276687 16204405
29. Saha S Fernandez A Language barriers in health care J Gen Intern Med 2007 22 Suppl 2 281 282 10.1007/s11606-007-0373-3 17924172
30. Norton B Language, identity, and the ownership of English TESOL Q 1997 31 409 429 10.2307/3587831
31. Park L Schwei RJ Xiong P Jacobs EA Addressing cultural determinants of health for Latino and Hmong patients with limited English proficiency: practical strategies to reduce health inequities J Racial Ethn Health Disparities 2018 5 3 536 544 10.1007/s40615-017-0396-3 28791616
32. Thornburn S Kue J Keon KL Lo P Medical mistrust and discrimination in health care: a qualitative study of Hmong women and men J Community Health 2012 37 4 822 829 10.1007/s10900-011-9516-x 22116737
33. Fang DM Stewart SL Social-cultural, traditional beliefs, and health system barriers of hepatitis B screening among Hmong Americans: a case study Cancer 2018 124 Suppl 7 1576 1582 10.1002/cncr.31096 29578596
34. Xiong P Westberg SM Perspectives from the Hmong population on Type 2 Diabetes Innov Pharm 2012 3 4 1 5 10.24926/iip.v3i4.280 22844651
35. Lor M Xiong P Park L Schwei RJ Jacobs EA Western or traditional healers? Understanding decision making in the Hmong population West J Nurs Res 2017 39 3 400 415 10.1177/0193945916636484 26941160
| 36512312 | PMC9746559 | NO-CC CODE | 2022-12-15 23:21:56 | no | J Racial Ethn Health Disparities. 2022 Dec 13;:1-8 | utf-8 | J Racial Ethn Health Disparities | 2,022 | 10.1007/s40615-022-01480-7 | oa_other |
==== Front
Br Dent J
Br Dent J
British Dental Journal
0007-0610
1476-5373
Nature Publishing Group UK London
5357
10.1038/s41415-022-5357-5
Research
A survey of mental wellbeing and stress among dental therapists and hygienists in South West England
Hallett Georgia [email protected]
4141524254001
Witton Robert 4141524254002
Mills Ian 4141524254002
4141524254001 grid.439442.c 0000 0004 0474 1025 Health Education England Southwest Dental Career Development Fellow, Special Care Dentistry, Torbay and South Devon NHS Foundation Trust, Castle Circus Health Centre, Torquay, TQ2 5YH, UK
4141524254002 grid.11201.33 0000 0001 2219 0747 Peninsula Dental School, University of Plymouth, Drake Circus, Plymouth, PL4 8AA, UK
13 12 2022
16
6 7 2022
15 8 2022
© The Author(s), under exclusive licence to the British Dental Association 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Introduction Mental health and wellbeing of the dental team has been brought into sharp focus during the COVID-19 pandemic. Despite this renewed interest, there has been longstanding issues with poor mental health and wellbeing in the dental profession for some time. While there is some evidence that documents poor mental wellbeing amongst dentists, there appears to be a lack of evidence concerning dental care professionals.
Aims To explore the level of mental wellbeing and stress amongst dental hygienists and therapists (DHTs) in South West England.
Method An online survey was distributed to DHTs in South West England via two professional networks.
Results A total of 129 surveys were completed. The mean levels of reported wellbeing were lower amongst DHTs than the general population and 45% of respondents reported high anxiety levels. Younger respondents reported lower levels of life satisfaction. Plus, 43.5% of dental therapists reported performing solely dental hygiene treatments, with those performing no dental therapy reporting lower happiness levels.
Conclusion Low mental wellbeing amongst DHTs in the South West has been identified in this survey and this is likely to impact negatively on the morale and motivation of the workforce, leading to increased levels of absenteeism and ultimately, loss of colleagues from the dental workforce. The stress encountered by DHTs is largely workplace-related and therefore, there is an increased need for team- and organisation-delivered interventions to improve mental wellbeing for this group.
Key points
Dental therapists and hygienists in South West England experience low levels of mental wellbeing compared to the general population.
The stress experienced by dental therapists and hygienists is predominantly workplace-centred.
A large proportion of the dental therapy profession in South West England are not working to their full scope of practice. Efforts to improve this may be beneficial for their mental wellbeing and would bring wider benefits to team-working, patient care and potentially NHS access.
==== Body
pmcIntroduction
It is widely accepted that dentistry is a stressful occupation,1,2,3,4 with a recent report suggesting that 'the high levels of self-reported stress, burnout and psychological distress…are a serious concern to the profession'.2 The recent COVID-19 pandemic has undoubtedly led to many additional challenges within the dental working environment.1,5 This has exacerbated existing stressors, such as financial pressures, NHS targets, staffing, time management and patient complaints.6,7 It has recently been reported that the impact of nation-wide increases in hospitalisations, death rates and tightening national restrictions during the course of the pandemic, in addition to personal COVID-19-related trauma, resulted in decreased psychological resources available to dental staff, which contributed to increased levels of fatigue, burnout and depressive symptoms in the workforce.8,9 These findings were not consistent across all dental staff and a great deal of variance in response to the pandemic was recorded, emphasising the importance of a tailored response to the emotional demands associated with dental care provision.9
Deterioration in mental health and wellbeing was evident before the COVID-19 pandemic, with a recent General Dental Council (GDC) report highlighting the increasing number of dentists demonstrating signs of burnout, poorer wellbeing and suicide ideation over the last decade.1,2Reports on stress among general dental practitioners have identified numerous systemic stressors, including time limitations, working environment, the NHS contract, fear of regulation and litigation, unrealistically high workload and patient issues.3,4,10,11The risk of litigation or regulatory chastisement have increasingly been recognised as major stressors for many, with dentists reported to be operating 'under constant fear of persecution'.2
Anxiety, stress or mental-health-related issues can impact significantly on clinical performance and although the evidence in dentistry is sparse, research in other areas of health care demonstrate a strong link between stress and impaired surgical competence and communication,12 with issues of burnout resulting in compromised work performance, absenteeism13 and worsening patient safety.14,15 Burnout can often lead to depression and stress increases the risk of developing mental health conditions, such as depression, alcoholism and drug addiction.3,4,16,17,18 Mental health and wellness of the dental team is critical in maintaining and retaining a healthy workforce. A recent report published by Dental Protection suggests that as many as half of dentists have considered leaving the profession due to concerns over their own personal wellbeing.19
Despite a growing evidence base relating to the mental health and wellbeing of dentists, the GDC report demonstrates a lack of evidence regarding the levels of mental wellbeing among other members of the dental team. Dental care professionals (DCPs) account for 63% of GDC registrants,20 encompassing a wide range of professional groups, each with a very individual skill set. In a study conducted among Northern Irish dentists and DCPs in 2011, it was reported that 20% of DCPs experienced psychological ill-health; however, this study fails to differentiate between the professional groups under the umbrella term of DCPs.21 A decade later, during the COVID-19 pandemic, mental wellbeing was again assessed in a UK dental hospital, which revealed that 53.3% of staff demonstrated symptoms of generalised anxiety.5 This study, undertaken in 2020, differentiates between DCPs, with dental nurses demonstrating the highest anxiety levels; however, no dental therapists or hygienists were recruited in this study.5
Dental therapists and hygienists (DHTs) account for a significant proportion of DCPs, yet this professional group would appear to be under-represented in the academic literature. An annual increase in the number of registrants and a desire for DCPs to contribute more effectively to NHS provision will reinforce the relevance of research across all dental professional groups to ensure their needs are met within future mental wellness strategies.
In order to provide safe and effective health care, organisations must protect their staff against burnout and emotional exhaustion through targeted intervention and prevention strategies.15 Within the dental sector, the Dental Professional Alliance has introduced the Mental wellness in dentistry framework, which enables dental workplaces to prioritise staff mental health.22 The framework provides guidance on how to promote mental wellbeing among the whole dental team and places emphasis on the importance of prevention and early intervention. This aligns with the NHS Long Term Plan, which aims to improve the NHS as a workplace and improve support offered to staff.23
This report aims to improve the evidence base, particularly in regards to prevalence, for mental wellbeing and stress among DHTs to better inform future practice and prevention strategies.
Methodology
Sample
An online survey was distributed to DHTs in the South West England branches of the British Association of Dental Therapists (BADT) and the British Society of Dental Hygiene and Therapy (BSDHT). The survey was distributed via a mailing list to the 207 members of the BADT (79 members) and BSDHT (128 members) in the South West region. The survey was also publicised on social media, including an open invitation on professional networking groups. Prior to distribution, the survey was piloted among DHTs practising in other regions of the UK.
The survey was launched at the start of December 2021 and was open for eight weeks, with a reminder email sent in the first week of January 2022. Details of the study, follow-up support, their right to withdraw at any time and a consent statement were included in the survey. Return of the survey was taken as consent to the process.
Participants
Survey participants were required to be registered DHTs currently providing dental services in the seven integrated care systems of South West England which comprises: Bristol, Cornwall, Dorset, Devon, Gloucestershire, Somerset and Wiltshire. Students and retired participants were excluded from the sample. Similarly, those delivering care outside of South West England were not included, regardless of home address.
Materials
The survey was distributed online using the Microsoft Forms platform and included questions related to:Basic demographics
Working patterns
Qualification status
Wellbeing
Wellbeing in response to COVID-19 pandemic stress.
The survey was designed to incorporate a similar structure and pre-validated domains used in previous surveys of dentists, with permission sought and granted from previous survey authors. The Office for National Statistics (ONS)-4 tool was used to measure wellbeing among the respondents.24 This is divided into four questions on a scale of 0-10, with questions relating to life satisfaction, worthwhileness, happiness and anxiety.
Questions related to life satisfaction, worthwhileness and happiness were scored as 'low' if respondents reported scores of 0-4.24 Anxiety was scored as 'high' if respondents reported scores of between 6-10.24
Following the questions relating to life satisfaction, worthwhileness and happiness, respondents were asked if they thought their answers would have been different before the COVID-19 pandemic. These were grouped into: 'yes - higher scores since COVID-19 pandemic'; 'yes - lower scores since COVID-19 pandemic'; 'no - the same'; and 'not sure'. The same question was asked following the anxiety section of the ONS-4 measure.
Stress
A single measure of stress was used, which has been utilised in many large-scale national surveys, which asked respondents: 'in general, how do you find your job?'.25 The results of which were recorded on a Likert-scale, ranging from 'not at all stressful' to 'extremely stressful'.
The work stress in dentistry (WSID) measure, developed by Cooper et al.,26 was also used to identify stressors in the workplace. Following the use of this measure by Collin et al., questions related to litigation and regulation were again utilised.2 The questions were answered on a Likert-scale from 'not at all stressful' to 'extremely stressful' and were grouped into the following categories: work pressure; pay related; work content; dealing with patients; and litigation and regulation. An additional response of 'not applicable' was included, as some areas may not be applicable to all respondents based on working patterns.
Finally, an open-ended free response question was included, asking: 'are there other areas of your work that you feel are stressful?'.
Completion of the Health Research Authority decision tool determined that ethical approval for the survey was not required. This was a cross-sectional survey design with recruitment via professional groups and no personal identifiable data collected. Appropriate consent was obtained from each participant for use of their anonymous data.
Results
Overall, there were 129 responses to the survey, representing a response rate of 62.3%.
Of the 129 responses, four respondents were not currently practising as dental therapists or hygienists and were therefore excluded from the final data analysis. Similarly, 11 respondents failed to complete the full survey and one respondent was working outside the area of interest and were similarly excluded. Therefore, 113 responses were used in the final analysis.
Frequency analyses were carried out to describe respondent characteristics and demographics using SPSS Statistics v28. Two-sided Pearson's chi-squared tests and Fischer's exact tests were used to compare variables. Statistical significance was set with a p-value of 0.05. Figure 1 is presented with error bars set at 95% confidence intervals.Fig. 1 Measures of wellbeing comparing dental therapists and hygienists, the general population from April to June 2021,27 and dentists2 (with 95% confidence intervals for dental therapists and hygienists)
Sample demographics are detailed in Table 1.Table 1 Characteristics of respondents
Respondent characteristics Frequency Percentage
Age Under 25 3 3%
25-34 29 26%
35-44 27 24%
45-54 31 27%
55-64 22 19%
65+ 1 1%
Sex Male (including transgender men) 2 2%
Female (including transgender women) 111 98%
Qualifications held BSc dental therapy and hygiene 26 23%
Diploma in dental therapy and hygiene 20 18%
Diploma in dental hygiene 64 57%
None of the above 3 3%
Years qualified Less than 5 years 26 23%
5-10 years 21 19%
10-20 years 20 18%
20-30 years 37 33%
40+ years 9 8%
Region of qualification Scotland 6 5%
Wales 15 13%
Northern Ireland 1 1%
North West England 4 4%
North East England 6 5%
Midlands 3 3%
South East England 38 34%
South West England 38 34%
Overseas 2 2%
With regards to field of practice, 85% of respondents worked primarily in general dental practice, with 8% based in teaching or research institutions, 2.7% in the community service, 1.8% in the armed forces, 1.8% in hospital services and 0.9% in specialist referral practices.
From the sampled dental therapists, 43.5% (n = 20) reported performing dental hygiene treatments only, with 56.6% performing any aspect of dental therapy in their role.
Wellbeing
The mean score reported by DHTs for life satisfaction was 6.56 (standard deviation [SD] = 1.81), worthwhile was 6.8 (SD = 1.93), happiness 6.52 (SD = 2.39) and anxiety 4.82 (SD = 2.91).
Overall, 14% (n = 16) of respondents reported low life satisfaction scores, 12% (n = 14) of respondents reported low worthwhile scores and 22% (n = 25) reported low happiness scores. Further, 45% (n = 51) of those surveyed reported high anxiety levels.
The number of respondents reporting low life satisfaction, worthwhileness and happiness scores were analysed by age, the number of years post qualification and the scope of practice delivered in their role. When comparing low life satisfaction scores against age, a significant difference was observed between those aged 45 or under and those aged above 45 (p = 0.049), indicating a higher frequency of younger DHTs reporting low life satisfaction compared with their older colleagues.
Dental therapists who routinely provided care within their scope of practice reported similar levels of wellbeing compared to those who provided hygiene treatments only. Where dental therapists only provided hygiene treatments, they reported statistically significant lower levels of happiness (p = 0.038).
COVID-19 pandemic
In total, 35.4% (n = 40) of respondents reported that their scores for life satisfaction, worthwhileness and happiness are lower since the COVID-19 pandemic. Similarly, 46% (n = 52) reported that their anxiety scores are higher since the COVID-19 pandemic.
Stress
On the single item of stress measure, 37.2% (n = 42) of respondents reported scores equating to high stress levels.
No significant difference was observed in occupational stress levels between those who worked exclusively privately and those who performed elements of NHS care (p = 0.128), or between therapists who performed solely hygiene procedures and those who performed elements of dental therapy (p = 0.555).
Sources of stress
Sources of stress were examined using the WSID measure and this was used to determine the most frequent sources of stress at work reported as 'very stressful' or 'extremely stressful'. The top ten sources of stress from the survey can be seen in Table 2.Table 2 Top ten stressors reported by dental therapists and hygienists
Stressor % of respondents who responded 'very stressful/extremely stressful'
Running behind schedule 73%
Striving for perfection 65%
Late patients 62%
Equipment malfunction 61%
Working quickly to see as many patients as possible 59%
Working under constant time pressure 53%
Risk of making a mistake 51%
Red tape and bureaucracy 50%
Dissatisfied patients 49%
The threat of complaints 49%
Discussion
The findings of this survey indicate low levels of life satisfaction, worthwhileness and happiness in DHTs compared to the general population (Fig. 1) and comparatively higher levels of reported anxiety.27 This may, in part, be attributed to the COVID-19 pandemic, which has had far reaching effects on the dental sector. In this survey, 46% of respondents reported that their anxiety levels were higher since the start of the COVID-19 pandemic, a trend that is consistent in the general population, with reported work-related stress and anxiety increasing on pre-pandemic levels.28 Work-related anxiety, stress and depression accounted for 50% of work-related ill-health in the general population in 2020/2021.28 This is likely to be similar, if not higher, in the dental sector, where wellbeing is comparatively lower. Higher levels of staff suffering from work-related ill-health is likely to have wider implications for patient care delivery through higher levels of absenteeism and sick leave.
This survey has revealed that almost half of dental therapists are not working to their full scope of practice, with 43.5% performing hygiene treatments only. These results demonstrate the underutilisation of dental therapists in the South West region, a finding that is consistent across the UK29 and internationally.30 With a lack of opportunities to deliver their full scope of practice, therapists risk losing confidence and competence in the skills they have acquired. This can have negative consequences at many different levels, including effective team-working, patient care, future career development, professional fulfilment and personal mental health and wellbeing. Dually qualified dental hygienist-therapists may choose to work as hygienists due to local market factors and systemic barriers within the existing NHS contract, which fails to incentivise the use of therapists.31
Our findings indicate that dental therapists who performed hygiene treatments only reported significantly lower levels of happiness than their colleagues who performed elements of dental therapy. It has been previously reported that among UK DHTs, the most important predictor for overall job satisfaction is the variety of clinical activity performed, with a lack of therapy procedures being a source of disappointment and frustration for many.32 In terms of care provision, it is reported that approximately 70% of routine care provided under the current general dental services NHS contract could be performed by dental therapists.33
At a time where access to NHS dentistry is under unprecedented pressure, it is unfortunate that we have a highly-skilled workforce who are being under-utilised due largely to failings within the current NHS dental contract. It is hoped that dental contract reform in England, when finally implemented, will support greater skill-mix, efficiency in care delivery and utilisation of all available workforce. There is clearly a sense of urgency to introduce change; however, there is some debate as to the extent to which this can be provided in the NHS without significant structural reform.34 Following the introduction of direct access for DHTs in 2013,35 two-thirds of those treating patients without prescription from a dentist felt that providing care in this manner increased their job satisfaction.34 It has also been reported that patients feel more satisfied when being treated by dental therapists as opposed to dentists36and full utilisation of a therapist's skill set is likely to bring improved satisfaction to not only the treating practitioner, but also to dentists, employers and patients.
While this survey demonstrates fewer DHTs suffer from high work-stress compared to dentists, the sources of stress encountered vary considerably. Collin et al.2 reported that four of the top five stressors experienced by dentists involved the fear of litigation and over-regulation: threat of complaints; risk of making a mistake; bureaucracy; and concern about the GDC. Comparing this to the top five stressors reported by DHTs in this survey (Table 2), it can be observed that these stressors are more focused within the workplace itself and are related to the structural and organisational factors experienced by therapists and hygienists. Gallagher et al. proposed a model for categorising wellbeing influences, including: macro-level factors relating to professional regulation and systems; meso-level factors related to workplace and job specification; and micro-level factors incorporating relationships and personal factors.37 Application of this model would indicate that the predominant influences on the wellbeing of DHTs occur at the meso-level and involve workplace factors, such as time pressures and patient factors, as opposed to dentists, whose predominant influences occur at macro-level (system pressure and risk of litigation). It might therefore be inferred that the high stress experienced by DHTs relates more to the working environment and practice owners and managers should consider which elements of support are likely to be the most effective.
Ensuring that staff within the profession have improved levels of mental wellbeing is vital for recruitment and retention. High levels of work stress and low wellbeing reported in this survey have the potential to cause recruitment and retention difficulties for key members of the dental team. The UK dental sector is facing significant challenges relating to staff retention, with 40% of dentists in a British Dental Association study anticipating a change in career or early retirement in the next 12 months.38 Similarly, it is anticipated that one-third of UK dental nurses also plan on leaving the profession in the next two years,39 resulting in a significantly depleted dental workforce.
In respect to this survey, younger DHTs reported significantly lower levels of life satisfaction than their older colleagues, which may reflect the lack of social interaction during the COVID-19 pandemic and the resultant isolation. These findings are mirrored in studies conducted over two decades ago, indicating there may be deeper seated reasons for this observation.40 Lower job satisfaction, disengagement and exhaustion are all factors that increase intent to leave the profession.41 This is already having implications for staff retention in the USA, where 8% of dental hygienists left the workforce at the start of the COVID-19 pandemic.42 Promoting wellbeing and improving job satisfaction is likely to slow the flow of dental professionals leaving the industry and this must be a priority for employers and commissioners.
There are limitations to this survey. The overall response rate of 62% was moderate, although this rate is consistent with other questionnaire-based studies of health professionals.43 The results may therefore be subject to selection bias and caution should be applied to their generalisability. There may also be some individuals who received the invitation to participate twice as they featured in both professional networks and although there is a theoretical risk of duplicate responses, we feel this is unlikely. There is no demographic data available locally to compare the profile of non-responders to the sampling frame and so the results must be interpreted with caution, as the issues identified here may not be representative of other DHTs locally or nationally in England. Nevertheless, this survey provides useful information on this important and underreported issue within the dental therapy and hygienist workforce.
Conclusion
This survey demonstrates that mental wellbeing among the dental therapy and hygiene community in South West England is poorer than the general population. Lower rates of wellbeing, higher stress levels and increased anxiety among the workforce is likely to result in increased rates of absenteeism and risks further loss of dental professionals from the workforce. The survey also highlights the underutilisation of dental therapists in the region, which is unlikely to improve without significant structural reform. Dentists' wellbeing is primarily influenced by regulatory pressures, while DHTs experience most stress in the workplace. Efforts must be concentrated to enhance wellbeing within the workplace to improve mental wellbeing at a practice level. The introduction of the Mental wellness in dentistry framework22 is seen as an excellent resource, which is likely to have benefits for clinicians, their employers and the communities they serve.
Acknowledgements
We would like to thank the British Dental Association for sharing their survey and the University of Plymouth Medical School for statistical support.
Author contributions
The study was conceptualised by Georgia Hallett, Robert Witton and Ian Mills. The survey was developed by Georgia Hallett and reviewed by Robert Witton and Ian Mills. Data analysis was performed by Georgia Hallett with assistance from the University of Plymouth medical statistics team. The draft manuscript was prepared by Georgia Hallett, with contributions from Robert Witton and Ian Mills.
Ethics declaration
The authors declare no conflicts of interest.
Completion of the Health Research Authority decision tool determined that ethical approval for the survey was not required. This was a cross-sectional survey design with recruitment via professional groups and no personal identifiable data collected.
Details of the study, follow-up support, the right to withdraw at any time and a consent statement were included in the survey and return of the survey was taken as consent to the process. Appropriate consent was obtained from each participant for use of their anonymous data.
==== Refs
References
1. General Dental Council. Mental Health and Wellbeing in Dentistry: A Rapid Evidence Assessment. 2021. Available at https://www.gdc-uk.org/docs/default-source/research/mental-health-and-wellbeing-in-dentistry27973e06-eb0f-4ee2-b92f-7fee3d2baf5b.pdf?sfvrsn=511f2ef9_5 (accessed June 2022).
2. Collin V, Toon M, O'Selmo E, Reynolds L, Whitehead P. A survey of stress, burnout and well-being in UK dentists. Br Dent J 2019; 226: 40-49.
3. British Dental Association. The Mental Health and Well-being of UK Dentists: A Qualitative Study. 2017. Available at https://bda.org/about-the-bda/campaigns/Documents/The%20Mental%20Health%20and%20Well-being%20of%20UK%20Dentists.pdf (accessed November 2022).
4. Myers H L, Myers L B. 'It's difficult being a dentist': stress and health in the general dental practitioner. Br Dent J 2004; 197: 89-93.
5. Mahendran K, Patel S, Sproat C. Psychosocial effects of the COVID-19 pandemic on staff in a dental teaching hospital. Br Dent J 2020; 229: 127-132.
6. Dental Protection. Covid-19: Dental Protection survey reveals top 5 worries for dentists. 2020. Available at https://www.dentalprotection.org/uk/articles/covid-19-dental-protection-survey-reveals-top-5-worries-for-dentists (accessed July 2022).
7. Bretherton R, Chapman H R, Chipchase S. A study to explore specific stressors and coping strategies in primary dental care practice. Br Dent J 2016; 220: 471-478.
8. Humphris G, Knights J, Beaten L et al. Exploring the Effect of the COVID-19 Padenmic on the Dental Team: Preparedness, Psychological Impacts and Emotional Reactions. Front Oral Health 2021; 2: 669752.
9. Freeman R, Knights J, Beaton L et al. Prediction of emotional exhaustion over the course of the COVID-19 pandemic in primary care dental staff: an intensive longitudinal study. BMJ Open 2021; DOI: 10.1136/bmjopen-2021-054827.
10. Chapman H, Chipchase S Y, Bretherton R. Understanding emotionally relevant situations in primary dental practice. 2. Reported effects of emotionally charged situations. Br Dent J 2015; DOI: 10.1038/sj.bdj.2015.852.
11. British Dental Association. Sources of Work-related Stress among Salaried Primary Care Dentists in the UK. 2014. Available at https://bda.org/about-the-bda/campaigns/Documents/Sources%20of%20stress%20among%20salaried%20dentists.pdf (accessed November 2022).
12. Arora S, Sevdalis N, Nestel D, Woloshynowych M, Darzi A, Kneebone R. The impact of stress on surgical performance: a systematic review of the literature. Surgery 2010; 147: 318-330.
13. Dyrbye L N, Shanafelt T D, Johnson P O, Johnson L A, Satele D, West C P. A cross-sectional study exploring the relationship between burnout, absenteeism, and job performance among American nurses. BMC Nurs 2019; 18: 57.
14. Garcia C L, de Abreu L C, Ramos J L S et al. Influence of Burnout on Patient Safety: Systematic Review and Meta-Analysis. Medicina (Kaunas) 2019; 55: 553.
15. Hall L, Johnson J, Watt I, Tsipa A, O'Connor D B. Healthcare Staff Wellbeing, Burnout, and Patient Safety: A Systematic Review. PLoS ONE 2016; DOI: 10.1371/journal.pone.0159015.
16. Denton D A, Newton J T, Bower E J. Occupational burnout and work engagement: a national survey of dentists in the United Kingdom. Br Dent J 2008; DOI: 10.1038/sj.bdj.2008.654.
17. Hill K B, Burke F J T, Brown J et al. Dental practitioners and ill health retirement: a qualitative investigation into the causes and effects. Br Dent J 2010; DOI: 10.1038/sj.bdj.2010.814.
18. Kay E J, Lowe J C. A survey of stress levels, self-perceived health and health-related behaviours of UK dental practitioners in 2005. Br Dent J 2008; DOI: 10.1038/sj.bdj.2008.490.
19. Dental Protection. Half of dentists in the UK have considered leaving dentistry for reasons of personal wellbeing. 2020. Available at https://www.dentalprotection.org/uk/articles/half-of-dentists-in-the-uk-have-considered-leaving-dentistry-for-reasons-of-personal-wellbeing (accessed July 2022).
20. General Dental Council. Registration report - January 2022. 2022. Available at https://www.gdc-uk.org/docs/default-source/registration-reports/registration-report_january-2022.pdf?sfvrsn=bd140c06_5 (accessed June 2022).
21. Gorter R C, Freeman R. Burnout and engagement in relation with job demands and resources among dental staff in Northern Ireland. Community Dent Oral Epidemiol 2011; 39: 87-95.
22. Mental Health Wellness in Dentistry. Framework. 2022. Available at https://mhwd.org/framework/ (accessed June 2022).
23. NHS. The NHS Long Term Plan. 2019. Available at https://www.longtermplan.nhs.uk/publication/nhs-long-term-plan/ (accessed May 2022).
24. Office for National Statistics. Personal well-being user guidance. 2018. Available at https://www.ons.gov.uk/peoplepopulationandcommunity/wellbeing/methodologies/personalwellbeingsurveyuserguide (accessed February 2022).
25. Houdmont J, Cox T, Griffiths A. Work-related stress case definitions and prevalence rates in national surveys. Occup Med (Lond) 2010; 60: 658-661.
26. Cooper C L, Watts J, Baglioni Jr A J, Kelly M. Occupational stress among general practice dentists. J Occup Psychol 1988; 61: 163-174.
27. Office for National Statistics. Quarterly personal well-being estimates - non-seasonally adjusted. 2022. Available at https://www.ons.gov.uk/peoplepopulationandcommunity/wellbeing/datasets/quarterlypersonalwellbeingestimatesnonseasonally adjusted (accessed February 2022).
28. Health and Safety Executive. Work-related stress, anxiety or depression statistics in Great Britain, 2022. 2022. Available at https://www.hse.gov.uk/statistics/causdis/stress.pdf (accessed November 2022).
29. Godson J H, Williams S A, Csikar J I, Bradley S, Rowbotham J. Dental therapy in the United Kingdom: part 2. A survey of reported working practices. Br Dent J 2009; 207: 417-423.
30. Chen D, Hayes M, Holden A. A global review of the education and career pathways of dental therapists, dental hygienists and oral health therapists. Br Dent J 2021; 230: 533-538.
31. Williams S A, Bradley S, Godson J H, Csikar J I, Rowbotham J. Dental therapy in the United Kingdom: part 3. Financial aspects of current working practices. Br Dent J 2009; 207: 477-483.
32. Turner S, Ross M K, Ibbetson R J. Job satisfaction among dually qualified dental hygienist-therapists in UK primary care: a structural model. Br Dent J 2011; DOI: 10.1038/sj.bdj.2011.50.
33. Gallagher J E, Lim Z, Harper P R. Workforce skill mix: modelling the potential for dental therapists in state-funded primary dental care. Int Dent J 2013; 63: 57-64.
34. Turner S, Ross M. Direct access: how is it working? Br Dent J 2017; 222: 191-197.
35. General Dental Council. Direct Access. 2019. Available at https://www.gdc-uk.org/standards-guidance/standards-and-guidance/direct-access (accessed June 2022).
36. Sun N, Burnside G, Harris R. Patient satisfaction with care by dental therapists. Br Dent J 2010; DOI: 10.1038/sj.bdj.2010.209.
37. Gallagher J E, Colonio-Salazar F B, White S. Supporting dentists' health and wellbeing - workforce assets under stress: a qualitative study in England. Br Dent J 2021; DOI: 10.1038/s41415-021-3130-9.
38. Westgarth D. The dental workforce: An assessment of the recruitment market. BDJ In Pract 2022; 35: 14-18.
39. Dentistry. Dentistry Census - dentistry risks losing one third of UK's dental nurses in next two years. 2022. Available at https://dentistry.co.uk/2022/04/18/dentistry-census-dentistry-risks-losing-one-third-of-uks-dental-nurses-in-the-next-two-years/ (accessed May 2022).
40. Gibbons D, Corrigan M, Newton J. The working practices and job satisfaction of dental therapists: findings of a national survey. Br Dent J 2000; 189: 435-438.
41. Patel B M, Boyd L D, Vineyard J, LaSpina L. Job Satisfaction, Burnout, and Intention to Leave among Dental Hygienists in Clinical Practice. J Dent Hyg 2021; 95: 28-35.
42. Gurenlian J R, Morrissey R, Estrich C G et al. Employment Patterns of Dental Hygienists in the United States During the COVID-19 Pandemic. J Dent Hyg 2021; 95: 17-24.
43. Edwards P J, Roberts I, Clarke M J et al. Methods to increase response to postal and electronic questionnaires. Cochrane Database Syst Rev 2009; DOI: 10.1002/14651858.MR000008.pub4.
| 36513756 | PMC9746560 | NO-CC CODE | 2022-12-15 23:21:56 | no | Br Dent J. 2022 Dec 13;:1-6 | utf-8 | Br Dent J | 2,022 | 10.1038/s41415-022-5357-5 | oa_other |
==== Front
Nat Rev Microbiol
Nat Rev Microbiol
Nature Reviews. Microbiology
1740-1526
1740-1534
Nature Publishing Group UK London
842
10.1038/s41579-022-00842-6
Research Highlight
Across the mucus
Taglialegna Agustina [email protected]
Nature Reviews Microbiology, http://www.nature.com/nrmicro/
13 12 2022
11
© Springer Nature Limited 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
This study reports that SARS-CoV-2 binds to cilia and reprogrammes microvilli to promote replication in the nasal airway.
Subject terms
SARS virus
Pathogens
==== Body
pmcSARS-CoV-2 enters the epithelial cells in the upper respiratory tract to begin replication and infection. Nasal airways are composed of stratified multiciliated epithelial cells and mucus-producing goblet cells. How the virus overcomes this mucus barrier to infect the nasal epithelium is not completely understood. In this new study, Wu et al. show that first, SARS-CoV-2 binds the ACE2 receptor present on airway motile cilia. Cilia then mediate the transport of the virus across the underlying mucus–mucin protective barrier. Once SARS-CoV-2 accesses the basal cell body, it manipulates the host cell machinery to induce the activation of p21-activated kinases 1 (PAK1) and 4 (PAK4). Such reprogramming results in the elongation and branching of microvilli, which enable the virus to reach the nasal airways and disperse via mucus flow. Finally, the authors show that Omicron variants bind with higher affinity to the cilia and show accelerated entry compared with other variants, which explains their higher transmissibility. In sum, SARS-CoV-2 interaction with cilia and microvilli is key for viral replication and spread in the airways.
==== Refs
References
Original article
Wu C-T SARS-CoV-2 replication in airway epithelia requires motile cilia and microvillar reprogramming Cell 2022 10.1016/j.cell.2022.11.030
| 36513767 | PMC9746561 | NO-CC CODE | 2022-12-15 23:21:56 | no | Nat Rev Microbiol. 2022 Dec 13;:1 | utf-8 | Nat Rev Microbiol | 2,022 | 10.1038/s41579-022-00842-6 | oa_other |
==== Front
Nat Rev Microbiol
Nat Rev Microbiol
Nature Reviews. Microbiology
1740-1526
1740-1534
Nature Publishing Group UK London
842
10.1038/s41579-022-00842-6
Research Highlight
Across the mucus
Taglialegna Agustina [email protected]
Nature Reviews Microbiology, http://www.nature.com/nrmicro/
13 12 2022
11
© Springer Nature Limited 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
This study reports that SARS-CoV-2 binds to cilia and reprogrammes microvilli to promote replication in the nasal airway.
Subject terms
SARS virus
Pathogens
==== Body
pmcSARS-CoV-2 enters the epithelial cells in the upper respiratory tract to begin replication and infection. Nasal airways are composed of stratified multiciliated epithelial cells and mucus-producing goblet cells. How the virus overcomes this mucus barrier to infect the nasal epithelium is not completely understood. In this new study, Wu et al. show that first, SARS-CoV-2 binds the ACE2 receptor present on airway motile cilia. Cilia then mediate the transport of the virus across the underlying mucus–mucin protective barrier. Once SARS-CoV-2 accesses the basal cell body, it manipulates the host cell machinery to induce the activation of p21-activated kinases 1 (PAK1) and 4 (PAK4). Such reprogramming results in the elongation and branching of microvilli, which enable the virus to reach the nasal airways and disperse via mucus flow. Finally, the authors show that Omicron variants bind with higher affinity to the cilia and show accelerated entry compared with other variants, which explains their higher transmissibility. In sum, SARS-CoV-2 interaction with cilia and microvilli is key for viral replication and spread in the airways.
==== Refs
References
Original article
Wu C-T SARS-CoV-2 replication in airway epithelia requires motile cilia and microvillar reprogramming Cell 2022 10.1016/j.cell.2022.11.030
| 0 | PMC9746563 | NO-CC CODE | 2022-12-15 23:21:56 | no | Padiatr Padol. 2022 Dec 13; 57(6):314-315 | latin-1 | Padiatr Padol | 2,022 | 10.1007/s00608-022-01041-5 | oa_other |
==== Front
J Soc Econ Dev
J Soc Econ Dev
Journal of Social and Economic Development
0972-5792
2199-6873
Springer India New Delhi
225
10.1007/s40847-022-00225-w
Research Paper
Disaster management in india: are we fully equipped?
Deshpande R. S. [email protected]
grid.464840.a 0000 0004 0500 9573 Institute for Social and Economic Change, Bangalore, India
13 12 2022
140
8 11 2022
© The Author(s), under exclusive licence to Institute for Social and Economic Change 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Disasters occur with almost unpredictable probability, even though some ideas about the regions of incidence and likely impact on likelihood are available in the scientific literature. In this lecture, I have taken a full view of six disasters that include hydro-geological, meteorological, climate based like floods and droughts as well as the biological holocaust of Covid-19 pandemic. The approach followed in this lecture is to analyse the occurrences, incidence, history and devastation caused by the disaster. The impact and policies to alleviate the effects are also discussed. The culture of disaster reliance is discussed at the end.
Keywords
Disaster management
Calamity
Vulnerability
Impacts of disaster
Policy
==== Body
pmcIntroduction
I feel quite honoured to be invited to deliver this prestigious lecture on a subject that is quite close to my academic work for decades. I was very fortunate having not only worked when Professor L S Venkataramanan was at the ISEC as senior Professor and later on as Director of ISEC. I joined his research team on the “Dynamics of Rural Transformation” a project sponsored by ICSSR, spear-headed by Prof C T Kurien. The project was already quite delayed and Prof LSV wanted me to join in order to finish the work. We worked together and completed the project out of which I could publish one or two papers. Unfortunately, his ambition that we should publish a book out of this work and on our joint works on supply & demand projections of food grains in Karnataka remained unachieved. He had negotiated the publications with Late Shri Tajeshwar Singh of Sage publications but unfortunately before it could take final shape that fateful accident took him away. That was a great loss for the academic fraternity and personally to me, as he was quite an affectionate person in my life. I specifically chose this topic to deliver this lecture as I have worked for quite some time in the areas covering many of the aspects which touch the core of disaster management. I am trying my best to bring together my thinking over years as well as the major hypothesis of institutional participation in disaster management.
What are disasters?
In the calculus of disaster management, it is essential to decipher the typology and intensity of the ensuing disasters. The Indian subcontinent distinguishes among others as one of the most disaster-prone area. More than 85% of India’s geographical area is prone to multiple hazards. Out of the total Indian states and union territories, almost three-fourth is disaster-prone (Shah 2011). More than 55% of India’s area falls under seismic vulnerability (high seismic zones III–V), about 65% face the occurrence of drought and then cyclones and floods threaten about 10% of the area. After the release of the IPCC reports, the link between disasters and climate change is being increasingly established (Shamsuddoha et al. 2013). The eastern coast is more vulnerable since the 2004 tsunamis in the Indian Ocean. The Government of India in the documents on disaster management in India noted that a disaster is an “event or series of events, which gives rise to casualties and damage or loss of properties, infrastructure, environment, essential services and means of livelihood on such a scale which is beyond the normal capacity of the affected community to cope with”. The UNISDR (2011) defined disaster as “a serious disruption of the functioning of a community or a society involving widespread human, material, economic or environmental losses and impacts, which exceeds the ability of the affected community or society to cope using its own resources”. Further, refining the concept, the Govt of India in 2011 defined it as “Disaster is also sometimes described as a catastrophic situation in which the normal pattern of life or eco-system has been disrupted and extra-ordinary emergency interventions are required to save and preserve lives and or the environment” (GoI 2011, p 4). A common understanding from the above is that a disaster disrupts normal human life without sufficient warning and causes huge economic and human losses. It resets the development clock in the affected regions a few years backwards.
Disasters certainly occur without much warning regarding their occurrence or the kind of havoc that they will leave around. It will certainly be futile to answer the question and reason for the causes of disasters since the advance prediction of disasters of all kinds has hoodwinked the best scientific brains of the world (Asimakopoulou et al. 2010). Certainly, disasters occur when there are significant changes in the holistic natural equation of the universe which includes biological entities, the earth, climatic factors and above all human interactions with many of the components of these elements. In the Indian Scriptures, it has been stated that there are five Maha-Bhootas (very crudely translated as five demons) that are responsible for cosmic creation (Gopal 1990, P.79). These five uncontrollable factors that dictate life on the earth are defined as Pruthvi (earth); Aaps (Water), Tej/Agni (Sunshine or Fire), Vayu (Air or Green House Gases), Akash/Dyaus (Space, Atmosphere, Universe) (indianscriptures.com). It has been stated in the scriptures that these are uncontrollable cosmic creations and hence need to be worshipped and pacified, (www.holybooks.com/wp-content/uploads/RigVeda.pdf). The interactions between these five and their subcomponents can further create abnormal natural events and human interaction with all these five uncontrollable factors sometimes increases the intensity of such abnormal events. These extreme events occur when humanity takes the naturalness out of nature (O’Keefe et al. 1976). Therefore, any event that occurs has a permutation of 120 multiplied by the innumerable geo-climatic zones and mega-human interactions in the world. This fact is truly multifarious and thus any catastrophe that occurs has a very thin, nearly unpredictable probability and is hence called a disaster. Mismanagement of these elements and disturbance to the base element equilibria are the main reasons behind every disaster. At times the disaster management bodies were not able to do anything to prevent deaths. Even a moderate earthquake of magnitude 6.0 on the Richter Scale in Sikkim in 2011 was a disaster, causing large-scale destruction and many deaths. There is little control that can be exercised over these events. The disturbance in the base element equilibria needs to be managed with human efforts. Even though forecasting the probability of occurrence is difficult, being prepared for extreme events is not impossible. Certainly, there have been many efforts towards preparedness for disasters and a list of many such requirements for avoidance and mitigation of disasters due to climatic changes are listed in the IPCC reports. In India, the Ministry of Home Affairs brought out an Act of 2005, a policy document in 2011 that was reviewed in 2013 suggesting a framework for the plan. Further, a Disaster Management Plan was brought out in 2016. However, it is a fact that disasters occur not simply due to disequilibria but more so due to under-preparedness and management failures of such events.
Understanding disasters
Understanding disasters, therefore, involves deciphering the complex interactions between a multiplicity of climatic and manmade factors with the environment and its components. Not all environmental and climatic variables need to behave normally across years and regions. There are always variabilities and these vacillations differ in their intensity across space and time. It is the intensity of these oscillations that distinguishes them from simple aberrations and normal behaviour to reach disaster or catastrophe. There is always a history of the probability of occurrence of most of these climatic events be it rainfall, temperature, pandemic, diseases, floods, droughts, tornados or any such things. The probability distribution of occurrences is shown in Fig. 1, which indicates the well-known normal probability distribution. This is a distribution of probabilities and standardized occurrences of the events. The most feasible probabilities are clustered in the centre having occurrences that are positive (favourable) and negative (unfavourable). The probability of the occurrences is spread on both sides and any events happening out of the normal (central bell shape) are not desired, whereas the events with probability falling within the central portion of the bell shape are desirable. The basic parameters of the normal distribution are that it has zero expected value and constant standard deviation.Fig. 1 Using normal distribution to understand occurrence of disaster. Source: https://images.app.goo.gl/EtA55CzBbwjdfLPF6
The figure shows a bell-shaped curve having four segments on the negative side (left side) and other four segments on the positive side (right side). If we take any event which may occur on either of the sides, they differ from normal occurrences to abnormal occurrences. The normal occurrences are shown in the central portion of the normal distribution. The central portion indicates the normal behaviour of any phenomenon. Initially, the first departure segments on either side beyond one standard deviation away from the mean can be called “Mild shock or aberrations”. The “aberrations” are events that go slightly away from the central portion of the bell-shaped normal distribution. The central portion which ranges from − 1 to + 1 standard deviation (SD or α) and covers about 68.26% area includes all the events that come within the one SD of the mean or normal occurrence. These could be called mild shocks or adverse situations that are usually in the range of routine management techniques. The second group comes on the right, and the left side includes the fluctuations which go beyond one SD and extend up to 2SD on either side. These should be called aberrations and occur in the probabilities falling in that range of two standard deviations below or above the normal. These events cover 27.18% of the area and therefore will occur approximately at that probability of occurrence. The severe fluctuations come in the range of 2 to 3 standard deviation either below or above the normal occurrences. The Agricultural Commission 1976 defined severe fluctuations in this range of probability and demarcated the regions as susceptible to droughts and floods based on the probability of occurrence in the regions between 2 and 3 standard deviations on either side, and severe fluctuations and up into calamity when the occurrence takes place in the range of 3 to 4 standard deviations from the normal. The severity depends on the closeness of the event towards the border of four standard deviations on either side of the distribution. If we take the example of drought and floods, the first group within one standard deviation show a mild impact depending on the vulnerability of the region. The second group falling beyond two standard deviations on either side will result in a drought or flood situation depending on the closeness of the event to the point mark at three standard deviations. These should be called severe fluctuations and require immediate management on behalf of the policymakers. The calamity occurs when the events take place marking the fluctuations beyond three standard deviations, and the severity depends on the closeness of the event near the four standard deviation point. Such situations are cataclysmic situations and have to be dealt with very carefully as calamities. Any event that occurs at any point beyond the four standard deviation limits on either side requires preparedness as well as skills to deal with catastrophic situations. Such events occur with very thin probability and hence can result in a holocaust. In all these events, the role of administrative response is quite important and necessary to avoid loss of assets and life as well as ameliorate the suffering of the population. Therefore, it is a battle of wits between the learner and scientists and the natural events that require preparedness and management of the calamity.
Disaster: typology and response
The Disaster Management Act, 2005, defined disaster as “a catastrophe, mishap, calamity or grave occurrence in any area, arising from natural or manmade causes, or by accident or negligence which results in substantial loss of life or human suffering or damage to, and destruction of, property, or damage to, or degradation of, environment, and is of such a nature or magnitude as to be beyond the coping capacity of the community of the affected area” (GoI 2005). It is a sudden occurrence with a severe jolt that disrupts and destroys normal life and brings economic activities to a standstill, destroying the normal functioning of society and resetting the development clock.
The administration of disasters is certainly a challenging job as the administrators do not get any warning and also have very little knowledge of the severity of the impact (Sharma and Kaushik 2012). The administrative response goes by the type of disaster, and it has been very clearly documented in the Disaster Management Manual of the Government of India, 2016. Broadly the type of disasters indicated there are shown in Box 1. These are classified according to the causes of disasters which is the usual way of classifying disasters across the world (UN and World Bank 2010; UNESCAP 2016). However, it is essential to understand that disasters do not strike uniformly across regions and humanity. The severe impact is felt in the most vulnerable regions, and hence, the impact gets weighed by the intensity of the disaster as also by the vulnerability of the region. The UNISDR/SDMC report (2014) gave a good framework for disaster risk reduction and climate change adaptation. The calculus of disaster management, therefore, works on three important dimensions. First, the type and probable intensity of desire. Second, the vulnerability and the population of the region. Third, the main economic base of the region and its dependent activities matter the most. It was in 2001 that a High Power Committee of the Ministry of Home prepared a report about disaster preparedness in India (GoI 2001). There was no institutional set up for managing disasters in India till the 1990s and all calamities were handled with a “firefighting” approach in a great hurry. A disaster management cell was placed under the Ministry of Agriculture after the declaration of “International Decade for Natural Disaster Reduction” (IDNDR) by the UN General Assembly and a series of disasters (like Latur Earthquake (1993), Malpa Landslide (1994), Orissa Super Cyclone (1999) and Bhuj Earthquake (2001)). The Government of India constituted a High Powered Committee under the Chairmanship of Mr. J.C. Pant in 2001.The Committee submitted a detailed report, and the recommendation was followed. Subsequently, the Disaster Management Division was placed in the Ministry of Home Affairs in 2002. Subsequently, the Disaster Management Act, 2005 (GoI 2005), was passed and the Act gave an elaborate framework for the Institutional set up to deal with such calamities.
India’s hazard-prone area has been mapped by the Building Material and Technology Promotion Council (BMTPC) in their “Vulnerability Atlas of India” presented in Map 1 showing the multi-hazard zones of India. We have placed this map here to give a glimpse of the problem and the dimension of the issues to be tackled. BMTPC has done significant work in mapping the hazard-prone areas of India and mapped the information in policy convenient maps. It has been well documented in the BMTPC work that 85% of India’s area faces multiple hazard vulnerabilities. Almost 22 states are under the continued shadow of disasters of one type or other. The eastern and western coasts are vulnerable to weather storms on both the Bay of Bengal and the Arabian Sea coast (Map 1). The eastern coast faces several sea initiated turmoil. About 8% of the coastal area is prone to cyclones. Droughts and floods are regular visitors in most of the states, especially in central India. It is seen that India has about 68% of its area categorized as “drought-prone”, and at the same time, about 12% of its area faces frequent floods. Acute seismicity resulting in earthquakes is confronted by almost 57% of the area that falls under the high seismic zone (World Bank 2010).Map 1 Multi-hazard-prone areas of India
Any disaster causes not only huge material losses but also human and animal lives. However, these are not uniform across various typologies of disasters. The predictability, damage, loss of lives and material differs across disasters (Table 1). In many cases, all the economic activities come to a standstill and the government has to spend additionally to recoup the losses and bring back economic activities to normalcy. The World Bank report documented 73 such instances of natural catastrophes in just four years between 1996 and 2000 and reported direct losses on public and private economic infrastructure in India (UN and World Bank 2010). Similarly, vulnerability across states also differs significantly. Some states are well prepared to deal with such situations (Karnataka, Odisha, West Bengal, Tamil Nadu and Andhra Pradesh) with little warning, but a few others squarely face destruction (Table 2).Table 1 Typology of disasters. Source: authors perceptions based on literature
Type of disaster Predictability Type of damage Preparedness
Cyclones Tornadoes, Hurricane, NEG
One week
S-LP L
Thunderstorms A Little M-S (LP) Low
Floods M-G M-S (LP) M
Earthquakes M S (LP) L-M
Landslides L-M S L
Landslides/Glacier Based NEG S L
Epidemics M M-S (Lives) L-M
Predictability: FG, fairly good; L, low, M, moderate; NEG, negligible; damage: LP, life and property; S, severe; NEG negligible
Table 2 Vulnerability of the states to various disasters
State Cyclone Flood Earthquake Landslide Drought 14th FC Allocations
Karnataka M M M M S 1527
Maharashtra M M M M S 8195
Madhya Pradesh(U) N M M N S 4747
Orissa S S M N M 4130
Andhra Pradesh S M N NA S 2430
Tamil Nadu S S N NA S 3751
West Bengal S S N N N 2854
Uttarakhand N S M S M 1158
North East States* M S S M N 731
S, Severe; M, Moderate; N, Negligible; NA, Not Available
North East: Manipur, Meghalaya, Mizoram, Nagaland, Sikkim & Tripura
Kamepalli Lenin (2019) and 14th Finance Commission p.466
It was noted that “Losses amounted to approximately $30 billion over the past 35 years (nominal values at then applying exchange rates)”. Since less than 25% of the registered loss events actually provide any loss estimates, the official numbers substantially understate the true economic impact of direct losses. A crude grossing up for reporting frequency indicates that direct natural disaster losses equate to up to 2% of India's GDP and up to 12% of federal government revenues” (World Bank 2010, Table 4). Another map from BMTPC clearly gives the geographical locations of climatic disasters. Even though most of the states in India are vulnerable, the Finance Commission has noted the severity in some selected states and accordingly sanctioned funds towards the purpose. The Fifteenth Finance Commission, after reviewing the earlier process under various Finance Commissions, submitted two reports on disaster management. The FC recommended a new methodology which stands as a combination of a) capacity or need of the state (based on past expenditure); exposure to disaster risk of the states (including area and population); and hazard & vulnerability (disaster risk index). The capacity or the need of states was to be prepared for disasters to deal with the situation fully and allocations were indicated to support SDMAs, SIDM, training and capacity building activities and emergency response facilities. The FC recommended funds under two headings namely: Corpus of Rs.1,60,153 crores for SDRMF for states (2021–26); and corpus of Rs.68, 463 crores for NDRMF for states (2021–26) (Finance Commission Report 2021).
It is known that the Indian subcontinent falls under severe disaster-prone regions and it has to confront various types of disasters. In the hundred years from 1900 to 2000 402 events could be termed as disasters and 354 in the two decades thereafter (2001–21). Each one of these has historically been subjected to such calamities. Based on the probability of occurrence of such events taken together with scientific data about the climatic factors and other geological formations disaster-proneness is arrived at. BMTPC and the Ministry of Home Affairs, Government of India have recorded the vulnerability maps (see Map 2). In about two decades after 2001, a total of 100 crore people have been impacted and nearly 83,000 lost lives due to these disasters. If the losses are adjusted with current prices, the losses come out to a staggering Rs. 13 lakh crore or 6 per cent of the GDP as estimated by the State Bank of India (SBI). The damages caused by the hydro-meteorological calamities were reported in the Indian Parliament recently under an un-starred question. The data from the Ministry of Home affairs are presented in Table 3Map 2 India climatic disaster risk map
Table 3 Details of damages due to hydro-meteorological calamities: 2001–2022. Source: Loksabha Un-starred Questions 1188 of 18–12-2018; No. 242, of 19 July 2022, and Rajyasabha 1974 of 03–08-2022
Losses due to disaster Total Average
Human lives (Thousands) 46.19 2.1
Livestock (Thousands) 1635.52 74.34
Houses damaged (Lakhs) 261 11.86
Cropped area affected (Lakh Ha) 891.61 40.5
It can be noted that in the last two decades more than 46.19 thousand lives, 16.35 lakh animals and 261 lakh houses have been damaged. Besides this, 89 million hectare crops have been destroyed (Table 3 and Fig. 2). Almost annually human and livestock lives that are lost are about 2 thousand and 74 thousand, respectively. The damaged houses count per year is 11.8 lakhs, and the cropped area damaged goes up to 40 lakh hectares a year. It is reported in the 15th Finance Commission report that the total expenditure on disaster response and relief across twenty-eight states between 2011 and 2019 was Rs. 1,66,702 crore (Table 8.1 of the report). A significant jump in the expenditure could be noted. Besides impromptu grants from the Government of India during calamities, successive Finance Commissions have followed different approaches to determine the allocation of funds for disaster management to the State Governments. Initially, from 1957 to 1984–89 it was called Margin Money Scheme to cover the ad-hoc expenditures incurred by the States. After 1995 and till 2010, the approach was based on providing the Calamity Relief Fund. That was changed during 2010–2015 to National Disaster Relief Fund (NDRF) and State Disaster Relief Fund (SDRF). Besides adding the National Calamity Contingency fund, the 15th Finance Commission brought in a new nomenclature of National Disaster Mitigation Fund (NDMF) and State Disaster Mitigation Fund (SDMF). The 15th Finance Commission Report also uses a new methodology, which is a combination of capacity (as reflected through past expenditure), risk exposure (area and population) and hazard and vulnerability (disaster risk index) for determining state-wise allocation for disaster management. This shall be continued for the five-year award period from 2021–22 to 2025–26. The mitigation funds at the state and country level were intended to aid the implementation of mitigation measures in states for the award period, as provided in the Disaster Management Act 2005.Fig. 2 Losses due to disasters in India: 2001–2022. Source: Based on the Emergency Events Database (2016), author’s work
Broadly, disasters can be placed into six groups: (I) Water and Climate Related: (a) Floods and drainage management; (b) Cyclones or Tsunami, Tornadoes and hurricanes; c) Hailstorm; d) Cloud burst; (e) Heat wave and cold wave; Snow avalanches; Droughts; Sea erosion; Thunder and lightning; (2) Land Related: Landslides and mudflows (b) Earthquakes (c) Dam failures/Dam bursts (d) Minor fires; (3) Natural Accident Related: (a) Forest fires (b) Urban fires (c) Mine flooding (d) Oil spills; (4) Accidents due to Human errors: (a) Major building collapse (b) Serial bomb blasts (c) Festival related disasters d) Electrical disasters and fires e) Air, road and rail accidents f) Boat capsizing g) Village fire; (5). Industrial: (a) Nuclear Leak; (b) Industrial Chemical leak; (c) Operational Negligence and the last one are: (6) Biological: (a) Biological disasters and epidemics; (b). Pest attacks; (c). Cattle epidemics; (d).Food poisoning. Usually, these originate as natural hazards or human-induced vulnerabilities. These can also result from a combination of any of these. As indicated earlier, it is the non-normal behaviour of the “Pancha Mahabhutas” coupled with human interactions that result in these disasters. Most human-induced factors can worsen the negative impacts of a natural disaster. The UN Inter-Governmental Panel on Climate Change (IPCC) also noted that human-induced climate change has a significant role in both the frequency and intensity of these episodes (Mall et al 2006, 2011a; b; Kelman 2008, 2009; Dhar 2010). At the policy level, the Government of India recognized the severity of the impact due to climate change even before 2001 and set in motion preparations accordingly (GoI 2008, 2012). IPCC (2013) brought forth the necessity of inter-governmental cooperation while keeping in view the probability of increasing disasters (Seidler et al (2018)). At about the same time, the Government of India appointed a task force to review disaster management in India and the task force gave clear guidelines which were followed in the Act (2005) and Disaster Management Plan that followed in 2016 (GoI 2013, 2016). While heavy rains, cyclones, or earthquakes are all extreme natural events, the impacts relate mainly to actions or inactions in human activity. Extensive industrialization and urbanization increase both the probability of human-induced disasters and the extent of potential damage to life and property from both natural and human-induced disasters. Climate change has also contributed to the severity and frequency of these events (UNISDR/SDMC (2014; Shamsuddoha et al 2013). In the first kind, we have hydro-meteorological disasters that cause significant damage to the ecology as well as human and animal lives. The economic impact of these disasters is quite significant. The impact of the four important components as recorded and analysed in the Ministry of Home Affairs of the Government of India, namely cyclones, tornadoes, whirl winds and floods as well as droughts, come under this severe most category. The second major cause is due to the changes in the geological patterns, especially reflected in earthquakes, landslides and avalanches are seemingly caused by geological causes, but the role of human activities in them cannot be undermined. The third typology of disasters is due to unpreparedness and failure to understand the pattern of natural events. Therefore, these are caused mainly due to the inaction of the State agencies and the response of the population. The last category includes biological disasters either due to human intervention or the intrusion of biological elements like viruses, bacteria or any such epidemiological events. These phenomena are of organic origin or carried by biological vectors, including exposure to pathogenic micro-organisms, toxins and bioactive substances. Major accidents due to the mishandling of chemicals, atomic energy or radioactive material also have a significant impact on human lives. It is, therefore, necessary to understand the mechanics of most of these disasters to equip and be prepared to meet the calamities for reducing the losses.
Mechanics of disasters
Disasters have different typologies. Among the disasters that we have taken for discussion in this lecture are cyclones or tornadoes, earthquakes, floods and droughts followed by the pandemic. It was necessary to keep the contours of this lecture limited to the available data and time at my disposal. Each one of the typologies mentioned above has a serious impact on human life and livelihood. Besides, most of them force to restart many developmental initiatives destroyed during the disaster. These include houses, roads, railway lines, bridges, factories, banking, trees, crops and many other things that come in the region.
Cyclones, tsunamis, thunderstorms and tornados
Cyclones, thunderstorms and tornadoes are among the worst impacting events that come with little warning but visit frequently. The eastern coastline covering about 8% of the land is extremely vulnerable. It is estimated that approximately three cyclones visit the coastal line every year with varying intensity. Therefore, regions from West Bengal to Tamil Nadu are highly vulnerable. The west coast is relatively less vulnerable due to its geographical formation, but any event on that side of the sea cannot be ruled out. If the west coast gets any visitations of events similar to that on the east coast of India, the devastation and loss of life will be unparalleled. In the event of a fully developed cyclone, the coastal line also faces gales and strong winds; torrential rain and high tidal waves. As the storm surges, it destroys human habitations. Large human casualties take place as many times it comes without sufficient warning (Table 4).Table 4 Typology of cyclones and impact. Source: India Meteorological Department, Public domain Information
Type of disturbance Associated wind speed Wind speed Kms per hour Impact
Low-pressure area Less than 17 Knots Less than 31 M; CL, WS
Depression 17 to 27 Knots 31 to 49 M to S; WS
Deep depression 28 to 33 Knots 50 to 61 WS-S-LAL
Cyclonic storm 34 to 47 Knots 62 to 88 S-LHL-LAL-WS-LI-CL
Severe cyclonic storm 48 to 63 Knots 89 to 118 MS-S-LHL-HD-LAL-WS-LI-CL
Very severe cyclonic storm 64 to 119 Knots 119 to 221 ES-S-LHL-LAL-HD-WS –LI-CL
Super severe cyclonic storm Above 119 Knots Above 221 ES-S-LHL-LAL-HD-WS–LI-CL
Impact severity is codes as: LHL, Loss of Human Life; LAL, Loss of Animal Life; HD, Houses Property Damaged; CL, Crop Loss; LI, Loss of Infrastructure; ES, Extremely Severe; MS, Most Severe; S, Severe; M, Moderate; WS, Wide Spread
The East Coast of India, along the Bay of Bengal, affects the population of West Bengal, Orissa, Andhra Pradesh and Tamil Nadu, but also parts of Karnataka, Pondicherry, and Telangana. A similar event on the West Coast adjoining the Arabian Sea will impact Maharashtra, Goa, Gujarat, Karnataka, and Kerala. India has a coastline of 7516 km, of which 5700 km is susceptible to cyclones and tornadoes. About eight per cent of the country’s area and one-third of the population living in the 13 coastal states and UTs are vulnerable to cyclone related disasters. 5700 km. of the coastline stays continuously exposed to tropical cyclones. Usually, these events originate in the deep side of the Bay of Bengal and south western side of the Arabian Sea.
The Indian Ocean is one of the six major cyclone-prone regions of the world. Cyclone incidence is very high during April–May and October–December. About 80% of total cyclones invariably touch the long eastern coastline. Cyclones of various intensities peak in October or November (Cyclone Phailin) and a few with low intensities strike in May (Cyclone Mahasen). Most of the cyclones have caused damage to the eastern states in India. It is noted that the ratio of cyclones occurring in the Bay of Bengal to the Arabian Sea is approximately 4:1. Gujarat is the state that faced a few cyclones originating from the Arabian Sea. A cyclone gets formed due to coriolis force and with variations in the vertical wind speed or upper divergences above the sea level systems. A mature tropical cyclone has strong spiralling wind around the centre called EYE, and this enters the region of calm subsiding air pocket. The diameter of any storm in the ocean is between 600 and 1200 km, and the system sometimes moves slowly but often with heavy speed. There are six main requirements for a tropical cyclone to start. These are sufficiently warm sea surface temperatures, atmospheric instability, and high humidity in the lower to middle levels of the troposphere; enough coriolis force to develop a low-pressure centre; a pre-existing low-level focus or disturbance; low vertical wind shear. The table above indicates the typologies of disturbances and associated wind speed along with the possible impact as seen with historical data. It was since the mid-90 s the practice of naming the storm has come into being and almost every storm is named by the World Metrological Department and it is followed by the metrological departments in the country of its origin. The worst tropical cyclone called Mitch visited in 1998 followed by hurricane Katrina in 2005 and Nargis in 2008. Phallin struck India in 2013 and Fani from 26 April 2019 to 5 May 2019. SCATSAT is the Indian satellite developed by the Indian Space Research Organisation to keep a close eye on cyclones and prepare an early warning system for the East Coast to help give warnings. However, such warnings come only a few days or a week before the actual strike. Besides, there are a few Doppler radars that have been placed at various places on the East Coast, but the density of radars in the Arabian Sea is comparatively less. That side depends more on the naval intelligence data (Table 5).Table 5 Visits of major cyclones in India. Source: Compiled from various reports of the Government of India
Andhra Pradesh Orissa West Bengal Tamil Nadu
1990; 1998; 2003; 2007 (Yemin);
2008 (KhaiMuk);
2010 (Laaila; 2012 (Nilam); 2013 (Helen &Leher); 2014 (HadHad);
2016 (Kyant);
2018 (Titli)
1999;
2013 (Phalin);
2014 (HadHad)
2018 (Titli);
2019 (Fani);
2020 (Amphan)
Partial list
1970;1981;1997; 1998;2000; 2002;
2007 (Sidr);
2008 (Rashmi);
2009 (Aila);
2015 (Komen);
2016 (Ronau);
2017 (Mora);
2019 (Fani & Bulbul); 2020 (Ambhau)
1991;1993;1996; 2000; 2005 (Fanoos);
2008 (Nisha);
2010 (Jai);
2011 (Thane);
2012 (Nilam);
2013 (Madi);
2016(Ronau, Kyant, Nada, Vardha);
2017 (Okchi)
2018 (Gaja)
12 6 15 17
The history of cyclones world over is quite distressing. On 7 October 1737, a Bay of Bengal-originated cyclone destroyed 20,000 ships and left 300,000 people dead. It is also recorded that on 25 November 1839, the City of Coringa in Andhra Pradesh, a harbour city was destroyed. In 1789, a different cyclone passed near the area, generating a large storm surge killing over 20,000 persons. The November 1839 cyclone killed 3,00,000 people and devastated the entire area. It is recollected that on 5 October 1864, most of Calcutta was denuded by a cyclone and 70,000 people drowned and the devastation of the region was unprecedented. There were many episodes of climatic strong aberrations after that but in October1967, a massive cyclone struck rural Orissa leaving behind a trail of deaths and destruction. The precise number of fatalities and destruction is unknown. After this, the next significant strike of the cyclone was in September 1971 along with huge tidal waves in the Bay of Bengal killing more than 10,000 people in Orissa. In Andhra Pradesh, cyclone and tidal waves claimed more than 20,000 lives in November 1977 and again in the same month in 1996 cyclone struck Krishna district. It took an unexpected turn towards the Godavari delta with high speed winds causing severe devastation and many deaths. Eastern Orissa confronted a super cyclone in October 1999 that recorded a wind speed of 190mph and sea waves which rose up to 15 feet high. All that caused devastation in the districts of Kendrapara, Jagatsinghpur, Puri, Cuttack and Jajpur. About 9,500 people died, and 2.5 million became homeless. Four hundred thousand heads of livestock were drowned, and damage estimation reached 3.5 US $ billion (cyclone https://en.wikipedia.org/wiki/Cyclone).
It can be seen from Table 5 that from1990 till recently almost 50 major cyclones visited the east coast of India, the maximum visitations being to Tamil Nadu followed by West Bengal. The severity was quite strong in Odisha and Andhra Pradesh. “The frequency of cyclones on the east and west coasts of India between 1877 and 2005 shows that nearly 283 cyclones occurred (106 severe) in a 50 km wide strip on the East Coast; comparatively the West Coast has had less severe cyclonic activity (35 cyclones) during the same period. More than a million people lost their lives during this period due to these cyclones” (National Disaster Management Plan 2016, P. 20).
Impact
Cyclones are one of the thin probability events, and predictions are hardly a few days to a week before. The event occurs, and the central spot as well as the starting point can be located. The Doppler radars and naval intelligence provide continuous signals, and the progress in intensity and direction could be tracked. Even then, preparedness to meet the effects of disasters does not totally eliminate the losses as the spillover effect is deep in the mainland too. The Super Cyclonic Storm Amphan 2020 in the Indian Ocean was one that created huge devastation. It impacted most of the eastern coast leaving behind a trail of severe damage. A large number of lives were lost, and as per records due to the cyclone about 98 persons died and 1167 km (725 mi) of power lines were damaged, 126,540 transformers and 448 electrical substations were destroyed, leaving 3.4 million without power. (The Times of India, 22 May 2020). The damage to the power grid reached ₹3.2 billion (US$42 million). Four people died in Odisha, two from collapsed objects, one due to drowning, and one from head trauma. Across the ten affected districts in Odisha, 4.4 million people were impacted in some way by the cyclone. At least 500 homes were destroyed, and a further 15,000 were damaged. Nearly 4000 livestock, primarily poultry, died (various news reports then). The cyclone was strongest in its northeast section. The next storm was a depression that did not affect India. Then, a severe cyclonic storm Nisarga hit Maharashtra, with significant damage. Nisarga caused 6 deaths and 16 injuries in the state. Over 5033 ha (12,435 acres) of land were damaged (Effects of the 2020 North Indian Ocean cyclone season in India, en.wikipedia.org and various news reports). Cyclones cause huge-scale devastation to society and life gets totally disrupted. Lakhs of livestock get inundated so also did the property and crops. All economic activities come to a standstill and even though rehabilitation works begin immediately, it takes a few years for the livelihood cycle to assume normality.
Policy response
India’s Disaster Management preparedness has improved substantially since 1996. Following the 2001 expert committee recommendation, in May 1916, the Government of India released a detailed Natural Disaster Management Plan. In chapter three of this report, a detailed framework of rehabilitation and recovery after the cyclone is systematically recorded. First, emphasis is laid on understanding the risk through observation networks, information systems and research forecasting. Zoning and mapping as well as monitoring the progress of cyclones need to be emphasized. Hazzard risk and relative vulnerability along with dissemination of the warning systems data and information are to be recognized as important aspects. The report further provided policy guidelines about the institutions involved and overall disaster governance, and response systems as well as responsibilities to be shared. It is recommended that multi-purpose cyclone shelters, social housing schemes and hazard-resistant constructions be built. Capacity building and legal systems needed for meeting emergencies during cyclones have also been stated. Pandey (2016) suggested revisiting the existing legal framework of disaster management in India. This report is quite holistic as far as policy is concerned. It is only to be seen how the implementation process takes it forward and that comes only during the next episode of the disaster.
Floods and water related disasters
The rainfall pattern in India is marked by its erratic fluctuations. The climatic conditions here range from snowy Himalayan heights in the north to the arid deserts of Rajasthan. The annual normal precipitation ranges between 83 and 4000 mm. Ruyli located in the Jaisalmer district of Rajasthan receives the lowest amount of rainfall measuring only 83mms, whereas Mawsynram in Khasi hills in East India gets 11,872 mm annual average rainfall. The two major problems associated with erratic rainfall are flood and drought. In unusual conditions, these turn into calamities, if accompanied by severe management and perception failures. Floods in the Indo-Gangetic-Brahmaputra basins have become almost an annual feature. On average, a few hundred lives are lost every year, some years even millions were rendered homeless and several hectares of crops were damaged.
More than 70% of the total annual precipitation occurs during pre-monsoon, south-west, north east and retreating monsoon seasons (De et al. 2005). It is estimated that roughly about 40 million hectares or 12% of Indian land is prone to flood hazards. Floods are very regular in Assam, Bihar, Orissa, Uttar Pradesh and West Bengal, but states like Maharashtra, Andhra Pradesh, Karnataka and Kerala also have faced sporadic inundation. The horrifying memory of the Morvi Floods in Gujrat which was described as the “city of dead bodies” cannot be wiped out of our memory (Noorani 1979). Map 3 shows the flood-prone areas of India and the regions with a high probability of catastrophe. Between 1998 and 2017, 10 of the 14 extreme weather-related disasters that hit the country were floods. The floods also claimed 3,396 human lives and 239,174 cattle, besides damaging 35,07,542 houses and affecting more than 370 million people (Central Water Commission 2022, Table 6). Floods reset the development clock in these regions: roads, bridges, dwelling houses on river banks; cattle, and whatever comes in the way of water flow is washed away. In the recent past, floods are noticed even in traditional, non-flood-prone areas like Rajasthan, Gujrat, or Kerala. Various reasons like a phenomenon connected to global climate change or El-Nino, La-Nina & ENSO could be the cause (Bhat et al. 2015). Annually 32 million people are vulnerable to floods in various regions. Largely in the Indo-Gangetic and Brahmaputra plains, the floods are annual devastating agents.Map 3 Flood-prone areas of India
Table 6 Average annual losses due to flood (1953 to 2020). Source: State-wise flood damage statistics, Central Water Commission, Ministry of Jalshakti, New Delhi dated 03 January 2022
Sl No Sector affected Total affected Average
1 Area Affected (Mill Ha) 492.557 7.243
2 Crop Area Affected (Mill Ha) 275.773 4.055
3 Crop Loss Value (Rs in Crores 131,462.177 1933.267
4 Damage to Houses (No in Mill) 82,525,198 57,017.903
5 Cattle Lost (Nos in Mill) 6,182,943 90,926
6 Human Lives lost (Nos) 113,943 1676
7 Public Utilities(Rs in crores) 234,149.322 3443.272
8 Total Damages (Rs in Crores) 437,149.710 6428.672
The incidence of floods has increased in both number and intensity during the twenty-first century as compared to the twentieth century. In the earlier century, the most devastating floods were recorded in 1943, 1979, 1987, 1988 and 1993; as per the records. However, during the twenty-first century when the recording became substantially good it is observed that we had more than 10 devastating floods after 2000. These were experienced in 2005, 2013, 2015 (3), 2016, 2017, 2018, 2019, 2020, 2021 and 2022 (Central Water Commission 2020). The worst floods recorded were in August 2018 in Kerala in Wayanad and Idukki districts causing 400 deaths and about Rs.30,000 crore in losses in terms of livestock houses and infrastructure. The urban floods in Mumbai (July 2005) caused thousands of deaths and more than 14,000 houses were inundated with a loss to the economy ranging up to Rs. 1000 crore. The Chennai flood of November 2015 caused 500 deaths and about 1.8 million displacements with 50,000 homes inundated. The total economic loss was in the range of Rs. 50,000 crore. The flood-prone areas of India as delineated by the National Disaster Management Authority are shown in the map.
Impact
Flood preparedness is one of the important areas discussed in most of the plans as certain areas are habitually flood-prone like the states of Assam, Bihar, West Bengal, Ganges, Yamuna, Brahmaputra basins, North Eastern regions, Godavari basin, Krishna and Cauvery basins. Every flood sets back the development clock in the region affected by the floods. The devastation begins at the banks of the rivers but moves inside the mainland as the feeder rivers also get flooded. Initially, the impact is reflected in the inundation of regions, crops, houses along with other economic activities. The NDMA has categorized eight groups of impacts due to floods. These include (1) The devastation of Property and Life; (2) Livestock, crops and agriculture; (3) Energy and Communication facilities; (4) Food, and Drinking Water; (5) Shortage of Basic Necessities; (6) Outbreak of Epidemics, (7) Health Hazards and viii. Economic Activities. The Central Water Commission in its report in 2019 reported the average losses due to flood which is presented in Table 6.
Policy
It is seen that the total losses to the economy from 1953 to 2020 are to the tune of 437 thousand crores in which only the damage to the crops, houses and public utilities are included. Cost relating to the loss of cattle and human lives and the cost to reset the development clock have not been calculated. The estimates of the loss provided in the table are reported losses, but then there are losses that happen in remote areas not covered by the agencies collecting statistics. These losses are quite huge in terms of all the sectors mentioned above. Floods have been our regular visitors and the region of the incidence as well as the likely intensity is easily understood from the studies conducted at the Central Water Commission of the Government of India. The important alleviation measures that have been suggested include (1) Early warning system based on the metrological data and forecasting of the rainfall in the catchment areas of the rivers; (2) The evacuation mechanism to be organized by NDRF and SDRF and keeping continuous alert in any situation; (3) The historical areas prone to floods have imbibed a perfect culture of accommodating the natural event and trying to adjust in the event of mild shocks but that does not help in unusual calamities. In the case of catastrophes, the losses even in the usual flood-prone areas are enormous despite experience & preparedness. More important is the identification of the new areas that are not usually affected by floods in the past. These regions report heavy damages both in economic as well as human and animal capital loss; iv. The embankment of the rivers every year and its maintenance needs to be high on the agenda. There are various flood control authorities for most of the rivers, however, the devastation of floods continues unabated. Either the authorities do not have a grasp of the real problem or the behaviour of the floods hoodwinks, the preparation of the authorities. v. For a long time now Inter Basin Transfer of Water Resources (IBTWR) and connectivity of rivers have been under discussion, but little has been achieved as yet. IBTWR is a project that envisaged connecting two or more rivers by creating a network of reservoirs and canals, expecting to alleviate the regular flooding of some rivers and using the water resources in the drought-prone areas. It is based on the assumption that surplus water in some rivers can be diverted to deficit rivers by creating a network. It all began with the Garland Canal Scheme proposed by an innovative Bombay-based consultant engineer Dinshaw Dastur in the year 1977. It is regarded by some scientists as a panacea for the country's ills, particularly the chronic flood problem and equally frustrating regular visitations of droughts. Under this (IBTWR) National Perspective Plan (NPP) for the interlinking of rivers, was prepared by the then Ministry of Irrigation (now Ministry of Jal Shakti) in 1980. Under NPP, 30 links were identified covering 14 links under Himalayan Rivers Component and 16 links under Peninsular Rivers Component for Inter Basin Transfer of Water based on field surveys and investigation and detailed studies. Precious little has happened after that.
Geological and land based disasters
India has a significant share of the Asian seismic zones and unlike cyclones, earthquakes occur without any prior warning. Therefore, the damage and loss of lives, as well as property, are huge and sudden. Asimakopoulou et al. (2010) reported that India’s 12% area is prone to “Very Severe” earthquakes, 18% to “Severe” earthquakes and 25% to “Damageable” earthquakes. The severe most quakes occurred in the Andaman and Nicobar Islands, Kutch, Himachal and the North-East. The Himalayan regions are particularly prone to earthquakes. The last two major earthquakes shook Gujarat in January 2001 and Jammu and Kashmir in October 2005. Beyond these, many smaller-scale quakes occurred in other parts of India in 2006. All 7 North East states of India—Assam, Arunachal Pradesh, Nagaland, Manipur, Mizoram, Tripura and Meghalaya; Andaman & Nicobar Islands; and parts of 6 other states in the North/North-West (Jammu and Kashmir, Uttaranchal, and Bihar) and West (Gujarat), fall in the Seismic Zone V. These confront higher probability of occurrence of quakes.
An earthquake is caused by a sudden change in the tectonic plates that are continuously moving, albeit at a snail’s pace. The plates that are deep below the earth’s surface are called “continental plates”. These plates move under a few kilometres in the deeper part of the earth (the mantle). They are always moving, bumping, or sliding past each other at a very slow speed but can brush against each other. Suddenly, when these plates brush strongly with their edges due to friction, a tremor eventuates. It is this stress on the edge that creates friction, which gives rise to an earthquake releasing huge energy waves and kilometres causing tremors through the earth's crust. This gets manifested into shaking of the earth’s surface which is experienced as an earthquake. The tectonic plates below the bottom of the sea also experience a similar phenomenon that gives rise to tremors. Such occurrence under the oceans also causes tremors and these are called oceanic quakes.
Measurement of the energy released that causes destruction is done with the help of seismographs. This is a scale-based measurement of the magnitude of earthquake given by Richter (e.g., Richter 1958). This is called “Richter scale”. The magnitude on the Richter scale is obtained through recordings of ground motion on seismographs through a seismometer. There are advances in the measurement of earthquakes that help to decipher tiny movements in the Earth's outermost layer that can provide a “Rosetta Stone”1 for deciphering the physics and warning signs of big quakes. New algorithms that work a little like human vision are now detecting these long-hidden micro-quakes in the growing mountain of seismic data (Tables 7, 8, Map 4).Table 7 Earthquakes that occurred in the Indian region between 1803 and 2022. Source: Based on https://en.wikipedia.org/wiki/List_of_historical_earthquakes
Sl No Range in Richter scale Number of episodes
1 Less than 4.5 6
2 4.5 to 5.0 4
3 5.0 to 6.0 20
4 6.0 to 7.0 16
5 Above 7.0 21
Table 8 History of earthquakes in India: 1950–2022
Date Area Magnitude Casualties Severity
15-08-1950 Assam-Tibet 8.6 3300 Mod
21-03-1954 India-Myanmar 7.4 NA Low
21-07-1956 Gujarat 6.1 115 Mod
09-02-1963 Kashmir 5.3 80 Mod
27-06-1966 Nepal/India 5.3 80 $1Mill
15-08-1966 North India 5.6 15 Low
12-11-1967 Maharashtra 6.6 2400 $0.4Mill
23-03-1970 Bharuch 5.4 226 Mod
19-01-1975 Himachal Pradesh 6.8 47 Mod
29-07-1980 Pithoragarh 6.5 200 $245 Mill
23-08-1980 Kashmir 4.9 55 Mod
20-01-1982 Little Nicobar 6.3 NA Mod
30-12-1984 Cachar 5.6 120 Severe
20-10-1991 Uttarakhand 6.8 3800 Severe
30-09-1993 Laatur 6.2 39,748 Severe
22-05-1997 Jabalpur 5.8 1556 $37143Mill
29-03-1999 Chamoli 6.8 103 Mod
26-01-2001 Gujarat 7.7 20,023 Severe $10Billion
25-09-2001 Tamil Nadu 5.2 3 Low
13-09-2002 Andaman 6.5 NA Tsunami
08-10-2005 Kashmir 7.6 87,351 2.8 Mill displaced
14-12-2005 Uttarakhand 5.1 4 Destruction
10-08-2009 Andaman 7.5 NA Tsunami
30-03-2010 Diglipur Andaman 6.6 10 Destruction
18-09-2011 Gangtok 6.9 111 Severe
21-05-2014 Odisha 6.0 252 Destruction
25-04-2015 Nepal Border 7.8 8964 Severe $10 Billion
12-05-2015 Nepal Border 7.3 218 Severe
26-10-2015 North West 7.7 399 Severe
04-01-2016 India Bangladesh 6.7 211 Severe
03-01-2017 India Bangladesh 5.7 11 Mod
06-02-2017 Uttarakhand 5.1 1 Mod
12-09-2018 Assam 5.3 26 Mod
24-07-2019 Maharashtra 4.1 2 Mod
28-04-2021 Assam 6.0 14 Low
28-07-2022 Chhattisgarh 4.6 5 Low
Impact
India has a long history of earthquakes. Increasing population and urbanisation with skyscrapers and indiscriminate apartments, factories, flyovers, gigantic malls, supermarkets and most of these being unscientific constructions, have all increased the frequency and magnitude of loss. Since 2001, the country has experienced 22 major and minor tremors that have resulted in over 30,000 deaths and large–scale devastation. The map of India’s seismic sensitive zone (IS 1893: 2002), indicates that about 60% of the land area is risk-prone due to seismic hazards. The entire Himalayan belt is prone to earthquakes of magnitude between 6 to 8 MSg on the Richter Scale. Geologists have indicated the likelihood of severe earthquakes in the Himalayan region that may endanger the lives of several million in that region. These events not only cause loss of life but also damage that includes: houses devastated, destruction of roads, bridges, breaches in barges and dams that can develop fissures, loss of livestock, an outbreak of diseases, and resetting all economic activity to the start. The cost of rehabilitation exerts strong pressure on the economy and many houses with families are totally destroyed. Latur earthquake caused many children to be orphans with all the elders of the hose getting crashed in the houses. One NGO named “My Home” (“आपलघर”) looked after 1650 orphaned children in Latur earthquake and educated them.
Policy
The Government of India through a special purpose vehicle namely the National Disaster Management Authority (NDMA) prepared earthquake disaster resources, shelf and capacity building programmes. The National Earthquake Risk Mitigation has also been ongoing as a Centrally Sponsored Plan Scheme with an outlay of Rs.24.87 crore, to be implemented between 2013 and 2015. The major components of National Earthquake Risk Mitigation Techno-legal Regime involving the adoption, enforcement and updating of the Techno-legal Regime in concerned Cities/States assumed prime importance as done in the Philippines (Britton 2006). This was allocated funds to the tune of Rs.8.20 Crore. Strengthening of institutional network and capacity building among citizens through colleges and educational centres was taken up at the cost of Rs.9.52 crore. This also involved practising engineers, civil contractors and works for which a separate fund of Rs 3.85 crore and for public awareness, a separate fund of Rs.1.88 Crore was allocated. A few important components were given priority in the action to mitigate the devastating effects of earthquakes in the NDMA (https://www.ndma.gov.in/Natural-Hazards):Seismic Vulnerability Assessment of Buildings typologies work was assigned to IIT Roorkee for North Zone, (2) IIT Kharagpur for East Zone, (3) IIT Guwahati for North East Region, (4) IIT Mumbai for West Zone, and (5) IIT Madras for South Zone. IIT Mumbai prepared the draft final report. However, the recommendations of these reports are awaiting implementation for want of funds.
Work has been undertaken for identifying and monitoring through the seismograph earthquake hazard zones.
NDMA has undertaken a project through Building Materials Technology Promotion Council (BMTPC) for the upgradation of Earthquake Hazards Maps for the country.
Project on research on soil piping in the highlands and foothills of Himalayas to avoid the disaster.
Soil piping2 is a recently noticed phenomenon in Kerala. It is a sub-surface soil erosion process which is a dangerous disaster since the soil erosion also takes place beneath the soil. The Centre for Earth Science studies (CESS) with financial assistance from NDMA is undertaking Soil Piping Project to study this phenomenon and suggest measures to avoid a disaster.
NDMA is financing the proposal of Mission for Geospatial Applications (MGA), Department of Science and Technology for River Monitoring, Modelling and development of Early Warning System.
Landslide Mitigation and Management in India, a technical committee has been constituted by the Ministry of Mines on the initiative of NDMA.
Flood protection, NDMA is coordinating with MoWR/CWC and Survey of India for steering and approval of the project for River Bathymetric Survey and Preparation of Digital Elevation Models.
In addition to these steps, every state government was advised to take steps that include undertaking training in the construction of earthquake proof housing, bridges and barges. Identifying and pin pointing seismically active regions and categorize them into acute, moderate and low risk categories so that the efforts could be accordingly directed. It is essential to establish a network of seismometers and undertake deep earth research on plate movements with expert geologists.
Droughts: that breaks the spine3
Drought is not simply a climatic phenomenon, but it also represents the failure of the human intellect to meet a situation that occurs due to climatic variability with a certain frequency. Over centuries human civilisation accommodated the climatic fluctuations of various magnitudes and evolved by imaginatively adjusting to them. There were also incidents of critical travails owing to the destructions created during some extreme events (Sen 1981; Dreze 1988; Sen and Dreze 1991; McMinn 1902). There are usual climatic aberrations that are regular micro climatic within one standard deviation of the bell shaped normal distribution. However, the extreme fluctuations in the climatic parameters crossing certain limits and extreme events like famines cause significant devastations resulting in human miseries.
Drought is an unescapable phenomenon which prevails in almost every region of the country with different intensities. Irrigated regions are insured with the availability of water but more frequently drought visits the rainfed regions of the country. It is a climatic event that intensifies the already fragile ecosystem of the rainfed regions and the extreme cases result in famine. It is only a climatic aberration of various magnitudes depending on the intensity of aggregate rainfall as well as the timelines of rainfall (Deshpande 2022; Nadkarni and Deshpande 1982). A devastating drought resulting in a famine fully nullifies the development efforts, besides inflicting serious miseries on the human population. The examples of these kinds of famines are many in history.
The phenomenon of droughts culminating into famines was quite pervasive in British India during the eighteenth and nineteenth centuries. Famine Commission’s Reports from 1870 to 1901 documented the travails and suffering of the population. Several researchers have analysed these reports of the Indian Famine Commission (IFC) from 1878 to 1901 (Romesh Dutt 1901; Loveday 1914; Bhatia 1967; Srivastava 1968; Baishya 1975; Brennan 1984) and brought out the failure of the state to ameliorate the impact. Therefore, the efforts in understanding the phenomenon of drought or for that matter famine could be traced back to the Famine Commission’s Reports of 1880 to 1901 as well as the First Irrigation Commission 1903, Royal Commission on Agriculture 1928 (GoI 1928).
As can be seen from Table 9, famines and droughts were certainly frequent during the British Raj. Despite their numbers being lower than that of independent India, the intensity was quite piercing. There is also a possibility of underreporting of drought miseries in smaller regions during British India. However, during those years famines were severe in impact and enveloped vast regions coupled with sluggishness in protection from the British state. In fact, the colonial government did not heed very easily into declaring a famine and bringing in any kind of amelioration measures. History of the British administration recorded that Lord Lytton (Viceroy of India between 1876 and 1880) remarked negatively to the British citizens urging relief for the suffering Indian population in 1877. Lord Lytton wrote back to London, “Let the British public foot the bill for its 'cheap sentiment’, if it wished to save life at a cost that would bankrupt India” and “there is to be no interference of any kind on the part of Government with the object of reducing the price of food," and he instructed his district officers to "discourage relief works in every possible way” (Osborn 1879).Table 9 History of droughts and famines in India: years 1800 to 2020. Source: Compiled in Deshpande (2022) from GoI (1901), Srivastava (1968), Murton (1984), Dreze (1988), Gupta (2011)
Period Years of drought and/famines
Nineteenth century first half 1801, 1803, 1804, 1806, 1812, 1818–19, 1822, 1825–26, 1832, 1833, 1837, 1839, 1845 {13}
Nineteenth century second half 1862, 1866–67, 1867–68, 1868–69, 1871–73, 1877–78, 1878–79, 1883,1891–92, 1896–97, 1898–99, 1899, 1900 {13}
Twentieth century first half 1900–01, 1904, 1905, 1907, 1908, 1911–12, 1912–13, 1916–17, 1918–19, 1921·22, 1934–35, 1939,1942–43, 1946, 1950, 1951 {16}
Twentieth century second half 1960, 1965–66, 1972–73, 1977, 1978, 1979, 1982, 1983, 1985, 1987, 1988, 1992 {12}
Twenty-first century 2002, 2009, 2014, 2015 {4)
This list does not include minor instances of droughts which were either very local, unreported widely or short-lived
Drought is a situation largely dictated by quite a few agro-climatic parameters. It must be noted that the announcement of drought till 1972–73 was based on what was commonly called the “Annawari” system, wherein the normal crop was taken as a quality to 12 anna (before the decimal system in the currency entered India, where a rupee was equivalent to 16 Annas). It was the duty of the village accountant to report to the “tehsil office” about the crop condition. Twelve Annas were considered equal to perfectly normal crop conditions and anything below eight Annas was to be reported to the Tahsildar (Government official in charge of a Tehsil). In addition to this, the activities of migrating cattle, human population, stoppage of certain usual village activities, and availability of food grains in the market were reported to the higher ups by the village official. There was a Famine Code prepared and implemented in British India after 1883. Drought as a failure of rainfall was certainly recognized in official circles; however, the declaration of a drought affected village was solely done by the revenue department through the tehsildar and the district collector (Table 10).Table 10 Meteorological sub-divisions having deficit rain fall in the crop growth period. Source: GoI (2016)
Drought year 1966 1972 1979 1987 2002 2009 2014 2015
Mid-July 19 (52.8) 19 (52.8) 19 (52.8) 19 (52.8) 19 (52.8) 19 (52.8) 19 (52.8) 19 (52.8)
Mid-August 14 (38.9) 16 (44.4) 14 (38.9) 16 (44.4) 14 (38.9) 16 (44.4) 14 (38.9) 16 (44.4)
Mid-Sept 21 (58.3) 21 (58.3) 21 (58.3) 21 (58.3) 21 (58.3) 21 (58.3) 21 (58.3) 21 (58.3)
MoA, Manual for drought management
There are a few other typologies of drought. In meteorological parlance broadly drought is viewed as a situation where the annual rainfall is less than 75% of the normal or there is a departure of − 25% from the normal. In addition to this, meteorological scientists have also developed criteria based on Moisture Index and Aridity Index (Aridity Index (AI) = {Rainfall/Potential Evapotranspiration} in mms), The Aridity Index (AI) classifies regions as Hyper-Arid: AI < , 0.05; Arid Area 0.05 < AI < 0.20; Semi-Arid 0.20 < AI < 0.50 and Dry Sub-Humid 0.50 < AI < 0.65. Another definition is that “Drought occurs at a period in a certain area when its rainfall is less than decile − 2 and severe drought occurs when rainfall is below decile-1”. Similarly, another method of Palmer’s Drought Index (1965), a two-layer approach is employed in arriving at water balance. The Palmer Drought Index includes the Palmer Drought Severity Index (PDSI), the Palmer Hydrological Drought Index (PHDI), and the Palmer Z (Moisture Anomaly Index) Index. Hydrological Drought is a result of the meteorological drought which puts stress on the surface and groundwater, thereby reducing the availability of water for different uses and arriving at the Surface Water Supply Index (SWSI). These measurements are used for the hydrological drought (Nagrajan 2010); (3) Agricultural Drought is a result of meteorological drought and hydrological drought as all the activities in the agricultural sector actually refer to the adequacy of the soil moisture during the growing season and increased aridity index. Even though it is very simple to understand the agricultural drought with satellite imagery, a Crop Moisture Index is utilized in order to declare an agricultural drought; (4) Ecological Wide Spread Drought: Ecological drought occurs when primary productivity of the natural ecosystem significantly goes down as an effect of reduced precipitation and availability of water in the ecosystem on which human as well as animal population is dependent; (5) Famine like conditions prevail when there is extreme aridity and the moisture index is at its lowest. Famine conditions are indicated by significant human and animal migration (GoI 1901; De Supriyo 2019), stress on the availability of water for drinking and other purposes, and deaths of animals and infants due to water shortages and drinking contaminated water (Fig. 3).
The early cautioning system is one concept that entered the lexicography of planning or amelioration of drought impact very late. It was till the 1987 drought, the approach to dealing with drought was only of the firefighting type. The Famine Code of 1883 continued in the books and from the declaration of drought to amelioration, works were handled in the same manner as in the fifties. The development of an early warning system was almost absent and the lag between the incidence and administrative action was quite huge. The first signals as well as the warning system should begin from the village. There is a large scope of handling it at the village level and the information seeking to be placed with the village panchayats, taluk panchayats and district panchayats. The management of the data and overseeing the data transfer from the village panchayat level was successfully carried out in Karnataka through the telemetric rain gauge stations at the village panchayat (Karnataka State Natural Disaster Management Centre). This initiated drought monitoring and preparations to meet the conditions as well as ameliorate the distress from the village level. However, the implementation of such systems needs to take precedence over the mere understanding of drought (Fig. 7, Map 5).Map 4 Seismic vulnerability Map of India
Fig. 3 Long run Rainfall behaviour. Source: Compiled by Author using data from Indian Meteorological Department
Map 5 Frequency of the occurrence of drought in16 years (2000–2016)
Among the essential steps to improve the monitoring of drought situation, the first step begins with the formation of a contingency plan from the micro level aggregating to the macro level. This plan should be further totalled at the state level and must include all the requirements of data at the village level to understand the intensity and spread of drought. Principally, the data have to be collected on the rainfall incidence, variability, moisture index of the soil with the well spread samples, crop conditions, and socio-economic responses of the people. The Soil Health Card scheme followed recently in many States provides only scientific information about the soil on the card supplied to the farmer, which the farmer neither understands nor can take corrective action and therefore, these are of little use other than keeping them in the cupboard. Recently, a World Bank-funded project on Watershed Development in Karnataka emphasized the Land Resource Inventory Card (LRI) with huge funds invested at the behest of the World Bank. This is not the only a case of putting investment down the drain on half-baked expert advice. Earlier on the advice of the World Bank experts, vetiver grass was planted as a vegetative bund with large investments in Manoli, Kabbalnala, Maheshwarnala and Puruanala watershed projects. Unfortunately, not a blade of that grass is available today in any of these watersheds and the borrowed funds invested certainly went down in the soil (Deshpande 2022).
As a submission to the Agricultural Commission 1976, the Indian Meteorological Department prepared a map of rainfall failure zones across meteorological sub-divisions. These maps were quite useful in drought planning and preparedness. The probabilities of rainfall failures in different ranges across states were arrived at and that served as a basis for future planning presented in Table 11. The states could be categorized into five levels, namely: very rare; rare, frequent, very frequent and severely drought-prone based on the probability of failure of rainfall.Table 11 Economic, hydrological social and environmental impact of drought. Source: Based on earlier studies (Deshpande 2022)
Economic Hydrological Environmental Social
Agricultural activity is suppressed and hence decline in area production and productivity Deficit in Reservoirs, Streams Tanks and Groundwater Wells
Shrinking Water Supply And Canals
Drying up of Forest and Vegetation
Shortage of fodder Availability
Shortages in other Forest Produce
Increased inequality as resourceful cultivators make good out of this opportunity
Large farmers provoke the purchase of water through tankers
Change in cropping pattern
Land degradation
Low fodder availability
Migration or sale of Animals
Drying of drinking water wells,
Impact on sources of drinking water
Non-availability of Non-Timber Forest Products
Impact on the livelihood of such persons who depend on this source
Money lenders and Informal money markets thrive on the high interest rate as demand for credit increases
Loss of Farm Employment
In search of other sources of employment
Reduced recharge of Groundwater Drinking water shortages for Animals in the Forest
Dried Water bodies
Liquor Trade and other clandestine trades proliferate
Reduced Wages
Compulsion to Migrate
Change in Vocation from Cultivator to Urban Labour
Drinking Water quality issues
Epidemics due to Drinking water pollution
Wild animals entering into human habitations and creating panic Dignity of work is given up and even a well to do farmer during non-drought year has to go for work as labour
Artisans’ loss of Work and thus changing to labour Proliferation of Tanker Water suppliers High Health Costs
Inadequate Health Infrastructure
Substantial increase in blind faith and sorcerers
Sale of Livestock and depressed allied agricultural activities Depletion in Watershed structures Difficult to rejuvenate the dried wells and tanks as the stakeholders have to survive and hence prefer other works Sale of livestock and hence the milk and dairy business is affected
Prices of all commodities increase substantially, supplies from PDS saves the affected population by providing essentials in food As the existing wells dry out, resourceful farmers dig more wells
Government also supports digging of wells or Bore wells
Depletion in the forest area and the bush forests
Needy villagers cut trees and sell the wood in local or other markets or to the contractors
Human dignity is negotiated with money
Instances of increased flesh trade and even children being abandoned
The system diagram (Fig. 4) represents three stages in the process of the impact of drought. In the first stage as soon as pre-sowing tillage is undertaken, the farmer expects the first monsoon shower that drenches the field. Sometimes, the pre-monsoon showers as well as the initial monsoon showers record failure below the normal rainfall of those three weeks. If protective irrigation is available, then the farmer continues in the first season of farming activities.Fig. 4 A system diagram of drought impact
The crops start growing, and this is the initial stage of the crops when at least 2 to 3 showers are required for healthy crop growth as well as the application of fertilizers and manure. In this mid-season of crop growth, if there is a dry spell and if protective irrigation is available only, then the farming activities continue. In the absence of a dry spell also farming and activities continue. In case these two conditions are not satisfied and if the areas are already sown, the farmer suffers the loss of seeds, labour charges utilized for tillage, sowing and other activities. Initially, the farmer can postpone sowing and may get connected to either the place indicated as (A) or (B) to reach the second stage in the system diagram. The impact on the large farmers (LF), small farmers (SF) and agricultural labourers (AL) is different, and the availability of resources only dictates the further journey of the farmer. At this stage, the farming activities get connected to (C) (D) or (G1), as depicted in stage I.
The second stage in the system diagram picks up from (A), (B), (C) from stage 1 and the farming activities continue under the condition of availability of resources to continue on the second season cropping. It is necessary to bear in mind that the loss in the kharif season makes it difficult for the farmer to have working capital in hand and thus continue cultivation in the rabi season cropping. In such situations, many farmers tend to borrow from the moneylenders as already the crop loan is used out and the usual sources of credit are also exhausted in the first season. An early dry spell in the rising rabi season puts two conditions, namely: 1. Availability of irrigation and if there was no drought in the kharif that signifies that the economic condition of the farmer is better and therefore the cultivator continues with farming. 2. In case there was a kharif drought confronted earlier by the farmer, the economic condition of the household gets deteriorated. However, in case there is protective irrigation available the farmer continues with farming, but in case it is not available, and Rabi sowing has already taken place then the farmer suffers rabi season drought and loss of seeds fertilizers and other costs of cultivation. Again, survival to the third stage depends on the resource position of the farmers and those who cannot afford to sustain the usual lifestyle either migrate out of the village or seek employment in their own village or outside. At this stage, the farmer tends to diversify their activities. The third stage of the system diagram begins with connectors (D), (E), (F), (G) and (GI) from the earlier stages. If the drought prevails and if the said position is quite strong, then the farmer continues with normal economic activities or diversifies into other economic activities. In this position, the farmer’s availability for employment decides quite a bit about future sustenance. Otherwise, if there are household assets for the purpose of sale and at times even land, the cultivator resorts to the sale of livestock, household assets, jewellery, and even land. At this point of time the cultivator is absolutely helpless and the entire household economy is in a wreck. The economics of drought directly hinges on three important components. The first one is the timeliness of rainfall at the proper time and the availability of resources for continuing the cultivation. The second is the capability to diversify and enter into the labour market or other economic activities ensures survival of the household from a complete wreckage. Third, many times extreme steps are taken by the farmers by either committing suicide or migrating to far off places.
Impact
Drought can lead to economic wreckage only when the cultivator fails in his/her ingenuity to adjust to the situation of drought at the earliest and has non-diversified economic activities. The non-availability of water leads to the collapse of many economic activities and gives rise to epidemics like cholera, influenza and many such diseases due to the use of contaminated water. The other economic activities also get impacted and the prices of usual household goods as well as agricultural necessities increase substantially. That puts heavy pressure on the cultivator households. As seen above in the system diagram, the impact of drought varies according to its severity and the point of its incidence. However, a wide spread drought causes damage not only to the crop economy in terms of productivity, production and quality of crop output but other economic activities also. It also impacts the non-recoverable cost as there is a total loss of production, loss of animals and the usual professional activities besides non-availability of the basic requirements like drinking water, milk and of all work for the workers. The impact of drought is all pervasive as it has a spillover effect on related economic activities. The loss of crop activities exerts pressure on connected production and service sector activities thereby slowing down the process of increasing GDP. This is visualized in the form of troughs observed in GDP data during drought years. All these are shown in Table 12.Table 12 Probability of occurrence of drought in the different meteorological subdivisions.
Source: Compiled by Author using Information from Indian Meteorological Department, Frequency of Rainfall in Indian Meteorological sub-divisions, IMD, Pune 2020
Sl No Meteorological sub-divisions Frequency of deficient rainfall below 75% of the normal
1 Assam Very rare, once in 15 years
2 West Bengal, Madhya Pradesh, Konkan, Bihar and Orissa Once in 5 years
3 South Interior Karnataka, Eastern Uttar Pradesh and Vidarbha Once in 4 years
4 East Rajasthan, Gujarat and Western Uttar Pradesh Once in 3 years
5 West Rajasthan, Tamil Nadu, Jammu & Kashmir and Telangana Once in 2.5 years
Policy
Drought has been one of the major despoilers of growth in the rainfed areas affecting economic activities. In the Indian context, drought strikes with a probability from as low as 10% to 50%. There are many regions which confront drought almost every second year like Rajasthan and a few other districts in the country. A complete review of the drought situation in the country as can be understood from various reports and academic researchers suggests failure of systematic efforts till the 1987 drought. There onwards, however, the preparedness has improved substantially.It is necessary to put in place an early warning system connected from taluk to district and through the State Government to NRAA. The best example of the early warning system is available in the Karnataka State Natural Disaster Management Centre’s program of establishing a network of Telemetric Rain Gauge Stations and obtaining the data on a real-time basis.
National Weather Watch Committee takes note of the process of drought in the country. The meetings of the Weather Watch Committee take place almost every week and almost daily during the drought situation.
Drought is an usual phenomena in rainfed areas, but it can also strike irrigated regions. A drought striking an unusual area is more devastating than its strike in the usual drought-prone areas. This possibility cannot be ignored, and hence, there should be preparation for drought striking even in the assured rainfall region or in non-rainfed areas.
Rural-based industries have not been proliferating as they should, in the rainfed areas these would employ the rural youth who prefer to work in industries rather than agriculture. This will also control the outmigration from rural India to urban centres.
MGNREGS is a flagship programme of the Government of India that employs a large number of agricultural labourers from rural India. Under this scheme, employment is provided for hundred days and the rest of the time the labourer is left to fend for his/her life as it goes. It will be possible to establish a Labour Supply Corporation (LSC), wherein the labourer should register with the corporation about their availability and time.
Watershed management practices are considered a panacea for the development of rainfed areas. Beginning with the initial projects in the domain of agricultural universities followed by a huge World Bank-funded Watershed Development Project at four important centres in the country has shown only the path. The technologies developed by the World Bank experts are pushed down the throat of the State governments, implementing agencies and stakeholder farmers. Therefore, these interventions disappear as soon as the World Bank team leaves the country.
There is a strong need of increasing public investment in rainfed areas as well as drought-prone areas taking up projects in rural industrialisation with the help of private industries supported by the government, wherever possible. With the availability of alternative employment, it is possible that the out-migration from rural India will reduce and the rural folks will find a better alternative for sustaining their livelihood.
Public awareness and drought proofing efforts have to be initiated and incentivized at the village level with the complete participation of the locals.
Crop insurance is not a fool proof remedy anymore due to faulty administration and manifesting as an institution governed solely by private interests. Private companies are given full authority to insure and pay the indemnity as they get State support in this operation. Area approach is being followed by these insurance companies, and this will depend on the crop cutting experiments or the metrological data which comes with a lag. It must be understood that if the farmers’ kharif crop has failed, she/he will not be able to undertake cultivation of Rabi crop immediately due to delay in the payment of indemnity.
More than anything, the “water, seed, fertilizer, pesticides” culture of technology is certainly not a solution towards drought proofing. It will not only increase the cost of cultivation and consequently net income of the farmer will fall, thereby making more farmers poorer, the farm unviable, pushing peasants out of agriculture. It was noted that between 1991and 2011 above 5 million cultivators have gone out of cultivation and that swelled the ranks of agricultural labourers moving to urban areas (Deshpande and Shaha 2021).
The best policy for drought-prone areas is diversification of rural industries along with agro processing.
COVID-19: pandemic: the biological holocaust
Covid-19 started in early 2020 with the first patient reporting from Trissur, Kerala. However, initially the Indian medical fraternity either had no knowledge of the potential threat or due to negligence did not note the severity of the threatening pandemic. Airports were not sealed immediately, and soon it spread like a whirlwind across the country. It needs no statement that Covid-19 spreads faster in high density regions and that explains why it was so severe in metropolitan cities in urban India, than in rural India. The initial wave of Covid-19 seemed to have subsided by October 2020 but immediately a second shock came in the summer of 2021. This was followed by a moderate amount of recovery in 2021–22. The first wave was at the peak in mid-September 2020 but slowed down thereafter till the end of that year. That allowed the Government to lift the nationwide restrictions were relaxed by June 2021. By the end of the second phase and till recently as of 30-09-2022, India reported officially4.46 crore cases along with 5.29 lakh deaths. The cumulative data are shown in Fig. 5 based on “ourworldindata.org” database.Fig. 5 Cumulative number of Covid-19 cases: 04/03/2020 to 30/09/2022. Source:https://ourworldindata.org/coronavirus/country/india?country=~IND#daily-confirmed-cases
One can see there are more than three spurts in the number of cases, the first came in October 2020 and steadily kept increasing. The second spurt in the cases was experienced in April 2021 and immediately by the end of May 2021 the country experienced another spurt in cases. The real jump came as the third push in April 2021 followed in July 2021. The next spurt in the cases appeared in Jan 2022, and the cases got in control from February 2022. It will be erroneous to generalize (fed by the news papers) that the pandemic had struck in two waves, actually 15 September 2020 saw a peak of 97,894 cases soon to reach 103,558 on 3 April 2021. It was assumed to be the peak but was negated by a new number of 414,188 on the 5 of May 2021. It is clear from Fig. 5 that there were three strong spurts but more fluctuations. It is argued that the recovery period was from August 2021 to January 2022, during which Covid-19 cases began declining rapidly, and the economy started recovering from the shock. By April 2022, however, again a spurt was seen on 4 August 2022. As the severity of the waves of the pandemic began subsiding, many of the nationwide mobility restrictions were gradually relaxed from June 2021 and again reintroduced in January 2022. The daily increments in the cases are shown in Fig. 6 starting from 4 March 2020 to 30 September 2022 (Fig. 11).Fig. 6 Daily increase in the Covid-19 cases in India. Source: as in Fig. 5
Fig. 7 Daily addition to the number of deaths. Data from: https://ourworldindata.org/coronavirus/country/india?country=~IND#daily-confirmed-cases
Map 6 Culture of resilience in India
The government initially responded with a partial lockdown and then clamped a full and strict lockdown. Almost all economic activities came to standstill. Economic activities in the informal sector and MSMEs (specifically dependent on casual workers) came to a complete cessation. The major dent was on the informal sector, out-sourced employees, the hospitality industry, travel, tourism, aviation, restaurants, entertainment, commercial real estate, small transport operators, etc. The workers dependent on these vocations suffered the real blow and unemployment shot up (CMIE 2020, 2021; Azim Premji University 2021). The economy underwent a severe contraction with severe suffering.
Indian economy had the first huge jolt from April to June 2020, with the country’s GDP declining by 24.4%. Followed by this in the second quarter of the 2020/21 (July to September 2020), the economy suffered another contraction of 7.4%. The mild reversal in the third and fourth quarters (October 2020 to March 2021) was not enough to cover the earlier loss. The contraction of GDP in India was (in real terms) 7.3% for the whole 2020/21 (RBI 2022a, b). The decline was the main reason for emboldening the picture of global inequality, which had been falling earlier but has started to widen again observed by some economists. While economies worldwide have been hit hard, India suffered one of the largest contractions. During the 2020/21 financial year, the rates of decline in GDP for the world were 3.3% and 2.2% for emerging markets and developing economies. Quarterly growth rates in Indian GDP in 2020–21 were − 24.4%, − 7.4%, 0.5% and 1.6%, respectively. The fact that India’s growth rate in 2019 was among the highest, caused the drop due to Covid-19 to be sharper and more noticeable. Unemployment has been one of the highly referred, important injuries inflicted by Covid-19. This was due to the lockdowns and the employers of the casual labourers did not care for them. Reverse migration was an inevitable outcome protected neither by their employer nor by the State Governments. Labour as a sector is under the concurrent list as given in the Constitution of India and, therefore, the State and Union Governments together carried the responsibility to ameliorate their livelihood shocks. India’s unemployment rate peaked at10.4% (CMIE, Jan 2020, p.5) indicating that the Govt of India could not manage this effectively compared to the reference group economies with similar per capita incomes. Possibly this was due to the composition of the Indian urban workforce where casualization of labour markets is pre-dominant. However, it should be noted that the data used for these rates were from CMIE and in the methodology of the survey, the CMIE report states “The sample size is 174,405 households (out of approx. 40 million households). Of this, 110,975 (out of approx. 11 million households) are urban households picked from 7920 CEBs of 322 towns, and, the remaining 63,430 (Approx. Out of 22 million households) rural households were picked from 3965 villages” (CMIE, Report, 2021, P. 193). This is certainly questionable, but probably at that time, there were no other sources available to get an idea of unemployment (Fig. 7).
Impact
The worst impact of the Covid-19 disaster was manifested in the number of deaths that occurred in this period. India lost 5, 28,655 persons up to 30 Sept 2022. This is from official records but many cases were not reported as death due to Covid. People carried a stigma attached to such declaring of death due to Covid and the neighbourhood stayed away from the entire family. Another severe agony was inflicted on the returning migrant labourers to their villages. There was a case reported from Kolhapur district that the migrants from the village were made to stay out of the village for a few days. During 1991 and 2011, there has been a significant increase in rural to urban migration from 21.2% in 1991–01 to 24.1% in 2001–11. This was due to the increase in economic activities at a very high speed in the urban locations. There were skyscrapers, new housing schemes, roads and other infrastructure, fly-over bridges, malls and many other casual employment opportunities that vindicated the Lewisian theory of rural–urban migration in search of higher wages. Unfortunately, this was without any preparations in urban areas to facilitate their stay in the urban cities.
The Government of India took quite a few steps to provide support to the affected population and these include some out of box measures such as direct spending and foregone/deferred revenue that included provision of in-kind (food; cooking gas) and in-cash transfers to lower income households; insurance coverage for workers in the healthcare sector; wage support and employment provision to low-wage workers, improving health infrastructure and increasing the number of hospital beds, ventilators, intensive care facilities and quarantine centres. In order to support businesses and shore up credit provision to several sectors of the economy and sections of the population, many of these measures are taken, but the implementation has many questions to answer.
Until August 2020, the Government of India had emphasized a supply-cantered strategy to boost GDP growth and the fall-out was squeezing the purchasing power needs. The employment losses due to the closure of the entire informal sector needed the generation of employment or providing some income support. In such a situation, it was certainly difficult to start any new employment programme and the MNREGS attendance also thinned down. As a result, the expenditure of the Union government on MNREGS declined between April and September 2020 as well as in April–September 2019. This was compensated by providing essential food articles to the poor and the reverse migrants. The government had little scope in expanding budgetary spending as the funds were directed towards the States as special grants.
Policy
Under the lockdown, complete disruption of all production activity was experienced with shrinking availability of capital, labour, and raw materials. All wholesale and retail markets as well as small shops were closed; even e-commerce was not operating. Under this pressure the MSMEs and all small businesses collapsed and these needed financing to restart, they downsized their businesses adding to unemployment. Along with this, the RBI announced a reduction in the policy rates and release of more liquidity and introduced a moratorium on term-loan repayments for six months. The Pradhan Mantri Kisan Samman Nidhi Yojana (PM-KISAN) providing Rs 6000 per year began with a total allocation of Rs. 160 billion. A new scheme of Direct Benefit Transfers (DBT) to old-age people, widows along with Ujjwala Yojana, and under Jan Dhan Yojana was speeded up amounting to Rs. 470 billion. On the employment front, MNREGA was extended to migrant workers and some workers in organized employment with an allocation of Rs. 922 billion. Besides, a special fund was created for construction workers of Rs. 310 billion. Direct food distribution through PDS was undertaken with an allocation of Rs. 35 billion. For the MSMEs, collateral-free bank loans of up to Rs. 3 trillion was announced. Along with this, aGovernment investment of Rs. 100 billion was initiated in funds that in turn will invest Rs. 500 billion in the equity capital of MSMEs. In addition to this and recognizing the stress on MSMEs a Rs. 200 billion subordinate debt was issued by banks and other financial institutions for strained MSMEs. Non-Banking Financial Companies were helped with Rs. 450 billion partial credit guarantee scheme, where the first 20% of the loss was guaranteed to be borne by the government. These were the new policy reforms including amendments to the Essential Commodities Act, liberalisation of investment norms for some sectors, etc.
Developing a culture of resilience
Disasters strike with a very thin probability and these events could be placed at the tail end of the normal distribution. Therefore, predictability of any disaster is a difficult process and even with historical experience, we understand the region of incidence with a rough probability of likelihood in about a decade. Most of the major disasters in India are now about too much or too little water and the meteorological forecasts are available to the population only a few days ahead of the event. There are short and focus which are given a few hours before the event and then medium and extended forecasts on the time scale. However, we must confess that our understanding of hydro-metrological or geophysical phenomena is still far away from advanced predictability. Therefore, disasters like earthquakes, landslides, flash floods and cloud bursts happen when the population is totally unprepared resulting in huge human and animal casualties besides the devastation of property, crops and infrastructure. At least since a few decades, the regions classified as most disaster-prone have been identified and maps are prepared accordingly.4 The Government of India has also opened offices of the National Disaster Management Authority along with the deployment of NDRF and SDRF in the regions. Even then, the administration fails to intake a rational decision while preparing for disaster management.
Certainly, disasters strike without sufficient warning and at times the events crawl into human life. Preparedness for the eventuality is one of the important steps before the disaster strikes. Even in Chanakya’s Arthshastra, it has been recorded that “There are eight kinds of providential visitations: These are fire, floods, pestilential diseases, earthquakes, famine, rats, tigers, serpents, and demons. From these shall the king protect his kingdom” (Shamasastry, p 294). Thus, the responsibility for ameliorating the conditions rests on the king or the government. It is necessary to note that the form of the government had changed from a kingdom to the Republic of India. It is not just a change of nomenclature, but such transformation enforces equal responsibility on the government and the people and that is most important to understand in today’s context. People’s participation in the preparedness as well as getting ideas from those who have been historically suffering disasters is essential. Our approach has always been top-down rather than learning from those who are at the bottom and suffer the most. This is expected in the republic of India (Map 6).
The map shows resilient, slightly non-resilient, moderately non-resilient and severely non-resilient districts in the country. This was prepared based on the perceptions of the disaster indicated by the population residing in the district. The level of resilience is largely decided by the probability of occurrence of extreme events and as some of the events occurred with very thin probability the perception of resilience differs significantly. Resilience is formed by three overlapping elements: (1) exposure (the shocks and stresses experienced by the system), (2) sensitivity (the response of the system), and (3) adaptive capacity (the capacity of the system with adaptive action) (Chakraborty and Joshi 2016). Like some of the coastal districts of Tamil Nadu and Odisha have a frequency of cyclones but their expectations are not formed based on the probability the population is not so much frightened of the event or rather thinks that precautionary steps can be taken at the time of the event itself. In the olden days, society was more cohesive and social interests preceded self-interests but during the commercialisation process when self-interest predominated, social insurance slowly vanished. Society as a whole is not taking collective decisions to protect from disasters. Panicky behaviour and herd mentality predominated in society and that increased the intensity of the damages. Information also plays a very important role in preparedness for such events. It will be essential to state here that every disaster causes huge devastation and loss of lives and livelihood. Many extreme disasters reset the development clock and every effort is forced to start right from the region. Infrastructure, houses, schools and roads and whatever comes in the way of the brute force of nature are destroyed and wrecked. It takes a long time to rebuild the physical assets but even longer to recoup the psychological impact on the population.
In the current situation, it should not be simply the responsibility of the government alone, but the entire rehabilitation and response work must be carried out by the people as well as the State. In India, we have accepted our Constitution and defined the nation as the Democratic Republic and therefore, the responsibility is equally on the people of the nation to meet the challenge thrown by the natural elements. Preparedness for the worst is the best kind of insurance besides the usual disaster insurance provided by financial intermediaries. The first step in such preparedness requires identifying the vulnerable regions at three levels, namely: (1) Seriously vulnerable, (2) Vulnerable and (3) Regions with low vulnerability. Similarly, in society also susceptibility differs according to the sensitivity of a social group and its social economic status. It is well known that disaster strikes fiercely at the most vulnerable. Therefore, poverty and destitution are quite common and located in these most vulnerable regions (Bhalla et al. 2022). The government has made significant efforts in putting several institutional frameworks to meet the disaster and ameliorate the after effects.5 It is necessary that these institutions work with horizontal coordination enhancing efforts of each other. Therefore, a coordination Body like NDC involving State Home Ministers with Non-Government Expert Participation is necessary to take a periodic overview. The coordination or joint meetings of NEC and NCMC are also necessary to discuss resilience. Actually, equal responsibility and liability also lie with the private sector operating in the region. Community participation; cooperation of local administration with NDRF and Permanent Disaster Management Funds and regular contributions from the States and Central Budget are essential components.
A final plan to meet the requirements of managing every type of disaster in India requires integration of disaster mitigation and preparedness in development programmes with multi-sectoral/inter-departmental coordination and multi-hazard approach in disaster management planning. An important component is the preparedness of society through public awareness and community capacity, to cope with the hazards, reduce dependence on the government and build a culture of self-help. It needs to be seen if the Task Force for Review of DMA Recommendations-2013 as well as the NDMA Plan of 2016 is implemented, and that step can go a long way if executed scrupulously. In all, full transparency of plans and actions needs to be maintained with clear documentation. Integration of State and National DM Plans for ensuring food security and establishing quick shelter facilities is a vital component. In any disaster, transportation gets impacted; therefore, establishing proper transport networks is crucial. Preparedness at the household level also needs to be looked into in the capacity building programmes. Finally, it is a battle of wits between human intelligence and Pancha Mahabhutas; therefore, disaster strikes without much warning throwing a challenge of ingenuity for defence with all the preparations to the human race.
Acknowledgements
I thank my guru Prof M V Nadkarni, Hon Professor, ISEC, Dr V S Prakash, former Director KSNDMC, and Dr Khalil Shaha for help as well as many inputs.
Funding
There is no funding provided for this article.
Declarations
Conflict of interest
There is no conflict of interest.
1 Rosetta Stone is a stele composed of granodiorite inscribed with three versions of a decree issued in Memphis, Egypt, in 196 BC during the Ptolemaic dynasty on behalf of King Ptolemy V Epiphanes.
2 Soil piping is a naturally occurring hydraulic process that leads to the development of macrospores (large air filled voids) in the sub-surface.
3 This section draws from my earlier work (Deshpande 2022).
4 Maps are the best tools which incorporate many dimensions of any concept in a simple two-dimensioned diagram.
5 National Safety Council of India-04–03-1966; National Centre for Disaster Management-1995; India Disaster Resource Network-2004; National Institute for Disaster Management-2004; National Disaster Management Act-2005; State Disaster Management Authorities-2005; National Disaster Response Force-2006; National Policy on Disaster Management 2009; National Disaster Management Plan-2016,18,19; National Emergency Operation Centre-2020.
L.S. Venkataramanan Memorial Lecture; This lecture was delivered by Prof. R. S. Deshpande at the Institute for Social and Economic Change, Bangalore (ISEC) as LSV Memorial Lecture on October 28, 2018. The article was submitted to ISEC for publication and circulation. This lecture is presented here in its original version (with minor modifications).
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Asimakopoulou E, Nik B (2010) Advanced ICTs for disaster management and threat detection: collaborative and distributed frameworks, information science reference, New York
Azim Premaji University (2021) State of working in India 2021: Report from Azim Premji University’s Centre for Sustainable Employment on the effects of Covid-19 on jobs, incomes, inequality and poverty
Baishya P Man-made famine Econ Pol Wkly 1975 10 21 821 822
Bhalla S, Bhasin K, Virmani A (2022) Pandemic, poverty and inequality: evidence from India. IMF Working Paper No. 2022/069, Washington DC
Bhatia BM Famines in India 1860–1965 1967 London Asia Publishing
Bhatt D Mall RK Banerjee T Climate change, climate extremes and disaster risk reduction Nat Hazards 2015 78 1 775 778 10.1007/s11069-015-1874-8
Brennan L (1984) The development of the India famine codes: personalities, policies and politics. In: Currey B; Hugo G (eds) Famine as a geographical phenomenon. Geo Journal library, vol 1. Springer, pp 91–110
Britton NR (2006) Getting the foundation right: in pursuit of effective disaster legislation for the Philippines. In: Proceedings of 2nd asian conference on earthquake engineering, March 10–11, Manila’
Chakraborty A Joshi PK Mapping disaster vulnerability in India using analytical hierarchy process Geomat Nat Haz Risk 2016 7 1 308 325 10.1080/19475705.2014.897656
De US Dube RK Prakasa Rao GS Disastrous weather events annual reports, IMD, 1968–77 extreme weather events over India in the last 100 years J Indian Geophys Union 2005 9 3 173 187
Deshpande RS Basu DN Kashyap SP Watershed development programme : a strategy under ACRP approach Agro climatic regional planning in India 1996 New Delhi Concept Publishing
Deshpande RS, Shaha K (2021) India’s political economy: genesis and changes in the last decade. In: Changing dynamics of the indian economy: the decade of twenty –tens and ahead, a book in commemoration of the birth centenary of Prof VM Dandekar, Symbiosis, Pune
Deshpande RS (2022) Under the shadow of development: rainfed agriculture and droughts in agricultural development of India, NABARD, Mumbai
Dev SM, Sengupta R (2020) Covid-19: impact on the Indian economy. Indira Gandhi Institute of Development Research, Mumbai Working Papers 2020-013, Indira Gandhi Institute of Development Research, Mumbai, India
Dhar CPG Integrating disaster risk reduction with climate change adaptation—recent initiatives in South Asia Disast Dev 2010 4 1 159 183
Drèze J (1988) Famine prevention in India, wider working paper no 45. World Institute of Development Economics Research, United Nations University, Helsinki
Dutt RC Indian famines, their causes and prevention 1901 London Springer 91 110
De Supriyo (2019) Internal migration in India grows, but inter-state movements remain low, December 18, 2019. https://blogs.worldbank.org/peoplemove/internal-migration-india-grows-inter-state-movements-remain-low. Accessed on 15 Nov 2021
Emergency Events Database (2016) The international disaster database. Natural Disasters Trends. Brussels, Belgium
Gopal M India through the ages 1990 Publication Division, Ministry of Information and Broadcasting, Government of India 79
Government of India (1901). Report of the Indian Famine Commission, 1901, Calcutta.
Government of India (1928) Report of the Royal Commission on Agriculture in India, (Presided by: Sir Flickr. Linlithgow), Reprinted by Agricole Pub. NewDelhi (1983)
Government of India (2001) The report of High Powered Committee on Disaster Management, (Chairman: J. C. Pant). National Centre for Disaster Management, Indian Institute of Public Administration, New Delhi
Government of India (2005) The Disaster Management Act, 2005; Act 53, Ministry of Home, New Delhi
Government of India (2008) National action plan on climate change. Ministry of Environment, Forest and Climate Change, New Delhi
Government of India (2011) Disaster management in India. Ministry of Home Affairs New Delhi
Government of India (2012) India second national communication to the United Nations framework convention on climate change. Ministry of Environment and Forest New Delhi
Government of India (2013) Report of the task force—a review of the Disaster Management Act, 2005, Ministry, New Delhi
Government of India (2016) National disaster management plan, a publication of the National Disaster Management Authority, Government of India, May 2016, New Delhi
Government of India (2019) Mainstreaming disaster risk reduction and climate change adaptation in the national flagship programmes, report prepared under the GOI-UNDP project on ’enhancing institutional and community resilience to disasters and climate change (2013–2020). Ministry of Home Affairs, New Delhi
Gupta AK Tyagi P Sehgal VK Drought disaster challenges and mitigation in India: strategic appraisal Curr Sci 2011 100 12
IPCC (2013) Climate change 2013: the physical science basis. Contribution of Working Group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change, Intergovernmental Panel on Climate Change, Report. In: Stocker TF, Qin D, Plattner GK, Tignor M, Allen SK, Boschung J, Nauels A, Xia, Y et al (eds) Cambridge University Press, Cambridge
Kamepalli LB Disaster relief financing: a journey from margin money to state disaster response funds Econ Polit Weekly 2019 1 4 16
Kelman I Gaillard JC Placing climate change within disaster risk reduction Disast Adv 2008 1 3 3 5
Kelman I Mercer J West JJ Integrating indigenous and scientific knowledge for community based climate change adaptation Particip Learn Action 2009 60 41 53
Loveday A (1914) The history and economics of Indian famines. A.G. Bell and Sons, London (reprinted by Usha Publications, New Delhi, 1985)
Mall RK Gupta A Singh R Rathore LS Water resources and climate change—an Indian perspective Curr Sci 2006 90 12 1610 1626
Mall RK Kumar R Bhatla R Climate change and disaster in India J South Asian Disast Stud 2011 4 1 27 76
Mall RK Attri SD Kumar S Extreme weather events and climate change policy in India J South Asian Disast Stud 2011 4 2 37 56
Mall RK Ravindra K Srivastava TB Mishra OP Bhatt D Sonkar G Disaster risk reduction including climate change adaptation over south asia: challenges and ways forward Int J Disast Risk Sci 2019 10 14 27 10.1007/s13753-018-0210-9
McMinn CW Famine: truth, half-truths, and untruths 1902 Calcutta Baptist Mission Press
Murton B Currey B Hugo G Spatial and temporal patterns of famine in Southern India, before the Famine codes Famine as a geographical phenomenon, GeoJournal library 1984 Springer 71 90
Nadkarni MV Deshpande RS Agricultural growth, instability in productivity and rainfall-case of Karnataka Econ Polit Weekly 1982 17 52 127 134
Nagarajan R Drought: assessment, monitoring, management and resource conservation 2003 New Delhi Capital Publishing Company 312
Nagarajan R Drought assessment 2010 Springer
Noorani AG The inundation of Morvi Econ Polit Weekly 1979 14 34 1454
O’Keefe P Westgate K Wisner B Taking the naturalness out of natural disasters Nature 1976 260 5552 566 567 10.1038/260566a0
Osborn RD (1879) The truth about the Indian famine of 1877–78. Contemp Rev
Pandey RK (2016) Legal framework of disaster management in India. ILI Law Review, Winter Issue 2016
Reserve Bank of India (2022a) RBI bulletin. Reserve Bank of India, Mumbai
Reserve Bank of India (2022b) Scars of the pandemic. Report on Currency and Finance, Mumbai
Richter CF Elementary seismology 1958 San Francisco W H Freeman & Co
Seidler R Dietrich K Schweizer S Bawa KS Chopde S Zaman F Sharma A Bhattacharya S Devkota LP Khaling S Progress on integrating climate change adaptation and disaster risk reduction for sustainable development pathways in South Asia: evidence from six research projects Int J Disast Risk Reduct 2018 31 92 101 10.1016/j.ijdrr.2018.04.023
Sen A Dreze J Hunger and public action (WIDER studies in development economics) 1991 Oxford Oxford University Press
Sen A (1981) Poverty and famines: an essay on entitlements and deprivation. Coromandal Lecture (1982), and book by Oxford Clarendon Press, Dreze 1988
Shah AJ (2011) An overview of disaster management in India, vol 119. Wessex Institute of Technology (WIT) Transactions on the Built Environment
Shamsuddoha M, Roberts E, Hasemann A, Roddick S (2013) Establishing links between disaster risk reduction and climate change adaptation in the context of loss and damage: policies and approaches in Bangladesh, Department for International Development Dhaka
Sharma VK, Kaushik AD (2012) Natural disaster management in India, Yojana, March 2012
Srivastava HS History of Indian famines and development of famine policy-1858-1918 1968 Agra Sri Ram Mehra & Co
UNESCAP (2016) Asia-pacific disaster report 2015: disasters without borders—regional resilience for sustainable development. United Nations Economic and Social Commission for Asia and the Pacific Bangkok
UNISDR Global assessment report on disaster risk reduction: revealing risk, redefining development—summary and main findings 2011 Geneva United Nations International Strategy for Disaster Reduction
UNISDR/SDMC (2014) Integration of disaster risk reduction and climate change adaptation in SAARC region: implementation of the Thimphu statement of climate change—a comprehensive study of the policy, institutional landscape and resource allocation for disaster risk reduction and climate change adaptation in South Asia (Disaster prevention, preparedness & management, linkages with CCA). United Nations International Strategy for Disaster Reduction/SAARC Disaster Management Centre, New Delhi
United Nations and World Bank Natural hazards, unnatural disasters economics of effective disaster prevention 2010 Joint Publication by UN and World Bank
Vij S Biesbroek R Groot A Termeer K Changing climate policy paradigms in Bangladesh and Nepal Environ Sci Policy 2018 81 77 85 10.1016/j.envsci.2017.12.010
World Bank India-financing rapid onset natural disaster losses in India: a risk management approach 2010 Washington The World Bank
www.indianscriptures.com Pancha Bhootas or the five elements. Retrieved 25 Mar 2015
www.undrr.org/terminology/disaster Accessed 06 July 2022
Zuccaro G Leone MF Building resilient cities: a simulation-based scenario assessment methodology for the integration of DRR and CCA in a multi-scale design perspective Proc Eng 2018 212 871 878 10.1016/j.proeng.2018.01.112
| 0 | PMC9746566 | NO-CC CODE | 2022-12-15 23:21:56 | no | J Soc Econ Dev. 2022 Dec 13;:1-40 | utf-8 | J Soc Econ Dev | 2,022 | 10.1007/s40847-022-00225-w | oa_other |
==== Front
Unfallchirurgie (Heidelb)
Unfallchirurgie (Heidelb)
Unfallchirurgie (Heidelberg, Germany)
2731-7021
2731-703X
Springer Medizin Heidelberg
1257
10.1007/s00113-022-01257-x
Leitthema
Digitalisierung und Telerehabilitation
Digitalization and telerehabilitationPförringer Dominik [email protected]
1PD Dr. Dominik Pförringer
ist Facharzt für Orthopädie und Unfallchirurgie. Seit 2018 ist er Co-Vorsitzender der AG Digitalisierung der DGOU sowie Organisator des Digital Health Summit im November in München: www.DigitalHealthSummit.de
AG Digitalisierung der DGOU
Back David 2PD Dr. David Back
ist Facharzt für Orthopädie und Unfallchirurgie. Er ist 2018 Vorsitzender der AG Digitalisierung der DGOU sowie Organisator des Kongresses „Digitalisation in Orthopaedics and Traumatology“ im Dezember 2022 in Berlin: https://otdigital.eu/.
1 grid.15474.33 0000 0004 0477 2438 Klinik und Poliklinik für Unfallchirurgie, Klinikum rechts der Isar der technischen Universität München, Ismaninger Str. 22, 81675 München, Deutschland
2 Klinik für Unfallchirurgie und Orthopädie, Bundeswehrkrankenhaus Berlin, Berlin, Deutschland
Redaktion Christoph Gutenbrunner, Hannover
13 12 2022
13
8 11 2022
© The Author(s), under exclusive licence to Springer Medizin Verlag GmbH, ein Teil von Springer Nature 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Digitalisierung nimmt einen kontinuierlich wachsenden Stellenwert in der Medizin ein. Prozesse werden optimiert, Daten digital erfasst, analysiert und archiviert. Wenngleich in Deutschland noch ein vergleichsweise großer Aufholbedarf existiert, befinden wir uns auf einem soliden Weg. Mit der Etablierung des European Health Data Space (EHDS) wird in Brüssel ein Meilenstein für den sicheren Austausch von Daten gesetzt. Digitalisierung beinhaltet das Potenzial für umfangreiche Prozessoptimierung. Während derzeit ein Gros der Arbeitszeit deutscher Ärzte von Bürokratie verschlungen wird, kann ein relevanter Anteil dieser Arbeit durch digitale Lösungen effizienter erledigt werden. Die Digitalisierung ersetzt nicht den Arzt, sondern unterstützt ihn, zum Wohle des Patienten. Zahlreiche Wege sowie damit verbundene Transport- und Logistikkosten sind vermeidbar bzw. durch digitale Ergänzungen und neue Darreichungsformen auf digitalem Wege adressierbar. Dies schont Ressourcen, spart Zeit und optimiert die Versorgung. Die Offenheit und Affinität von Ärzten und Patienten für das Thema hängen bedeutend von der Digital Health Literacy, d. h. dem Verständnis und Wissen zum Thema, ab. Kontinuierlich Ängste abzubauen und Akzeptanz zu steigern, ist das Ziel der kommenden Jahre. Zudem bedarf es relevanter Investitionen in die technische Grundausstattung auf Soft- und Hardwareseite.
Digitalization and digitization are becoming increasingly more important in medicine. Processes are being optimized and data are being digitally recorded, analyzed and archived. Although there is still a comparatively large need to catch up in Germany, we are on a solid transformation path. The establishment of the European Health Data Space (EHDS) in Brussels represents a milestone for the secure exchange of data. Digitalization holds the potential for extensive process optimization. While a large part of the working time of physicians in German is currently consumed by bureaucracy, a relevant part of this can be solved digitally. The digitalization does not replace the physician but plays a supporting role for the benefit of the patient. Numerous routes and the associated transport and logistics costs can be avoided or addressed digitally through digital supplementation and new forms of treatment administration. This conserves resources, saves time and optimizes the care of patients. The openness and affinity of physicians and patients towards the topic significantly depends on digital health literacy, i.e. the understanding and knowledge on the topic. The goal for the coming years is to continually reduce fears and increase acceptance. In addition, relevant investments are needed for the basic technical equipment on the software and hardware side.
Schlüsselwörter
E‑Health
Telekommunikation
Telemedizin
Gesundheitskompetenz
Metaphylaxe
Keywords
E‑health
Telecommunications
Telemedicine
Health literacy
Metaphylaxis
==== Body
pmcDie Digitalisierung der Medizin gewinnt laufend an Relevanz und hat durch zahlreiche Gesetze seit 2016 einen entsprechenden Rahmen erhalten. Digitale Technologien offerieren neue Ansätze und Herangehensweisen hinsichtlich Prävention, Diagnostik, Therapie und Metaphylaxe. Im Klinik- und im Praxisalltag kann die Digitalisierung eine signifikante Unterstützung bieten, gestaltet sich aber oft immer noch auch als Herausforderung, ein Versprechen, das es zu erfüllen gilt [4].
Die COVID-19-Pandemie hat sich als zusätzlicher Katalysator erwiesen. Digitale Suchen nach medizinischer Information haben zugenommen, wobei die medizinische Zuverlässigkeit der gefundenen Informationen nach wie vor einen Schwachpunkt darstellt [3].
Digitalisierung kann und wird den Arzt im Klinik- und Praxisalltag unterstützen und Zeit einsparen
Digitale Innovationen wurden ausgegründet, Produkte bzw. Anwendungen eingeführt oder verbessert – was ohne die politischen Weichenstellungen der Vorjahre nicht möglich gewesen wäre. Neben der klinischen Praxis wurde der wissenschaftliche Output in der Orthopädie und Unfallchirurgie seit 2019 im Bereich digitaler Themen deutlich gesteigert.
Orthopädie und Unfallchirurgie und die Telerehabilitation
Die zunehmende Digitalisierung in Kombination mit dem demografischen Wandel beflügelt die Entwicklung neuartiger Versorgungskonzepte [15]. Orthopädie und Unfallchirurgie als mechanisch fokussierte Disziplinen sind nicht primär prädestiniert, um digitale Versorgung zu ermöglichen, da diese beiden Fächer oft physische Untersuchungsmethoden nötig machen. Des Weiteren erfordern die chirurgischen Fächer in einer Vielzahl der Fälle eine physisch basierte Therapie. Aus diesem Grund sind beispielsweise die bildgebenden Fächer den chirurgischen logischerweise im Thema Digitalisierung von Datensätzen und Datenanalyse weit voraus. Gleichzeitig wird es zunehmend klar und immer weiter wissenschaftlich untersucht, welchen Einfluss eine koordinierte physiotherapeutische Nachbehandlung auf das Outcome von Patienten mit orthopädischen Eingriffen hat [5, 10]. Diese kann und wird zunehmend auch digital unterstützt und damit ohne laufenden persönlichen Kontakt durchgeführt bzw. begleitet. Die Effizienz dieser Telerehabilitation steht konventionellen Programmen gemäß der aktuellen Studienlage nicht nach [8], wobei es noch an Langzeitergebnissen zur Verifizierung dieser Ergebnisse fehlt.
Die Kombination aus analoger und digitaler Behandlung bietet dem Patienten das Beste aus zwei Welten
Eine primäre physisch präsente Unterweisung des Patienten in sein physiotherapeutisches Nachsorgeschema in der erstbehandelnden Institution erscheint als Ausgangspunkt zur Einleitung der postoperativen Nachbehandlung nach einem Gelenkersatz sinnvoll [6]. Die Kombination aus analoger und digitaler Nachsorge scheint als ideale Ausgangsposition ein Zukunftspotenzial zu offerieren.
Die Ergebnisse der Studie von Tahami et al. deuten hingegen an, dass Telerehabilitation nach arthroskopischen Meniskusrekonstruktionen noch nicht mit konventioneller persönlicher Therapie auf Augenhöhe steht [14]. Fokussierte Studienprotokolle für geplante oder bereits begonnene Beobachtungsstudien versprechen Aufschluss über den Vergleich der beiden Behandlungswege [16].
Mobile Apps geben Patienten die Möglichkeit, diverse Übungen ortsunabhängig, jedoch kontrolliert und medizinisch begleitet durchzuführen. Dies ermöglicht die kontinuierliche Fortführung der Rehabilitationsmaßnahme über den Zeitraum der ambulanten oder stationären Behandlung hinaus. Inwieweit die Initiative der digitalen Gesundheitsanwendungen (DiGA) der Bundesregierung auch im Bereich der Telerehabilitation nachhaltige Erfolge erzielen kann, ist im Moment noch nicht abzuschätzen und Gegenstand verschiedener Studien und Analysen [9]. Bäcker et al. konnten bereits Vorteile der App-gesteuerten Reha-Unterstützung nachweisen [2].
Forschungsanalysen legen für Gesundheitsfachberufe wie Physiotherapeuten bereits heute nahe, dass telerehabilitative digitale Ansätze, nicht nur zu Zeiten der COVID-19-Pandemie, sondern auch in Zukunft die diagnostische oder therapeutische Entscheidungsfindung und damit die Patientenbehandlung nachhaltig beeinflussen dürften [12].
Soft- und Hardwarelösungen werden kontinuierlich verbessert, ebenso wie die Verfügbarkeit des zuverlässigen Datenaustausches auf mobilem Wege. Technische Geräte wie Digitalkameras sind, beschleunigt durch die pandemische Situation, weitreichend verfügbar, und ihr Nutzen steht großen Teilen der Bevölkerung offen. Die Vorzüge innovativer Kommunikationswege wurden von Back et al. anhand der Videosprechstunde ausgewertet [1].
Ausschlaggebend für Erfolg und Akzeptanz ist die gemeinsame Entwicklung innovativer Lösungen
Ausschlaggebender Faktor für Erfolg und Akzeptanz neuartiger Lösungsansätze ist die gemeinsame Entwicklung innovativer Lösungen in Kooperation zwischen Medizinern, auch Physiotherapeuten und Technikern, die Bedürfnisse und Anforderungen von Patienten und Heilberufen fachlich und technisch begreifen und somit adäquat adressieren können [7]. In vielen Fällen sind sowohl Techniker allein als auch medizinische Dachberufe allein nur inadäquat in der Lage, die Thematik umfassend zu lösen, da diesen Teams das relevante Schnittstellen-Know-how fehlt.
Um Vorurteile und Ängste zu reduzieren, bedarf es zudem der Aus- und Weiterbildung im Bereich digitaler Gesundheit zur Steigerung der sog. Digital Health Literacy. Nur, wer die Technik begreift und nicht fürchtet, wird gewillt und in der Lage sein, sie einzusetzen, sie zu verordnen und sie weiterzuempfehlen [11, 13].
Fazit für die Praxis
Die Grundsteine sind gelegt; die technischen Möglichkeiten, um die Telemedizin in Deutschland auf dem Gebiet der muskuloskeletalen Fächer auszubauen und weiterzuentwickeln, existieren großteils.
Telerehabilitation kann und wird in Zukunft flächendeckend als Ergänzung und erweiternde Option zu konventionellen Anschlussheilbehandlungen angeboten werden. Effizienz und Effektivität auf lange Sicht werden weiterhin wissenschaftlich begleitend untersucht werden.
Die enormen Potenziale zur Reduktion von Bürokratie durch den Einsatz digitaler Technologien können dem Gesundheitswesen auf lange Sicht auf Mikro- und auf Makroebene zugutekommen.
Einhaltung ethischer Richtlinien
Interessenkonflikt
D. Pförringer, AG Digitalisierung der DGOU und D. Back geben an, dass kein Interessenkonflikt besteht.
Für diesen Beitrag wurden von den Autoren keine Studien an Menschen oder Tieren durchgeführt. Für die aufgeführten Studien gelten die jeweils dort angegebenen ethischen Richtlinien.
QR-Code scannen & Beitrag online lesen
==== Refs
Literatur
1. Back DA Estel K Pforringer D Implementation of online video consultations in a regional health network: a management feasibility analysis from an orthopedic perspective BMC Health Serv Res 2022 22 1029 10.1186/s12913-022-08352-0 35962358
2. Bäcker HC Wu CH Pförringer D A review of functional outcomes after the app-based rehabilitation of patients with TKA and THA J Pers Med 2022 12 1342 10.3390/jpm12081342 36013291
3. Dadaczynski K Okan O Messer M Digital health literacy and web-based information-seeking behaviors of university students in Germany during the COVID-19 pandemic: cross-sectional survey study J Med Internet Res 2021 23 e24097 10.2196/24097 33395396
4. Dunn P Hazzard E Technology approaches to digital health literacy Int J Cardiol 2019 293 294 296 10.1016/j.ijcard.2019.06.039 31350037
5. Fortier LM Rockov ZA Chen AF Activity recommendations after total hip and total knee arthroplasty J Bone Joint Surg Am 2021 103 446 455 10.2106/JBJS.20.00983 33337819
6. Greiner JJ Drain NP Lesniak BP Self-reported outcomes in early postoperative management after shoulder surgery using a home-based strengthening and stabilization system with Telehealth Sports Health 2022 10.1177/19417381221116319
7. Hil/Aerzteblatt.De (2017) Mehr ärztliches Know-how bei der Entwicklung von Gesundheits-Apps nötig. In:Aerzteblatt.de. https://www.aerzteblatt.de/nachrichten/80601/Mehr-aerztliches-Know-how-bei-der-Entwicklung-von-Gesundheits-Apps-noetig. Zugegriffen: 02.10.2022
8. Hirohama K Tamura H Hamada K Effects of non-face-to-face and noncontact interventions on knee pain and physical activity in older adults with knee osteoarthritis: a systematic review and meta-analysis J Aging Phys Act 2022 10.1123/japa.2022-0037
9. Lobker W Bohmer AC Hofgen B Support for innovation at the BfArM-experiences from the consultations on digital health applications (DiGA) Bundesgesundheitsblatt Gesundheitsforschung Gesundheitsschutz 2021 64 1241 1248 34519834
10. Moyer R Ikert K Long K The value of preoperative exercise and education for patients undergoing total hip and knee arthroplasty: a systematic review and meta-analysis JBJS Rev 2017 5 29232265 e2 10.2106/JBJS.RVW.17.00015
11. Nutbeam D Lloyd JE Understanding and responding to health literacy as a social determinant of health Annu Rev Public Health 2021 42 159 173 10.1146/annurev-publhealth-090419-102529 33035427
12. Seron P Oliveros MJ Gutierrez-Arias R Effectiveness of telerehabilitation in physical therapy: a rapid overview Phys Ther 2021 10.1093/ptj/pzab053Free
13. Smith B Magnani JW New technologies, new disparities: the intersection of electronic health and digital health literacy Int J Cardiol 2019 292 280 282 10.1016/j.ijcard.2019.05.066 31171391
14. Tahami M Vaziri AS Tahmasebi MN The functional impact of home-based self-rehabilitation following arthroscopic meniscus root repair BMC Musculoskelet Disord 2022 23 753 10.1186/s12891-022-05662-6 35932028
15. Vogt F Seidl F Santarpino G van Griensven M Emmert M Edenharter G Pförringer D Eur Surg Res 2018 59 1 100 113 10.1159/000490241 30048992
16. Wang Q Hunter S Lee RL Mobile rehabilitation support versus usual care in patients after total hip or knee arthroplasty: study protocol for a randomised controlled trial Trials 2022 23 35804429 553 10.1186/s13063-022-06269-x 35804429
| 36512038 | PMC9746568 | NO-CC CODE | 2022-12-15 23:21:56 | no | Unfallchirurgie (Heidelb). 2022 Dec 13;:1-3 | utf-8 | Unfallchirurgie (Heidelb) | 2,022 | 10.1007/s00113-022-01257-x | oa_other |
==== Front
Spinal Cord
Spinal Cord
Spinal Cord
1362-4393
1476-5624
Nature Publishing Group UK London
867
10.1038/s41393-022-00867-x
Article
Developing spinal cord injury physiotherapy clinical practice guidelines: a qualitative study to determine how physiotherapists and people living with spinal cord injury use evidence
Nunnerley Joanne L. [email protected]
12
Glinsky Joanne V. 34
http://orcid.org/0000-0002-2894-7533
Dunn Jennifer A. 2
Stavric Verna A. 5
Haber Amanda 6
Denis Sophie 7
Ben Marsha 6
Chen Lydia W. 8
Harvey Lisa A. 34
1 Burwood Academy, Christchurch, New Zealand
2 grid.29980.3a 0000 0004 1936 7830 University of Otago, Christchurch, New Zealand
3 grid.1013.3 0000 0004 1936 834X Kolling Institute, Faculty of Medicine and Health, The University of Sydney, Sydney, NSW Australia
4 grid.482157.d 0000 0004 0466 4031 John Walsh Centre for Rehabilitation Research, Northern Sydney Local Health District, Sydney, NSW Australia
5 grid.252547.3 0000 0001 0705 7067 Auckland University of Technology, Auckland, New Zealand
6 grid.419366.f 0000 0004 0613 2733 Royal Rehab, Sydney, NSW Australia
7 grid.415193.b Prince of Wales Hospital, Sydney, NSW Australia
8 grid.412703.3 0000 0004 0587 9093 Royal North Shore Hospital, Sydney, NSW Australia
13 12 2022
19
7 7 2021
4 10 2022
9 11 2022
© The Author(s), under exclusive licence to International Spinal Cord Society 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Study design
Generic qualitative design.
Objectives
Australian and New Zealand SCI physiotherapists are developing clinical practice guidelines for the physiotherapy management of people living with spinal cord injury. To guide the development of the guidelines it was important to understand how physiotherapists and people living with spinal cord injury use evidence to choose interventions and the potential barriers and facilitators to the uptake of the clinical practice guidelines.
Setting
Spinal Cord Injury Centres in Sydney, Australia and New Zealand.
Methods
Focus groups and interviews with physiotherapists and people living with spinal cord injury were recorded, transcribed, and subjected to thematic analysis.
Results
A total of 75 participants took part in the study, 45 physiotherapists and 30 people living with spinal cord injury. Three main themes were identified from the data: (1) Types and sources of evidence that influence treatment choices, (2) the many factors determining treatment choices, and (3) ways in which clinical practice guidelines could influence treatment.
Conclusions
Clinical practice guidelines have the potential to reduce the barriers identified by physiotherapists in accessing and interpreting research evidence on interventions for people living with spinal cord injury. Supported implementation of guidelines is required to demonstrate their benefit and encourage physiotherapists to factor in evidence when balancing the multiple factors influencing choice of physiotherapy intervention.
Subject terms
Health care
Spinal cord diseases
Physiotherapy New Zealand Neurology Special interest groupiCare Australia
==== Body
pmcIntroduction
Physiotherapy is an integral part of the rehabilitation of people living with SCI (PLwSCI) and crosses an extensive scope of practice and specialty [1]. There are various physiotherapy interventions available to PLwSCI [1]. A recent mapping review published in Spinal Cord identified over 450 randomised controlled trials involving PLwSCI. Nine of the top ten topic areas were relevant to physiotherapists, with an estimated 300 plus randomised controlled trials relevant to physiotherapy practice [2]. However, surprisingly few high-quality and conclusive randomised controlled trials support the effectiveness of physiotherapy interventions [3].
Physiotherapists are often conflicted when deciding on treatments knowing the treatments they provide, or patients request, are not always supported by high-quality evidence [3]. Even when research evidence is available, many physiotherapists report lacking the skills needed to interpret the evidence [4–6]. The increased access to unregulated information available through the internet and social media sites [7] may give PLwSCI expectations about treatments that physiotherapists also need to take into consideration.
Clinical practice guidelines (CPGs) are widely accepted as providing a summary of research that can be implemented in clinical settings [8]. CPGs have been identified as facilitating physiotherapists’ use of research in clinical practice [9]. While there are already a range of CPGs for SCI [10–12], few are physiotherapy specific [3], and many do not include the physiotherapy interventions commonly provided to PLwSCI [3]. Therefore, specific physiotherapy CPGs, which summarise and interpret evidence from the existing trials and make clear recommendations based on the best available evidence are needed. When there is insufficient evidence about a treatment, then a recommendation for or against a treatment is based on a consensus opinion of a group of experts. The consensus opinion provides clinicians with clear guidance in the absence of evidence.
Australian and New Zealand SCI physiotherapists are developing CPGs for the physiotherapy management of PLwSCI. This is a 3-year project funded by four different organisations. As a first step to the development of these CPGs, we set out to better understand how physiotherapists and PLwSCI use evidence and potential barriers and facilitators to the future rollout and uptake of our CPGs.
Methods
A study with a generic qualitative design within a paradigmatic framework of interpretivism and constructivism was undertaken based on interviews with physiotherapists and PLwSC. The project was approved by the Northern Sydney Local Health District Human Research Ethics Committee (2019/ETH00589) and University of Otago Health Ethics Committee (H19/076). Lived experience consultation was provided by the Burwood Academy Consultation Network [13, 14].
Recruitment
Physiotherapists from specialist SCI centres in Sydney, Australia (Prince of Wales Hospital, Royal North Shore Hospital and Royal Rehab), Auckland (Auckland Spinal Rehabilitation Unit) and Christchurch (Burwood Spinal Unit) New Zealand and working in the community across New South Wales, Australia and New Zealand were invited to participate. Physiotherapists were eligible to participate if they had provided physiotherapy (or similar) services to PLwSCI in a hospital or community setting. The experience level of physiotherapists was purposely kept open to ensure breadth.
PLwSCI were invited to participate by physiotherapists or doctors based in hospital or community settings. They were eligible to participate if they had a SCI of any level or severity, and had received physiotherapy services related to their SCI. Efforts were made to ensure diversity in ages, ethnicity, gender, and level of impairment. Participants were excluded if they had a serious medical condition, cognitive impairment, drug dependency, psychiatric illness, or behavioural problems, or did not speak English sufficiently well to provide informed consent.
Data collection
A pragmatic approach to data collection was used to accommodate the range of participants, maximise participation, and accommodate restrictions imposed as a result of the COVID-19 pandemic. Information was collected through interviews and focus groups either in person (face to face), by telephone or virtually (via Zoom™) (Table 1).Table 1 Data collection methods.
Participants Country Focus groups N (no. of participants) Interviews N (no. of participants) Total N (no. of participants)
in person Zoom in person Zoom/phone
Physiotherapists Australia 3 (15) 1 (2) 4 (17)
New Zealand 2 (19) (spinal unit) 2 (9) (community) 4 (28)
People living with SCI Australia 12 (12) (in-patients) 8 (8) (community) 20 (20)
New Zealand 2 (7) (in patients) 3 (3) (community) 5 (10)
33 (75)
Two interview schedules of semi-structured questions were developed by LH and JG to address the aims of the study in consultation with an independent commercial organisation experienced in qualitative research. One schedule was developed for physiotherapists and the other for PLwSCI, with the same issues raised in both focus groups and interviews. These were then reviewed by the wider research team and by PLwSCI (via The Burwood Academy Consultation network). Example questions are shown in Table 2.Table 2 Example interview questions.
People living with SCI (PLwSCI)
1. Why do people with SCI ask for certain types of therapies?
2. What sort of therapies do PLwSCI want and why?
3. On what basis do PLwSCI decide on the types of therapies they are happy to receive/participate in?
4. How important is it to PLwSCI to know that a therapy has a good evidence base?
5. How much time/effort are PLwSCI willing to devote to therapies?
6. What sorts of benefits do PLwSCI expect/want to see in return for spending time doing therapy?
7. How satisfied are PLwSCI with the therapies they have received to date? If so why, and if not, why not?
8. Where do PLwSCI get information about the types of therapies that are best for them?
9. How useful do PLwSCI believe evidence-based guidelines for treatments would be?
10. How likely are PLwSCI to accept evidence-based guidelines for treatments, and if not, why not (and vice versa)?
Physiotherapists
1. Why do physiotherapists administer certain types of therapies, particularly if there is little evidence to support decisions?
2. How much are physiotherapists’ decisions guided by what people with PLwSCI want? And how important a consideration should this be?
3. How much benefit do physiotherapists need to see from a therapy to justify its use?
4. What sorts of therapies do physiotherapists currently provide that may not be justified on current evidence?
5. What sorts of therapies don’t physiotherapists provide, that they believe they should provide?
6. How likely are physiotherapists to follow evidence-based guidelines for treatments, and if not, why not (and vice versa)?
The interviews and focus groups in Australia were conducted by an independent commercial organisation experienced in qualitative research, and in New Zealand by JN, JD and VS. All sessions were audio recorded and transcribed verbatim. NVivo 12 software was used to store and manage data.
Data analysis
A thematic analysis of the data was used following the six-step process suggested by Braun and Clarke [15]. Inductive coding was performed by separate coders in Australia and NZ. Codes and relevant data were collated in an iterative process of returning to the original data. Secondary coding was conducted by two authors (VS and JD) to check for consistency. The initial themes were established by JN that reflected the data from all participants across both countries. These were refined collaboratively with authors from Australia and NZ (VS, JD, JG and LH).
Results
A total of 75 participants took part in the study: 45 physiotherapists and 30 PLwSCI. Physiotherapists were predominantly female with a range of experience. PLwSCI had sustained their injuries between 2 months and 16 years prior and had received physiotherapy in a range of settings. They were predominantly male, with a range of impairment levels (Table 3).Table 3 Characteristics of the physiotherapist and PlwSCI participants.
Physiotherapist participants PlwSCI participants
NZ (n = 28) AU (n = 17) NZ (n = 10) AU (n = 20)
Sex
Male/female 4/24 2/15 8/2 14/6
Years of SCI physiotherapy experience
<1 year 5 0 – –
1–5 13 4 – –
6–10 5 6 – –
11–15 3 3 – –
>15 2 4 – –
Workplace
Inpatient/Community 16/12 9/8 – –
ASIAa Impairment Scale
C1–4 AIS A, B or C – – 1 3
C5–8 AIS A, B or C – – 4 5
T1-S5 AIS A, B or C – – – 3
AIS D any injury level – – 5 9
Time post injury
<1 year – – 7 4
1–5 – – – 13
6–10 – 1 2
11–15 – – 2 1
aAmerican Spinal injuries Association.
Focus groups were conducted separately for the physiotherapists (8 focus groups) and PLwSCI (2 focus groups). Focus groups lasted between 40 and 90 min. Twenty-three interviews with PLwSCI were performed which lasted between 20 and 60 min (Table 4). Separate schedules were developed for physiotherapists and PLwSCI, with the same issues raised in both focus groups and interviews.Table 4 The three themes with the key ideas from physiotherapists and people living with SCI (PLwSCI) summarised.
Themes Participants
Physiotherapists People living with SCI (PLwSCI)
Types and sources of evidence that influence treatment choices - Personal clinical expertise
- Physiotherapy peers
- Published research
- Broad evidence such as Google Scholar
- Physiotherapists or doctors
- Resources on the internet including peer-based sites
- Friends and family
- Peers living with SCI
The many factors that influence treatment choices - Evidence from wide range of sources
- Clinical expertise
- Patient goals
- External factors such as funding, availability of equipment, treatment location (hospital or community), staffing levels, time available for treatment or length of stay
- Personal and injury characteristics such as age, level of injury, and comorbidities
- Personal goals
- Physiotherapistsʼ opinions
- External factors such as funding
Ways in which clinical practice guidelines could influence treatment choices
Usefulness - Most likely for students, those less experienced - Most likely for less experienced staff
- Provides accountability, safety
- Concerns may restrict treatments options
Willingness to adopt - Variable willingness, especially if recommendations contradict current views
- Need to have agency for therapists to make own clinical judgement
Operationalisation - Want evidence available as part of CPGs
- Need to reflect complexities of SCI
- Could limit treatment options
Deliverability - Need to be freely accessible
- Need to be updated regularly
- Need to be advertised and promoted widely
- Need to be advertised and promoted widely
Three themes were identified: (1) Types and sources of evidence that influence treatment choices, (2) The many factors that influence treatment choices, and (3) Ways in which CPGs could influence treatment choices. Themes 1 and 2 describe factors that influence how physiotherapy interventions are selected. Theme 3 describes specific facilitators and barriers to physiotherapy specific CPGs for SCI rehabilitation. The overall themes are described in Table 4 with illustrative quotes provided for each theme in Figs. 1–3.Fig. 1 Theme 1 Quotes from physiotherapists and people living with spinal cord injury (PLwSCI).
Fig. 2 Theme 2 Quotes from physiotherapists and people living with spinal cord injury (PLwSCI).
Fig. 3 Theme 3 Quotes from physiotherapists and people living with spinal cord injury (PLwSCI).
Theme 1: Types and sources of evidence that influence treatment choice’s
Both physiotherapists and PLwSCI talked about how evidence influenced their treatment choice. In both groups, evidence was broad ranging and not necessarily considered just research evidence. The physiotherapists explained the diverse and complex presentation of SCI meant sample sizes in research were often too small to provide conclusive results. They also expressed concern about the methodology and potential for bias in studies. They believed that many common interventions used in their practices were not backed by strong scientific evidence. Consequently, they relied heavily on their clinical expertise when making treatment decisions. Physiotherapists who had worked in the SCI units for some time felt they had gained specialist or ‘expert’ knowledge in providing SCI treatment, which gave them more confidence in using clinical experience to justify a treatment option.
Most physiotherapists identified peers with equal or greater experience, as their first source of evidence. Peers were perceived as a source of clinical and research evidence or were relied upon to direct them to appropriate research evidence. Community physiotherapists in NZ described SCI unit physiotherapists as ‘experts’ and perceived them to be more up to date with the research evidence. However, they were considered difficult to access for those external to the SCI centre. Community physiotherapists also described uncertainty knowing where to look for evidence and a lack of access to databases that enabled article retrieval.
When asked about looking for ‘evidence’ to support treatment choices, most physiotherapists acknowledged they didn’t routinely search for the latest research evidence, but relied on research they had accessed previously (sometimes many years earlier). Most physiotherapists indicated they only looked for the latest evidence for treatments they were not familiar with, or when treatments were not working. When searching for evidence online, Google Scholar was often a primary source of information, especially for physiotherapists with limited access to academic databases or for those that didn’t know where else to look. Other sources of evidence included collated information such as SCIRE (SCI Rehabilitation Evidence https://scireproject.com/), PEDro (www.pedro.org.au), conferences, company representatives and members of the multi-disciplinary team.
PLwSCI reported multiple sources for obtaining evidence about physiotherapy treatments. The most frequently mentioned sources (from most to least) were clinicians (physiotherapists, medical staff), the internet (including peer-based sites), friends and family, and other PLwSCI. While receiving inpatient treatments, PLwSCI were more likely to rely on their clinical team, particularly their physiotherapists, for information. PLwSCI living in the community relied more on the internet and advice from other PLwSCI. This aligned with the observations of the physiotherapists.
PLwSCI described accessing specific websites they felt were useful for learning about exercises and recovery, including muscle growth and nerve regeneration. Websites were considered more credible if they were affiliated with a university, new research, or had an extensive resource library e.g., Spinal Cord Injuries Australia, the Mayo Clinic, and Project Edge. PLwSCI most commonly accessed social media sites, such as chat forums or closed Facebook groups, run by or heavily featuring PLwSCI. These were described as easily accessible and providing useful information about post injury life with a community feeling. However, for some the information could be a bit overwhelming.
Physiotherapists acknowledged the value of online sites with PLwSCI sharing their experiences, and some reported directing PLwSCI to these sources. However, physiotherapists had reservations as to whether PLwSCI could fully appreciate that treatments effective and suitable for one type of SCI may not be suitable for them.
Theme 2: The many factors that influence treatment choices
Most physiotherapists stated that ‘evidence’ was not the only consideration in clinical decision making. They agreed that evidence was important but they felt that evidence often failed to reflect the complexity of SCI and the limitations of clinical practice. They expressed differing opinions on the level of evidence needed to support a treatment. When treatments lacked evidence, both physiotherapists and PLwSCI were willing to try the treatments especially if there was any suggestion that that they could improve psychological wellbeing or facilitate recovery, provided there was assurance that the treatments were safe. Physiotherapists described using a complex process of weighing up evidence, clinical expertise, patient goals, and other external factors to select and plan for treatment. External factors included practical considerations such as funding, availability of equipment, treatment location (hospital or community), staffing levels, and time available for treatment or length of stay. In addition, personal and injury characteristics such as age, level of injury, and comorbidities were considered. While all physiotherapists agreed evidence was important, they felt that evidence often failed to reflect the complexity of SCI and the limitations of clinical practice. For most physiotherapists interviewed, ‘evidence’ was not the only consideration in clinical decision making.
In some instances, decisions over treatment choice were perceived as a negotiation or compromise, in which physiotherapists would explain treatment options and then PLwSCI would have the final say.
Physiotherapists observed that PLwSCI generally accepted their advice particularly during their initial inpatient stay. PLwSCI paid particular attention when advised not to use a certain treatment although the physiotherapists noted that some PLwSCI would get frustrated if the physiotherapists did not provide a treatment that the PLwSCI believed would benefit them. Some PLwSCI would seek these treatments elsewhere. A small number of PLwSCI spoke of questioning their physiotherapists’ recommendations or pushing for a particular treatment. For many, the turning point of control in decision-making came once they were in the community, having learnt more about themselves, their injuries and feeling more confident in voicing their opinions. All acknowledged that funding was a barrier to accessing more treatment options.
Theme 3: Ways in which clinical practice guidelines could influence treatment choices
Overall physiotherapists welcomed the development of the SCI Physiotherapy CPGs, but listed substantial expectations about what CPGs should deliver. They felt CPGs should summarise effectiveness of treatments, contain practical information and recommend treatments compatible with current practice expertise. They suggested the CPGs be digestible and brief, but also wanted clear, comprehensive information that was accessible and easy to navigate, but catered to different experience and knowledge levels.
Usefulness
Physiotherapists perceived the CPGs would be most useful for student physiotherapists and, those with less experience or who treated few PLwSCI. The CPGs were considered potentially useful for deciding the appropriateness of treatments outside usual practice, to support equipment funding applications, to use as a discussion point with PLwSCI, or to justify treatment options. They were also seen as a way of improving equity across services and funding streams.
PLwSCI living in the community perceived CPGs as being useful for guiding generalist physiotherapists who may not be specifically trained or experienced in SCI. Some also saw CPGs as a form of accountability, ensuring that a PLwSCI received best practice physiotherapy and safe treatment thereby avoiding risky or costly treatments.
Several PLwSCI raised concerns that CPGs could reinforce treatments based on a person’s classification of neurological/impairment level, which they felt could be horizon-limiting for people during recovery. Many PLwSCI did not want treatments to be limited by current evidence. A small number felt anecdotal evidence and the benefits of a positive mind-set for treatments could outweigh evidence. Many PLwSCI voiced a willingness to try anything if it could help them stand one day.
Willingness to adopt
Despite positive responses to the development of CPGs in principle, physiotherapists expressed different willingness to adopt them. Physiotherapists with more SCI experience and those working in SCI centres anticipated the CPGs would reflect their views about treatments. However, if contradicted, they admitted they may find the CPG recommendations difficult to adopt. Many hoped CPGs would not be imposed on them, rather they would still have professional agency to make their own clinical judgements.
Operationalisation of CPGs
Physiotherapists questioned how the CPGs would be operationalised. For example, there was a desire for the actual evidence to be made available to confirm the rigour behind the recommendations. They were curious how the recommendations would reflect the complexity of SCI and how a lack of evidence would be addressed. Some physiotherapists expressed concern that CPGs may reduce the range of treatment options if there was limited evidence available about current treatments. The idea of a separate CPG for PLwSCI was considered useful but physiotherapists expressed doubts that a standalone PLwSCI specific CPG would be sufficient, believing that a PLwSCI would still need a physiotherapist’s guidance to interpret the information.
Deliverability of CPGs
Physiotherapists recommended CPGs be freely accessible in different formats (e.g., online resource, hard copy). It was important to all participants that CPGs were regularly updated to maintain their relevance and usefulness. Both groups felt the CPGs needed to be advertised in multiple ways including promotion through funding organisations, mentors for physiotherapists in rural communities, support coordinators, conference presentations, and by the Australian and New Zealand physiotherapy professional organisations. Some recommended the CPGs be embedded across university curricula to encourage awareness in early-career physiotherapists. This would involve working together with key academics to design and develop approaches to embed CPGs into undergraduate and postgraduate programmes.
Discussion
The first aim of this study was to understand what factors influence how physiotherapy interventions for PLwSCI are selected and the role of evidence in decision making. The results show that although physiotherapists currently use evidence to make treatment decisions, it is only one part of a complex balance of many factors [16]. Physiotherapists in this study identified difficulties incorporating evidence in their practice. Organisational barriers to using evidence were identified as reported previously. Many found journal articles difficult to find, to understand or to relate to their own clinical practice [9, 17, 18]. Community-based physiotherapists in particular reported a lack of access to expert peer support and felt that their isolation limited their abilities to implement research into their practice [6, 9, 19, 20]. Because of these difficulties many physiotherapists relied on dated literature accumulated over their working career, or quick Google or Google Scholar searches. Consequently, there is a risk that physiotherapists are not keeping up to date with the evidence. CPGs could reduce these barriers if they provide a concise synthesis and interpretation of the current available research on a treatment intervention, which could guide a decision about whether to choose a treatment.
Although it was anticipated that physiotherapists and PLwSCI would have different knowledge requirements of CPGs, this research also illustrated distinct requirements between hospital and community settings. Our findings indicate that physiotherapists working within the SCI centres were more likely to prioritise clinical experience and were more likely to seek guidance from peers than use evidence [9].
Community physiotherapists not linked to a hospital or spinal injury service potentially face additional challenges when deciding on treatment options. They have reduced access to expert peers and research evidence, and the PLwSCI they work with are likely to have increased expectations and more confidence to articulate their views. As a result, community physiotherapists might have greater incentive to use CPGs to help with decisions making [4].
PLwSCI identified their physiotherapists as their main source of evidence to justify physiotherapy interventions while in hospital, and generally, did not question treatment choices during their early rehabilitation. Once in the community, PLwSCI gained confidence in making decisions, and began to look for information themselves and explore intervention choices. PLwSCI were more likely to use information from peers or online peer-based information, than look for scientific evidence. They wanted to try new interventions irrespective of the evidence provided they would not cause harm. This attitude was acknowledged by the physiotherapists but did also have the potential to cause internal conflict for the physiotherapists striving to ensure the treatment they provided were evidence based. Both the physiotherapists and PLwSCI felt that having information on best practice in a CPG would be reassuring and provide support for physiotherapists, particularly those with less SCI experience.
The second aim of this study was to understand facilitators and barriers to using physiotherapy CPGs for SCI. Although all participants were in favour of CPGs, physiotherapists working in the SCI units felt they were less likely to need to refer to the CPGs because of their experience and expertise. They accepted CPGs were likely to become part of policy and practice because of the systems, procedures and accountability within the SCI centres or hospital setting. There was a perception that using CPGs may inhibit physiotherapists using their professional judgement and could even reduce the range of treatment options if there were limitations on the evidence available. This is similar to findings exploring CPGs use in the stroke population [17]. However, the CPGs may provide an easy resource to access up to date evidence for newer treatment options.
Community physiotherapists appeared more likely to refer to the CPGs as part of their professional practice, especially if they had limited expertise in SCI management. However, they identified specific barriers to implementing the CPGs in a community setting. The biggest perceived barrier to organisations providing community care implementing the CPGs was funding restrictions which dictate the number and type of treatments that can be prescribed. Other constraints included staffing levels and access to equipment, especially when working with people in their own homes.
This study illustrates the benefits of early stakeholder consultation. Through this process we have developed an appreciation of the requirements of CPGs before they are developed. Careful consideration needs to be given to the implementation phase, as our results indicate that we may encounter some resistance to adopting the CPGs into clinical practice. Some strategies to help overcome barriers to the use of CPG include education, the inclusion of all stakeholders during the development of the CPG, the use of clinical champions in workplaces and the widespread embedding of the CPG into university curriculums [9, 17, 18, 21]. Evidence indicates that education [22], opinion leaders [23] and to a lesser degree printed material [24] are more effective than nothing at changing clinical practice. Combining these strategies has been shown to be effective at increasing adherence to clinical guidelines in a SCI setting [25]. Such strategies will be needed to convince physiotherapists of the benefits of the CPGs and to encourage them to use the CPG to make clinical decisions. Any implementation strategy will need to be tailored [26] differently for community and hospital-based physiotherapists as well as for PLwSCI, considering the barriers to implementation identified in this study. For example, clinical champions can be more readily used in the hospitals than community. In the community it may be necessary to work with the funders of services to help encourage the use of the CPGs.
The design of CPGs has been shown to influence likelihood of implementation [27, 28], and in an environment of rapidly evolving technology, the platforms that CPGs are provided on become increasingly relevant. Our study participants identified the need for the CPGs to be widely accessible and to be available in multiple formats. To maximise uptake, the CPGs will need to be available on a number of platforms, readily accessible, and free to download.
This study is not without its limitations. Importantly, some planned face-to-face interviews needed to be performed over the telephone due to COVID-19 restrictions. The pragmatic approach to data collection and multiple people involved in data collection may be seen as a limitation of the study. In addition, a different set of interviewers were used for Australia and New Zealand, and the coding and themes were independently extracted for the Australian and New Zealand data. However, the inclusion of data from the two countries is a strength of the study because it makes the findings more generalisable.
Conclusion
CPGs have the potential to reduce the barriers physiotherapists identified in accessing and interpreting research evidence on interventions for PLwSCI. If CPGs are to be used as a tool in deciding on clinical interventions, supported implementation of CPGs is required to demonstrate their benefit in providing collated evidence and encouraging physiotherapists to factor in evidence when balancing the multiple factors influencing the choice of physiotherapy interventions. Physiotherapists and PlwSCI have high expectations of easily accessible CPGs in a variety of formats that will meet differing information needs for community and inpatient settings while addressing the complexity in presentation of PlwSCI.
Acknowledgements
We would like to thank all the participants, and Maree Walters for coordinating recruitment in Auckland. ARTD Consultants conducted the interviews in Australia, transcribed the interviews, entered the data onto NVivo and extracted the themes.
Author contributions
LAH and JVG: initiated and secured the funding for the CPG, designed the study and interview guide, interpreted the results, wrote the report. JLN: NZ data collection, primary analysis of NZ data, synthesis of NZ and Australian data, preparation of first draft of report. JAD and VAS: NZ data collection, secondary analysis of NZ data, wrote the report. AH, MB, LWC, and SD: designed the study, coordinated recruitment, commented on the report.
Funding
iCare Australia; Physiotherapy New Zealand Neurology Special Interest Group grant for JLN to guide the development and subsequent rollout of clinical practice guidelines.
Data availability
The datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request.
Competing interests
The authors declare no competing interests.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Harvey LA Physiotherapy rehabilitation for people with spinal cord injuries J Physiother 2016 62 4 11 10.1016/j.jphys.2015.11.004 26701156
2. McIntyre A Benton B Janzen S Iruthayarajah J Wiener J Eng JJ A mapping review of randomized controlled trials in the spinal cord injury research literature Spinal Cord 2018 56 725 32. 10.1038/s41393-018-0155-2 29904188
3. Harvey LA Lin CW Glinsky JV De Wolf A The effectiveness of physical interventions for people with spinal cord injuries: a systematic review Spinal Cord 2009 47 184 95 10.1038/sc.2008.100 18725889
4. Grimmer-Somers K Lekkas P Nyland L Young A Kumar S Perspectives on research evidence and clinical practice: a survey of Australian physiotherapists Physiother Res Int 2007 12 147 61 10.1002/pri.363 17624895
5. Iles R Davidson M Evidence based practice: a survey of physiotherapists’ current practice Physiother Res Int 2006 11 93 103 10.1002/pri.328 16808090
6. Salbach NM Jaglal SB Korner-Bitensky N Rappolt S Davis D Practitioner and organizational barriers to evidence-based practice of physical therapists for people with stroke Phys Ther 2007 87 1284 303 10.2522/ptj.20070040 17684088
7. Ventola CL Social media and health care professionals: benefits, risks, and best practices P T 2014 39 491 520 25083128
8. Hill J Bedford J Houston D Reid DA Baxter GD Ellis R Exploring physiotherapists’ use of clinical practice guidelines, screening, and stratification tools for people with low back pain in New Zealand NZ J Physiother 2020 48 59 69
9. Dannapfel P Peolsson A Nilsen P What supports physiotherapists’ use of research in clinical practice? A qualitative study in Sweden Implement Sci 2013 8 31 10.1186/1748-5908-8-31 23497502
10. Eng JJ Teasell R Miller WC Wolfe DL Townson AF Aubut JA Spinal cord injury rehabilitation evidence: methods of the SCIRE systematic review Top Spinal Cord Inj Rehabil 2007 13 1 10 10.1310/sci1301-1 22767989
11. Fehlings MG Tetreault LA Aarabi B Anderson P Arnold PM Brodke DS A clinical practice guideline for the management of patients with acute spinal cord injury: recommendations on the type and timing of rehabilitation Glob Spine J 2017 7 231S 8S 10.1177/2192568217701910
12. Guy SD Mehta S Casalino A Cote I Kras-Dupuis A Moulin DE The CanPain SCI clinical practice guidelines for rehabilitation management of neuropathic pain after spinal cord: recommendations for treatment Spinal Cord 2016 54 S14 23 10.1038/sc.2016.90 27444715
13. Bourke JA. The Burwood Academy consultation committee scoping review. Christchurch: Burwood Academy of Independent Living; 2016.
14. Bourke JA, Snell DL, Sinnott KA, Cassidy B. A user‐led consultation model: making inclusion of service users in research a reality. Ethn Inequal Health Soc Care. 2012;5:61–9.
15. Braun V, Clarke V. Successful qualitative research: a practical guide for beginners. London: Sage; 2013.
16. Vingerhoets C Hay-Smith J Graham F Intersection of the elements of evidence-based practice in interdisciplinary stroke rehabilitation: a qualitative study NZ J Physiother 2020 48 148 54.
17. Donnellan C Sweetman S Shelley E Health professionals’ adherence to stroke clinical guidelines: a review of the literature Health Policy 2013 111 245 63 10.1016/j.healthpol.2013.05.002 23727250
18. Grimshaw JM Eccles MP Lavis JN Hill SJ Squires JS Knowledge translation of research findings Implement Sci 2012 7 50 10.1186/1748-5908-7-50 22651257
19. Francke AL Smit MC de Veer AJ Mistiaen P Factors influencing the implementation of clinical guidelines for health care professionals: a systematic meta-review BMC Med Inf Decis Mak 2008 8 38 10.1186/1472-6947-8-38
20. Nilsagård Y Lohse G Evidence-based physiotherapy: a survey of knowledge, behaviour, attitudes and prerequisites Adv Physiother 2010 12 179 86 10.3109/14038196.2010.503812
21. Goetz LL Nelson AL Guihan M Bosshart HT Harrow JJ Gerhart KD Provider adherence to implementation of clinical practice guidelines for neurogenic bowel in adults with spinal cord injury J Spinal Cord Med 2005 28 394 406 10.1080/10790268.2005.11753839 16869086
22. Forsetlund L O’Brien MA Forsen L Reinar LM Okwen MP Horsley T Continuing education meetings and workshops: effects on professional practice and healthcare outcomes Cochrane Database Syst Rev 2021 9 CD003030 34523128
23. Flodgren G O’Brien MA Parmelli E Grimshaw JM Local opinion leaders: effects on professional practice and healthcare outcomes Cochrane Database Syst Rev 2019 6 CD000125 31232458
24. Giguere A Zomahoun HTV Carmichael PH Uwizeye CB Legare F Grimshaw JM Printed educational materials: effects on professional practice and healthcare outcomes Cochrane Database Syst Rev 2020 8 CD004398 32748975
25. van der Wees PJ Jamtvedt G Rebbeck T de Bie RA Dekker J Hendriks EJ Multifaceted strategies may increase implementation of physiotherapy clinical guidelines: a systematic review Aust J Physiother 2008 54 233 41 10.1016/S0004-9514(08)70002-3 19025503
26. Baker R, Camosso-Stefinovic J, Gillies C, Shaw EJ, Cheater F, Flottorp S, et al. Tailored interventions to address determinants of practice. Cochrane Database Syst Rev. 2015:CD005470.
27. Mudge S Hart A Murugan S Kersten P What influences the implementation of the New Zealand stroke guidelines for physiotherapists and occupational therapists? Disabil Rehabil 2017 39 511 8 10.3109/09638288.2016.1146361 26940156
28. Reynolds SS Murray LL McLennon SM Ebright PR Bakas T Implementation strategies to improve knowledge and adherence to spinal cord injury guidelines Rehabil Nurs 2018 43 52 61 10.1002/rnj.304 27781292
| 36513762 | PMC9746570 | NO-CC CODE | 2022-12-15 23:21:56 | no | Spinal Cord. 2022 Dec 13;:1-9 | utf-8 | Spinal Cord | 2,022 | 10.1038/s41393-022-00867-x | oa_other |
==== Front
J Racial Ethn Health Disparities
J Racial Ethn Health Disparities
Journal of Racial and Ethnic Health Disparities
2197-3792
2196-8837
Springer International Publishing Cham
1466
10.1007/s40615-022-01466-5
Article
Motivators and Barriers to COVID-19 Research Participation at the Onset of the COVID-19 Pandemic in Black Communities in the USA: an Exploratory Study
Barre Iman 1
Cunningham-Erves Jennifer 2
Moss Jamal 1
Parham Imari 1
Alexander Leah R. 3
http://orcid.org/0000-0002-5078-1306
Davis Jamaine [email protected]
4
1 grid.259870.1 0000 0001 0286 752X School of Medicine, Meharry Medical College, 1005 Dr. D.B. Todd Jr. Blvd, Nashville, TN 37208 USA
2 grid.259870.1 0000 0001 0286 752X Department of Internal Medicine, School of Medicine, Meharry Medical College, 1005 Dr. D.B. Todd Jr. Blvd, Nashville, TN 37208 USA
3 grid.259870.1 0000 0001 0286 752X Division of Public Health Practice, School of Graduate Studies and Research, Meharry Medical College, 1005 Dr. D.B. Todd Jr. Blvd, Nashville, TN 37208 USA
4 grid.259870.1 0000 0001 0286 752X Department of Biochemistry, Cancer Biology, Neuroscience & Pharmacology, School of Medicine, Meharry Medical College, 1005 Dr. D.B. Todd Jr. Blvd, Second Floor of Harold West Basic Science Building, Suite 2007, Nashville, TN 37208-3599 USA
13 12 2022
110
31 3 2022
16 11 2022
21 11 2022
© W. Montague Cobb-NMA Health Institute 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Introduction
Black individuals in the USA continue to be underrepresented in clinical trials with low participation rates in COVID-19 research studies. Identifying participation barriers is necessary as we develop more vaccines and other treatments to address SARS-CoV-2 and associated sequelae. The purpose of this explorative, qualitative study is to apply the theory of planned behavior to understand motivators and barriers to COVID-19 research participation at the early stages of the COVID-19 pandemic. Understanding these factors is important to ultimately lead to increased vaccination rates among Black individuals, especially in strategies that increase preparedness in response to public health emergencies.
Methods
A phenomenological qualitative study design was conducted between May and September 2020 among 62 Black participants. The participants were purposefully selected from vulnerable subgroups of the Black population: essential workers, young adults, parents, and individuals with underlying medical conditions. An inductive-deductive content analysis approach was used to analyze the interview data.
Results
Majority (54.8%) reported willingness to participate in COVID-19 research. The following themes emerged from the interviews: (1) positivity toward research exists yet fear and distrust remain; (2) views toward COVID-19 research vary; (3) motivators to COVID-19 research participation; (4) barriers to COVID-19 research participation; and (5) potential strategies to increase COVID-19 research participation.
Conclusions
Based on our findings, majority of the participants reported willingness to participate in research with observational research being the most commonly cited type of research. Providing data on the attitudes and perspectives of Black individuals and their intentions for COVID-19 research participation using TPB informs intervention targets for healthcare providers and policy makers for an equitable emergency response. Our results suggest improved communication on the research process, research opportunities, and participant testimonial through trusted sources could increase the likelihood of participation. This is especially important as we continue through the pandemic and new treatments for COVID-19 vaccines become readily available.
Keywords
COVID-19
Vaccination
Research participation
Theory of planned behavior
http://dx.doi.org/10.13039/100006108 National Center for Advancing Translational Sciences 5UL1TR0002243-03
==== Body
pmcIntroduction
The global crisis caused by the coronavirus disease 2019 (COVID-19) pandemic has affected every country in the world. The USA, however, has steadily experienced the highest incidence and mortality rates associated with COVID-19. As a result, COVID-19 exacerbated the underlying social and economic health inequities within the USA. Since the beginning of the pandemic in 2020, federal, state, and local data have shown that low-income and racial/ethnic minority populations disproportionately bore the brunt of the morbidities and mortalities due to COVID-19 [1]. More than 2 years later, this trend has continued due to the emergence of several SARS-CoV-2 variants [2] and low vaccination rates. The incidence rates for COVID-19 have been two times higher, and mortality rates almost three times higher in Black Americans compared to White Americans [3]. Over time, cumulative data has shown persisting disparities in COVID-19 deaths for Black people, particularly in rural areas [4]. Preventative strategies are urgently needed to be implemented to reduce the disproportionate transmission, health complications, and mortalities associated with this ongoing pandemic, particularly among Black individuals.
To be most effective at achieving positive health outcomes for all, research should mirror the population most affected by the disease [5]. The race toward a COVID-19 vaccine, however, highlighted the lack of representation of Black Americans in clinical trials, both in enrollment and outreach at the onset. Black Americans make up 13% of the US population but accounted for 21% of COVID-19 deaths, but only 3% of enrollees in vaccine trials [6]. Efforts were implemented to increase participation of underrepresented groups such that in the phase 2/3 testing of Pfizer’s COVID-19 BNT162b2 vaccine, the safety population included 9.1% Black or African American compared to 83.1% White [7]. Moderna’s mRNA-1273 vaccine reported 10.2% Black or African American participants compared to 79.2% White [8], although both contained a smaller percentage of Black participants relative to the 13% of the US population. For any disease, low enrollment rates in clinical trials correlate with limitations in generalizability in findings, minimizing the development and/or improvement in drugs and treatments and, ultimately, positive outcomes [9]. While there are several efforts at increasing recruitment of Black participants in clinical trials, barriers to participation persist and need to be better understood.
For Black Americans especially, barriers to clinical trial participation exist on many levels and vary in complexity [10–12]. At the individual level, major obstacles include a lack of information about various types of research opportunities (surveys vs. clinical trials), fear of the research process, and logistical barriers (e.g., issues of transportation, availability of childcare) that can result in problems with longitudinal retention. At the community level, mistrust of academic and research entities is the most significant attitudinal barrier to participation reported [12, 13]. Identifying barriers to participation specific to COVID-19-related clinical trials is necessary for researchers across the translational research continuum (from bench to bedside, then to the community) as we learn more about the evolution of the virus, and the vaccines, antivirals, and other treatments. Furthermore, this knowledge lays the foundation to understand how to inform and engage Black individuals in research overall, and more importantly aid in the event of future pandemics.
Theoretical Framework
The theory of planned behavior (TPB) has been extensively applied by researchers to understand individuals’ beliefs and reasons for performing a specific behavior [14, 15] including research participation [16]. According to the TPB, an individual’s intention to perform a behavior is the best predictor of performing a behavior. There are several factors that influence an individuals’ intentions to participate in COVID-19 research: (a) socio-demographics; (b) beliefs about a COVID-19 research and associated consequences (i.e., attitude); (c) beliefs about normative expectation of others to engage in COVID-19 research and likelihood to comply (i.e., subjective norms); and (d) beliefs about the factors that facilitate or hinder performance in COVID-19 research (i.e., self-efficacy) [14]. Several studies used the TPB to understand intentions and behaviors related to COVID-19 prevention including social distancing, PCR testing, compliance with health protocols, and vaccine uptake [17–20]. However, no studies to our knowledge have applied the TPB to explore intentions and actual COVID-19 research participation. Hence, application of TPB will allow us to better understand the rationale and decision-making process for COVID-19 research participation among Black Americans.
The purpose of this study is to apply the theory of planned behavior to understand motivators and barriers to COVID-19 research participation and how they differ across the study at the initial stages of the COVID-19 pandemic. These results will serve to better inform the development of new strategies aimed at increasing research participation among Black Americans in high-risk subgroups. Findings from this study can be incorporated into recruitment strategies across the translational continuum.
Methods
Study Design
A phenomenological, qualitative study design was chosen to explore motivators and barriers to COVID-19 research participation among Black individuals living in the USA at the early stages of the COVID-19 pandemic [21]. We conducted semi-structured interviews among Black individuals who identified as essential workers, young adults, parents, or individuals with underlying medical conditions (UMC), uniquely vulnerable subgroups. This study was approved by Meharry Medical College Institutional Review Board (IRB).
Sampling and Recruitment
A purposive, criterion sampling method [21] was used to recruit participants using existing databases of past research participants, word of mouth, and social media. The participants were formally recruited via telephone and email where they were given information on the goals of the study and their rights as participants, and agreed then subsequently provided informed consent. These methods were chosen to achieve diversity and navigate pandemic restrictions on social gatherings and mobility. The inclusion criteria for the participants were (1) Black, (2) English-speaking, (3) aged 18 and over, and (4) a member of one of the following categories: an essential worker, a parent, a young adult (aged 18–35), or an individual with UMC.
Interview Protocol and Training of Facilitators
The interview guide was developed to seek views of research participation both generally and specific to COVID-19, and types of research (clinical trials, such as treatment and prevention trials, compared to observational research including surveys and interviews), along with motivators and barriers to participation (Table 1). Each question in the questionnaire tied with components of the TPB, i.e., their attitude toward research generally (“What are your overall views of research?”), subjective norms (social approval or disapproval) of participating in research (“What do you know of COVID-19 research?” and “What do you perceive are the benefits of COVID-19 research?”), and perceived behavioral control and barriers (“Why would you/would you not participate in COVID-19 research?”). Three medical students were trained to assist in conducting the study prior to initiation. Training included (1) a 1-h training session on qualitative research; (2) a 1-h training session on how to conduct interviews (i.e., reflective techniques to encourage discussion); and (3) a 1-h training session on how to use REDCap [22], a secure web application for building and managing online surveys and databases, to obtain sociodemographic information.Table 1 Interview protocol questions
1. What are your overall views toward research? (attitudes)
2. What do you know about COVID-19 research and your overall view of it? (modifying factor)
a. Can you elaborate more on…?
3. Why would you/would you not [choose one based on the screening eligibility form] participate in research about COVID-19?
a. Unaware of research opportunities
b. Lack of access to research opportunities
c. Lack of trust in physician
d. Lack of trust in a pharmaceutical company
e. Family/friend/physician/significant other influence (subjective norms)
f. Concerns about research (long-term effects, use of data)
g. Cultural experience
h. Type of clinical trials (i.e., treatments and prevention trials) versus observational research (i.e., surveys, interviews)
i. Prior research experiences
j. Other: Are there other reasons you would or would not participate in research that we have not discussed?
4. What do you perceive are the benefits of COVID-19 research? (motivators)
a. SARS-CoV-2 and COVID-19 prevention
b. End pandemic
c. Please explain more on ___________________________________
d. Could you please provide an example of _____________________?
e. Why do you think so?
Procedures
Data collection took place between May and September 2020. Three medical students and two researchers conducted the interviews. Prior to the interview, the participants were sent a link using REDCAP [22], and, on the day of the interview, the participants were read an information sheet and provided verbal consent to participate. The interviews lasted between 45 and 90 min, depending on the amount of information the participant provided. The participants were compensated a $30 electronic gift card. All interviews were audio-recorded and transcribed verbatim using Zoom and verified by the members of the research team to ensure accuracy.
Analysis
Survey data were analyzed with SPSS version 26. Descriptive statistics (i.e., frequencies and percentages, and chi-square) were used to analyze the data. Microsoft Excel 2016 was used to manage qualitative data. An inductive, thematic analysis approach was used for interview data. First, the lead qualitative researcher (co-author JCE) developed a priori concepts using the theory of planned behavior, literature, and past research experience. Then, two researchers and one medical student used the priori codes to code the transcripts, and the new concepts that emerged were assigned codes. Each code(s) was placed into a category (i.e., axial coding) until saturation was met. To assess coding consistency, the codes and their assignment to text were checked and rechecked in their patterns and explanation. If discrepancies arose, researchers discussed the codes until consensus was reached. The codes were compared and queried to identify emerging themes within and across groups. The verification procedures of the interview data were done using triangulation by comparing the participants’ views within and across subgroups, peer debriefing among two researchers and one medical student, and rich, thick description of study findings [21, 23].
Results
The semi-structured interviews were conducted among 62 Black participants during the COVID-19 pandemic across four groups: (1) 16 parents, (2) 16 essential workers, (3) 15 individuals with UMCs, and (4) 15 young adults between the age of 18 and 35 years. Majority (54.8%) reported willingness to participate in COVID-19 clinical trial research across all socio-demographics. See Table 2 for socio-demographics by willingness to participate in COVID-19 research. Statistically, there were no significant differences in willingness to participate by socio-demographics.Table 2 Socio-demographics by willingness to participate in COVID-19 clinical trial research
Yes No/I do not know
Mean SD Mean SD r p value
Age 38.74 14.40 42.18 13.67 0.040 0.755
Yes No/I do not know
N % N % Χ2 p value
Total 34 54.8 28 45.2
Gender 0.501 0.479
Male 12 63.2 7 36.8
Female 22 51.2 21 48.8
Education 0.050 0.822
Some college or less 13 59.1 9 40.9
Associates degree or higher 21 52.5 19 47.5
N % N % F p value
Category 1.886 0.125
Essential worker 9 56.3 7 43.7
Underlying medical condition 7 46.7 8 53.3
Young adults 10 66.7 5 33.3
Parents 8 50.0 8 50.0
Income 1.387 0.250
Less than $40,000 13 65.0 7 35.0
$40,001–$80,000 8 47.1 9 52.9
Over $80,000 9 56.3 7 43.7
Do not want to answer 4 44.4 5 55.6
The TPB argues that attitudes toward and beliefs about health behaviors shape people’s intention to adopt them [18]. The following themes emerged directly from the interviews and corresponded with elements of the TPB as they relate to the perceived motivators and barriers to COVID-19 research participation in early pandemic stages: (1) positive attitudes toward research exist yet fear and distrust remain; (2) attitudes toward COVID-19 research vary; (3) physical and psychological capability to participate in COVID-19 research; (4) motivation to participate in COVID-19 research; and (5) potential opportunities to increase COVID-19 research participation.
Theme 1: Positive Attitudes Toward Research Exist Yet Fear and Distrust Remain
The participants in all groups expressed positivity toward research. Many participants perceived research helps to understand and extend the knowledge base of science and technology in health and healthcare. The participants further described how research helps to identify the best strategies to address health issues and confirm progress and improvements in health and healthcare.One parent stated, “It’s important. It validates experiences for people. I think it helps us to find better solutions to problems to give others a better quality of life.”
Parents acknowledged the importance of children’s participation in research with the caveat that they (Erves, Mayo-Gamble, Hull et al.) were the primary decision-makers for their children’s participation. Some parents did express there was room for older children to play a role in their own decision to participate.
Despite positive views toward research, fear and distrust in the research process persist. The participants continuously identified the historical context of research abuse particularly among Black Americans.One participant stated, “You know, just the history with the Tuskegee experiment. The history with slavery. How people used to practice maternity procedures on Black women without any type of anesthesia. I’m not 100% confident that that mentality is extinct in this country. So, I’m very hesitant to contribute and volunteer my body.”
Another stated, “We can’t forget our history. We can’t forget Henrietta Lacks. We can’t forget them gynecological exams where Black women were not anesthetized. We can’t forget the Tuskegee experiment…”.
Hence, the history of multiple discriminatory health interventions has hampered willingness to participate in research.
Theme 2: Attitudes Toward COVID-19 Research Vary
Almost all participants supported COVID-19 research. They deemed research necessary to learn more about the virus and the disease process. The participants stated research would help manage and eliminate the pandemic via identifying strategies (e.g., drugs, treatments, vaccines) to prevent and/or manage COVID-19 and its severity. While research was unanimously expressed as essential, many participants believed Black people in the USA have long served as “guinea pigs” in this process and should not volunteer first.One participant stated, “Well just that I think we’ve been experimented on enough and, you know, marginalized communities have always been the guinea pigs when it comes to something like this. I understand somebody has to, but it doesn’t have to be us this time because, look, we’ve had our turn.”
Theme 3: Physical and Psychological Capability to Participate in COVID-19 Research
There was consensus across all groups of the physical and psychological factors that discourage research participation or make it difficult for interested candidates to participate. We briefly describe these barriers below.
Limited Awareness and Access to Research Opportunities
The participants commonly described limited awareness and access to information related to COVID-19 and other health topics, both physical and psychological barriers to participation. A few participants further stated if they knew how to access clinical trials that were available, then they could become familiarize with the different types of trials and.“I do not know a lot about research. I know that they are using human DNA. Testing human beings in this research. And that’s pretty much all I know about the research.”, a participant said.
Furthermore, the participants lacked understanding of types of research opportunities available along with the requirements and the process for enrolling into a research study, which were potential physical and psychological barriers to COVID-19 research participation. A few participants stated if they gained an understanding of the types of research and the process, then they could see benefit and potentially motivate them to participate.
Limited Health Literacy
Poor health literacy was another identified barrier.
One participant stated, “There are certain areas, certain pockets throughout the country that have poor health literacy rates. If my health literacy is piss poor, do you really think I know something about research? And do you really think I’m gone trust it.”
This was highlighted with some participants’ lack of information or misinformation on vaccines, dosage, and their roles in the body. Other participants used the term “vaccines” interchangeably with the concept of antiviral medication.
For example, a parent stated, “So if we had Corona, and we needed a vaccine, I would get it but to get the vaccine to prevent corona, no, but a pain relief for the kids. Yes, like a pain reliever…No preventive care, no. I don’t want to fix nothing that ain’t broke so I have to have it in order to get the vaccine.”
Influence of Political Climate
The perceived role of the federal government in the research process was a major contributor to participants’ unwillingness to accept current treatments for COVID-19.
One individual with an UMC said, “And now I have to get political. Under this administration [Trump], I would not take a vaccine, if I was the last person on earth, that came from this administration.”
Some further reported knowing the role of the government in the research process would help to make a decision on whether or not to participate in the process. The participants were also skeptical of the rapidness of research trial completion, especially vaccine trials that appeared to be “pushed” by the government and developed at warp speed.“I think when things are rushed, they can be sloppy. And the reasons I love research are exactly because you have to be very meticulous and like thought through with how you’re going to do something, how it can be replicated, this that and the other. And so, if we’re kind of rushing to put something together, are we really like…I just don’t know.”, stated a young adult.
History of Unethical Research Studies in the USA
The participants across all groups acknowledged that historical and current research abuses within Black communities contributed to their negative perceptions of COVID-19 research. As indicated in Theme 1, many participants believed Black people in the USA have long served as “guinea pigs” in the research process. This historical context also influenced how participants with children were doubtful of the process. Parents were especially uncomfortable with their children participating in COVID-19 research.One parent said, “No, I don’t even fool with that flu shot. So, but no, because if I go the rest of this year without corona I ain’t fooling with no [COVID-19] vaccine. I’ll let everybody else get it. Would I let my kids get it? No. Unless it was absolutely mandatory.”
The Unknown Side Effects
All participants cited they were concerned of the unknown side effects of participating in COVID-19 research trials, including the side effects of vaccines and related treatments. Individuals with UMC in particular worried how the vaccine would interact with their current medications.One individual with UMC said, “That’s a good question because my first concern would be how is that going to interact with what I’m already taking? How is that going to interact with the current medication that I’m on that’s keeping me alive?”.
Many young adults stated they preferred natural medicines rather than risk experiencing side effects of the COVID-19 vaccine and related treatments.
One young adult said, “I don’t want to do no vaccine or anything. I’m not taking a medication. Because like I said I was a fan of Dr. Sebi. And doing his research and listening to interviews and things like that. I don’t feel like medication is necessary because it’s all chemical and some of those things, if you look at the ingredients on medications, a lot of more are pretty much poison.”
Theme 4: Motivation to Participate in COVID-19 Research
All groups expressed the significance of research and its benefits across many levels. We describe each below.
Advancing Science
The participants expressed how research would help to manage and alleviate the pandemic as well as provide insight into future pandemic management. So, understanding the research and how it could potentially advance science increased the likelihood of many participants. This is a form of reflective motivation.One participant stated, “Once we’ve completed the research, we’ll have a better understanding of what’s going on and how to not only combat it now, but how to make preparations so that it’s not this bad in the future.”
Altruism
Many participants described the social benefit of research participation at the individual and community levels from a sense of altruism. So, these individuals were more motivated to participate if there was a positive outcome for self and/or others, a form of automatic motivation. This theme was particularly highlighted among individuals with UMC, many of whom recognized the historical injustices of research malpractice against Black communities while also acknowledging that their own health status was being maintained by treatments that came from research.
One participant said, “The example that you set for the next generation is important, and if making sure that they live a COVID-19-free life is something that will be beneficial to my grandchildren’s children, then I think that’s an important thing.”
Study Type
Participants across all groups expressed preference for participating in studies such as surveys and interviews, rather than those that involved the introduction of foreign material into the body, such as in treatment trials (i.e., drugs) and prevention trials (e.g., COVID-19 vaccine trials). This means participants were motivated to avoid a negative consequence.
One parent said, “And so if it’s similar to this, just asking them questions or what not, I don’t mind. But that’s not what came to mind. I was thinking shots. I was thinking you being isolated and so many different tests are ran on you and stuff like that. So nah, I wouldn’t be down for that.”
Theme 5: Potential Opportunities to Increase COVID-19 Research Participation
Many participants offered social opportunities to increase research participation. One of the strongest suggestions was to increase the ability of researchers, both from the clinical sciences and particularly basic sciences, to engage with and educate communities on research without shame or judgment. Additional strategies included past research participants sharing testimonials on clinical research experiences and recommending research participation, improving communication around research opportunities along with requirements, and more details regarding the research process. Proposed communication social and physical opportunities included making better use of webpages of academic institutions and non-profit organizations as well as improving communication between providers and patients along with researchers and the participants they recruit.“I’m not aware of opportunities. I would feel more comfortable if I could go on Meharry’s website or you know Vanderbilt website or even if it was a hospital website, you know, St. Thomas Midtown or whatever and see what’s going on. And if there was something I wanted to sign up for I would feel comfortable doing that as long as the information was, you know, laid out.”, a participant stated.
Lastly, the participants highlighted the importance of using social media as a tool to improve communication around research.
One participant said, “Flat out, flat out social media, especially, especially, especially websites. Like from the USDA and other known credible websites. Now, I’m not sure this is part of the question but of course, paid incentives, especially during this time.”
Ultimately, these recommended strategies are likely to increase research participation.
Discussion
Research participation is often understood and discussed as a binary choice (i.e., acceptance or rejection). Using the TPB, our study provides new perspectives of a complex issue by offering context and nuance to what motivates research participation in general, and COVID-19 research more particularly in the early pandemic stages. While there will always be a portion of the population that outright refuses to participate in research, our study explores underlying reasons for those rejections as well as barriers that prevent interested individuals from research participation.
Our study shows that essentially all study participants understood the value and benefits of research to improve human health. Altruism and social benefits were motivating factors, particularly for participants with UMC who expressed the value that research has had on their ability to care for themselves. This finding is congruent with previous studies that have shown that intrinsic motivation to contribute to a broad social benefit is associated with research participation [17]. This finding is also consistent with the TPB which posits that attitudes, social norms, and perceived control influence health behavior. Individuals with a strong sense of perceived benefit are likely to participate in research. A significant motivator for research acceptance from our study was study type. All participants, regardless of study group, expressed a preference for research studies that were quantitative surveys or qualitative in nature, such as interviews and storytelling as they were perceived as low or zero risk to the participant. Such perceived low risk is conceptualized as perceived behavioral control in TPB and is positively related to engaging in a health behavior. However, most participants associated research with clinical trials and did not perceive that their ideas and observations contributed to research participation. In addition, these results suggest that bidirectional communication, which clearly defines the research subject and the format, as well as options for participating, including incentives, should be clearly communicated in research recruitment to potentially increase research participation for Black individuals.
A lack of participation, on the other hand, manifested as either outright rejections to research or barriers that made it challenging for individuals to participate regardless of interest. The most strongly associated barrier to research participation was mistrust of academic and medical institutions, with many participants citing the human rights violations of the Tuskegee Syphilis Study [24] as the root for mistrust. This finding is in accordance with existing literature that has shown racial minority participants, relative to White participants, were found to be persistently less positive about the use of medical information for research [25]. Similarly, in the context of the TPB, we can understand that negative attitudes toward research and researchers correspond to a negative association with research participation. However, rejections to research participation do not entirely account for the disproportionate rates seen among racial minorities. Frequently discussed barriers to participation were low levels of health literacy, a lack of information regarding clinical trials and various research platforms, and limited awareness of opportunities for participation, as well as logistics of the research process. Addressing these barriers requires intentional effort on the part of public health officials, and basic and clinical scientists especially, to communicate information effectively, as well as being culturally sensitive [26]. While communication was noted to be an area for improvement to increase research participation in our study, the mode of communication is also an important factor [27]. Information dissemination has been confusing and inconsistent throughout the entire pandemic [28]. For example, several new terms were presented to the public [29], which may have added to an individual’s undecidedness to participate in the COVID-19 vaccine clinical trial and even hesitation to getting the vaccine after approval. Rather than credible sources defining these terms, many communities relied on social media as their primary source of information, particularly in areas where there is a dearth of healthcare providers [30]. While it has its uses, social media can subject the public to misinformation, and deliberately false information, that can negatively influence decision-making [31]. The participants clearly stated the need to understand more about mRNA and DNA, in addition to how viruses and vaccines work. These findings illustrate the importance of improving communication about the research process to effectively recruit marginalized and underrepresented participants for future research.
Limitations
This study was conducted toward the beginning of the pandemic from May to September of 2020. Our results may not reflect current perceptions of COVID-19 research that may have been influenced by increased information and the change in federal administration. Additionally, the results of this study may not be generalizable to other underrepresented research populations beyond the USA. However, we purport the findings will provide valuable guidance for future communication of research opportunities. Future studies are needed to better understand how intentions and ultimate participation in COVID-19 research are influenced by evolving attitudes toward COVID-19 as well as potential feelings of burn out from taking precautions against multiple strains, adhering to changing mandates, and research concepts (e.g., socioeconomic status, informed consent, transportation, childcare). Additional studies should also explore if and how strategies may differ to recruit and retain African Americans in studies, especially longitudinal studies where many individuals are lost to follow-up.
Conclusions
This study aimed to use the TPB to explore the views toward COVID-19 research among Black communities in the USA to better understand the nuances around hesitation and the scope of willingness to participate in COVID-19 research in early pandemic stages. This is especially important due to the surge of evolving strains of SARS-CoV-2, such as the Omicron and its subvariants. Globally, we have witnessed multiple SARS epidemics; the first was reported in Asia in 2003 caused by the original strain, SARS-CoV, which spread to more than two dozen countries causing 8000 illnesses and 700 deaths before it was contained [32]. The second epidemic, which was found in Saudi Arabia, called the Middle East respiratory syndrome coronavirus (MERS-CoV) [33], occurred in 2012 and resulted in about 2500 cases and 800 deaths. These observations suggest a high probability that future epidemics will emerge. Whether these epidemics progress to pandemics is contingent on our preparedness and response. Involving a representative proportion of racial/ethnic minority participation in clinical trials that target infectious agents like viruses contributes to informing the continued evaluation of the safety and efficacy of FDA-approved products. Our findings aim to support improved interventions and communication strategies, such as the efficient use of social media by credible sources, providing incentives for research participation, and using clear language when describing each component of research participation. Understanding what motivates marginalized and historically excluded people to participate in research is critical for the development of best research recruitment and implementation practices, systemic changes in healthcare delivery, and improved health outcomes. With the potential shifting of the pandemic to becoming endemic, adequate representation in future clinical trials for vaccines and novel treatments against SARS-CoV-2 are equally important.
Acknowledgements
The authors would like to thank the research participants for their valuable insights, views, and experiences as they relate to the COVID-19 pandemic and research participation. This work was supported by the National Center for Advancing Translational Sciences (NCATS) Clinical Translational Science Award (CTSA) (award number: 5UL1TR0002243-03).
Author Contribution
Iman Barre: validation, formal analysis, data curation, writing original draft, review, and editing; Jamal Moss: investigation, data curation, visualization, writing, review, and editing; Imari Parham: investigation, data curation, visualization, writing, review, and editing; Taneisha Gillyard: data curation, writing, review, and editing; Leah Alexander: conceptualization, writing, review, editing, supervision, planning administration, visualization, and funding acquisition; Jennifer Cunningham-Erves: conceptualization, writing original draft, writing, review, editing, visualization, supervision, planning administration, and funding acquisition; Jamaine Davis: conceptualization, methodology, validation, formal analysis, resources, data curation, writing original draft, review, editing, visualization, supervision, planning administration, and funding acquisition.
Data Availability
Research data are not shared due to ethical restrictions.
Materials Availability
Research data are not shared due to ethical restrictions.
Code Availability
Not applicable.
Declarations
Ethics Approval
This study was approved by Meharry Medical College Institutional Review Board (IRB).
Consent to Participate
Informed consent was obtained from all study participants.
Consent for Publication
Not applicable.
Conflict of Interest
The authors declare no competing interests.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Lopez L Hart LH Katz MH Racial and ethnic health disparities related to COVID-19 JAMA 2021 325 8 719 720 10.1001/jama.2020.26443 33480972
2. Pan D COVID-19 and the new variant strain in England - what are the implications for those from ethnic minority groups? EClinicalMedicine 2021 33 100805 10.1016/j.eclinm.2021.100805 33754139
3. Yancy CW COVID-19 and African Americans JAMA 2020 323 19 1891 1892 10.1001/jama.2020.6548 32293639
4. Sood L Sood V Being African American and rural: a double jeopardy from COVID-19 J Rural Health 2021 37 1 217 221 10.1111/jrh.12459 32362036
5. Frerichs L Integrating systems science and community-based participatory research to achieve health equity Am J Public Health 2016 106 2 215 222 10.2105/AJPH.2015.302944 26691110
6. Warren RC Trustworthiness before trust - COVID-19 vaccine trials and the Black community N Engl J Med 2020 383 22 e121 10.1056/NEJMp2030033 33064382
7. Pfizer V, Vaccines and related biological products Advisory Committee Meeting, BNT162b2, FDA, Editor. 2021.
8. Moderna V, Vaccines and related biological products Advisory Committee Meeting, m RNA-1273, FDA, Editor. 2021.
9. Unger JM The role of clinical trial participation in cancer research: barriers, evidence, and strategies Am Soc Clin Oncol Educ Book 2016 35 185 198 10.1200/EDBK_156686 27249699
10. Rollins L, et al., Learning and action in community health: using the health belief model to assess and educate African American community residents about participation in clinical research. Int J Environ Res Public Health, 2018. 15(9).
11. Shah M Sachdeva M Dodiuk-Gad RP COVID-19 and racial disparities J Am Acad Dermatol 2020 83 1 e35 10.1016/j.jaad.2020.04.046 32305444
12. Corbie-Smith G Attitudes and beliefs of African Americans toward participation in medical research J Gen Intern Med 1999 14 9 537 546 10.1046/j.1525-1497.1999.07048.x 10491242
13. Smirnoff M A paradigm for understanding trust and mistrust in medical research: the Community VOICES study AJOB Empir Bioeth 2018 9 1 39 47 10.1080/23294515.2018.1432718 29368998
14. Ajzen I The theory of planned behavior Organ Behav Hum Decis Process 1991 50 2 179 211 10.1016/0749-5978(91)90020-T
15. Fishbein M. and Ajzen I, Belief, attitude, intention, and behavior: an introduction to theory and research. 1975, Reading, MA\: Addison-Wesley\.
16. Quinn GP Lung cancer patients’ decisions about clinical trials and the theory of planned behavior J Cancer Educ 2011 26 4 641 648 10.1007/s13187-010-0169-8 20949381
17. Aldalaykeh M COVID-19 help-seeking behaviors: application of the theory of planned behavior Psychol Belg 2021 61 1 391 400 10.5334/pb.1034 35070346
18. Chu H Liu S Integrating health behavior theories to predict American’s intention to receive a COVID-19 vaccine Patient Educ Couns 2021 104 8 1878 1886 10.1016/j.pec.2021.02.031 33632632
19. Cordina M Lauri MA Lauri J Attitudes towards COVID-19 vaccination, vaccine hesitancy and intention to take the vaccine Pharm Pract (Granada) 2021 19 1 2317 10.18549/PharmPract.2021.1.2317 33828623
20. Wolff K COVID-19 vaccination intentions: the theory of planned behavior, optimistic bias, and anticipated regret Front Psychol 2021 12 648289 10.3389/fpsyg.2021.648289 34220620
21. Creswell J, Research design: qualitative, quantitative, and mixed methods approaches/J.W. Creswell. 2009.
22. Harris PA Research electronic data capture (REDCap)–a metadata-driven methodology and workflow process for providing translational research informatics support J Biomed Inform 2009 42 2 377 381 10.1016/j.jbi.2008.08.010 18929686
23. Morse JM Critical analysis of strategies for determining rigor in qualitative inquiry Qual Health Res 2015 25 9 1212 1222 10.1177/1049732315588501 26184336
24. Alsan M Wanamaker M Hardeman RR The Tuskegee study of untreated syphilis: a case study in peripheral trauma with implications for health professionals J Gen Intern Med 2020 35 1 322 325 10.1007/s11606-019-05309-8 31646456
25. Erves JC Needs, priorities, and recommendations for engaging underrepresented populations in clinical research: a community perspective J Community Health 2017 42 3 472 480 10.1007/s10900-016-0279-2 27812847
26. Barrett NJ Engaging African Americans in research: the recruiter’s perspective Ethn Dis 2017 27 4 453 462 10.18865/ed.27.4.453 29225447
27. Krieger J.L, Neil J.M, Communication and recruitment to clinical research studies. 2016, Oxford University Press.
28. Toney J.H. and Ishack S, A pandemic of confusion: conflicting messages have characterized not just COVID-19, but also many past disease outbreaks. American Scientist, 2020. 108: p. 344+.
29. Asif M Linguistic analysis of neologism related to coronavirus (COVID-19) Soc Sci Humanit Open 2021 4 1 100201 34490418
30. Wang D Coping with and adapting to COVID-19 in rural United States and Canada Fam Soc 2021 102 1 78 90 10.1177/1044389420960985
31. Depoux A, et al., The pandemic of social media panic travels faster than the COVID-19 outbreak, in J Travel Med. 2020.
32. Thomas PA Severe acute respiratory syndrome Indian J Med Microbiol 2003 21 3 152 160 10.1016/S0255-0857(21)03064-4 17643010
33. (CDC), C.f.D.C.a.P., Update: severe respiratory illness associated with Middle East respiratory syndrome coronavirus (MERS-CoV)--worldwide, 2012–2013. MMWR Morb Mortal Wkly Rep, 2013. 62(23): p. 480–3.
| 36512311 | PMC9746576 | NO-CC CODE | 2022-12-15 23:21:56 | no | J Racial Ethn Health Disparities. 2022 Dec 13;:1-10 | utf-8 | J Racial Ethn Health Disparities | 2,022 | 10.1007/s40615-022-01466-5 | oa_other |
==== Front
Russ J Gen Chem
Russ J Gen Chem
Russian Journal of General Chemistry
1070-3632
1608-3350
Pleiades Publishing Moscow
6029
10.1134/S1070363222110366
Article
Effect of a Substituent in the Fourth Position on the Optical Properties of 2-Oxonicotinonitriles
https://orcid.org/0000-0003-2959-9101
Sorokin S. P.
https://orcid.org/0000-0002-0749-2613
Fedoseev S. V.
https://orcid.org/0000-0002-0938-4659
Ershov O. V. [email protected]
grid.411669.d 0000 0001 0664 3937 I.N. Ulyanov Chuvash State University, 428015 Cheboksary, Russia
13 12 2022
2022
92 11 25002506
20 7 2022
20 7 2022
18 8 2022
© Pleiades Publishing, Ltd. 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Based on six representatives of 2-oxonicotinonitriles, the effect of the nature of the substituent in the fourth position of the pyridine system on the photophysical characteristics was studied. The role of the donor/acceptor nature of the substituent and the solvent nature in the absorbing and fluorescent properties of the compounds was shown.
Keywords:
pyridines
nicotinonitrile
fluorescence
nitriles
trifluoromethyl
isonicotinic acid
issue-copyright-statement© Pleiades Publishing, Ltd. 2022
==== Body
pmcINTRODUCTION
2-Oxonicotinonitrile derivatives (2-oxo-1,2-dihydropyridine-3-carbonitrile, 3-cyanopyrid-2-one) are of considerable interest due to their diverse applications in various fields of science and technology. Among them, substances were found that are used in pharmaceuticals [1–14], agrochemistry [15], as agents of reducing steel corrosion [16], and in the creation of organic functional materials [17–28]. Substituted 2-oxonicotinonitriles are well known for their versatile biological activity. For example, they exhibit antitumor [3,4], anti-tuberculosis [5], anti-inflammatory [6], antipyretic [7], cytotoxic [8], and antimicrobial activity [9]. The possibility of their use as inhibitors of SARS-CoV-2 protease [10], aggregation of α-synuclein [11],
In addition, 2-oxonicotinonitrile derivatives are known for their unique photophysical properties and a wide range of potential applications based on them, for example, as dyes and pigments [17–21], nonlinear optical (NLO) and photorefractive materials [22], dye-sensitized solar cells (DSSC) [23, 24], an emitter in a device with a host–guest configuration in the manufacture of OLED displays [25], a fluorescent probe for visualizing lipid droplets to distinguish between dead and live zebrafish [26], a fluorescent dye for visualizing latent fingerprints and detecting nitrite ion (NO2–) [27], a multisensitive sensor for Ru3+, Fe3+, CrO42–, Cr2O72– and MnO4– ions [28]. Many of the above properties are based on the luminescence phenomenon.
The pyridone fragment, in particular, the 2-oxonicotinonitrile one, causes the appearance of fluorescent properties in the molecule [25–36]. However, there is practically no information in the literature on the systematic study of the effect of an individual substituent on the photophysical properties of such compounds, despite the fact that the optical properties can be finely tuned by introducing individual functional groups. In this regard, this work was devoted to comparing the fluorescent properties of 6-methyl-2-oxo-1,2-dihydropyridine-3-carbonitriles 1а–1f differing by a substituent in the fourth position of the pyridine system (Scheme 1). Pyridone 1а was chosen as the model structure. Molecule 1b contains an electron-donating methyl group. Perfluoroalkyl groups (compounds 1c and 1d) were studied as electron-withdrawing substituents with a strong negative inductive effect. The ester group (1e) and cyano group (1f) were studied as substituents with the conjugation effect.
Scheme 1.
RESULTS AND DISCUSSION
Compound 1a was synthesized by reacting enaminoketone 2 with cyanoacetamide in acetonitrile (Scheme 2). Compounds 1b–1e were obtained from the corresponding β-diketones 3 and cyanoacetamide in refluxing ethanol in the presence of 1,4-diazobicyclo[2.2.2]octane (DABCO) (Scheme 3).
Scheme 2.
Scheme 3.
Compound 1f was prepared according to a previously developed procedure [37] based on the intramolecular cyclization of 4-oxopentane-1,1,2,2-tetracarbonitrile 4 in the presence of pyruvic acid in acetone at room temperature (Scheme 4).
Scheme 4.
Compounds 1a–1c, 1e, and 1f have been reported earlier, their structure was confirmed by IR and mass spectrometry data, as well as by 1H, 13C, and 19F NMR for pyridone 1e, which was not previously described.
Photophysical properties of compounds 1a–1f were studied in three solvents of different nature: acetonitrile, acetic acid and pyridine (Table 1, Figs. 1–3). It was found that, as a rule, for all the studied compounds, changing the solvent from acetonitrile to acetic acid leads to a hypsochromic shift of the absorption maxima by an average of 12 nm, but has almost no effect on the location of the emission maximum. The use of pyridine leads to a bathochromic arrangement of the absorption band maxima by 6 nm on average, and the fluorescence maxima are shifted to the red region relative to acetonitrile.
Table 1. Spectral data for compounds 1a–1f
Compound Solvent λabs (max), nma εmax, L mol–1 cm–1 λfl (max), nmb ΦFc
1a MeCN 340 9542 385 0.18
AcOH 329 8891 385 0.21
Pyridine 346 9891 390 0.13
1b MeCN 334 9254 382 0.07
AcOH 323 5082 379 0.13
Pyridine 338 9068 386 0.06
1c MeCN 360 10369 410 0.66
AcOH 347 9353 410 0.71
Pyridine 365 8280 409 0.35
1d MeCN 360 11116 411 0.69
AcOH 349 9787 410 0.70
Pyridine 365 9045 419 0.32
1e MeCN 372 7545 439 0.71
AcOH 360 7202 439 0.74
Pyridine 380 4068 447 0.39
1f MeCN 378 9171 434 0.77
AcOH 366 7030 431 0.85
Pyridine 384 4948 437 0.29
a Absorption spectra of solutions with a concentration of 10–4 M.
b Fluorescence spectra of solutions with a concentration of 10–4 M. The excitation wavelength corresponds to the absorption maximum.
c Relative quantum yield was measured using 7-hydroxy-4-methylcoumarin in phosphate buffer pH = 10 (ΦF = 0.7, λex 330 nm).
Fig. 1. Absorption spectra of compounds 1a–1f in acetic acid solution (c = 10–4 M).
Fig. 2. Fluorescence spectra of compounds 1a–1f in acetic acid solution (c = 10–4 M).
Fig. 3. Fluorescence quantum yield of compounds 1а–1f.
For all the compounds obtained, with the exception of 1a and 1b, there is a tendency to a decrease in the value of the molar light absorption coefficient when pyridine or acetic acid is used instead of acetonitrile.
The maxima of the absorption and fluorescence bands for compounds substituted with acceptor substituents with the conjugation effect in all the studied solvents are bathochromic relative to all others and are in the range of 360–380 and 439–447 nm, respectively, for compound 1e, 366–384 and 431–437 nm for 1f. Compounds 1a and 1b have the most hypsochromic shifts in the electronic spectra, with absorption and emission band maxima in the range of 329–346 and 385–390 nm, respectively, for 1a, 323–338 and 379–386 nm for 1b. The maxima of the absorption and photoluminescence bands for fluoroalkyl-substituted derivatives 1c and 1d have intermediate values relative to those described above.
The maximum quantum yield for all the tested compounds is observed in acetic acid solution. Pyridine, having a basic nature, can lead to deprotonation of the NH acid center and the process of salt formation [38]. Apparently, therefore, the minimum values of the fluorescence quantum yield are observed in it. In contrast, acetic acid suppresses the dissociation process, which affects the efficiency of the radiative process of pyridones 1 in its solution.
It was found that the introduction of an electron-withdrawing substituent into the molecule of 2-oxonicotinonitrile 1a leads to a significant increase in the quantum yield (more than 3 times), the leader is compound 1f containing a cyano group, ΦF(AcOH) = 85%. On the contrary, if the fourth position of the pyridine ring contains an electron-donating substituent (compound 1b), then a sharp decrease in the efficiency of the radiative process is observed in all the studied solvents (Table 1, Fig. 3).
EXPERIMENTAL
IR spectra were recorded in a thin layer (suspension in mineral oil) on an FSM-2201 IR Fourier spectrometer. NMR spectra were recorded on a Bruker DRX-500 spectrometer, operating frequency 500.13 (1H), 125.76 (13C), 470.59 MHz (19F), using DMSO-d6 as a solvent and TMS as an internal standard. Mass spectra were taken on a Shimadzu GCMS-QP2020 spectrometer (energy of ionizing electrons was 70 eV). Elemental analysis was performed on a FlashEA 1112 CHN analyzer. Progress of reactions and purity of the synthesized substances were monitored by TLC on Sorbfil PTSKh-AF-A-UV plates (detecting with UV irradiation, iodine vapor, thermal decomposition). Melting points of the substances were determined on an OptiMelt MPA100 instrument. Absorption spectra were recorded on an Agilent Cary 60 UV-Vis Spectrophotometer. Fluorescence spectra were recorded on an Agilent Cary Eclipse spectrometer. Fluorescence quantum yield for all solutions was measured relative to 7-hydroxy-4-methylcoumarin in phosphate buffer with pH = 10 (ΦF = 0.7) [39]. The excitation wavelength is 330 nm.
6-Methyl-2-oxo-1,2-dihydropyridine-3-carbonitrile (1a). Cyanoacetamide (0.372 g, 4.4 mol) was added to solution of enaminoketone 2 (0.5 g, 4.4 mmol) in 15 mL of acetonitrile, and the mixture was refluxed for 4 h with stirring. The formed precipitate was filtered off and recrystallized from a mixture of ethanol–water (1 : 1). Yield 83%, mp 294–296°C. IR spectrum, ν, cm–1: 3291 (N–H), 2223 (C≡N), 1666 (C=O). Mass spectrum, m/z (Irel, %): 134 (100) [M]+, 119 (4) [M – CH3]+, 106 (30) [M – CO]+, 105 (73) [M – CO – H]+. Found, %: C 62.76; H 4.55; N 20.79. C7H6N2O. Calculated, %: C 62.68; H 4.51; N 20.88.
General procedure for the synthesis of 6-methyl-2-oxo-1,2-dihydropyridine-3-carbonitriles 1b–1e. The corresponding diketone 3 (4.4 mmol) was dissolved in 15 mL of propanol-2, then 0.372 g (4.4 mol) of cyanoacetamide and 1 g of 1,4-diazabicyclo[2.2.2]octane (8.9 mmol) were added. The mixture was stirred at reflux for 2–4 h (monitoring by TLC). After completion of the reaction, the reaction mixture was cooled to room temperature, poured into cold water (30 mL), and acidified with 2 M HCl solution until acidic. The resulting precipitate was filtered off, washed with water, recrystallized from the appropriate solvent, and dried in a vacuum desiccator over CaCl2 to constant weight.
4,6-Dimethyl-2-oxo-1,2-dihydropyridine-3-carbonitrile (1b). Yield 90%, mp 293–295°C. IR spectrum, ν, cm–1: 3291 (N–H), 2219 (C≡N), 1664 (C=O). Mass spectrum, m/z (Irel, %): 148 (100) [M]+, 133 (2) [M – CH3]+, 120 (35) [M – CO]+, 119 (80) [M – CO – H]+. Found, %: C 65.01; H 5.40; N 18.82. C8H8N2O. Calculated, %: C 64.85; H 5.44; N 18.91.
6-Methyl-2-oxo-4-(trifluoromethyl)-1,2-dihydropyridine-3-carbonitrile (1c). Yield 92%, mp 234–236°C. IR spectrum, ν, cm–1: 3324 (N–H), 2226 (C≡N), 1675 (C=O). Mass spectrum, m/z (Irel, %): 202 (100) [M]+, 187 (2) [M – CH3]+, 174 (45) [M – CO]+, 173 (67) [M – CO – H]+, 105 (26) [M – CF3 – CO]+. Found, %: C 47.47; H 2.50; N 13.79. C8H5F3N2O. Calculated, %: C 47.54; H 2.49; N 13.86.
6-Methyl-2-oxo-4-(pentafluoroethyl)-1,2-dihydropyridine-3-carbonitrile (1d). Yield 84%, mp 241–243°C. IR spectrum, ν, cm–1: 3306 (N–H), 2233 (C≡N), 1677 (C=O). 1Н NMR spectrum (DMSO-d6), δ, ppm: 2.37 s (3H, CH3), 6.56 s (1H, Py), 13.38 br. s (1H, NH). 13C NMR spectrum (DMSO-d6), δC, ppm: 19.9 (СН3), 85.7 (β-Pyr), 103.1 (β-Pyr), 112.1 t. q (CF, 1JCF 285, 39 Hz), 113.9 (CN), 118.6 q. t (CF, 1JCF 287, 37 Hz), 144.8 q (γ-Pyr, C–CF3, 2JCF23 Hz), 156.7 (α-Pyr), 160.9 (α-Pyr). 19F NMR spectrum (DMSO-d6), δF, ppm: –82.2, –112.7. Mass spectrum, m/z (Irel, %): 252 (100) [M]+, 225 (5) [M – HCN]+, 224 (39) [M – CO]+,223 (12) [M – CO – H]+, 155 (94) [M – CF3 – CO]+. Found, %: C 42.96; H 1.98; N 11.16. C9H5F5N2O. Calculated, %: C 42.87; H 2.00; N 11.11.
Methyl 3-cyano-6-methyl-2-oxo-1,2-dihydropyridine-4-carboxylate (1e). Yield 88%, mp 242–244°C. IR spectrum, ν, cm–1: 3306 (N–H), 2221 (C≡N), 1746 (C=O). Mass spectrum, m/z (Irel, %): 192 (100) [M]+, 161 (52) [M – OCH3]+, 134 (58) [M – OCH3 – HCN]+, 133 (73) [M –COOCH3]+, 132 (18) [M – COOCH3 – H]+. Found, %: C 56.36; H 4.17; N 14.66. C9H8N2O3. Calculated, %: C 56.25; H 4.20; N 14.58.
6-Methyl-2-oxo-1,2-dihydropyridine-3,4-dicarbonitrile (1f). To a solution of 1 g (5.4 mmol) of 4-oxopentane-1,1,2,2-tetracarbonitrile 4 in 5 mL of acetone was added a solution of 0.47 g (5.4 mmol) of pyruvic acid in 3 mL of water. The mixture was stirred at room temperature for 6 h, the formed precipitate was filtered off, washed with water, and dried in a vacuum desiccator over CaCl2 to constant weight. Yield 40%, mp 247–249°C. IR spectrum, ν, cm–1: 3329 (N–H), 2225 (C≡N), 1661 (C=O). Mass spectrum, m/z (Irel, %): 159 (83) [M]+, 144 (4) [M – CH3]+, 132 (6) [M – HCN]+, 131 (50) [M – CO]+, 130 (100) [M – CO – H]+. Found, %: C 60.29; H 3.14; N 26.46. C8H5N3O. Calculated, %: C 60.38; H 3.17; N 26.40.
CONCLUSIONS
In conclusion, six derivatives of 2-oxonicotinonitrile were obtained. The effect of the nature of the substituent in the fourth position of the pyridine system on the photophysical properties was studied. It was shown that the introduction of a cyano group leads to the maximum values of the fluorescence quantum yield among the studied functional groups and reaches 85%.
FUNDING
This work was financially supported by the Russian Science Foundation (grant no. 22-13-00157, https://rscf.ru/project/22-13-00157/).
CONFLICT OF INTEREST
No conflict of interest was declared by the authors.
==== Refs
REFERENCES
1 Litvinov V.P. Krivokolysko S.G. Dyachenko V.D. Chem. Heterocycl. Compd. 1999 35 509 10.1007/BF02324634
2 Dolganov A.A. Levchenko A.G. Dahno P.G. Guz’ D.D. Chikava A.R. Dotsenko V.V. Aksenov N.A. Aksenova I.V. Russ. J. Gen. Chem. 2022 92 185 10.1134/S1070363222020074
3 El-Sayed H.A. Moustafa A.H. Hassan A.A. ElSeadawy N.A.M. Pasha S.H. Shmiess N.A.M. Awad H.M. Hassan N.A. Synth. Commun. 2019 49 3465 10.1080/00397911.2019.1672747
4 Saleh, N.M., Abdel-Rahman, A.A.-H., Omar, A.M., Khalifa, M.M., and El-Adl, K., Arch. Pharm., 2021, vol. 354, no. 8, p. e2100085. 10.1002/ardp.202100085
5 Jian Y. Hulpia F. Risseeuw M.D.P. Forbes H.E. Munier-Lehmann H. Caljon G. Boshoff H.I.M. Calenbergh S.V. Eur. J. Med. Chem. 2020 201 112450 10.1016/j.ejmech.2020.112450 32623208
6 Singh V.P. Dowarah J. Marak B.N. Sran B.S. Tewari A.K. J. Mol. Struct. 2021 1239 130513 10.1016/j.molstruc.2021.130513
7 Fayed E.A. Bayoumi A.H. Saleh A.S. AlArab E.M.E. Ammar Y.A. Bioorg. Chem. 2021 109 104742 10.1016/j.bioorg.2021.104742 33647742
8 Tadić J.D. Lađarević J.M. Vitnik Z.J. Vitnik V.D. Stanojković T.P. Matić I.Z. Mijin D.Z. Dyes Pigm. 2021 187 109123 10.1016/j.dyepig.2020.109123
9 Ragab A. Fouad S.A. Ali O.A.A. Ahmed E.M. Ali A.M. Askar A.A. Ammar Y.A. Antibiotics 2021 10 162 10.3390/antibiotics10020162 33562582
10 Fischer C. Vepřek N.A. Peitsinis Z. Rühmann K.-P. Yang C. Spradlin J.N. Dovala D. Nomura D.K. Zhang Y. Trauner D. Synlett 2022 33 458 10.1055/a-1582-0243
11 Mahía A. Peña-Díaz S. Navarro S. Galano-Frutos J.J. Pallarés I. Pujols J. Díaz-de-Villegas M.D. Gálvez J.A. Ventura S. Sancho J. Bioorg. Chem. 2021 117 105472 10.1016/j.bioorg.2021.105472 34775206
12 Cheney I.W. Yan S. Appleby T. Walker H. Vo T. Yao N. Hamatake R. Hong Z. Wu J.Z. Bioorg. Med. Chem. Lett. 2007 17 1679 10.1016/j.bmcl.2006.12.086 17251021
13 Yuldashev P.K. Chem. Nat. Compd. 2001 37 274 10.1023/A:1012534527414
14 Fleming F.F. Yao L. Ravikumar P.C. Funk L. Shook B.C. J. Med. Chem. 2010 53 7902 10.1021/jm100762r 20804202
15 Soliman N.N. Salam M.A.E, Fadda, A.A. Abdel-Motaal M. J. Agric. Food Chem. 2020 68 5790 10.1021/acs.jafc.9b06394 32343563
16 El Basiony N.M. Tawfik E.H. El-Raouf M.A. Fadda A.A. Waly M.M. J. Mol. Struct. 2021 1231 129999 10.1016/j.molstruc.2021.129999
17 Lađarević J. Božić B. Matović L. Nedeljković B.B. Mijin D. Dyes Pigm. 2019 162 562 10.1016/j.dyepig.2018.10.058
18 Mijin D. Ušćumlić G. Perišić-Janjić N. Trkulja I. Radetić M. Jovančić P. J. Serb. Chem. Soc. 2006 71 435 10.2298/JSC0605435M
19 Alimmari A.S.A. Marinković A.D. Mijin D. Valentić N.V. Todorović N. Ušćumlić G. J. Serb. Chem. Soc. 2010 751 1019 10.2298/JSC091009074A
20 Kirchner E. Bialas D. Wehner M. Schmidt D. Würthner F. Chem. Eur. J. 2019 48 11285 10.1002/chem.201901754
21 Zhao X.-L. Geng J. Hu B. Xu D. Huang W. Dyes Pigm. 2018 155 1 10.1016/j.dyepig.2018.03.014
22 Würthner F. Yao S. Debaerdemaeker T. Wortmann R. J. Am. Chem. Soc. 2002 124 9431 10.1021/ja020168f 12167038
23 Eltoukhi M. Fadda A.A. Abdel-Latif E. Elmorsy M.R. J. Photochem. Photobiol. (A) 2022 426 113760 10.1016/j.jphotochem.2021.113760
24 Khalifa M.E. Almalki A.S.A. Merazga A. Mersal G.A.M. Molecules 2020 25 1813 10.3390/molecules25081813 32326499
25 Vinayakumara D.R. Ulla Н. Kumar S. Pandith A. Satyanarayan M.N. Rao D.S.S. Prasad S.K. Adhikari A.V. J. Mater. Chem. (C) 2018 6 7385 10.1039/C8TC01737A
26 Cui W.-L. Wang M.-H. Chen X.-Q. Zhang Z.-H. Qu J. Wang J.-Y. Dyes Pigm. 2022 204 110433 10.1016/j.dyepig.2022.110433
27 Manjunatha B. Bodke Y.D. Mounesh Nagaraja, O. Navaneethgowda P.V. New J. Chem. 2022 46 5393 10.1039/D1NJ04751E
28 Wu Y. Liu D. Lin M. Qian J. RSC Adv. 2020 10 6022 10.1039/C9RA09541A 35497449
29 Nawel Mehiaoui N. Kibou Z. Gallavardin T. Leleu S. Franck X. Mendes R.F. Paz F.A.A. Silva A.M.S. Choukchou-Braham N. Res. Chem. Intermed. 2021 47 1331 10.1007/s11164-020-04373-8
30 Chernov N.M. Shutov R.V. Sipkina N.Y. Krivchun M.N. Yakovlev I.P. ChemPlusChem 2021 86 1256 10.1002/cplu.202100296 34472730
31 El-Sayed H.A. Moustafa A.H. El-Salam A.E.A. J. Heterocycl. Chem. 2020 57 2738 10.1002/jhet.3982
32 Hamid A.M.A. El-Sayed H.A. Mohammed S.M. Moustafa A.H. Morsy H.A. Russ. J. Gen. Chem. 2020 90 476 10.1134/S1070363220030226
33 Hagimori M. Nishimura Y. Mizuyama N. Shigemitsu Y. Dyes Pigm. 2019 170 107705 10.1016/j.dyepig.2019.107705
34 Darehkordi A. Hosseini M. Rahmani F. J. Heterocycl. Chem. 2019 56 1306 10.1002/jhet.3501
35 Vinayakumara D.R. Kesavan R. Kumar S. Adhikari A.V. Photochem. Photobiol. Sci. 2019 18 205 10.1039/C9PP00046A
36 Ershov O.V. Fedoseev S.V. Belikov M.Y. Ievlev M.Y. RSC Adv. 2015 5 34191 10.1039/C5RA01642H
37 Nasakin O.E. Nikolaev E.G. Terent'ev P.B. Bulai A.K. Zakharov V.Y. Chem. Heterocycl. Compd. 1985 21 1019 10.1007/BF00515027
38 Belikov M.Y. Ershov O.V. Eremkin A.V. Kayukov Y.S. Nasakin O.E. Russ. J. Org. Chem. 2010 46 615 10.1134/S1070428010040378
39 Chen R.F. Anal. Lett. 1968 1 423 10.1080/00032716808051147
| 0 | PMC9746578 | NO-CC CODE | 2022-12-15 23:21:56 | no | Russ J Gen Chem. 2022 Dec 13; 92(11):2500-2506 | utf-8 | Russ J Gen Chem | 2,022 | 10.1134/S1070363222110366 | oa_other |
==== Front
Metabolomics
Metabolomics
Metabolomics
1573-3882
1573-3890
Springer US New York
1957
10.1007/s11306-022-01957-w
Review Article
Providing metabolomics education and training: pedagogy and considerations
http://orcid.org/0000-0002-1648-5255
Winder Catherine L. 1
http://orcid.org/0000-0002-1462-4426
Witting Michael 23
http://orcid.org/0000-0001-8172-6599
Tugizimana Fidele 45
http://orcid.org/0000-0001-6924-0027
Dunn Warwick B. 1
http://orcid.org/0000-0002-0758-0330
Reinke Stacey N. [email protected]
6
the Metabolomics Society Education and Training Committee
1 grid.10025.36 0000 0004 1936 8470 Analytical & Clinical Metabolomics Group and Liverpool Training Centre for Metabolomics, Department of Biochemistry and Systems Biology, Institute of Systems, Molecular and Integrative Biology, University of Liverpool, Biosciences Building, Crown Street, Liverpool, L69 7ZB UK
2 grid.4567.0 0000 0004 0483 2525 Metabolomics and Proteomics Core, Helmholtz Zentrum München - German Research Center for Environmental Health (GmbH), Neuherberg, Germany
3 grid.6936.a 0000000123222966 Chair of Analytical Food Chemistry, TUM School of Life Sciences, Technical University of Munich, Freising-Weihenstephan, Germany
4 grid.412988.e 0000 0001 0109 131X Department of Biochemistry, University of Johannesburg, Auckland Park, Johannesburg, 2006 South Africa
5 International Research and Development Division, Omnia Group Ltd., Bryanston, Johannesburg, 2021 South Africa
6 grid.1038.a 0000 0004 0389 4302 Centre for Integrative Metabolomics & Computational Biology, School of Science, Edith Cowan University, Joondalup, 6027 Australia
13 12 2022
2022
18 12 10623 6 2022
11 11 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Background
Metabolomics is a highly multidisciplinary and non-standardised research field. Metabolomics researchers must possess and apply extensive cross-disciplinary content knowledge, subjective experience-based judgement, and the associated diverse skill sets. Accordingly, appropriate educational and training initiatives are important in developing this knowledge and skills base in the metabolomics community. For these initiatives to be successful, they must consider both pedagogical best practice and metabolomics-specific contextual challenges.
Aim of review
The aim of this review is to provide consolidated pedagogical guidance for educators and trainers in metabolomics educational and training programmes.
Key scientific concepts of review
In this review, we discuss the principles of pedagogical best practice as they relate to metabolomics. We then discuss the challenges and considerations in developing and delivering education and training in metabolomics. Finally, we present examples from our own teaching practice to illustrate how pedagogical best practice can be integrated into metabolomics education and training programmes.
Keywords
Metabolomics
Education
Training
Pedagogy
http://dx.doi.org/10.13039/501100000265 Medical Research Council MR/S010483/1 issue-copyright-statement© Springer Science+Business Media, LLC, part of Springer Nature 2022
==== Body
pmcIntroduction
The term ‘metabolomics’ was first coined in 1998 (Oliver et al., 1998). Since then, substantial technological, methodological, and computational advancements have led to the maturation of the field (Kell & Oliver, 2016). Over the last 20 years, metabolomics has demonstrated utility in a wide range of contexts, including human health and disease (Beger et al., 2016; Rattray et al., 2018), nutrition (Brennan, 2017), environmental science (Bundy et al., 2008), plant and agriculture science (Fraser et al., 2020; Rai et al., 2019; Tinte et al., 2021), microbiology (Mashego et al., 2007), and synthetic biology (Hollywood et al., 2018). For a substantial period, the application of metabolomics was limited to academic researchers who had direct access to the required analytical and computational facilities. However, the arrival of analytical service providers has led to metabolomics becoming much more accessible to a wider range of applied life science researchers, further accelerating the growth of the field. Today, metabolomics is positioned as one of the key pillars in the systems biology (omics) sciences.
Despite the increased accessibility of metabolomics, two unique challenges remain for researchers. First, metabolomics is a highly multidisciplinary science (Fig. 1). To complete a metabolomics study from beginning to end requires expertise in the applied biological context, analytical and physical chemistry, quality assurance/control, cheminformatics, statistics and data science, bioinformatics, and biochemistry. Second, the metabolome itself is chemically highly diverse (in terms of physicochemical properties, concentration range and stability) and complex in that metabolites are highly interdependent and exhibit correlated dynamics. As a result, metabolomics analytical and data science workflows are difficult to standardize as often there is no single correct approach and each analysis is context specific. Accordingly, effective metabolomics researchers require a diverse set of knowledge in all areas of the workflow as well as specific skills in some parts of the workflow, not only cross-disciplinary objective knowledge but also a high level of subjective experience-based judgement.Fig. 1 Metabolomics workflow steps. The metabolomics workflow, from experimental design to biological interpretation, is shown. Different levels of (1) expertise and (2) awareness of high-level concepts are required for each step of the workflow
The development of knowledge and skills can be supported through two non-mutually exclusive learning processes: training and education (Table 1). Training is a short-term process that facilitates the development of skills needed to perform a particular task (for example, how to use a particular software package). Education is a longer-term learning process that focusses primarily on the development of theories and concepts and in some cases on their application. This type of learning typically occurs at schools, colleges, and universities. In an attempt to identify training needs in metabolomics, the Metabolomics Society and ELIXIR-UK jointly conducted and published a survey in 2015 (Weber et al., 2015). The key recommendations from this survey included: (i) development of face-to-face and online training courses in analytical metabolomics and bioinformatics, (ii) creation of funding opportunities to develop (inter-)national training networks, (iii) improvements in advertising and accessibility of training courses, and (iv) considerations for formal accreditation. Since 2015 the provision of metabolomics training has advanced considerably, by dedicated academics providing short course training and several centres specialising in the provision of both face-to-face and online training including the Liverpool Training Centre for Metabolomics, EMBL-EBI, Imperial International Phenome Training Centre, Birmingham Metabolomics Training Centre, the Metabolomics Workbench, the West Coast Metabolomics Centre, and Metabolomics South Africa (this is not a complete list of training centres available). These centres have the facilities and infrastructure via state-of-the-art instrumentation and computational infrastructure to provide a range of hands-on and online training and fill training gaps highlighted in the survey. Geographical location of the trainer and trainees may limit accessibility of face-to-face training. Other training opportunities include conference workshops (for example, at Metabolomics Society conferences). While workshops provide an excellent opportunity to learn about focussed topics, such as new software developments, the time available and scale of the workshops may not provide sufficient opportunity to develop a higher-level knowledge and skill set.Table 1 A comparison of training and education
Factor Training Education
Provider Primarily training centres and scientific companies which require no official registration (as would be required for education at a school or university) Officially registered school, college, or university
Focus Typically, hands-on training of skills to allow trainee to perform a task or set of tasks in their jobs Provision of understanding concepts as primary focus with hands-on skill provision as a secondary focus
Course duration Short-term (days) Long-term (months to years)
Quality Assurance Not required to operate Normally required to operate
Students/trainees Young adults and adults Children, young adults, adults
Formal assessments No Yes
Informal assessments Yes Yes
Qualification No (continuing professional development points possible) Yes
One of the major challenges in delivering metabolomics training is to consider the varying levels of knowledge of participants. Researchers originate from varying backgrounds and although it may be expected that metabolomics teaching would be embedded in some undergraduate and post-graduate programmes (for example, biology degrees), this cannot be guaranteed. To the authors’ knowledge, metabolomics teaching in education is emerging but remains limited; the authors are aware of only a small number of undergraduate and postgraduate programmes which include metabolomics in the curricula. Metabolomics-focussed papers has also recently started to appear in the educational literature (Boyce et al., 2019; La Frano et al., 2020; Sandusky, 2017). A search for metabolomics in postgraduate Master’s programs across the globe also returned limited results; of 24,646 Master’s programmes available globally and listed on the www.findamasters.com portal, only 50 programmes were listed using the search term metabolomics. Moreover, no single programme was focused on metabolomics only; instead, metabolomics was included as only one of multiple modules taught (FindAMasters, 2022). Indeed, the authors believe that a Master’s course focused only on metabolomics is not appropriate because the volume of materials available is less than would be expected in a Master’s course and that such a course would not be financially rewarding to the academic institutions as a low number of student registrations would be expected. In 2019 the Metabolomics Society Strategy Task Group published a survey of the membership which reported that 47% of respondents had been working in metabolomics for less than 5 years and 44% of respondents in trainee positions including undergraduate, Master’s or PhD students and postdoctoral fellows (Zanetti et al., 2019). Owing to this relative inexperience of metabolomics researchers and lack of fundamental metabolomics teaching in the education setting, training and education remains a top priority for the field. The approach and mechanism of this training requires consideration to support all researchers working in the field.
Educating the scientific community in the application of metabolomics shares the same challenges as other disciplines such as bioinformatics, where potential users have different educational backgrounds and skills sets. The International Society of Computational Biology (ISCB) Educational Committee addressed the challenge by developing a framework for training needs and curricula using a set of bioinformatics core competencies and mapping the competencies to typical user profiles (Mulder et al., 2018). The competencies and mapping to user profiles were refined via several workshops; however, the scoring approach was too ambiguous to allow meaningful classification until the competency levels were aligned to the pedagogical hierarchical cognitive model of Bloom’s taxonomy (described in the next section). This illustrates the necessity of integrating pedagogical best practice when designing training and educational frameworks in contemporary research practice. To effectively support the development of knowledge and skills in metabolomics, pedagogical best practice should be implemented when designing and delivering education and training curricula; however, there is a lack of consolidated pedagogical guidance for the metabolomics community, including how to overcome metabolomics-specific challenges. In this manuscript we will describe key pedagogical principles as well as other practical considerations which we recommend should be applied in education and/or training. Pedagogical principles are reviewed from the education perspective and then adopted in the development of training courses. We also discuss additional considerations for developing and providing training courses focused on metabolomics workflows. Finally, we will provide case study examples of educational and training curricula from our own practice. As such, this manuscript will provide future and new teachers with a perspective on pedagogical and practical considerations, researchers with a deeper understanding of pedagogy as it relates to metabolomics, and experienced teachers with creative examples of integrating pedagogy with metabolomics education and training.
Principles of pedagogical best practice—constructive alignment
Before discussing considerations that are specific to the field of metabolomics, it is important to understand key foundational pedagogical principles. In this section we will discuss and provide examples of pedagogical principles applied in education, though many of the principles are equally important in training programmes. It is outside the goals of this manuscript to discuss in detail the general principles of pedagogy and we recommend the following resources to provide more detailed information (Anderson et al., 2001; Biggs, 1996; Hoque, 2017; Jarvis, 2012; Print, 1993).
Constructive alignment
A traditional view is that teachers often plan the content of their lecture(s), but students plan according to their exam requirements. This might lead to a misalignment of teaching provided by the educator and learning required by the student. To avoid this, the concept of constructive alignment has been developed (Biggs, 1996). In this concept, learning outcomes, assessment, and teaching and learning activities are aligned. The three points to consider in constructive alignment are:Learning Outcomes—What should students be able to do after the lecture/course?
Teaching and Learning Activities—What content and methods shall be used to achieve the learning outcomes?
Assessment—Will and how will achievement of the learning outcomes be assessed?
Bloom’s taxonomy—cognitive domain
It is important to define the competencies that students should have after successfully completing a teaching programme and to which level of expertise each competency can be acquired. A tool to order these levels is Bloom’s taxonomy, which comprises three different hierarchical models (cognitive, emotional, psychomotor). The cognitive domain model [first reported in 1956 (Bloom, 1956) and revised in 2001 (Anderson et al., 2001)] consists of six different levels and is widely applied in curriculum development (Fig. 2). The model is hierarchical, with each level building off the last. Importantly, all learners, regardless of age or educational/training level, must start at the lowest level and progressively move upward as their learning evolves.The first three levels are often referred to as the lower-order thinking skills. The lowest cognitive level is ‘Remember’. This involves rote memory and is evidenced by activities such as recognition or recall of facts, definitions, and lists.
The second level is ‘Understand’, which learners evidence by organising, summarising, generalising or describing concepts.
The third level is ‘Apply’, where learners start to use their knowledge in a new way. Application is evidenced by modelling, solving, or choosing.
At this stage, learners enter the higher-order thinking skills. In the fourth level, ‘Analyse’, learners can dissect concepts and ideas into parts and are able to compare them against other concepts and ideas.
In the fifth level, ‘Evaluate’, decisions made can be justified. That is, learners defend their choices, and they can deliver arguments.
The sixth and highest level in the taxonomy is ‘Create’. At this level, learners develop new ideas, concepts, or products. Since learners have evidenced all the lower levels, they can also defend the decisions they made in terms of design and conception. Without any doubt this is the level to which students should reach and it is the formal objective of many undergraduate programmes though is often only truly achieved for postgraduate PhD students.
Fig. 2 Bloom’s taxonomy levels for A cognitive domain and B emotional domain. Evidence of each domain level is shown on the left, with indicative examples of learning verbs shown on the right. In both domains, learners must start at the bottom level and work through the respective levels to develop expertise in each level
Learning outcomes
In educational development, the first question that needs to be asked is: ‘What skills or knowledge should the learners be able to evidence upon successful completion of the teaching programme?’ Statements of this evidence are articulated through learning outcomes. To be most effective, learning outcome statements should integrate verbs that directly align to the Bloom’s taxonomy levels being targeted. By doing so, the learning can be planned to target an intended cognitive level. Learning outcome statements are most effective when they are clear and observable (Print, 1993). For example, learning verbs such as explain, describe, or choose clearly communicate expectations to learners, and completion of these outcomes can be measured. Contrarily, the verbs know, understand, and demonstrate are vague and cannot be directly observed.
In any educational programme, it is important to align the learning outcomes with the intended Bloom’s taxonomy levels. Learners must always progress through the cognitive levels, starting at the bottom. Thus, consideration must be made for what cognitive level is realistically achievable based on prior competencies of the students and the time available for learning. For example, in a short course, it may be realistically achievable for a novice to learn how to perform Quality Assurance (QA) checks on a Liquid Chromatography–Mass Spectrometry (LC–MS) instrument, apply existing acquisition methods, or use existing data analysis tools. It may not be achievable for a novice learner to be able to troubleshoot LC–MS problems, create novel acquisition methods, or create bespoke data analysis methods in the same timeframe. Contrarily, these outcomes might be achievable for learners who have evidenced progression through the Bloom’s levels prior to the body of learning.
Teaching and learning activities
Once the intent of the learning (i.e. learning outcomes and cognitive levels) is defined, the next step is to consider what content will support the intent and evidence of learning as well as the sequence and mechanism of delivery. There are many ways to deliver content, which all fall under two overarching approaches (Jarvis, 2012): teacher-centred and student-centred learning (Fig. 3). Teacher-centred learning views students as ‘empty vessels’ and teachers as the source of knowledge. In this approach, teachers control the learning process and convey information to the students (for example, lectures and demonstrations). Teacher-centred approaches are effective for contexts where learning factually correct information such as rules, processes, and procedures is required. They are also appropriate in content-heavy curricula where a large volume of information needs to be delivered efficiently and often to many students. Student-centred learning approaches shift the focus away from the teacher and towards the student. Learning becomes less of a transmission of knowledge and more about an active process that the learner participates in (for example, workshops and project-based learning). This approach places much more accountability on the learner and it promotes autonomous learning. Typically, teacher-centred and student-centred approaches are used in combination to support both knowledge dissemination and active skills development. This is particularly relevant for metabolomics, where learners often come from diverse discipline backgrounds. An example of applying this dual approach is teaching metabolite annotation/identification. Students listen to an online pre-recorded lecture (teacher-centred) on the various physicochemical properties that support metabolite identification decisions. In class, students complete a student-centred activity called ‘Put Yourself on the Line’. The teacher draws an imaginary line along the length of the classroom. One end of the line represents ‘I am very confident in the identification of this metabolite’, the other is ‘I have no confidence in the identification of this metabolite’, and the rest of the line is a continuum between the two choices. Students are then presented an example metabolite peak with evidence to support its annotation/identification. They decide how confident they are in the identification and stand at the corresponding place along the line. Once all students are standing on the line, they are invited to articulate their decision for choosing their position. This process may be repeated with several examples to challenge students and develop higher level ‘analyse’ and ‘evaluate’ skills.Fig. 3 Methods of delivering teaching and training material
A range of teaching and learning tools can be applied to facilitate content delivery and supplement more traditional teaching tools and also maintain interest, accommodate different learning styles, and improve the learning experience. Specific examples that are relevant to metabolomics include:Applying live polls during teaching/training sessions to allow (1) students to assess their level of knowledge and compare this to their peers and (2) teachers to determine the current level of learning achieved. For example, a live poll may ask students to report on the criteria they would use for data cleaning. The results would provide different perspectives on what the learners find important (example: peak shape, quality metrics such as relative standard deviation, a confident identification).
Providing online lectures and/or recordings of classroom sessions allows students to watch in their own time and at their own speed with the ability to rewind and fast forward. The typical 50-min classroom lecture can be divided in to smaller and multiple lectures with linking text which can improve students’ interest and learning. This can be particularly useful for a metabolomics audience, where learners have different levels of prior knowledge. Recordings allow novices to spend more time on learning the content.
Providing links to audio-visual (AV) resources to supplement teaching. For example, providing a link to a YouTube video focused on the learning objective for which a lecture was provided by the teacher allows the same concept to be described by two different teachers and most probably in two different ways. If the first teacher’s approach was not best suited to the learning style of the student, then the YouTube approach may be successful. Again, this can be particularly useful for novice learners.
Computational notebook-style live scripts (example: Jupyter Notebooks, R markdown, MATLAB live scripts) can help students develop their coding capabilities. With live scripts, teachers can provide background on the algorithms and functions (including links to external resources) as well as suggestions for changing input arguments to promote experiential learning. Mendez et al (2019) provide a tutorial review on using Jupyter Notebooks to support metabolomics.
Online tools and resources for hands-on teaching. For example, web-based software requiring no downloading and installation, such as MetaboAnalyst, and which can easily be applied by one or one hundred students. Microsoft Azure Labs (https://labs.azure.com/) offers a virtual machine (VM) solution, where teachers create a master VM with all required software and files loaded. The template is then published and a private VM clone is created for each student. This enables students to access the required computational resources at any time and from anywhere in the world. This is particularly relevant for teaching metabolomics in contemporary contexts. Metabolomics requires access to and installation of specialised software, and large volumes of data, which is practically difficult to support in remote teaching contexts. For an example that illustrates how Microsoft Azure Labs has been used in teaching, see the ‘Case Studies’ section below.
Open access data. Metabolomics data collected applying NMR spectroscopy and LC–MS is now available in metabolomics data repositories [e.g. MetaboLights (Haug et al., 2012), Metabolomics Workbench (Sud et al., 2015) and GNPS (Wang et al., 2016)] and can be used for hands-on teaching. The reanalysis of deposited data will be applied more frequently in future research and so teaching the next generation in its use is important.
Assessment
Assessment is a critical component of the learning process in educational settings. It not only evidences achievement of learning outcomes (Biggs, 1996) but it is a critical component in supporting student learning (Jarvis, 2012). There are three main reasons why assessment is included in educational curricula.Assessment for learning. This provides the teacher diagnostic information which they can use to inform, improve, and direct their teaching.
Assessment of learning, which evidences learners’ achievements of learning outcomes.
Assessment as learning or self-assessment. This provides students an opportunity to not only monitor their own progress but reflect on their learning to direct future activities.
There are two primary types of assessment: formative and summative. Formative assessment occurs before and during the learning process and provides students with feedback to improve their learning during the learning cycle. Formative assessments are typically low stakes (or not marked) and include activities such as in-class or online discussions, small and frequent quizzes, and short writing assignments. Summative assessments occur at a later stage or the end of the teaching programme and are used as evidence that learning outcome have been achieved and to award achievement. These assessments are typically high stakes and include assessments such as exams, project reports, literature reviews and portfolios. There is also a third type of assessment, self-assessment. A range of tools are available to embed self-assessment into teaching and training to facilitate self-assessment and self-direction of learning. The range of assessments are shown in Table 2 along with a classification of the assessment type, and examples when typically used. Specific examples of assessments used in metabolomics curricula are described in the Case Studies section below.Table 2 Types of assessments that are routinely used across the teaching and training environments in metabolomics
Assessment/activity Classification* Description Examples
SA FA S TR
Examination x Formal procedure may employ MCQs, short answer and essay questions. Applied to award qualifications Final year examinations in undergraduate teaching
Test x x Tool to measure knowledge levels and adjust teaching material/delivery accordingly. Various formats may be employed including MCQs, short and long answer questions Mid or end of module assessment in undergraduate and post-graduate teaching
Thesis/Dissertation x An extended written document focussed on a specific subject, demonstrates research and contribution of new knowledge, theories or practice to a specific field BSc or MSc thesis
PhD dissertation
Viva/Oral Defence x x Oral examination a body of work in front of a panel of experts and chair. Frequently used in combination with a written dissertation or report as part of a final assessment or to assess progress at specific milestones A PhD viva
Project report x A structured report to present the strategy, results, and conclusions of a project, may include a laboratory, computational or problem-solving project Laboratory project
Portfolio x x Collection of student work completed over time. Learning is structured such that students complete various tasks throughout the course. Evidence of task achievement is then collated into a portfolio, may include a reflective element Undergraduate/masters portfolio to document completion of modules in data processing, analysis and metabolite identification
Reflective assignment x x Assignment that guides students to consider, articulate their experience, and inform future practice. May be a stand-alone assignment or part of another assignment (such as a portfolio) Undergraduate assignment to define and consider how a work-based placement has improved their skill-set for future employment
Presentation (oral, poster, elevator pitch) x x Can be used in a variety of different ways. Facilitates development of oral and visual communication skills Oral presentation at end of project to present project findings, often applied in undergraduate and post-graduate taught courses
Group task x x A collaborative activity where students work in groups to perform tasks or solve problems, may involve a short task or project over several days or weeks Problem-based activity to design a metabolomics experimental plan
Laboratory/skills assessment x x x A laboratory-based skills assessment to test the practical skills and competence level of students. The assessment will be based on a set of transparent criteria outlined in advance Undergraduate laboratory practical exam in range disciplines including biology, chemistry, biochemistry
Quiz x x x x Low-stakes assessment. Options include short answers or input the missing word (cloze question). If implemented using an online platform, auto-feedback and auto-grading can be applied, providing students immediate feedback upon completion (or deadline) Cloze questions to add in missing word as part of an online quiz
Multi-choice question (MCQ) quiz or test x x x x Tests knowledge and understanding, applied on online platform to support self-assessment in education and training sections. Implement auto feed-back to direct students back to specific areas of course. Question design important to test students’ knowledge, not providing obviously incorrect answers MCQ's embedded in online course employs auto-marking and auto-feedback
Discussion activity x x Face-to-face or asynchronous online activities. Questions are posed by the facilitator. Can be used for a variety of reasons, including to promote inclusion between students, extension of learning, development of critical thinking skills, and compilation of resources The use of discussion boards in data analysis courses to explore and evaluate the use of data analysis approaches across the field
Polls x x Live or asynchronous survey. Typically, multiple choice or true/false answer options. Results can be conveyed back to learners Live poll during lectures to stimulate thought and discussion, a good option when considering quality assurance and quality control in metabolomics. Facilitates interactive learning in lectures
Directed exercise x x Data processing activity Workshops on how to code or data analysis tutorial using web-based software such as MetaboAnalyst
*Classification of assessment abbreviations: SA summative assessment, FA formative assessment, S self-assessment, TR Tutor review to direct learning
Principles of pedagogical best practice—other considerations
Understanding prior knowledge
“To teach is to learn twice” (Joseph Joubert)—A prominent fact to consider when teaching is the “expert blind spot”. Metabolomics teachers are often experts who have years or even decades of experience in their field. It can be easy to take this knowledge and experience for granted and forget how difficult it is to learn something new. As metabolomics is a multidisciplinary field involving aspects from biochemistry, analytical chemistry and bioinformatics, heterogeneity of the students’ knowledge and experiences is an important factor. It is important to not only factor this in during planning but also to regularly check with students that the pace and depth is suitable. Providing additional resources (for example: readings, online resources) can help students to close the gap and consolidate their knowledge away from teaching/training sessions.
Bloom’s taxonomy—emotional domain
A more recent development of Bloom’s taxonomy has considered the emotional levels of the students and their ability to learn, specifically that the student must be emotionally involved and motivated to participate in the learning process (Krathwohl et al., 1964). The teaching and learning activities and tools can be developed to facilitate the emotional involvement and motivation of the student. There are five levels and as with the cognitive levels, each new level builds on the previous levels (Fig. 2).Receiving: the student passively approaches the teaching process by paying attention to the teaching process. This level is about the student’s memory and recognition of materials.
Responding: the student actively participates in the learning process through reaction to the teaching stimulus.
Valuing: the student understands the value of learning to increase their knowledge.
Organizing: the student can integrate different information, ideas, and values to allow comparison and elaboration of what has been learnt.
Characterising: the student can build abstract knowledge.
The challenges and considerations in education and training development and operation
Pedagogical best practice must be considered within the constraints of the teaching context. In this section we will discuss challenges and considerations that are relevant to the metabolomics context. Although both education and training will be covered, we will focus more on training course development and operation. Training courses form an important part of Continuing Professional Development (CPD). As metabolomics is a relatively new discipline, most of the formalised metabolomics education/training to date will have been performed via post-graduate education or CPD-based training. Undertaking CPD within your role ensures your knowledge stays relevant and current. Short courses are used as an intermittent booster to fast-track knowledge in a specialised subject but also provide access to experts. With new approaches such as metabolomics, CPD-based training courses provide a mechanism to teach new knowledge and skills to those who completed their formal education many years ago. Many of the concepts discussed above for education can also be related to training, for example, Bloom’s taxonomy and teaching tools and the teaching approaches summarised in Fig. 3. However, the goals of education and training are different. When developing a training course, the content will typically include a narrower syllabus than in the education course to impart specialised knowledge, skills, and direct students to resources. When developing such courses considering the audience and challenges specific to the subject area and considering what cognitive level is realistically achievable within the course are very valuable and are discussed below.
Challenges and considerations for education and training
All omics are not the same
Many different omics research areas and tools are available for ‘discovery-based’ biological research including genomics, transcriptomics, proteomics, and metabolomics. The general research goal of these omics techniques is to identify biochemicals which are involved in biological mechanisms or have the potential to act as biomarkers. Importantly, and often forgotten, is that different tools are applied in the research laboratory and computational office for different omics disciplines. For example, the biological roles of metabolites and speed of metabolism can be different to other types of biochemicals, therefore sampling and storage methods can be different for metabolomics compared to other omics techniques and so an expert in say proteomics is not an expert in metabolomics (and vice versa). This is an example of how users require specialised knowledge in for example, sample collection, storage, and preparation. Scientists who transition from another ‘omics disciplines will have transferable skills, such as the ability to operate a mass spectrometer or experience of working with large data sets but will require metabolomics-specific training to understand the challenges of studying the metabolome and develop an appreciation of how the multiple omics approaches differ. For new users to the field, who are not working in an established metabolomics research group and therefore cannot learn from their peers, hands-on laboratory, bioinformatics, and data analysis training is an invaluable part of their professional development.
Research involving metabolomics is multidisciplinary
A wide range of different scientific fields can be involved in a study which applies metabolomics research tools including biologists, clinicians, plant biologists, analytical chemists, physicists, bioinformaticians and computational biologists. Therefore, one unique challenge to consider when developing training is the diverse group of scientists working in or applying metabolomics in their research, understanding the varying background knowledge and skills. Whether a scientist is a specialist in one part of the workflow or is involved in the entire pipeline, they require an appreciation of all steps in the metabolomics workflow. For example, when a statistician is reviewing the outputs of the data analysis it is useful to have an appreciation of the experimental pipeline and steps which may introduce sources of variation. A clinician will often collaborate with a core metabolomics group to perform the experimental and data analysis work. So, whilst an in-depth understanding of some skills (for example, operating a mass spectrometer) is not required, a good background knowledge of the subject, experimental design and an understanding of the challenges is essential. Accordingly, consideration must be made about whether one trainer/educator can reasonably cover all aspects of the course content. The authors recommend that multiple trainers with different expertise areas are involved in a single training course and that trainers limit their teaching scope to content that they have deep expertise in.
Educating and training scientists with no or limited expertise requires specialised courses that:Consider and accommodate different scientific backgrounds and levels of knowledge
Address the needs of the trainee at different stages of their careers
Provide training that is relevant to the needs of the scientist
Supports transitioning between areas of specialisation, for example, an experimental scientist moving from a biological focus to analytical chemistry or learning to develop computational tools.
We provide two examples in the case study section to demonstrate how to overcome the diverse backgrounds of students in (1) a teaching-focussed postgraduate course and (2) supporting research-based students with minimal knowledge in metabolomics.
Feasibility
Learning is a complex process. It can be easy to overestimate what can be achieved both by the teacher and the learners. Two primary factors that are often underestimated by inexperienced teaching staff are time and cognitive load. Effective scaffolding and development of learning activities takes a substantial amount of time. It is important to consider what can realistically be achieved in development from a workload standpoint. Every iteration of delivery will identify new areas of improvement (through trainee feedback and trainer reflection), which can be developed incrementally. It is also easy to underestimate the amount of time it will take trainees to perform the learning activities, and this should be accounted for in the planning stage. Cognitive load is also important to consider. According to cognitive load theory (Sweller, 1988), there is a limited capacity of the working memory at any given time. If the working memory is overloaded with too many simultaneous tasks, learning is impaired. This is important for metabolomics, as students are often required to learn both concepts and applications. For example, students learning about statistics must learn both statistical concepts and how to calculate statistics using software. Learning to use statistical software with a graphical user interface (GUI) such as SPSS is less cognitively demanding (at least for beginners) than using a command-line interface (CLI) software such as R. Therefore, efforts to reduce cognitive load will allow learners to focus on the key learning objectives. The use of MetaboAnalyst in the delivery of a data analysis course is an example where participants can focus on the reasons and correct use of the processing steps without having to learn how to programme.
Is the infrastructure available?
Anyone can theoretically operate a training course. However, appropriate infrastructure is required to deliver high quality training and learning. For online courses a training platform with the capabilities to present training material delivery in different formats and facilitate interaction between trainees and trainers; for example, providing only a series of YouTube videos with no interaction will not facilitate engagement and higher-level learning (as defined by Bloom) in all participants. The delivery of face-to-face courses will require a training room and possibly a laboratory or computational cluster. The training environment has to be appropriate (e.g. comfortable, appropriate temperature, no loud noises) with access to state-of-the art instrumentation and software. Many users would like to gain hands-on sample preparation and data acquisition training. Metabolomics training is frequently provided by experts working in the field however, the ability to operate hands-on laboratory courses is somewhat limited and this type of metabolomics training are limited to a small number of centres and organisations; examples include the Imperial International Phenome Training Centre, Birmingham Metabolomics Training Centre, UC Davis West Coast Metabolomics Center, the Liverpool Training Centre for Metabolomics and Metabolomics South Africa (MSA). Providing access to state-of-the-art instrumentation is a challenge. Instruments in research laboratories are extensively used to generate data for multiple research projects and so are not always readily available to use in training courses. In addition, highly trained operators who have other commitments are required to set-up and run the hands-on training. The provision of infrastructure to provide lecture-based or computer workshops is easier to facilitate than laboratory space. The ideal scenario is to mirror the real-life metabolomics workflow and deliver face-to-face training via a combination of lectures, lab classes and workshops allowing attendees to gain practical experience. Providing dedicated time for questions and discussion helps the trainee to reinforce learning objectives and receive advice and feedback on specific problems. In the author’s experience, having time for trainees to discuss their research with the trainers so to provide advice and support is essential in any training course.
Course development and revision
Metabolomics, as with the other ‘omics is a rapidly developing field. A finding of the survey by Weber et al. (2015) reported that researchers wish to receive training in key areas of scientific development within the field and would like hands-on experience. Developing and updating courses that include hands-on training, laboratory or data workshops will require a greater time investment. In addition to methodology and skills-based activities, CPD-based training provides a platform to communicate and foster an understanding of key issues in the field and share experience from more experienced users. For example, the tools and techniques applied in metabolite annotation/identification have undergone rapid development over the last few years. Trainees benefit from understanding how to use the new approaches but to also understand and communicate the limitations and confidence in metabolite annotation/identification to the wider audience. Other key issues to share with new users are the importance of open data sharing and data repositories such as the ELIXIR resource MetaboLights (Haug et al., 2012) or Metabolomics Workbench (Sud et al., 2015), and an understanding of the minimum reporting standards conceived by the Metabolomics Standards Initiative (MSI) (Fiehn et al., 2007). The adoption of open data standards assists scientist to better share and re-use data. The FAIR (Findable, Accessible, Interoperable, Re-useable) guiding principles for data management and stewardship were developed to support both data producer and publisher. Development of metabolomics data repositories has provided the community with a new resource to reuse data in the development of bioinformatics and data analysis tools. Good research data management enhances the reusability and data integration and is an important area of training within the metabolomics community. Within the UK the EMBL-EBI metabolomics team promote the importance of open data sharing and data repositories through numerous training activities. The global success of these repositories will only be maintained by continued support across the metabolomics community to promote open science and data sharing. Encouraging existing and new users to dedicate time in submitting data using guiding principles such as FAIR and educating users in the tools developed to facilitate the process.
Course impact and evaluation
Although most training courses do not contribute towards a formal qualification it is desirable for attendees to receive a certificate of attendance or completion to record course attendance and is often required by the attendee to document their professional development and justify course fees. A European Credit Transfer System (ECTS) operates for students from the European Higher Education Area whereby they can gain credits towards qualification from attending short training courses. Accreditation of training courses by professional bodies provides official recognition that a course reaches a particular standard. It is a mechanism to demonstrate the value of the courses and is useful in course marketing. The Royal Society of Chemistry, Royal Society of Biology and Royal Society of Medicine in the UK are examples of professional bodies that accredit courses within the remit of metabolomics. For example, courses provided by the Imperial International Phenome Training Centre are accredited by the Royal Society of Medicine. The final consideration when performing any type of training is how to evaluate the success. Criteria may include the level of participation and popularity. Feedback surveys are often used but are provided at the end of the course, so feedback is usually focussed on course logistics rather than the longer-term impact of the training. Trainers can also inform training development and revision through their own observations, sometimes enabling real-time adjustments to be made. Finally, if formal assessments are included in the programme, they can also offer insight on gaps between curriculum intent and achievement against the learning outcomes. While deciding on adjustments to make, it is important to consider how the adjustments will enable achievements of the learning outcomes and what is practical within the context of time and resources.
Challenges and considerations specific to training courses
Trainees have different levels of metabolomics expertise
When undertaking any type of teaching, training or even a scientific presentation you must know your audience and consider the limitations of what can be achieved. To address the training needs of the scientific research community a range of courses from introductory, through intermediate to advanced levels are required. As discussed in the Principles of Pedagogical Best Practice section the starting point of any course development is the learning objectives. Introductory courses start from the beginning and have the objective to define and explain concepts to scientists who have not applied or have minimal exposure to metabolomics in their research. For example, a one-day lecture-based course which defines the basics of when metabolomics should and shouldn’t be applied, what metabolomics is, the experimental workflow, and appropriate considerations for the trainees to develop a core understanding of the topic is a good starting point. It is appropriate for those who are interested in applying metabolomics either in their own research groups or in core metabolomics groups and applies the two lower levels of Bloom’s cognitive domain (Fig. 2), ‘remember’ and ‘understand’. More specialised courses can be operated at intermediate to advanced levels addressing specific areas such as mass spectrometer maintenance and operation, data analysis or metabolite identification, and vocational based training to cater for specific groups of trainees generally interested in applying metabolomics in their own research groups. These courses look to develop learning at all levels of Bloom’s cognitive domain including ‘create’. Multiple learning opportunities are required to develop a complete understanding of the approach, be it via self-centred learning, learning from experts in 1:1 settings or more advanced CPD-based training. With practical components such an instrument operation, or data analysis hands-on testing of the approaches outside of the teaching environment in-between formalised training are essential to maximise the learning experience during the training courses and develop a deep understanding of the subject. In our (CW and WD) experience of operating metabolomics-based training courses since 2015 attendees benefit from the opportunity to ask questions months to years after attending the course. Vocational training is targeted towards the requirements of a specific group. For example, both clinical and environmental scientists benefit from dedicated sessions to address particular challenges they encounter when collecting their different types of samples in different environments (clinic, bench, remote field site). They may share the overarching challenge of how to quench or cool and transport samples during collection, but specific conditions (for example a clinic or a remote field site) will vary and influence procedures that may be applied. Training by the community is content focused and is typically delivered by a small number of centres or groups specialising in the field, at conference workshops or by instrument manufacturers. This specialised training can also be supplemented by other skills training to allow a researcher to develop the diverse range of skills required. For example, skills training such as software carpentry courses in the programming languages R or Python are a great resource and a valuable part of the training portfolio.
The number of trainees registered on a course
For efficiency and to minimise the number of training courses needed to be operated, each training course ideally would be available to many trainees. However, there are several reasons that influence the maximum number of trainees on a particular course. In laboratory-based training, where a demonstration of a technique or instrument use is required then having many people trying to view this process will result in some trainees not being able to view fully and learn appropriately. The authors (CW and WD) have found that a maximum number of four trainees per instrument is appropriate when viewing the instrument operating computer to allow each trainee good vision and the time to practice the process during the training session. This may vary for different training environments. Applying such limits will reduce the number of trainees than can attend each course; however, multiple different processes can be operated at the same time with different trainees working on one process and then moving on to the next process. In the authors experience (CW and WD) running face-to-face courses with a maximum of 8 trainees works well in relation to high-level visual training but also in group dynamics and interactions. For online courses a larger number of trainees can attend, and the authors (CW and WD) typically encourage more than 50 trainees for a scientific online training course. With online courses such as Massive Open Online Courses (MOOC), a key part of the learning experience is through the discussion forums and peer learning, it is therefore beneficial for large cohorts to enrol on the course to facilitate discussions. A high number of trainees may not be beneficial in all circumstances for example, if a live component is included in the training, then appropriate time or training support will be required to ensure all questions are answered during the session. Even with courses that are not operated in real time sufficient resources are required to answer questions and facilitate large cohort courses, this can include PhD students working in the field to work alongside the more experienced trainers.
Should education and training be provided face-to-face or online?
Whether to operate a course face-to-face or online is an important consideration and may be influenced by the type of training to be delivered. Different online approaches can be applied in the learning environment including webinars, distance learning as either live or recorded videos and MOOC’s that balance good content with context. The use of online approaches in the training environment has increased in popularity from 2011 with the widespread development of MOOC’s. Delivering online teaching rapidly increased in 2020 when due to necessity online distance learning was employed for most undergraduate and post-graduate courses. The rapid development of online tools to attend meetings, learn and network in addition to social media provides new mechanisms that have been embraced by the scientific community to develop global networks and can be applied to enhance the online training experience.
Online training can reach a larger audience and is a useful format although not appropriate for all steps in the metabolomics pipeline. The face-to-face format is more appropriate for laboratory-based activities where specialised facilities and equipment are required and results in a high level of interaction between trainers and trainees. Trainees often state that one of the reasons for attending training courses is to develop professional networks. Live discussions, question and answer panels and social events in face-to-face courses allows trainees to network with experts and build networks. In addition to laboratory processes, training in bioinformatics tools is frequently performed during face-to-face training but could also be provided via online videos with the potential to reach more trainees. With this approach, interaction and discussion must be facilitated through other mechanisms. In live online sessions, interaction can be synchronous with verbal discussions taking place at the whole group level or in virtual breakout rooms. Asynchronous online courses do not provide synchronous interaction observed in face-to-face courses. Alternative approaches can be employed to ask and answer questions, for example, the training material can be hosted on a suitable platform with discussion boards, comment sections and tasks to facilitate discussion between trainees.
Developing online courses require a greater time investment than producing face-to-face material. It is estimated that for every hour of training it takes 6 h to develop face-to-face training and 15 h to develop online training. However, online courses are accessible to a wider global audience, as attendees do not need to travel, and many online courses are uploaded to a learning platform so the user can complete the course in their own time frame. The intensity of the training can vary between the face-to-face and online format. Face-to-face training courses and some short online courses are operated over a small number of hours (for example, a scientific webinar) or over a small number of days with training for several hours each day. Online training courses allow for the training materials to be provided over a longer period (multiple weeks) with fewer hours per week required by the trainees and so can be accommodated into the routine working pattern rather than taking several days away from work.
The authors apply both synchronous and asynchronous in their course delivery and find blended approaches using online and face-to-face formats useful. Flip-style teaching where participants can view recorded videos of introductory material and concepts in advance of attending a tutorial or short face-to-face training course is useful when training participants from diverse scientific and knowledge backgrounds. A range of training materials can be provided in advance to allow all attendees to start for a common point of understanding. The face-to-face time can then be spent on learning outcomes at the higher level of Blooms cognitive domains. We provide two examples in the case study section to demonstrate the application of online training. One example demonstrates a blended course delivery employing both face-to-face and distance learning on a Master’s courses and the second example uses an online platform to deliver asynchronous training.
The costs of developing and operating training courses
Training courses, like any educational process, requires (dedicated) resources to develop and operate. These resources include one or more of the following: trainer time/salary, training room and laboratory space, laboratory consumables and scientific instrumentation, training materials (written or online), software, PCs/laptops, access to data servers and catering requirements. High quality training courses require trainers that are not only experts in metabolomics but also experts in course design and provision. The environment should be conducive to learning so with minimal background noise and availability of state-of-the art instrumentation and software. A cost recovery model is often the route to support training centres with a fee charged to attend courses, this may be offset for the trainee when bursaries are available from research council funding. Some courses can be operated for free, typically short courses operating up to a single day. Registration fees can be paid by attendees and is a common approach for laboratory-based or online courses. However, this can be limiting to early-career scientists for example where the finances are not available. A third option is for the trainers to apply to research councils for bursaries which can be applied to trainees who meet specific criteria. Costs of a training course are dependent on the resources required and length of course; face-to-face courses generally have higher registration fees compared to online courses which require fewer resources though costs related to the training platform applied must be considered.
Case studies
In this section, the authors provide examples from our own training and educational practice to illustrate how pedagogical best practice can be applied to metabolomics whilst considering practical challenges.
Developing face-to-face taught postgraduate university course to introduce the metabolomics workflow to students from diverse educational backgrounds
The School of Life Sciences at the Technical University Munich (Author MW) offers an elective course on metabolomics in the international Master of Science, Nutrition and Biomedicine course. Since metabolomics is only partially covered in other course components, this elective course aims to close this gap for interested students. This course covers all elements of the metabolomics workflow from study design, analytical technologies and data analysis and interpretation and aims at higher cognitive domains according to Bloom’s taxonomy. On completing the course students should be able to compare analytical methods used in metabolomics and evaluate the advantages and limitations of each method to solve a specific scientific question. Students design their own metabolomics experimental plans and evaluate their peers’ experimental plans. In order to enable students to develop the required skills this course is split into a one-hour lecture and two hours project work session each week.
In developing the course contextual challenges were considered due to the high heterogeneity in student background (international students from different countries with different bachelor studies and degrees). The diversity in the background of the students is taken as an advantage for this course. The entire course work is based on the concept of problem-based/problem-oriented learning. This type of learning is well suited for higher cognitive domains and allows students to develop independent and critical thinking. The course is structured into 3 different sections and different aspects of metabolomics are always discussed in the frame of a specific problem. Each section covers about 4–5 weeks. In the first section the lecturer uses the first problem to introduce basic concepts of metabolomics mostly on its own with partial input from the students. After this introduction, the second section is used to allow students to study a common problem and start to develop their own ideas for metabolomics workflows with partial input from the lecturer. Students work on different aspects in groups and present their results and solutions to the entire audience for open discussion. In the last section, groups of 2–3 students work on individual topics given by the lecturer. These topics also serve as an assessment for the students. It was not possible to offer a laboratory component in the course, this leads to difficulties in explaining specific parts of the workflow. At least laboratory visits of a running metabolomics laboratory are conducted to get an impression on the laboratory work. However, selected interested student are offered an internship or a master thesis position afterwards. Presentations and discussions throughout the lectures enable students to develop presentation skills, which is required for the assessment as well in future settings, e.g. during defence of a Master’s thesis or other project work. The assessment of this module is conducted in the form of project group work. It consists of an oral presentation of 5–7 min per person and submission of a maximum 6-page long research paper. The group selects from different possible topics introduced by the lecturer to the audience. The research paper is a method to measure the overall understanding of the stated problem and their ability to solve complex problems, analyze the current state-of-art and develop novel solutions. The oral presentation allows students to present their results to a wider auditory and subsequent discussion is a mean to measure their understanding of the scientific subject. Through the selected modes of assessment students can develop important skills such as literature search, extraction, and condensation of information.
Developing education and training to support metabolomics research-based programmes with minimal education in the metabolomics discipline
Metabolomics, as a multidisciplinary science, is not yet featured in the curricula of high educational institutions and universities in South Africa. Metabolomics research, in South African academia, is conducted at and through the Master’s and PhD levels. Thus, the education and training in metabolomics are part of these research-based programmes. The candidate (a postgraduate student) and his/her supervisor (PI or mentor) share responsibilities. To illustrate this, the case of metabolomics (part of the research performed in the Department of Biochemistry) at the University of Johannesburg (Author FT) is highlighted.
Modules such as advanced analytical techniques, taught at the BSc Honours level (a 1-year postgraduate degree, before a Master’s course) introduces students to a range of analytical techniques some of which are used in metabolomics studies, e.g. chromatographic techniques (LC and gas chromatography), mass spectrometry, NMR and other spectroscopic techniques. This module on advanced analytical techniques aims at cognitive domains of Bloom’s levels from understanding to evaluate. After the course, the student should demonstrate an advanced understanding of principles of different analytical techniques, and their applications and limitations. The course runs throughout the first semester and uses both teacher-centred and student-centred approaches. The module is designed to have a 2-h session each week of a lecture (theory and demonstration) and hands-on training on instruments (e.g., liquid/gas chromatography-mass spectrometry). Assessments are performed throughout the course via assignments, tests, exam (with theory and application questions) and a competence-based report on hands-on training tasks. The first year of the master’s (or/and PhD) programme is used to train and teach new concepts and aspects of the metabolomics workflow, from study/experimental design through to the data interpretation step. The training and teaching are carried out through various activities and approaches. The training aims to develop both cognitive and emotional domains of Bloom’s taxonomy levels. Assessments and progress evaluations are conducted through different forms, some of which include (i) regular progress meetings (one-on-one and in a group), (ii) group discussions, (iii) progress reports (twice a year), and (iv) a 15-min oral presentation to the Department (at least twice a year), during which the student presents the progress on his/her work, concepts learned and applied, and receives constructive feedback through a Q&A session. This training/teaching is carried out through different activities and tasks: studying the literature, continuous one-to-one and group discussions, and demonstrations.
The Metabolomics South Africa (MSA), an affiliate to the Metabolomics Society, organizes a series of workshops annually with both introductory and advanced workshops that cover different aspects in the metabolomics workflow. These workshops are presented by senior SA scientists involved in metabolomics research or international collaborators (mostly leading scientists in the field). Students are encouraged to attend and actively participate in these workshops. Some of these workshops provide hands-on training; for example, in data processing, data mining, spectral pre-processing and feature annotation/identification. The PI also invites overseas collaborators (experts in the field) to be involved in research projects (Master’s and/or PhD’s), and interact with the students, training them on particular aspects, for example, the use of computational tools for metabolite annotation via online platforms. Outside of the core teaching and training students are encouraged to attend international webinars (for example, those delivered by the Metabolomics Society Early Members Network). To assist students who are struggling in grasping certain concepts or practical aspects the PI assigns a ‘mentor’, these mentors are often senior students who have already demonstrated advanced knowledge and skills. Throughout the year, the PI pays attention to (each) individual student, and sets regular progress meetings, providing guidance and additional training/teaching where necessary. To improve presentation and communication skills students are provided with opportunities to attend and present at conferences (local and international, and both poster or oral presentation); and encouraged to develop and form networks with their peers or senior scientists in the field. At the end of their master’s and PhD programme, the student should have at least one research article published; and the student submits a dissertation (master’s) or a thesis (PhD), which is evaluated by external assessors (generally, experts in the field).
Delivering computational education in a taught postgraduate university course via face-to-face teaching and distance learning
Edith Cowan University (Author SR) offers a Master of Bioinformatics programme. This programme includes a course called Mass Spectrometry in Systems Biology that covers metabolomics and proteomics spectral pre-processing, feature annotation/identification, and quality assessment/data cleaning processes. This course has been offered yearly since 2020.
The learning outcomes are to (1) apply the theory of high-resolution mass spectrometry to metabolomics and proteomics by (2) critically analysing methods and guidelines on the above-mentioned tasks, (3) performing the above-mentioned tasks, and (4) defending decisions that informed performing the tasks. The learning outcome verbs align to Bloom’s levels 3–5 (apply, analyse, evaluate). The course was designed to accommodate students from diverse discipline backgrounds (including no prior chemistry knowledge) and is offered via a dual delivery pattern (both on-campus and online enrolment modes are offered). Students are required to use spectral processing software to achieve the learning outcomes. This, in turn, required computers with sufficient compute power, software to be installed, transfer of large raw spectral files, and the ability to accommodate off-campus software use. Microsoft Azure Labs was used to overcome computational challenges. A virtual machine (VM) template was created, all software was pre-installed, and all files were uploaded. The template was cloned to create an identical private VM for each student. Students were able to access their VM from anywhere and at any time, and they could save their work each week. Thermo Fisher Scientific’s Compound Discoverer was used for the spectral processing software. It reduced cognitive load by providing (1) a graphical-user interface (GUI), (2) visually intuitive presentation of concepts such as LC–MS peaks and MS/MS spectral matching, (3) an all-in-one software package to achieve all required content topics, and (4) the ability to process workflow steps in isolation.
To accommodate the two parallel study modes, the flipped classroom approach was used. Students watched a series of short (15-min) lecture recordings and completed required readings ahead of the 3-h weekly on-campus workshop. On-campus students participated in student-centred learning activities (such as Put Yourself on the Line mentioned earlier). Recordings were made of software demonstrations and relevant activities, then posted on the learning management system (LMS) for online students. Weekly discussion board activities were used to build inclusion between on-campus and online students, enable students to extend their own learning, and to promote critical thinking of metabolomics topics. Engagement in the discussion board activities was assessed as part of the portfolio assessment (see below) and offers on opportunity for formative feedback throughout the course.
Learning was assessed through two assessments. (1) Upon completing the content, students had an opportunity to explore and analyse how spectral processing, metabolite ID, and quality assessment/data cleaning are reported in the literature. In a 10-min oral presentation, students presented a published LC–MS metabolomics study, and critically assessed the reported workflow. This assessment piece was mapped against learning outcomes #1 and 2, demonstrating up to the ‘analysis’ level of Bloom’s taxonomy. (2) In a summative portfolio, students completed, presented, and reflected on their own spectral processing, metabolite ID, and quality assessment/data cleaning workflow. The portfolio mapped against learning outcomes #1, 3, and 4, demonstrating Bloom’s levels 2 – 5 (understand through evaluate).
In summary, contemporary research-focused educational criteria can pose unique design and delivery challenges. Research and business solutions can help to overcome these challenges and support learning.
The application of asynchronous online course delivery in the training environment
The Birmingham Metabolomics Training Centre (Authors CW, WD) developed the small private online course (SPOC) to provide access to online courses for 50–100 trainees per course run. From 2017 to June 2021 a total of 384 trainees participated in the courses. These courses were typically operated twice per year though were operated more frequently in 2020 and 2021 as the desire for online training increased during the Covid-19 epidemic.
Learning objectives were designed across all levels of Blooms cognitive domain to remember and understand knowledge and to apply and evaluate data processing/analysis methods that are applied in metabolomics. An asynchronous online format was used to deliver the course. This allowed a high number of trainees to attend the courses; typically, 30–100 trainees were present on each course. The course was delivered using a professional online course platform (FutureLearn) with multiple delivery tools. Each week focused on specific learning objectives and included a variety of delivery methods (short videos provided by the trainer, articles, exercises, polls and quizzes to allow the trainees to assess their knowledge). The training format applied both teacher-centred and student-centred learning. Teacher-centred activities included short video lectures whereas, student-centred tasks included workshops with step-by-step protocols to perform the analysis of data. Courses were operated over 3 or 4 weeks with an estimated learning times of 4 h per week. Background reading was provided for those who did not have the background knowledge of key concepts, however, to undertake all background reading would require a greater time commitment than 4 h per week. Trainees retained access to the course material after the course finished so could revisit the course material at any time. Links to papers published in the scientific literature were included to allow trainees to observe experimental processes and research performed by other experts in metabolomics.
The application of this asynchronous training method allows participants to complete the course when they have time available. It is advantageous to complete within the designated weeks to receive support from the trainers and fellow participants but is not essential. The low time commitment (4 h per week) also allows participants without the required background knowledge to undertake extra reading during the course. The use of technology enhanced the training experience, in particular discussion forums supported trainer-to-trainee and trainee-to-trainee interaction and sharing of knowledge. In addition, a one-hour live Q&A session was operated via the Zoom platform at a time most suitable for the trainees. Trainees were encouraged to post questions for the Q&A from Week 1 of the course. The most frequently asked questions and those deemed as important by the trainers were answered in an end-of-course video provided by one of the trainers. If all questions could not be answered during the live session, questions were answered on the discussion boards. The Q&A was also recorded and uploaded to the FutureLearn platform for those who could not join live or to re-visit the discussion. Self-assessment steps were included via quizzes or tests, if an incorrect answer was submitted participants were direct back to particular sections of the course to review the course material. Participants’ completion of the course tasks was logged in the online platform and a Certificate of Achievement was provided for those who completed the course.
In summary, the SPOC courses provided flexibility in the provision of training materials, flexibility in the time when trainees access the training materials, online support of trainees and the ability to train many trainees.
Summary
The field of metabolomics is complex, multidisciplinary, and not standardized. Accordingly, researchers working in metabolomics must possess in-depth cross-disciplinary content knowledge, a high level of subjective experience-based judgement knowledge, and the associated complementary skill sets. To develop this knowledge-base and skill sets necessitates appropriate training and educational initiatives within the metabolomics community. Initiatives must aim to be effective, but also feasible. As such, both pedagogical best practice and metabolomics-specific contextual challenges must be considered with equal weight. In this paper, we have presented consolidated pedagogical guidance for educators and trainers in metabolomics including considerations for addressing metabolomics-specific challenges. We provided case studies from our own practice to illustrate how this can be achieved. As metabolomics research grows in importance and diversity it is essential that best practice guidelines are put in place to ensure maximally effective learning for future metabolomics scientists. Although no community currently exists to discuss best practices and develop training materials, the authors recommend that a community focused on metabolomics teaching and training is developed and the Metabolomics Society should lead in such an effort.
Acknowledgements
The authors wish to thank Michelle Reid for her valuable feedback as an end-user. SNR wishes to thank Ralf Tautenhahn (Thermo Fisher Scientific) for licensing and technical support in using Compound Discoverer and Paul Haskell-Dowland (Edith Cowan University) for technical support in using Microsoft Azure Labs in her teaching practice. The authors wish to thank the Medical Research Council (MRC) in the UK for funding CW and WBD (MR/S010483/1).
Author contributions
All authors contributed to conceptualising and writing the manuscript. All authors read and approved the final manuscript.
Data availability
Data sharing not applicable to this article as no datasets were generated or analysed during the current study.
Declarations
Conflict of interest
The authors have no disclosures of potential conflicts of interest related to the presented work.
Research involving human and/or animal participants
No research involving human participants or animals was performed for this work.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Anderson LW Krathwohl DR Bloom BS A taxonomy for learning, teaching, and assessing : A revision of Bloom's taxonomy of educational objectives 2001 Abridged Longman
Beger RD Dunn W Schmidt MA Gross SS Kirwan JA Cascante M Brennan L Wishart DS Oresic M Hankemeier T Broadhurst DI Lane AN Suhre K Kastenmuller G Sumner SJ Thiele I Fiehn O Kaddurah-Daouk R for "Precision Medicine and Pharmacometabolomics Task Group"-Metabolomics Society Initiative Metabolomics enables precision medicine: “A White Paper, Community Perspective” Metabolomics 2016 12 149 10.1007/s11306-016-1094-6 27642271
Biggs J Enhancing teaching through constructive alignment Higher Education 1996 32 347 364 10.1007/BF00138871
Bloom BS Bloom BS Taxonomy of educational objectives: The classification of educational goals Handbook 1, Cognitive domain; by a committee of college and university examiners 1956 1 London Longman Group
Boyce MC Lawler NG Tu Y Reinke SN Introducing undergraduate students to metabolomics using liquid chromatography-high resolution mass spectrometry analysis of horse blood Journal of Chemical Education 2019 96 745 750 10.1021/acs.jchemed.8b00625
Brennan L The nutritional metabolomics crossroads: How to ensure success for dietary biomarkers American Journal of Clinical Nutrition 2017 105 293 294 10.3945/ajcn.116.150847 28100510
Bundy JG Davey MP Viant MR Environmental metabolomics: A critical review and future perspectives Metabolomics 2008 5 3 10.1007/s11306-008-0152-0
Fiehn O Robertson D Griffin J van der Werf M Nikolau B Morrison N Sumner LW Goodacre R Hardy NW Taylor C Fostel J Kristal B Kaddurah-Daouk R Mendes P van Ommen B Lindon JC Sansone S-A The metabolomics standards initiative (MSI) Metabolomics 2007 3 175 178 10.1007/s11306-007-0070-6
FindAMasters. (2022). https://www.findamasters.com. Accessed 1 January 2022
Fraser PD Aharoni A Hall RD Huang S Giovannoni JJ Sonnewald U Fernie AR Metabolomics should be deployed in the identification and characterization of gene-edited crops The Plant Journal 2020 102 897 902 10.1111/tpj.14679 31923321
Haug K Salek RM Conesa P Hastings J de Matos P Rijnbeek M Mahendraker T Williams M Neumann S Rocca-Serra P Maguire E González-Beltrán A Sansone S-A Griffin JL Steinbeck C MetaboLights—An open-access general-purpose repository for metabolomics studies and associated meta-data Nucleic Acids Research 2012 41 D781 D786 10.1093/nar/gks1004 23109552
Hollywood KA Schmidt K Takano E Breitling R Metabolomics tools for the synthetic biology of natural products Current Opinion in Biotechnology 2018 54 114 120 10.1016/j.copbio.2018.02.015 29567581
Hoque M Three domains of learning: Cognitive, affective and psychomotor The Journal of EFL Education and Research 2017 2 45 51
Jarvis P Adult education and lifelong learning : Theory and practice 2012 Taylor and Francis
Kell DB Oliver SG The metabolome 18 years on: A concept comes of age Metabolomics 2016 12 148 10.1007/s11306-016-1108-4 27695392
Krathwohl DR Bloom BS Bertram BM Taxonomy of educational objectives: The Classification of educational goals. Handbook II: Affective domain 1964 Longmans
La Frano MR Amin S Fanter RK Participation of undergraduate students in a controlled feeding study with metabolomics analysis to enhance learning of metabolism Journal of Chemical Education 2020 97 1595 1603 10.1021/acs.jchemed.0c00028
Mashego MR Rumbold K De Mey M Vandamme E Soetaert W Heijnen JJ Microbial metabolomics: Past, present and future methodologies Biotechnology Letters 2007 29 1 16 10.1007/s10529-006-9218-0 17091378
Mendez KM Pritchard L Reinke SN Broadhurst DI Toward collaborative open data science in metabolomics using Jupyter Notebooks and cloud computing Metabolomics 2019 15 125 10.1007/s11306-019-1588-0 31522294
Mulder N Schwartz R Brazas MD Brooksbank C Gaeta B Morgan SL Pauley MA Rosenwald A Rustici G Sierk M Warnow T Welch L The development and application of bioinformatics core competencies to improve bioinformatics training and education PLoS Computational Biology 2018 14 e1005772 10.1371/journal.pcbi.1005772 29390004
Oliver SG Winson MK Kell DB Baganz F Systematic functional analysis of the yeast genome Trends in Biotechnology 1998 16 373 378 10.1016/S0167-7799(98)01214-1 9744112
Print M Curriculum development and design 1993 2 Allen & Unwin
Rai A Yamazaki M Saito K A new era in plant functional genomics Current Opinion in Systems Biology 2019 15 58 67 10.1016/j.coisb.2019.03.005
Rattray NJW Deziel NC Wallach JD Khan SA Vasiliou V Ioannidis JPA Johnson CH Beyond genomics: Understanding exposotypes through metabolomics Human Genomics 2018 12 4 10.1186/s40246-018-0134-x 29373992
Sandusky PO Introducing undergraduate students to metabolomics using a NMR-based analysis of coffee beans Journal of Chemical Education 2017 94 1324 1328 10.1021/acs.jchemed.6b00559
Sud M Fahy E Cotter D Azam K Vadivelu I Burant C Edison A Fiehn O Higashi R Nair KS Sumner S Subramaniam S Metabolomics Workbench: An international repository for metabolomics data and metadata, metabolite standards, protocols, tutorials and training, and analysis tools Nucleic Acids Research 2015 44 D463 D470 10.1093/nar/gkv1042 26467476
Sweller J Cognitive load during problem solving: Effects on learning Cognitive Science 1988 12 257 285 10.1207/s15516709cog1202_4
Tinte MM Chele KH van der Hooft JJJ Tugizimana F Metabolomics-guided elucidation of plant abiotic stress responses in the 4IR Era: An overview Metabolites 2021 11 445 10.3390/metabo11070445 34357339
Wang M Carver JJ Phelan VV Sanchez LM Garg N Peng Y Nguyen DD Watrous J Kapono CA Luzzatto-Knaan T Porto C Bouslimani A Melnik AV Meehan MJ Liu WT Crusemann M Boudreau PD Esquenazi E Sandoval-Calderon M Kersten RD Pace LA Quinn RA Duncan KR Hsu CC Floros DJ Gavilan RG Kleigrewe K Northen T Dutton RJ Parrot D Carlson EE Aigle B Michelsen CF Jelsbak L Sohlenkamp C Pevzner P Edlund A McLean J Piel J Murphy BT Gerwick L Liaw CC Yang YL Humpf HU Maansson M Keyzers RA Sims AC Johnson AR Sidebottom AM Sedio BE Klitgaard A Bandeira N Sharing and community curation of mass spectrometry data with Global Natural Products Social Molecular Networking Nature Biotechnology 2016 34 828 837 10.1038/nbt.3597
Weber RJ Winder CL Larcombe LD Dunn WB Viant MR Training needs in metabolomics Metabolomics 2015 11 784 786 10.1007/s11306-015-0815-6 26109924
Zanetti KA Hall RD Griffin JL Putri S Salek RM Styczynski MP Tugizimana F van der Hooft JJJ The metabolomics society-current state of the membership and future directions Metabolites 2019 9 89 10.3390/metabo9050089 31058861
| 36512139 | PMC9746579 | NO-CC CODE | 2022-12-15 23:21:56 | no | Metabolomics. 2022 Dec 13; 18(12):106 | utf-8 | Metabolomics | 2,022 | 10.1007/s11306-022-01957-w | oa_other |
==== Front
Chirurgie (Heidelb)
Chirurgie (Heidelb)
Chirurgie (Heidelberg, Germany)
2731-6971
2731-698X
Springer Medizin Heidelberg
1788
10.1007/s00104-022-01788-4
Originalien
Die Auswirkungen der COVID-19 Pandemie führten zu signifikanten Veränderungen der Operations- und Liegezeiten bei Patienten nach Cholezystektomie
The COVID-19 pandemic had significant impact on duration of surgery and hospitalization time for patients after cholecystectomyFischer L. [email protected]
1
Iber T. [email protected]
2
Feißt M. [email protected]
3
Huck B. [email protected]
1
Kolb G. [email protected]
1
Huber B. [email protected]
1
Segendorf C. [email protected]
1
Fischer E. [email protected]
1
Halavach K. [email protected]
1
1 grid.506801.a 0000 0004 0411 7927 Abteilung für Allgemein‑, Viszeral- und Metabolische Chirurgie, Klinikum Mittelbaden, Balger-Str. 50, 76532 Baden-Baden, Deutschland
2 grid.506801.a 0000 0004 0411 7927 Abteilung für Anästhesie und Intensivmedizin, Klinikum Mittelbaden, Balger-Str. 50, 76532 Baden-Baden, Deutschland
3 grid.5253.1 0000 0001 0328 4908 Institut für Medizinische Biometrie, Universitätsklinikum Heidelberg, Im Neuenheimer Feld 130.3, 69121 Heidelberg, Deutschland
13 12 2022
16
29 11 2022
© The Author(s), under exclusive licence to Springer Medizin Verlag GmbH, ein Teil von Springer Nature 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Hintergrund
Die COVID-19 Pandemie (Pandemie) hat erhebliche Veränderungen der medizinischen Versorgung notwendig gemacht. Ziele dieser Studie waren herauszufinden, welchen Einfluss die Pandemie auf den perioperativen Verlauf bei Patienten mit Cholezystektomie (CHE) hatte und mögliche bleibende Schlussfolgerungen aufzuzeigen.
Methodik
Vom 01.07.2018 bis 31.12.2021 wurden 735 Patienten mit CHE analysiert. Dabei wurden die Patienten bis zum 21.03.2020 als reguläre Patientengruppe (Reg, n = 430), die Patienten danach (1. Lockdown 22.03.2020) als Cov19-Patientengruppe (Cov19, n = 305) bezeichnet und miteinander verglichen.
Ergebnisse
Das Durchschnittsalter aller Patienten betrug 59 Jahre, 63 % der Patienten waren Frauen. Die durchschnittliche Krankenhausverweildauer (KrVD, Zeitraum zwischen Operation und Entlassung) betrug 4,4 Tage. Die Patientengruppen Reg und Cov19 unterschieden sich nicht bez. Alter, Geschlecht oder KrVD. Die Gesamtzahl der durchgeführten CHE hat sich durch Cov19 um 21,4 % reduziert. Dies betraf gleichermaßen elektive sowie auch notfallmäßige CHE. Die Operationsdauer stieg in der Gruppe Cov19 signifikant von 64 min (SD 34 min) auf 71 min (SD 38 min) an (p < 0,05). Die Anzahl der Kurz- und Langlieger (KrVD 2 bzw. > 4 Tage) stieg in der Cov19-Gruppe signifikant von 4 % auf 20 % (Kurzlieger, p < 0,01) bzw. von 23 % auf 27 % (Langlieger, p < 0,01) an. Dies war vor allem bei über 70-jährigen Patienten mit einem Anstieg der Langlieger von 43 % auf 56 % in der Cov19-Gruppe zu beobachten.
Diskussion
Die COVID-19-Pandemie führte zu einer deutlichen Abnahme der CHE sowohl bei Elektiv- als auch bei Notfalleingriffen. Des Weiteren konnte eine signifikante Verlängerung der Operations- und Liegedauer bei älteren Patienten beobachtet werden. Bleibende Konsequenzen der Pandemie könnten Liegezeitverkürzungen nach unkomplizierten CHE und mehr interventionelle Therapieverfahren in komplexen Fällen sein.
Background
The COVID-19 pandemic made substantial changes in medical care necessary. The aims of this study were to find out what influence the pandemic had on the perioperative course in patients with cholecystectomy (CHE) and to highlight possible residual consequences.
Method
From 1 July 2018 to 31 December 2021 a total of 735 patients with CHE were analyzed. Up to 21 March 2020 patients were assigned to the regular patient group (Reg, n = 430), patients after this date (first lockdown 22 March 2020) to the Cov19 patient group (Cov19, n = 305) and the 2 groups were compared.
Results
The average age of all patients was 59 years and 63% were women. The average length of hospitalization (KrVD, time period between surgery and discharge) was 4.4 days. The patient groups Reg and Cov19 did not differ with respect to age, gender or KrVD. The total number of CHEs carried out was reduced by 21.4% in the Cov19 group. This affected elective and emergency CHE to the same extent. The length of surgery significantly increased in the Cov19 group from 64 min (SD 34 min) to 71 min (SD 38 min). The number of short and long hospital stays (KrVD 2 or >4 days) significantly increased in the Cov19 group from 4 % to 20 % (short stay, p < 0.01) and from 23 % to 27 % (long stay, p < 0.01). This was particularly observed for patients >70 years old with an increase in long stays from 43 % to 56 % in the Cov19 group.
Conclusion
The COVID-19 pandemic led to a clear reduction in CHE both for elective and emergency interventions. Furthermore, a significant lengthening of the surgery and hospitalization times could be observed for older patients. The residual consequences of the pandemic could be shortened hospitalization times after uncomplicated CHE and more interventional treatment procedures in complex cases.
Schlüsselwörter
COVID-19
Pandemie
Operation
Cholezystektomie
Liegezeit
Keywords
COVID-19
Pandemic
Surgery
Cholecystectomy
Hospitalization time
==== Body
pmcDie COVID-19-Pandemie (Pandemie) hat einen erheblichen Einfluss auf die medizinische Versorgung weltweit. Die Pandemie hat auch im Mikrokosmos Krankenhaus deutlich gemacht, wie abhängig die einzelnen Abteilungen voneinander sind. Die schiere Anzahl an COVID-19 erkrankten Patienten und COVID-19-bedingten Personalausfällen haben zu erheblichen Leistungsverschiebungen in der Chirurgie bis hin zum kompletten Stopp elektiver Operationen geführt. Diese retrospektive Studie untersucht den Einfluss der COVID-19-Pandemie auf die Operationsindikation und den postoperativen Verlauf bei Patienten mit Notwendigkeit zur Gallenblasenentfernung.
Hintergrund
Die COVID-19 Pandemie stellt für unsere Gesellschaft eine bisher einmalige Herausforderung dar, deren gesamtgesellschaftliches Ausmaß noch nicht absehbar ist [1, 2]. Die Auswirkungen der Pandemie betreffen nicht nur die onkologische und Notfallversorgung [1, 3], sondern auch elektive Therapieverfahren der konservativen und operativen Fächer in verschiedenen Schweregraden [4, 5]. Die Auswirkungen in den chirurgischen Abteilungen waren nicht nur infrastrukturell bedingt (COVID-bedingte Reduktion der Kapazitäten im Operationssaal, auf Intensiv- und Normalstation), sondern hatten auch personelle Konsequenzen. So kam es in unserer Klinik durch starke Belegung von COVID-Patienten und/oder durch COVID-bedingte Personalausfälle immer wieder zu Schließungen ganzer (chirurgischer) Stationen, Stationszusammenlegungen, Umzügen ganzer Stationen und Entsendung chirurgischer Ärzte in fachfremde Abteilungen. Dies alles führte nicht nur bei uns teilweise zu einem kompletten Erlahmen des elektiven Operationsprogramms und zu Verzögerungen selbst bei onkologischen Operationen [1, 3–6].
Die laparoskopische Gallenblasenentfernung (lap. CHE) ist einer der häufigsten Eingriffe weltweit und in Deutschland [6, 7]. Die Indikation für die Entfernung der Gallenblase deckt viele Facetten vom reinen Elektiveingriff bis hin zum akuten Notfalleingriff ab. Die nationalen und internationalen Empfehlungen für die Entfernung der Gallenblase während der COVID-19-Pandemie waren differenziert und rieten von einem pauschalen Verbot der CHE ab [8, 9]. Konservative Therapieoptionen mittels Antibiose wurden gehäuft angewendet. Bei erwartbar schweren Operationen rückten interventionelle Therapieverfahren mehr in den Vordergrund [10, 11]. Es gibt eine Anzahl an Publikationen [5, 6, 12], die den Einfluss der COVID-19-Pandemie auf die operativen Fächer, speziell auch auf die Gallenblasenchirurgie untersucht haben. Die Ergebnisse sind recht unterschiedlich. Steffanie et al. zeigten [12], dass die Zahl der Cholezystektomien während der COVID-19-Pandemie um 30 % abfiel, in einer anderen Arbeit punktuell sogar um 50 % [6]. Die Operations‑, Krankenhausverweildauer und der klinische Verlauf der Patienten unterschieden sich in dieser Analyse nicht signifikant zu den Vorjahreszeiträumen. Rahimi et al. [13] konnten an 182 Patienten präsentieren, dass die histologische Untersuchung von Gallenblasen bei älteren Patienten während der COVID-19-Pandemie mehr akute und gangränöse Befunde aufwiesen. Dass der Schweregrad der Erkrankung im Rahmen der COVID-19-Pandemie angestiegen ist, wird von Koch et al. [6] unterstützt. Hier wurde eine vermehrte Anzahl an Konversion, ein längerer stationärer Aufenthalt, eine erhöhte Komplikationsrate und Mortalität an 8561 Patienten mit CHE beschrieben.
Das Ziel dieser retrospektiven Untersuchung war es, zu zeigen, welchen Einfluss die COVID-19-Pandemie auf Operationsindikation und den perioperativen Verlauf bei Patienten mit Notwendigkeit zur Gallenblasenoperation hatte. Des Weiteren wurden im Kontext der COVID-Pandemie Veränderungen im perioperativen Verlauf analysiert, die langfristig Bestand haben könnten.
Methodik
Im Rahmen dieser retrospektiven Studie wurden die Operationszahlen für die operative Entfernung der Gallenblase (CHE) vom 01.01.2018 bis zum 31.12.2021 gezählt. Eine detaillierte Analyse für Patienten mit CHE fand für den Zeitraum vom 01.07.2018 bis 31.12.2021 statt. Diese schloss folgende Parameter ein: Alter, Geschlecht, Operationsverfahren (offene Operation, laparoskopische Operation, Konversionsoperation), Durchführung einer ERCP (präoperativ), Operationszeitpunkt (Schnittbeginn im Tagesprogramm oder außerhalb), Operationsindikation (elektiv, Notfall), Krankenhausverweildauer (Zeitraum zwischen Aufnahme- bzw. Operationstag und Entlassung) und postoperative Komplikationen. Alle Patienten, die nicht über die reguläre Sprechstunde zur Operation geplant, Patienten, die im Dienst operiert wurden oder Patienten, die bei Versagen der konservativen Therapie operiert werden mussten, wurden als Notfalloperationen deklariert. Die Analyse zwischen den Gruppen nutzt den 22.03.2020 als Beginn des 1. Lockdowns als Schnittpunkt. Die Patienten davor werden als reguläre Patientengruppe (Reg), die danach als COVID-19-Patientengruppe (Cov19) bezeichnet. Im Rahmen der elektronischen Dokumentation wurden die Patienten pseudonymisiert erfasst, auf die elektronische Erfassung von Name und Geburtsdatum zu Auswertungszwecken wurde verzichtet. Bei der Krankenhausverweildauer wurden Aufnahme und Entlasstag als je 1 Tag gewertet. Der Aufnahmetag war bei den meisten elektiven Patienten, aber auch bei vielen Notfallpatienten in der Regel auch der Operationstag. Falls die stationäre Aufnahme vor der Operation stattfand, wurde die Krankenhausverweildauer als Zeitraum zwischen Operations- und Entlassungstag definiert.
Die Operationsindikationen wurden entsprechend der Leitlinie bzw. den aktuellen Empfehlungen umgesetzt. Dies schloss das komplette Stoppen elektiver CHE ein. Patienten mit akutem Abdomen konnten unabhängig von der Ätiologie immer zeitnah operiert werden. Ebenso wurden Patienten mit Cholezystitis, bei denen die konservative Therapie nicht anschlug, operiert. Wann immer möglich fanden die Operationen nach negativer COVID-Testung statt.
Statistik
Stetige Parameter wurden als Mittelwert mit Standardabweichung (SD) angegeben, kategoriale Parameter wurden in absoluten und relativen Häufigkeiten beschrieben. Um Unterschiede zwischen verschiedenen Subgruppen zu untersuchen, wurden entweder der χ2-Test nach Pearson oder der Zweistichproben t‑Test angewendet. Als statistisch signifikant wurden in dieser explorativen und retrospektiven Studie p-Werte < 0,05 bewertet. Die statistische Analyse wurde mit dem Programm Microsoft Excel 2016 (Microsoft Corporation, Redmond, USA) und der statistischen Programmiersprache R durchgeführt.
Ergebnisse
Vom 01.07.2018 bis 31.12.2021 wurden 735 Patienten mit Gallenblasenoperationen (CHE) erfasst. Dabei wurden Patienten, die vom 01.07.2018 bis 21.03.2020 operiert wurden, als reguläre Patientengruppe (Reg) definiert (n = 430 Patienten). Patienten, die zwischen dem 22.03.2020 (Beginn 1. Lockdown) und dem 31.12.2021 operiert worden sind, wurden als COVID-19-Patientengruppe (Cov19) definiert (n = 305 Patienten). Die Patienten beider Gruppen (Reg vs. Cov19) unterschieden sich nicht signifikant bez. Alter (Gesamt 59 Jahre, SD 17 Jahre; Gruppe Reg 60 Jahre und Gruppe Cov19 58 Jahre, SD jeweils 17 Jahre) und Geschlecht (Gesamt: 460 Frauen, 275 Männer; Gruppen Reg und Cov19 mit jeweils 63 % Frauen). Bei 3 Patienten erfolgte eine Nachresektion bei nachgewiesenem Gallenblasenkarzinom.
Die Auswirkungen der COVID-19-Pandemie führten zu einer Reduktion der CHE um 21,4 % (Abb. 1). Die quartalsmäßige Gesamtzahl der CHE zusammen mit der prozentualen Häufigkeit der notfallmäßigen CHE ist in Abb. 2 dargestellt. Der 1. Lockdown ab 22.03.2020, vor allem aber die Einschränkungen des öffentlichen Lebens ab November 2020 (Lockdown light) mit dem 2. Corona-Lockdown ab 06.01.2021, führten zu erheblichen und längerfristigen Einschränkungen der Operationen. Erst ab dem 3. Quartal 2021 zeigt sich eine beginnende Normalisierung der operativen Versorgung bei Gallenblasenerkrankungen.
Die Analyse der Gruppen Reg vs. Cov19 zeigte bez. Alter, Geschlecht, Operationsindikation (elektiv vs. Notfall), Operationsverfahren (laparoskopische Operationen – Konversionsoperation – primär offene Operation), präoperative ERCP-Häufigkeit, Krankenhausverweildauer (KrVD) und postoperative Komplikationen keine signifikanten Unterschiede (p > 0,05; Daten nicht gezeigt). Lediglich die Operationsdauer stieg in der Gruppe Cov19 signifikant von 64 min (SD 34 min) auf 71 min (SD 38 min) an (p = 0,01). Dies ist vor allem auf die verlängerte Operationsdauer bei Frauen zurückzuführen (Anstieg von 58 min, SD 25 auf 67 min, SD 35, p < 0,01). Bei den männlichen Patienten konnte ein nicht signifikanter Anstieg der Operationsdauer (Anstieg von 73 min, SD 43 min auf 77 min, SD 42 min, p = 0,46) festgestellt werden.
Die Analyse zeigte keinen signifikanten Unterschied in der Krankenhausverweildauer zwischen den Patientengruppen Reg und Cov (KrVD, 4,3 Tage, 3,6 SD vs. 4,6 Tage, 4,8 SD, p = 0,359). Zusätzlich wurde eine Kategorisierung der Patienten in Kurzlieger (KrVD = 2 Tage), Mittellieger (KrVD = 3–4 Tage) und Langlieger (KrVD > 4 Tage) vorgenommen. Hier zeigt sich ein signifikanter Unterschied in den Verteilungen der Kategorisierung (p < 0,001). So stieg bspw. der Anteil der Kurzlieger für Patientinnen und Patienten jeweils signifikant an (Frauen: 4 % auf 22 %, Männer 4 % auf 16 %). Dieser Anstieg betrifft vor allem die Patienten, die bisher bei komplikationsloser CHE 3 bis 4 Tage im Krankenhaus lagen. Hier zeigte sich in der Gruppe Cov19 eine signifikante Reduktion im Vergleich zur Gruppe Reg (Tab. 1). Gruppe Reg Gruppe Cov19
Zeitraum 01.07.2018 bis 21.03.2020 22.03.2020 bis 31.12.2021
Gruppengröße n = 430 n = 305
KrVDa, b 4,3 Tage (SD 3,6) 4,6 Tage (SD 4,8)
KrVDb 2 Tage 3–4 Tage > 4 Tage 2 Tage 3–4 Tage > 4 Tage
Alle Patienten
n 18 (4 %) 315 (73 %) 97 (23 %) 60 (20 %) 163 (54 %) 81 (27 %)
Männer
n 7 (4 %) 102 (63 %) 52 (32 %) 18 (16 %) 52 (46 %) 43 (38 %)
Frauen
n 11 (4 %) 213 (79 %) 45 (17 %) 42 (22 %) 111 (58 %) 38 (20 %)
Reg Patientengruppe vor COVID vom 01.07.2018 bis 21.03.2020, Cov19 Patientengruppe während COVID-19-Pandemie vom 22.03.2020–31.12.2021, KrVD Krankenhausverweildauer (Intervall zwischen Operation und Entlassung), SD Standardabweichung
aAngabe als Mittelwert
bAufnahme- bzw. Operationstag und Entlassungstag wurden als jeweils 1 Tag gezählt
Die Anzahl der Langlieger (mehr als 4 Tage KrVD nach CHE, Tab. 1) hingegen stieg in der Cov19-Gruppe von 23 % auf 27 % an. Dies war vor allem in der Gruppe der über 70-jährigen Patienten mit einem Anstieg der Langlieger von 43 % auf 56 % in der Cov19-Gruppe zu beobachten. In der Gruppe der bis 50-jährigen Patienten bzw. der 50- bis 70-jährigen Patienten betrug dieser Anstieg zwischen den Gruppen Reg und Cov19 lediglich 2 % bzw. 3 %.
Diskussion
Die COVID-19-Pandemie hat zu erheblichen Veränderungen auch im Gesundheitswesen geführt [1]. Obwohl die akute Situation mit den extrem virulenten und krankheitserregenden ersten COVID-Varianten überstanden zu sein scheint, sind die gesamtgesellschaftlichen langfristigen Konsequenzen noch nicht absehbar [1]. Das deutsche Gesundheitssystem wurde in den Infektionsspitzen der COVID-19-Pandemie an den Rand seiner Funktionsfähigkeit getrieben [1]. Aktuell hat sich die Situation erfreulicherweise stabilisiert, die derzeitigen Regeln sind gut verständlich und werden weiter bedarfsadaptiert angepasst [1, 2, 6, 14, 15]. Eine der Regeln [1, 9, 16] besagt, dass das Screening für COVID-19 vor einer Operation sinnvollerweise durchzuführen ist. Zumindest für die früheren COVID-19-Varianten konnte gezeigt werden [16], dass eine asymptomatische COVID-19-Infektion mit einem 2fach erhöhten Mortalitätsrisiko, die symptomatische COVID-19-Infektion mit einem 10fach erhöhten Mortalitätsrisiko einhergeht. Ob das für die nachfolgenden Virusvarianten noch gilt, bleibt abzuwarten. Aber die Operations- und Intensivbettenkapazität sind teilweise weiterhin erheblich eingeschränkt. Eine suffiziente Aufklärung von Patienten über Operationsverschiebungen oder sogar notwendige Verlegungen von Patienten bei dringlicher Operationsindikation und nicht absehbarer Operationsmöglichkeit (je nach Klinik des Patienten als Notfalloperation oder Operation innerhalb von 12 oder 24 h) oder bei komplexen Fällen mit fehlender Intensivbettenkapazität sollte in derartigen Situationen bedacht werden [21].
Bei Erkrankungen der Gallenblase bildet das Spektrum der Therapieoptionen konservative Verfahren wie Antibiose bis hin zur Operation ab [22]. Die Indikation zur Operation reicht dabei vom reinen Elektiveingriff bis hin zum Notfalleingriff z. B. bei peritonitischem Abdomen. Aufgrund seiner Häufigkeit ist die CHE besonders geeignet, die Auswirkungen der COVID-19-Pandemie zu untersuchen. Die hier gezeigten Resultate erlauben drei wesentliche Schlussfolgerungen:Die COVID-19-Pandemie mit dem 1. und vor allem mit dem 2. Lockdown hat zu einer erheblichen Reduktion der Gallenblasenoperation geführt. Erst mit Beginn des 2. Quartals 2021 trat zumindest für die Gallenblasenoperation eine leichte Erholung ein. Dies hat seine wahrscheinliche Ursache in der niedrigeren Erkrankungsschwere der COVID-19-Omikron-Variante.
Die COVID-19-bedingte Ressourcenknappheit (Personal, Betten, Operationskapazität) hat zu einer signifikanten Steigerung der Kurzlieger nach CHE geführt. Kurzlieger sind Patienten, die bereits am 1. postoperativen Tag nach CHE entlassen wurden. Dass die frühe Entlassung von Patienten nach CHE nicht zu einer erhöhten Komplikationsrate führt, war nicht nur in dieser Untersuchung zu beobachten, sondern konnte auch in anderen Studien gezeigt werden [17, 18]. Teilweise wurden sogar ambulante Gallenblasenoperationen verstärkt durchgeführt, um Patienten mit Notwendigkeit zur CHE die Operation während der COVID-19-Pandemie zu ermöglichen [19].
Vor allem ältere Patienten mit CHE schienen mehr von den Verzögerungen, die im Rahmen der Diagnostik und Therapie während der COVID-19-Pandemie auftraten, betroffen zu sein. Die hier präsentierten Daten der Cov19-Gruppe suggerieren, dass bei diesen Patienten die technisch offenbar schwierigeren CHE zu längeren Operationszeiten und längeren Krankenhausverweildauer (Dauer zwischen Operations- und Entlassungstag) geführt haben. So stieg die Anzahl der Patienten, die mehr als 4 Tage nach CHE im Krankenhaus bleiben mussten, bei den über 70-jährigen Patienten in der Gruppe Cov19 signifikant an. Ein Anstieg der Konversion oder der postoperativen Komplikationen in der Gruppe Cov19 konnte, im Gegensatz zu anderen Publikationen [6], nicht beobachtet werden. Im Gegenteil, die postoperative Komplikationsrate fiel in der Gruppe Cov19 eher ab (Daten nicht gezeigt). Die Beobachtung, dass auch ältere Patienten während der COVID-19-Pandemie öfter mit schweren Verläufen zur Operation kamen, wurde in anderen Untersuchungen bestätigt [6, 13].
Auch in der Viszeralchirurgie hat die Pandemie ihre Spuren hinterlassen, die bis heute nachwirken. Die starke Belastung der Krankenhäuser mit COVID-19-Patienten und COVID-19-bedingte Personalengpässe haben zu signifikantem und andauerndem Abbau der Bettenverfügbarkeit auf Normal- und Intensivstation und von Operationsressourcen geführt. Es ist sicher nicht übertrieben, zu sagen, dass durch die COVID-19-Pandemie einige chirurgische Standards in ihren Grundfesten erschüttert wurden. Die Verstärkung konservativer Therapieverfahren bei der akuten Appendizitis und der akuten Cholezystitis sind als Beispiel zu nennen; die Ergebnisse sind für Patienten und Chirurgen mit gemischten Erfolgen verbunden [3, 5, 10, 19, 20]. Einige dieser Veränderungen werden sich wahrscheinlich dennoch durchsetzen und sich ggf. sogar in den Leitlinien [22] oder den ggf. folgenden angepassten Verordnungen zur Pandemieversorgung niederschlagen [1]. Das viszeralchirurgische Team am Klinikum Baden-Baden hält z. B. bei der unkomplizierten laparoskopischen Cholezystektomie unter Beachtung wesentlicher perioperativer Parameter [7, 18] weiterhin an den verkürzten postoperativen Liegezeiten fest; auch wenn diese teilweise unterhalb der unteren Grenzverweildauer liegen. Auch außerhalb der COVID-Pandemie haben sich die in wenigen Fällen notwendigen interventionellen Verfahren (anstelle einer eher technisch aufwendigen und riskanteren Operation bei älteren und komorbiden Patienten) bewährt. Diese Faktoren sind, neben dem Wegfallen des Händeschüttelns zumindest in unserem Krankenhaus, einige der wenigen positiven Effekte, die sich durch die COVID-19-Pandemie manifestiert haben.
Fazit für die Praxis
Die Auswirkungen der COVID-19-Pandemie auf die operative Versorgung von Patienten mit Gallenblasenerkrankungen waren relevant und sind persistierend. Vor allem ältere Patienten scheinen unter den Folgen der pandemiebedingten Verzögerungen in Diagnostik und Therapie zu leiden. Dies drückt sich in einer verlängerten Operationszeit und in einer Zunahme der Langlieger nach CHE aus.
Die Anpassung der verminderten Ressourcen erfolgte an unserer Klinik hauptsächlich über eine weitere Reduktion der Kurzlieger bei Patienten mit unkompliziertem Verlauf nach CHE.
Es zeichnet sich eine langsame Erholung der Operationszahlen bei der CHE ab, die wahrscheinlich auf der geringeren Erkrankungsstärke der COVID-19-Folgevarianten (z. B. Omikron) basiert.
Einhaltung ethischer Richtlinien
Interessenkonflikt
L. Fischer, T. Iber, M. Feißt, B. Huck, G. Kolb, B. Huber, C. Segendorf, E. Fischer und K. Halavach geben an, dass kein Interessenkonflikt besteht.
Dies ist eine retrospektive Untersuchung mit den damit verbundenen Konsequenzen bez. der Datenqualität. Diese retrospektive Studie erfolgte nach Konsultation der zuständigen Ethikkommission und im Einklang mit nationalem Recht.
Alle Geschlechtsbezeichnungen im Artikel sind neutral, schließen also weibliches, männliches und diverses Geschlecht ein.
QR-Code scannen & Beitrag online lesen
==== Refs
Literatur
1. Klauber J Wasem J Beivers A Mostert C Krankenhaus-Report 2022 Patientenversorgung während der Pandemie 2022 Berlin, Heidelberg Springer
2. Chilamakuri R Agarwal S COVID-19: characteristics and therapeutics Cells 2021 10 206 10.3390/cells10020206 33494237
3. Lenzen-Schulte M Pandemiefolgen in der Onkologie: Einbruch der Krebsdiagnosen alarmierend Dtsch Ärztebl 2022 119 24 A1089 A1090
4. Karaca O Rüggeberg A Bialas E Schuster M Critical operations during the SARS-CoV-2 pandemic—an analysis of surgical cases with subsequent need for intensive care Dtsch Ärzteblt Int 2022 10.3238/arztebl.m2022.0225
5. von Dercks N Seehofer D Steinert M Wie stark trifft die Corona-Pandemie die chirurgische Klinik eines universitären Maximalversorgers? Eine Analyse der ersten 7 Wochen Chirurg 2020 91 755 761 10.1007/s00104-020-01255-y 32793988
6. Koch F Hohenstein S Bollmann A Cholecystectomies in the COVID-19 pandemic during and after the first lockdown in Germany: an analysis of 8561 patients J Gastrointest Surg 2022 26 408 413 10.1007/s11605-021-05157-0 34671914
7. Fischer L Kolb G Segendorf C Welcher Patient benötigt Laborwertkontrollen nach elektiver laparoskopischer Cholezystektomie? – Kann ein Score helfen? Chirurg 2021 92 369 373 10.1007/s00104-020-01258-9 32757046
8. Campanile FC Podda M Arezzo A Acute cholecystitis during COVID-19 pandemic: a multisocietary position statement World J Emerg Surg 2020 15 38 10.1186/s13017-020-00317-0 32513287
9. Coimbra R European Society of Trauma and Emergency Surgery (ESTES) recommendations for trauma and emergency surgery preparation during times of COVID-19 infection Eur J Trauma Emerg Surg 2020 46 3 505 510 10.1007/s00068-020-01364-7 32303798
10. Shakir T Matwala K Vasan A Karamanakos S Percutaneous cholecystostomy for acute cholecystitis: a three-year single-centre experience including during COVID-19 Cureus 2021 13 12 e20385 10.7759/cureus.20385 35036216
11. Loozen CS Laparoscopic cholecystectomy versus percutaneous catheter drainage for acute cholecystitis in high risk patients (CHOCOLATE): multicentre randomised clinical trial BMJ 2018 363 k3965 10.1136/bmj.k3965 30297544
12. Steffani M Merz C Stöß C Auswirkungen der erstenCOVID-19-Welle auf die Viszeralchirurgie. Ein retrospektiver Fallzahlenvergleich an einem Universitätsklinikum und einem Krankenhaus der Grund- und Regelversorgung Chirurg 2021 92 559 566 10.1007/s00104-021-01434-5 34009441
13. Rahimli M Wex C Wiesmueller F Laparoscopic cholecystectomy during the COVID-19 pandemic in a tertiary care hospital in Germany: higher rates of acute and gangrenous cholecystitis in elderly patients BMC Surg 2022 22 168 10.1186/s12893-022-01621-z 35538571
14. Becke-Jakob K Zeitpunkt elektiver Eingriffe nach SARS-CoV-2-Infektion und Impfung Dtsch Ärztebl 2022 119 21 A964 A965
15. Lenzen-Schulte M Chirurgie in Zeiten der Pandemie Dtsch Ärztebl 2020 117 18/1 A940 A944
16. Challine A Dousset B de’Angelis N Impact of coronavirus disease 2019 (COVID-19) lockdown on in-hospital mortality and surgical activity in elective digestive resections: A nationwide cohort analysis Surgery 2021 170 1644 1649 10.1016/j.surg.2020.12.036 33597086
17. Friedlander DF Krimphove MJ Cole AP Where is the value in ambulatory versus inpatient surgery? Ann Surg 2021 273 5 909 916 10.1097/SLA.0000000000003578 31460878
18. Fischer L Watrinet K Kolb G Patienten nach unauffälliger elektiver laparoskopischer Cholezystektomie können ohne Laborwertkontrollen entlassen werden – Ergebnisse einer prospektiven Studie Chirurg 2022 10.1007/s00104-022-01713-9
19. Hinchcliffe Z Mohamed I Lala A Day case laparoscopic cholecystectomy: Identifying patients for a ‘COVID-Cold’ isolated day-case unit during the pandemic J Perioper Pract 2021 31 3 62 70 10.1177/1750458920977418 33544661
20. Ma JLG Yogaraj V Siddiqui M The impact of COVID-19 on emergency cholecystectomy ANZ J Surg 2022 92 409 413 10.1111/ans.17406 34859559
21. Hess R Patientenrechte im Krankenhaus Chefarzt Aktuell 2022 03 22 25 26
22. Gutt C Jenssen C Barreiros A-P Updated S3-guideline for prophylaxis, diagnosis and treatment of gallstones. German Society for Digestive and Metabolic Diseases (DGVS) and German Society for Surgery of the Alimentary Tract (DGAV)—AWMF Registry 021/008 Z Gastroenterol 2018 56 912 966 30103228
| 36512029 | PMC9746580 | NO-CC CODE | 2022-12-15 23:21:56 | no | Chirurgie (Heidelb). 2022 Dec 13;:1-6 | utf-8 | Chirurgie (Heidelb) | 2,022 | 10.1007/s00104-022-01788-4 | oa_other |
==== Front
Padiatr Padol
Padiatr Padol
Padiatrie Und Padologie
0030-9338
1613-7558
Springer Vienna Vienna
1033
10.1007/s00608-022-01033-5
Editorial
Jahresrückblick 2022 – Jahresausblick 2023
Review of 2022—Outlook for 2023Kerbl Reinhold [email protected]
grid.508273.b Abteilung für Kinder- und Jugendheilkunde, LKH Hochsteiermark, Standort Leoben, Vordernberger Str. 42, 8700 Leoben, Österreich
13 12 2022
2022
57 6 267269
4 11 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Austria, ein Teil von Springer Nature 2022
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
issue-copyright-statement© Springer-Verlag GmbH Austria, ein Teil von Springer Nature 2022
==== Body
pmcZum dritten Mal in Folge geht ein Jahr zu Ende, in dem „Corona“ (nicht nur) den medizinischen Alltag bestimmt hat. Es waren einerseits direkte Auswirkungen durch COVID-19-Erkrankungsfälle, aber auch indirekte Auswirkungen im Sinn sogenannter Kollateralschäden. Eine Zunahme von Übergewicht und psychischen Problemen, eine Infektionswelle mit humanem Respiratorischen Synzytial-Virus (RSV) zu untypischer Zeit und Verzögerungen von Diagnosestellung und Therapiebeginn sind nur einige Beispiele. Die weitgehende Normalität im Sommer 2022 hat zwar die Situation für Kinder und Jugendliche verbessert, gleichzeitig aber sind mit Ukrainekrieg, (behaupteter) Energiekrise und den Sorgen um den Klimawandel neue Belastungen für unsere Jugend hinzugekommen.
Als Kinder- und Jugendärzt*innen müssen wir uns mit diesen Belastungsfaktoren auseinandersetzen und Wege aufzeigen, wie die Folgen dieser Bedrohungen möglichst geringgehalten werden können. Es gilt also, Prävention zu stärken und zu fördern, sowohl im somatischen als auch im psychosozialen Bereich.
Leider gab es – wohl auch durch häufige Wechsel der Zuständigen – von politischer Seite wenig Verständnis und Unterstützung; als Beispiel sei der langjährige Stillstand beim „Mutterkindpass neu“ genannt. Es gibt aber auch zahlreiche andere Baustellen.
Wo sind die alle?
Bettensperren, nicht besetzbare Nachtdienste, Terminverschiebungen, überlange Wartezeiten und dergleichen gehören heute zum Alltag unseres Gesundheitssystems, von dem wir so lange behauptet haben, es sei das beste der Welt. Auch vor der Pädiatrie hat diese Entwicklung (leider) nicht Halt gemacht. Mängelverwaltung bzw. „management by chaos“ ist nicht mehr Ausnahme, sondern Regel. Warnungen vor dem zu erwartenden Mangel an Pflegefachkräften wurden (viel zu) lange in den Wind geschlagen; jetzt fehlen Lösungsstrategien.
Dabei ist der Mangel an Fachkräften neuerdings nicht einmal auf das Gesundheitswesen beschränkt – Gastronomie, Pädagogik, Facharbeit und viele andere Bereiche sind ähnlich betroffen. Im Gesundheitssystem sind die Defizite allerdings deshalb kritischer, weil sie zu einer Gefährdung der Gesundheit und manchmal auch des Lebens führen können.
Die bedauerliche Entwicklung ist neben der (politisch zu verantwortenden) Fehlplanung auch auf verschiedene weniger oder nicht vorhersehbare Umstände zurückzuführen. Dazu zählen die Pandemie und deren Maßnahmen, Pensionierungen der Babyboomer-Generation, der immer stärker werdende Wunsch nach „life balance“, die fehlende Attraktivität von Mehrleistung(en), zuletzt vielleicht auch Ukrainekrieg und (behauptete) Energiekrise.
COVID-19 in der Pädiatrie
Auch wenn COVID-19 Kinder und Jugendliche in der Regel deutlich leichter erkranken lässt als ältere Personen, gab es doch im vergangenen Jahr auch zahlreiche Spitalsaufenthalte: einerseits im Rahmen der Akuterkrankung vor allem bei Säuglingen und Kleinkindern, wesentlich seltener im Rahmen des hyperinflammatorischen Verlaufs durch Pädiatrisches Multiorgan-Immun-Syndrom (PIMS) bzw. Multisystem-Entzündungssyndrom bei Kindern (MIS-C). Soweit erhebbar, gibt es in Österreich bisher 10 COVID-Todesfälle im Kindes- und Jugendalter, wobei mit großer Mehrheit Kinder und Jugendliche mit Vorerkrankungen betroffen waren. Leider ist aber insgesamt die österreichische Datenlage zu COVID-19 bei Kindern und Jugendlichen sehr unbefriedigend. Zwar wird die Zahl der Neuinfektionen in der Datenbank der Österreichischen Agentur für Gesundheit und Ernährungssicherheit GmbH (AGES) erfasst, bezüglich Hospitalisierungsrate, Kausalität (mit oder wegen COVID) und Verlauf liegen aber – auch wegen großer Zeitverzögerung bei der Datenerfassung – kaum belastbare Ergebnisse vor. Dies betrifft insbesondere auch die Prävalenz von Long COVID, die im Kindes- und Jugendalter deutlich niedriger ist als im Erwachsenenalter. Aktuell ist in der Zeitschrift Children [1] eine Spezialausgabe zu diesem Thema geplant, um etwas mehr Licht ins Dunkel zu bringen.
Pro futuro ist davon auszugehen, dass uns in der Pädiatrie COVID-19 immer wieder betreffen wird
Pro futuro ist davon auszugehen, dass uns auch in der Pädiatrie COVID-19 immer wieder (wahrscheinlich wellenförmig) betreffen bzw. belasten wird. Aus aktueller Sicht wird sich allerdings die SARS-CoV-2-Infektion letztlich in die Reihe der anderen respiratorischen Viren einordnen.
Primärversorgung – kleine Schritte
Der Mangel an § 2-Kassenfachärzt*innen ist seit mehreren Jahren manifest, österreichweit sind 11–15 % der Kassenfacharztstellen unbesetzt. Dabei bestehen deutliche regionale Unterschiede, das größte Defizit wird im Bundesland Niederösterreich registriert. Die Österreichische Gesellschaft für Kinder und Jugendheilkunde (ÖGKJ) hat im Rahmen der Versorgungsforschung die Gründe für diese Mangelversorgung analysiert [2] und Lösungsstrategien vorgelegt. Seitens der gesundheitspolitisch Verantwortlichen (Gesundheitsministerium, Länder, ÖGK) zeigen sich zuletzt auch deutlich wahrnehmbare Bemühungen, die Situation zu verbessern. Durch die Zulassung von Gruppenpraxen, Job Sharing, Anstellung von Ärzten bei Ärzten, Übernahme einer Praxis im Angestelltenverhältnis etc. wurde den Wünschen der jungen Kolleginnen und Kollegen teilweise entsprochen. Noch nicht möglich ist derzeit allerdings die pädiatrische Primärversorgungseinheit (PVE), gegenwärtig ist noch immer die Leitung einer PVE durch einen Allgemeinmediziner vorgesehen. Es besteht allerdings Aussicht, dass auch in diesem Punkt noch Beweglichkeit besteht und das PVE-Gesetz entsprechend adaptiert wird.
Mutterkindpass – keine Schritte
Weiterhin (Stand Ende Oktober 2022) keine Bewegung gibt es beim „Mutterkindpass neu“. Nach Abschaffung der Mutterkindpass-Kommission im Jahr 2011 und der Erarbeitung neuer Inhalte durch eine Facharbeitsgruppe (FAG) in den Jahren 2015–2018 herrscht seit nunmehr 4 Jahren Stillstand, einerseits bezüglich neuer Inhalte, andererseits aber auch bezüglich elektronischer Umsetzung (e-Mutterkindpass) und Valorisierung der Tarife. Letztere sind seit 1994 und somit seit nunmehr 28 Jahren nicht angepasst – mit Sicherheit mit ein Grund für die geringe Attraktivität der § 2-Kassenpraxis. Die unbefriedigende Situation hat die Ärztekammer zuletzt auch dazu veranlasst, Vertragskündigungen anzudrohen. Ob dies der richtige Weg ist, wird sich zeigen, die Notwendigkeit der Valorisierung steht jedoch außer Zweifel – insbesondere auch in Anbetracht der allgemeinen Teuerung und Inflation.
Ausbildung: 2023 wird vieles anders
Im Oktober 2022 haben 70 Kolleginnen und Kollegen ihre Facharztausbildung mit der Facharztprüfung positiv abgeschlossen – wir begrüßen diese sehr herzlich als „frisch gebackene“ Fachärztinnen und Fachärzte. Die Ausbildung in Kinder- und Jugendheilkunde mit einer Ausbildungsdauer von 63 Monaten ist in Österreich insgesamt auf hohem Niveau, auch wenn in Teilbereichen manches noch verbesserungswürdig erscheint (Sonographie, Intensivmedizin, Psychosomatik u. a.). Seitens der ÖGKJ gibt es seit mehreren Jahren Bestrebungen, die Ausbildung auch in der Niederlassung zu forcieren und öffentlich zu finanzieren. Diesbezüglich erfolgten wiederholt Vorsprachen im Gesundheitsministerium. Der Kostenaufwand dafür würde etwa € 800.000 pro Jahr betragen – ein vergleichsweise geringer Betrag. Es ist leider derzeit nicht absehbar, ob bzw. wann diesem Ansinnen nachgekommen wird und damit eine Gleichstellung der Kinderärzte mit den Allgemeinmedizinern als „Primärversorger“ erfolgt.
Bezüglich Ausbildungsstellen wurde von verschiedenen Seiten wiederholt der Wunsch vorgebracht, die Pädiatrie als Mangelfach zu deklarieren. Dies wurde bisher seitens Österreichischer Ärztekammer (ÖÄK) und Bundesministerium abgelehnt. Einerseits, weil Österreich über etwa 2000 Pädiaterinnen und Pädiater verfügt, andererseits aber auch weil derzeit ohnehin fast jedes medizinische Sonderfach als Mangelfach bezeichnet werden muss.
Eine Änderung ergibt sich ab Januar 2023 dadurch, dass die Verantwortung für Ausbildung und Ausbildungsstätten bzw. -stellen ab 01.01.2023 bei den jeweiligen Ländern bzw. explizit bei den Landeshauptleuten verankert ist. Die Zahl von Ausbildungsstellen wird dann voraussichtlich nicht mehr zentral festgelegt, sondern auf Länderebene. Es bleibt abzuwarten, ob dies für einzelne Bundesländer einen Vorteil darstellt oder – wie von der Ärztekammer vermutet – zu einem Qualitätsverlust führen wird.
Fortbildungen und Tagungen
Nach 2‑jähriger Pause wurden im Jahr 2022 endlich wieder vorübergehend ausgesetzte Fortbildungstagungen als Präsenzveranstaltungen angeboten. Die pädiatrischen Fortbildungen in Obergurgl, in Seggau, die Pädiatrietage in Venedig und der pädiatrische Samstag in Linz waren gut besucht und dokumentierten den Wunsch der Kollegenschaft nach hochwertiger Präsenzfortbildung.
Die von Peter Voitl und Susanne Diesner-Treiber in Wien organisierte Jahrestagung war ein besonderes Highlight im pädiatrischen Fortbildungsjahr. Die historischen Veranstaltungsorte (Josephinum, van Swietensaal und Billrothhaus) gaben der Jahrestagung eine besondere Note. Das Hauptthema „Auf den Schultern von Riesen“ sollte auch die historische Bedeutung der österreichischen Medizin/Pädiatrie zum Ausdruck bringen, in zahlreichen hochwertigen Beiträgen ist dies auch sehr gut gelungen.
Die nächste Jahrestagung wird Ende September 2023 an der Montanuniversität in Leoben stattfinden, voraussichtlich unter dem Leitthema „Vernetzte Pädiatrie“. Wir freuen uns schon jetzt auf Deine/Ihre Teilnahme!
Die „Richtigen“ finden …
In einer insgesamt schwierigen Zeit ist es unsere Aufgabe, die „Richtigen“ für die Pädiatrie zu gewinnen und zu begeistern. Die Breite unseres Sonderfachs macht dieses hochinteressant und anspruchsvoll, gleichzeitig müssen wir aber auch die berechtigten Ansprüche von Kindern, Jugendlichen und deren Eltern im Auge behalten. Der pädiatrische Nachwuchs soll daher nicht nur fachlich gut qualifiziert sein, sondern sich insbesondere auch durch einen humanitären und empathischen Zugang auszeichnen.
Verpflichtet sind wir den Kindern und Jugendlichen
Die Tätigkeit als Kinderärztin/als Kinderarzt ist ein besonderes Privileg. Auch in einer Zeit von Life Balance und Selbstverwirklichung sollte unsere fachliche, aber auch moralische Verpflichtung gegenüber den Kindern und Jugendlichen im Vordergrund bleiben. Berufliche Zufriedenheit ergibt sich letztlich wahrscheinlich aus der Summe von fachlicher Qualifikation, anspruchsvoller Tätigkeit, kontinuierlichem Lernen, Teamarbeit, vor allem aber auch aus der Hilfestellung, die wir Kindern, Jugendlichen und deren Eltern geben können.
Danksagung
Abschließend bedanke ich mich wieder bei der Chefredakteurin der Pädiatrie und Pädologie, Frau Dr. Lessky-Höhl. Sie bemüht sich mit großem Einsatz um qualitativ hochwertige Beiträge zu aktuellen und wichtigen pädiatrischen Themen. Und es gelingt ihr auch im Zeitalter der Life Balance, hochqualifizierte Autor*innen trotz knapper Ressourcen für Beiträge zu gewinnen. Auch diesen möchte ich ganz besonders danken – sie gewährleisten die anhaltend hohe Qualität von Pädiatrie und Pädologie.
Schließlich gilt mein Dank unseren Leserinnen und Lesern. Ihr Interesse und Feedback sind für uns als Verantwortliche die Motivation für die Fortsetzung unserer Arbeit.
Mit den besten Wünschen für 2023,
Ihr Reinhold Kerbl
Herausgeber Pädiatrie & Pädologie
Interessenkonflikt
R. Kerbl gibt an, dass kein Interessenkonflikt besteht.
Hinweis des Verlags
Der Verlag bleibt in Hinblick auf geografische Zuordnungen und Gebietsbezeichnungen in veröffentlichten Karten und Institutsadressen neutral.
==== Refs
Literatur
1. https://www.mdpi.com/journal/children/special_issues/1JJ41U998K. Zugegriffen: 22. Nov. 2022
2. Kohlfürst DS Zöggeler T Karall D Kerbl R Workload and job satisfaction among Austrian pediatricians: gender and generational aspects Wien Klin Wochenschr 2022 134 13–14 516 521 10.1007/s00508-022-02050-x 35739286
| 0 | PMC9746583 | NO-CC CODE | 2022-12-15 23:21:56 | no | Padiatr Padol. 2022 Dec 13; 57(6):267-269 | utf-8 | Padiatr Padol | 2,022 | 10.1007/s00608-022-01033-5 | oa_other |
==== Front
SN Comput Sci
SN Comput Sci
Sn Computer Science
2662-995X
2661-8907
Springer Nature Singapore Singapore
1526
10.1007/s42979-022-01526-x
Original Research
A Precise Method to Detect Post-COVID-19 Pulmonary Fibrosis Through Extreme Gradient Boosting
http://orcid.org/0000-0003-0211-942X
Jha Manika [email protected]
Gupta Richa
Saxena Rajiv
grid.419639.0 0000 0004 1772 7740 Department of Electronics and Communication Engineering, Jaypee Institute of Information Technology, 201309 Noida, India
13 12 2022
2023
4 1 8925 4 2022
17 11 2022
© The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
The association of pulmonary fibrosis with COVID-19 patients has now been adequately acknowledged and caused a significant number of mortalities around the world. As automatic disease detection has now become a crucial assistant to clinicians to obtain fast and precise results, this study proposes an architecture based on an ensemble machine learning approach to detect COVID-19-associated pulmonary fibrosis. The paper discusses Extreme Gradient Boosting (XGBoost) and its tuned hyper-parameters to optimize the performance for the prediction of severe COVID-19 patients who developed pulmonary fibrosis after 90 days of hospital discharge. A dataset comprising Electronic Health Record (EHR) and corresponding High-resolution computed tomography (HRCT) images of chest of 1175 COVID-19 patients has been considered, which involves 725 pulmonary fibrosis cases and 450 normal lung cases. The experimental results achieved an accuracy of 98%, precision of 99% and sensitivity of 99%. The proposed model is the first in literature to help clinicians in keeping a record of severe COVID-19 cases for analyzing the risk of pulmonary fibrosis through EHRs and HRCT scans, leading to less chance of life-threatening conditions.
Keywords
Pulmonary fibrosis
COVID-19
Medical diagnosis
Clinical decision support
Machine learning
Extreme gradient boosting
Tree boosting
issue-copyright-statement© Springer Nature Singapore Pte Ltd 2023
==== Body
pmcIntroduction
The global pandemic due to SARS-CoV-2 novel coronavirus was first reported in 2019. WHO has reported 486,761,597 infections and 6,142,735 deaths till 1, April, 2022 September 2021 [1]. A wide spectrum of health complications has been noticed in patients with severe COVID-19 patients. Diseases after severe COVID-19 have been shared with black fungus, cardiac arrest, temporary paralysis, joint pain and respiratory disorders [2]. The SARS-CoV-2 virus has been seen responsible to cause an incidence of acute respiratory distress in a huge volume of COVID-19 cases [3]. Clinical and radiographic reports suggest the onset of pulmonary fibrosis, a common course after SARS infection. It is known as a sequela of persistent damage to the lung or acute respiratory distress syndrome (ARDS). Pulmonary fibrosis turns out as a serious complication of lung pneumonia, which leads to impaired lungs or dyspnea [4]. Various clinical studies indicate the link of COVID-19 patients with respiratory disorders, which sometimes lead to mortality. It has been noticed that the inflammatory mechanism starts around 60–90 days of hospital discharge of severe COVID-19 cases. The lungs become scarred over time, and symptoms like dry cough, tiredness, shortness of breath, nail clubbing and weight loss have been noticed in the patients [5, 6].
Pulmonary fibrosis gradually converts the normal parenchyma of lungs into fibrotic tissues. Those scarred tissues decrease the oxygen capacity, which leads to stiffness and restrictive lungs. There are various risk factors associated with the development of fibrosis after COVID-19 [7, 8]. The first risk factor is an extended stay in the ICU and the use of mechanical ventilation. While the severity of the condition is linked to the amount of time spent in the ICU, mechanical ventilation increases the risk of ventilator-induced lung injury (VILI). This injury is caused by abnormal pressure or volume settings, which cause the production of pro-inflammatory modulators, exacerbating acute lung injury, and higher mortality or pulmonary fibrosis in survivors [9]. Increased disease severity is the second risk factor, which includes comorbidities like hypertension, diabetes, and coronary artery disease, as well as lab abnormalities like lymphopenia, leukocytosis, and high lactate dehydrogenase (LDH). Following acute lung injury, the level of serum LDH has been utilized as a measure of disease severity. It's a marker for lung tissue loss that’s linked to a higher risk of death [10]. According to the World Health Organization, 80 percent of SARS-CoV-2 infections are mild, 14 percent causes severe symptoms, and 6% result in death. Third risk factor includes smoking and drinking alcohol. When compared to non-smokers, smokers are 1.4 times more likely to develop severe COVID-19 symptoms, 2.4 times more likely to require ICU admission and mechanical ventilation, and 2.4 times more likely to die [11, 12].
There are currently no fully proven methods for treating post-inflammatory COVID-19 pulmonary fibrosis. Various therapeutic options are being considered. It has been proposed that long-term use of antiviral, anti-inflammatory, and anti-fibrotic medications reduces the risk of lung fibrosis. However, it is yet unknown whether early and prolonged use of antiviral medications can prevent lung remodeling or which antiviral is the most beneficial. Anti-fibrotic medications like pirfenidone and nintedanib also have anti-inflammatory properties. Therefore, they can be utilized even during the acute phase of COVID-19 pneumonia. Pirfenidone works as an anti-fibrotic, anti-oxidant, and anti-inflammatory agent. Pirfenidone may minimize ARDS-induced lung injury by decreasing NLRP3 inflammasome activation, which reduces LPS-induced acute lung injury and eventual fibrosis. 26 Anti-fibrotic therapy has a few drawbacks in the acute phase. Hepatic dysfunction is common in COVID-19 patients, as evidenced by elevated transaminases, and the anti-fibrotics pirfenidone and nintedanib cause hepatotoxicity. Because most COVID-19 patients are taking anticoagulants, nintedanib is linked to an increased risk of bleeding. Anti-fibrotic therapy should be started within the first week of ARDS onset to avoid complications of lung fibrosis [13–15]. As a result, identifying people who are at risk of developing pulmonary fibrosis is critical. The rationale for employing medication should be customized, and precision medicine's role assumes the prediction of high-risk populations, a better understanding of pathophysiology, and the avoidance of disease deterioration or the formation of lung fibrosis. Analysis of COVID-19 patients after discharge from hospitals could only decrease the risk of developing fibrotic abnormalities. The majority of diagnostic procedures are based on various symptoms, medical imaging (mainly High-Resolution computed tomography), Lung function tests (LFT) and biopsy. As these procedures take longer interpretation time, cause discomfort in patients, expose patients to radiations, clinicians and radiologists are more inclined toward computer-aided diagnosis [16]. Therefore, this work presents an efficient model to detect pulmonary fibrosis in severe COVID-19 patients after 90 days of discharge from the hospital, by analyzing EHRs and HRCT scans.
The paper is organized as follows: “Literature Review” and “Motivation” presents the existing literature review and motivation for research, respectively. “Materials and Methods” includes dataset description, model development and evaluation metrics, while “Experimental Results” provides the experimental setup and comparison results with other machine learning models. “Discussion” discusses the significant findings and “Conclusion” concludes the paper by emphasizing the salient points and future perspectives.
Literature Review
To find relevant research work, a literature review was conducted utilizing several databases (PubMed, Scopus, Science Direct, and Google Scholar). Coronavirus, severe acute respiratory syndrome coronavirus 2, COVID-19, post-COVID fibrosis, and anti-fibrotic were among the search phrases. The search yielded around many articles, comprising review articles, case studies and reports. Few of them have been discussed in this section. Carfi et al. evaluated 143 patients who were discharged from the hospital after recovering from COVID-19 with ongoing symptoms. Only 18 (12.6 percent) of patients were completely free of any COVID-19-related symptom at the time of evaluation, whereas 32 percent had one or two symptoms and 55 percent had three or more. There was no fever or other indications or symptoms of acute illness in any of the patients. In 44.1 percent of patients, their quality of life had deteriorated. They also discovered that fatigue (53.1%), dyspnea (43.4%), joint pain (27.3%), and chest pain were the most common symptoms that persisted after discharge (21.7%) [17]. Zhao et al. evaluated COVID-19 survivors’ pulmonary function and related physiological features three months after recovery, enrolling 55 patients and finding varying degrees of radiological abnormalities in 39 of them. The presence of CT abnormalities was linked to a high blood urea nitrogen content upon admission [18]. A chest CT scan was obtained on the last day before discharge, two weeks after discharge, and four weeks after discharge in the study of Liu et al. The anomalies in the lungs (including focal/multiple GGO, consolidation, interlobular septal thickening, sub-pleural lines, and irregular lines) were gradually absorbed in the first and second follow-ups after discharge, compared to the previous CT scan before discharge. After a 4-week follow-up, 64.7 percent of released patients had their lung lesions completely absorbed. It suggested that COVID-19-induced lung tissue damage might be reversible in most COVID-19 patients. It was also proposed that non-severe patients had a good prognosis, and that clinical intervention should be done early to prevent common COVID-19 individuals from becoming severe [19].
A recent study by Yasin et al. [20], showed the age of the patients, CT severity score, consolidation score, and admissions in ICU were identified as the independent risk variables related with the occurrence of post-COVID-19 fibrosis after a multivariate analysis. At a cut-off point of 10.5, the chest CT severity score has a sensitivity of 86.1%, a specificity of 78%, and an accuracy of 81.9%. Another study from ‘The Lancet’ had confirmed a few risk factors, such as age, hypertension, the severity of COVID-19 and diabetes, as important indicators of developing fibrosis [21]. Li referred post-COVID-19 pulmonary fibrosis as a worrisome sequela in surviving patients [22]. The study also presented the significance of early detection of fibrosis in high-risk patients through appropriate CT scans. According to a review of Spagnolo et al. biomarkers of susceptibility could help identify patients with a higher risk and could be used to personalize COVID-19’s long-term effects treatment. It emphasizes the importance of patient and illness-related contributing risk factors for pulmonary fibrosis in COVID-19 survivors, as well as the potential utility of acute phase and follow-up biomarkers for identifying patients most at risk of developing the disease [23]. Another research article described the correlation of risk factors, such as leukocyte count, lactate dehydrogenase, the severity of COVID-19 and duration of mechanical ventilation, with the development of fibrotic abnormalities [24]. Chen et al. [25] analyzed 169 autopsies of patients with ARDS caused by a variety of causes and found that fibrosis was present in three (4%) out of 82 patients with a disease duration of less than one week, 13 (24%) out of 54 patients with a disease duration of one to three weeks, and 14 (61%) out of 23 patients with a disease duration of more than three weeks. Das et al. investigated 27 patients who had put on ventilation for ARDS and found that 23 (85%) of them had symptoms of fibrosis 110–267 days after extubating, with a strong link to the length of the pressure-controlled inverse-ratio ventilation [26]. Yu et al. divided patients in to two groups—early fibrosis and severe fibrosis, based on post-COVID-19 follow-ups. On preliminary CT imaging, the fibrosis group had a higher prevalence of the irregular interface (57.1%) and parenchymal band (50.0%). On the worst-state CT imaging, the fibrosis group had a higher prevalence of parenchymal band (92.9%), interstitial thickening (786%) air bronchogram (571%), uneven interface (85.7%) and coarse reticular pattern (28.6%) [27].
The literature indicates a strong link between fibrotic abnormalities and COVID-19 for around 15–20% of recovered patients. Considering millions of cases of COVID-19 over the world, even a small percentage of post-COVID lung fibrosis is concerning. The research articles also specify the importance of blood investigations and HRCTs of recovered COVID-19 patients to analyze the risk of developing fibrosis in the lungs. Recently, EHRs have been considered as a critical tool of patient data collection. At the time of care, EHR delivers accurate, up-to-date, and full information about patients. It also aids in the accurate diagnosis of patients, the reduction of medical errors, the provision of safer care, and quick access to patient records necessary for more coordinated and efficient care. HRCT, on the other hand, is a more precise radiological examination than a chest X-ray for the diagnosis and monitoring of lung tissue and airway illnesses. A volume HRCT scan of the entire lung tissue is possible with modern CT equipment. Contrast-enhanced CT scans of the chest or the entire body can also be used to create HRCT slices. Idiopathic interstitial pneumonias and pulmonary fibrosis, are among the most well-known indications for HRCT. These diagnostic tools have been considered beneficial in finding the abnormalities present in lungs after discharging COVID-19 patients.
Motivation
After the COVID-19 pandemic, an increasing number of individuals worldwide who have survived the sickness are still suffering from its symptoms, even though they have been clinically tested negative for the virus. As we fight this pandemic, the most difficult part will be figuring out how to deal with COVID-19 sequelae, which can range from mild fatigue and body aches to severe forms requiring long-term oxygen therapy and lung transplantation due to lung fibrosis, significant cardiac abnormalities, and stroke, all of which lead to a significant reduction in quality of life. Various studies have found that 70–80 percent of COVID-19 patients still have at least one or more symptoms after being declared COVID-free. Existing literature indicates the lack of detection or prediction model for Post COVID-19 pulmonary fibrosis. Thus, there is an urgent need of developing computer-aided diagnostic models to help the healthcare sector in detecting the fibrotic abnormalities before its onset time. Recently, many artificial intelligence-assisted systems based on EHRs and CT scans have been reported for diagnosing diseases. Powerful models based on machine learning assists clinicians and medical practitioners to diagnose the abnormalities effectively. The health reports indicate important risk factors that could be a vital diagnostic indicator for the early detection of pulmonary fibrosis.
As, computer-aided diagnostic models are now a great alternative to human experts due to their speed, accuracy and decreased false positive rates, an effective model for detecting early onset of pulmonary fibrosis could help in decreasing mortality due to the severe scaring of lungs. In this proposed work, clinical characteristics and chest HRCT data of patients were collected, with follow-up studies on the evolution of pulmonary fibrosis, who returned to the hospital for chest HRCT re-examinations 90 days after hospital discharge. In the case of pulmonary fibrosis, the major risk factors that were reported, are age, symptoms like cough, cold, fever, chest tightness, IL-6 levels, WBC counts, Lymphocytes, Albumin, Creatinine, CRP, D-dimer and humoral immunity-related indexes (IgG). The chest CT image analysis included the spread of the lesions, the position of the lesions, lobes affected, features of the lesions and external immersion. For each patient, the CT presentation was described according to the parameters of Lesion degree, Quantitative scoring of pulmonary fibrosis and Inflammation score. These risk factors are obtained after stratifying COVID-19 patients (with and without pulmonary fibrosis). A statistically significant difference has been acquired in most of the risk factors. For the technical aspect, the model has gone through optimum algorithm selection procedures and hyper-parameter tuning. The overall architecture of the proposed pulmonary fibrosis detection system is present in Fig. 1. The major inputs are highlighted as follows:A dataset of 1175 severe COVID-19 patients has been created using EHRs and corresponding HRCT scans that includes general clinical data, such as sex, age, main clinical symptoms and radiological images.
Statistical analysis has been performed to evaluate the clinical characteristics and to make a comparison between patients with or without pulmonary fibrosis. Feature importance has also been obtained to acquire the most prominent indicator of fibrosis in COVID-19 patients.
After pre-processing, statistical analysis, null value assessment and feature scaling, training of various machine learning algorithms have been executed to achieve the classification of patients into fibrosis cases and normal cases.
Several machine learning algorithms are then compared on the considered dataset, in which Extreme Gradient Boosting (XGBoost) provides the best performance in terms of performance metrics, such as accuracy, precision, recall and specificity. The XGBoost is thus used as the base model and is optimized for the application by tuning the major hyper-parameters, such as learning rate, gamma rate and regularization lambda.
The improved XGBoost model is then trained with different training testing splits of the same dataset. The model is then tested for the prediction of pulmonary fibrosis and normal lungs. This novel approach exhibits potency and thus can be embedded in clinical diagnosis systems to provide fast, reliable and low-cost results.
Fig. 1 The overall system architecture of the proposed system for pulmonary fibrosis detection
Materials and Methods
The paper aims to propose a machine learning-based diagnostic system to automatically detect pulmonary fibrosis by evaluating a patient’s risk factors and HRCT scans.
Dataset
As the emergence of post-COVID-19 complications is recent, none of the large data repositories contain any labeled data for pulmonary fibrosis, thereby leading us to rely on chest examination reports, EHRs and CT scan interpretations of Centre Theatre General Hospital, China for the training proposed model [28, 29]. The clinical characteristics and HRCT scans were collected at the time of follow-ups of COVID-19 patients after 90 days of hospital discharge. The dataset includes single comma separated values (csv) file with 32 risk factors and HRCT scans for all 1175 patients with their labels as Normal lungs or Fibrosed Lungs. In the acquired dataset, 725 patients have developed pulmonary fibrosis while 450 patients did not develop pulmonary fibrosis after COVID-19 recovery. A statistical analysis has been done to evaluate the relationship between fibrosis progression and related risk factors in all 1175 COVID-19 patients. The analysis of EHR was carried out in SPSS (version 26.0) software. The result showed a significant relationship between pulmonary fibrosis with levels of Interleukin-6 (IL-6), albumin and cellular immunity-related indexes in patients through analyzing the ϰ2 values and P values by Fisher’s exact test [30, 31]. Details of the dataset have been presented in Table 1. The HRCT scans were pre-processed using resizing, normalizing de-noising filters to be classified through the proposed model. Figure 2 depict the samples of Normal and Fibrosed HRCT scans after recovering from COVID-19.Table 1 The clinical risk factors of severe COVID-19 patients
Patients with pulmonary fibrosis Patients without pulmonary fibrosis ϰ2 value P value
Numbers 725 450 – –
Age (mean) 53.82 45.64 3.109 0.002
IL-6 31.32 4.33 3.566 0.001
WBC, × 109 per L 5.46 5.22 0.120 0.911
Lymphocyte % 24.56 31.22 3.324 0.002
Platelet, × 109 per L 197.22 211.32 1.056 0.297
HB (g/L) 125.87 132.87 1.768 0.077
CRP 27.50 8.65 3.544 0.001
PCT (ng/mL) 0.33 0.04 0.897 0.412
ALT, U/L 30.40 25.58 1.235 0.223
AST, U/L 32.14 26.33 2.987 0.011
Albumin 38.45 43.09 5.879 0.001
Creatinine (umol/L) 67.44 67.88 0.165 0.877
Glucose (mmol/L) 6.50 6.13 1.543 0.155
Potassium (mmol/L) 3.88 3.80 0.654 0.546
CK (U/L) 62.37 53.90 0.543 0.544
Myoglobin (ug/L) 49.42 42.11 1.098 0.317
hs-cTnT, pg/mL 0.045 0.01 0.657 0.600
Prothrombin time, s 13.34 15.33 3.031 0.303
D-Dimer 386.78 165.98 1.435 0.157
APTT 32.86 33.14 1.877 0.080
CD3 943.87 1432.11 5.655 0.003
CD4 557.04 897.22 5.765 0.001
CD8 320.94 654.32 3.566 0.008
CD19 211.71 422.26 5.433 0.002
CD16 + 56 228.40 287.65 0.765 0.556
IGG 9.15 9.08 0.122 0.878
IGM 5.54 7.09 0.854 0.543
IGA 2.42 1.87 0.344 0.662
AST/ALT 1.36 1.32 − 0.057 0.987
CRP/Albumin Ratio 0.55 0.17 − 4.809 0.000
Platelet/lymphocyte ratio 185.41 150.22 − 3.432 0.021
Fig. 2 Samples of HRCT scans a fibrosed lungs, b normal lungs
Development of XGBoost model
In this study, 32 features in the form of risk factors have been fed into the XGBoost model to automatically perform classification between patients with normal lungs and patients with fibrosed lungs [32]. XGBoost is a mathematical technique based on sequence ensemble. It evaluates the second-order partial derivative of the loss function to get the gradient patterns. These patterns then obtain the minimum loss function which eventually optimizes the model. In comparison with conventional gradient boosting, XGBoost uses regularization to improve the speed, parallelization and generalization of the model [33]. While implementing XGBoost, competent models are built from a collection of weak learners iteratively. The algorithm of XGBoost works on Newton–Raphson optimization in function space [34]. The generic version of XGBoost is stated below:Input: training set {(xi,yi)}i=1N, loss function L (y, F(x)), weak learners M and Learning rate of α.
To train the model, loss function has to be optimized by obtaining gradient descent and second-order Taylor approximation, represented in Eqs. (1) and (2).1 g^mxi=∂Lyi,fxi∂f(xi)fx=f^m-1`x,
2 h^mxi=∂2Lyi,fxi∂f(xi)2fx=f^m-1`x,
where g^mxi is the gradient and h^m(xi) is the hessian for m = 1 to M.
Fit the base learners (or weak learners) using the updated training set xi,-g^m(xi)h^m(xi)i=1N and solving the optimization problem stated below in Eq. (3):3 ∅m=argmin∑i=1N12h^mxi-g^mxi-∅xih^mxi2.
Updating model has been done as indicated in Eqs. (4) and (5):4 f^mx=α∅mx,
5 f^mx=f^m-1x+f^mx.
Loss function6 yi=f^x=f^(M)x=∑m=0Mf^mx.
The final loss function shown in Eq. (6) is then adjusted by taking the best values of parameters and input function to gain the optimum result.
In comparison to conventional gradient boosting, there are few in-built algorithm enhancement methods present in XGBoost. It reprimands complex models by applying regularization that avoids overfitting. It also handles sparsity patterns in the dataset more efficiently by learning automatically from the missed values while training. The XGBoost algorithm does cross-validation at each iteration of its own and employs a weighted Quantile algorithm to find the optimal points of the split [35]. For regularization of the model, another term known as regularization term is added to the cost function, as stated below:Objective function=Loss function+Regularization term.
Regularization term = λ2m * ∑|w|2, where λ is the regularization parameter, that is optimized to obtain the best results, m is the number of weak learners and w is the leaf weight matrix [36].
The hyperparameters present in the XGBoost model can be grouped as general, command line, booster and learning task. To achieve optimal performance, the model must be tuned carefully. Tuning the model is an unsettling task due to the number of parameters it has. The proposed model has used random search on a few important parameters for the tuning of the model. These tuned parameters provided exceptional results with less computational complexity for the proposed application. Table 2 shows the details of XGBoost model hyperparameters that are adjusted to make the model more efficient.Table 2 Optimal values for important hyperparameter
Hyperparameter Value
Learning rate 0.3000
n_estimator 100
Gamma 0
Regularization lambda 1
Subsample 1
Min_child_weight 1
Max_depth 6
Max_delta_step 0
Performance Evaluation Metrics
After the training process, predictions have been made on the test data. Total training samples taken under consideration are 881 and testing samples are 294. To obtain the performance, metrics, such as accuracy, precision, recall, F1 score, specificity, Matthew’s correlation coefficient (MCC), Youden Index (YI) and Cohen Kappa score, have been obtained through confusion matrix. The Matthews correlation coefficient (MCC) is a more reliable statistical rate that only yields a high score if the prediction performed well in all four confusion matrix categories (true positives, false negatives, true negatives, and false positives), proportionally to the size of positive and negative elements in the dataset. The Youden index assesses a diagnostic test’s ability to strike a balance between sensitivity (detection of disease) and specificity (detecting health or no disease). Diagnostic model's sensitivity is added to the specificity percentage, and the sum is deducted from 100. If the Youden index is less than 50%, the model does not meet the empirical criteria for being used for diagnostic purposes. The mathematical representation of all these metrics have been stated below:True Positive (TP): number of fibrosed lung cases that are correctly predicted as fibrosed.
False Positive (FP): number of normal lung cases that are wrongly predicted as fibrosed.
True Negative (TN): normal cases that are correctly predicted as normal.
False Negative (FN): number of fibrosed cases that are wrongly predicted as normal.
The above terms are utilized to form several performance measures present in Eqs. (7) to (10):7 Accuracy=∑TPTotal Samples,
8 Precision=TPTP+FP,
9 Sensitivity=TPTP+FN,
10 Specificity=TNTN+FP,
11 F1-score=2∗Recall∗PrecisionRecall+Precision,
12 MCC=TP∗TN-FP∗FNTP+FP∗TP+FN∗TN+FP∗(TN+FN),
13 YI=Sensitivity%+Specificity%-100,
14 Cohen Kappa Score=po-pe1-pe,
where po is the empirical probability of agreement on the label assigned to the sample and pe is the predictable agreement when both annotators assign labels randomly.
The implementations also obtain Receiver Operating Characteristic (ROC) curve, which is the graph between True Positive Rate (TPR) and False Positive Rate (FPR). It also displays the indicative ability of the model. Another performance metric Area Under Curve (AUC) is present under the ROC curve. AUC provides the sum of evaluated performance across thresholds of all possible classification.
Experimental Results
Experimental Setup
As mentioned earlier, the dataset is split into different sets for training and testing, respectively. After implementing the XGBoost model with tuned hyperparameters, performance metrics have been obtained. For the simulation, Python packages and Keras libraries with Tensorflow 1.7 have been used on an Intel Core (TM) i5-2.2 GHz processor.
Result Analysis of EHR File
As various literature related to machine learning discusses the dependence of performance results of a model on the size of dataset considered, an analysis by taking different sets of dataset has been performed to rule out the possibility of overfitting or over constraint conditions [37]. The dataset has been divided into three different sets of train-test data to obtain the performance metrics by implementing the XGBoost model. The details have been discussed in Table 3. This analysis indicates the efficiency of the proposed model by acquiring satisfying results in all the training–testing sets. Figure 3 depicts the confusion matrix of the test phase of XGBoost architecture for pulmonary fibrosis classification with 80% training data. Fibrosed cases were labeled as 1, while normal cases were labeled as 0. Among the 1170 patient data, only 1 was misclassified as false positives and 1 was misclassified as false negatives. Furthermore, in Fig. 4, the ROC curve is plotted between true positive rate and false positive rate for the 80:20 split of dataset to compare the overall performance of the model. The AUC was calculated to be 1.00.Table 3 Performance of the XGBoost model for different sets of training and testing data from EHR dataset
Train data (%) Accuracy (%) Precision (%) Recall Specificity F1 score MCC κ YI
70 99.11 ± 0.32 99.16 ± 0.25 99.42 ± 0.28 99.79 ± 0.27 99.79 ± 0.27 99.06 ± 0.25 99.58 ± 0.24 99.21 ± 0.47
75 99.28 ± 0.37 99.36 ± 0.36 99.58 ± 0.39 99.78 ± 0.46 99.80 ± 0.36 99.27 ± 0.38 99.49 ± 0.36 99.36 ± 0.12
80 99.37 ± 0.83 99.54 ± 0.35 99.85 ± 0.22 99.94 ± 0.26 99.82 ± 0.57 99.36 ± 0.27 99.70 ± 0.46 99.44 ± 0.50
Fig. 3 Confusion matrix of the proposed system with 80% training data
Fig. 4 ROC analysis of the proposed system with 80% training data
The hyperparameter of XGBoost scale_pos_weight is used to tune the behavior of the algorithm for an imbalanced dataset with great efficiency. In default conditions, the parameter scale_pos_weight is set to 1.0 and has the significance of keeping the balance of positive examples, relative to the negative examples when boosting model’s decision trees. A feature importance graph, presented in Fig. 5, was also plotted to recognize the significant features of clinical data. Importance delivers a specific score that specifies how beneficial each feature was in the building of decision trees based on boosting, within the model. It is calculated explicitly for each attribute of dataset, allowing attributes to be compared and ranked accordingly. The performance measure used to select the split points is the Gini index. Importance of all the features is then averaged across all decision trees within the XGBoost model.Fig. 5 Feature importance graph of clinical dataset
Result analysis of HRCT-Scans
The XGBoost model was used to classify Fibrosed and normal lungs from the HRCT dataset. Again, three independent sets of train–test data have been considered to acquire the performance metrics. The details have been discussed in Table 4. By obtaining satisfactory results in all the training–testing sets, this analysis demonstrates the efficacy of the proposed model with HRCT scans as well. The confusion matrix for the test phase of the XGBoost architecture for pulmonary fibrosis classification with 80% training data is shown in Fig. 6. The area under ROC curve, obtained in Fig. 7, for the case of HRCT images has also attained the value of 1.00.Table 4 Performance of the XGBoost model for different sets of training and testing data
Train data (%) Accuracy (%) Precision (%) Recall Specificity F1 score MCC κ YI
70 98.01 ± 0.37 98.16 ± 0.35 98.02 ± 0.48 98.70 ± 0.37 98.79 ± 0.15 98.08 ± 0.35 98.54 ± 0.25 96.72 ± 0.34
75 98.08 ± 0.56 98.26 ± 0.58 98.06 ± 0.25 98.81 ± 0.26 98.88 ± 0.58 98.60 ± 0.65 98.75 ± 0.66 96.87 ± 0.25
80 98.48 ± 0.81 99.09 ± 0.84 99. 13 ± 0.83 98.64 ± 0.81 98. 35 ± 0.91 98.02 ± 0.90 99.35 ± 0.91 97.77 ± 0.15
Fig. 6 Confusion matrix of the proposed system with 80% training data
Fig. 7 ROC analysis of the proposed system with 80% training data
Comparison with Other Machine Learning Models
There has been boundless development in machine learning over the decades. The recent inclination of researchers is toward deep learning that desires a dataset comprising a huge number of attributes. For clinical cases like the pulmonary fibrosis dataset associated with COVID-19, the dataset is much smaller, particularly after deleting the invalid data points. In the available literature, it is known that tree-based algorithms, SVM and regression models perform well with small datasets. Thus, few standard machine learning algorithms are considered to do a performance comparison with the proposed methodology. Models, such as Support vector machine (SVM), Naïve Bayes, Decision Tree, Random Forest, Logistic Regression, XGBoost and the proposed optimized XGBoost, have been implemented on the pulmonary fibrosis patient dataset and their performances have been evaluated [38–45]. Among all, the optimized XGBoost presented pleasing results because its tuned parameters and chosen for the final classification task. The detailed metrics comparison based on EHR data has been presented in Table 5, and analysis based on HRCT scans has been presented in Table 6.Table 5 Comparison of machine learning models with EHR dataset
Model Accuracy Precision Recall Specificity F1 score MCC κ YI
SVM 84.48 ± 0.81 84.56 ± 0.82 86.80 ± 0.81 85.23 ± 0.80 84.87 ± 0.02 84.84 ± 0.73 84.49 ± 0.73 72.03 ± 0.25
Logistic regression 87.39 ± 0.83 87.49 ± 0.84 87.75 ± 0.82 88.13 ± 0.83 88.88 ± 0.02 88.43 ± 0.80 88.40 ± 0.74 76.05 ± 0.59
Naïve bayes 90.28 ± 0.87 90.35 ± 0.91 90.65 ± 0.87 90.02 ± 0.90 90.35 ± 0.02 90.64 ± 0.81 90.29 ± 0.75 80.67 ± 0.08
Decision tree 92.59 ± 0.51 92.70 ± 0.47 92.92 ± 0.46 92.28 ± 0.45 92.59 ± 0.02 92.05 ± 0.66 92.60 ± 0.65 85.2 ± 0.24
Random forest 95.94 ± 0.02 94.06 ± 0.25 94.11 ± 0.26 95.58 ± 0.24 95.16 ± 0.25 95.42 ± 0.28 95.79 ± 0.27 89.69 ± 0.06
XGBoost 97.56 ± 0.65 97.08 ± 0.44 97.61 ± 0.89 97.06 ± 0.28 97.78 ± 0.63 97.85 ± 0.36 87.79 ± 0.27 94.67 ± 0.55
Optimized XGBoost (proposed) 99.37 ± 0.83 99.54 ± 0.35 99.85 ± 0.22 99.94 ± 0.26 99.82 ± 0.57 99.36 ± 0.27 99.70 ± 0.46 99.79 ± 0.36
Table 6 Comparison of machine learning models with HRCT scan dataset
Model Accuracy Precision Recall Specificity F1 score MCC κ YI
SVM 77.55 ± 0.02 77.56 ± 0.82 76.50 ± 0.68 77.25 ± 0.80 77.87 ± 0.12 76.74 ± 0.23 77.69 ± 0.13 53.75 ± 0.36
Logistic regression 80.23 ± 0.72 80.35 ± 0.91 80.65 ± 0.87 80.02 ± 0.90 80.35 ± 0.02 80.64 ± 0.81 80.29 ± 0.75 60.67 ± 0.23
Naïve bayes 87.94 ± 0.02 87.06 ± 0.25 87.11 ± 0.26 87.58 ± 0.24 87.16 ± 0.25 87.42 ± 0.28 87.79 ± 0.27 74.69 ± 0.15
Decision tree 89.18 ± 0.27 90.05 ± 0.19 90.05 ± 0.10 89.02 ± 0.11 89.15 ± 0.42 89.34 ± 0.41 89.09 ± 0.05 79.07 ± 0.55
Random forest 92.83 ± 0.24 92.75 ± 0.62 92.70 ± 0.25 92.09 ± 0.86 92.75 ± 0.55 92.13 ± 0.80 92.88 ± 0.22 85.45 ± 0.30
XGBoost 96.39 ± 0.34 96.05 ± 0.12 97.13 ± 0.33 96.88 ± 0.62 96.35 ± 0.47 96.12 ± 0.50 96.05 ± 0.42 94.18 ± 0.07
Optimized XGBoost (proposed) 98.48 ± 0.81 99.09 ± 0.84 99. 13 ± 0.83 98.64 ± 0.81 98. 35 ± 0.91 98.02 ± 0.90 99.35 ± 0.91 97.77 ± 0.35
Discussion
At the time of pandemic, a severe insufficiency of diagnostic resources has been reported, even in the developed part of world. Timely detection of diseases could result in better treatment and could save lives. EHR is the digital version of patient’s reports. It contains patient’s medical history, diagnoses, medications, treatment plans, immunization dates, allergies, radiology images, and laboratory and test results. EHRs could easily be used to automate the diagnostic process as they are already in the digitized version. It can make the whole process efficient, fast and precise than other diagnostic methods. A classification study with 1175 High-resolution Computed Tomography (HRCT) scans has also been included, to add an additional analysis to the results acquired from EHRs. The progression of pulmonary fibrosis in COVID-19 patients is a difficult classification problem that necessitates the application of a powerful optimization algorithm and an efficient feature extraction process. Treatment decisions, prognostication, and research into the pathogenesis of pulmonary fibrosis can all be aided by automated diagnosis. The work is primarily an application of Extreme Gradient Boosting algorithm for detecting pulmonary fibrosis. The technical contribution includes statistical analysis of the dataset and hyperparameter tuning of the algorithm using grid search. The results attained were sufficient to develop a diagnostic tool for detecting pulmonary fibrosis. Effective pre-processing and statistical analysis have been implemented on the dataset to obtain consistent and uniform values.
The modified XGBoost approach was chosen for this study because it has outstanding scalability and a fast-running speed, making it an effective ML method. Furthermore, machine learning methods allow for the simultaneous assessment of several variables and their complex interactions, as well as nonlinearity in the development of predictive models. This strategy has been used to solve a variety of machine learning issues. XGBoost has been used to classify cancer patients, epilepsy patients, and to diagnose chronic renal disease in biomedical domains. The results suggest that utilizing our tuned XGBoost classification system, it is possible to discriminate between normal patients and fibrosed lung with high accuracy. Our proposed framework was found to have a maximum classification accuracy of 99 percent, suggesting the potential clinical utility of EHR and HRCT data to categorize pulmonary fibrosis lung patients. Tables 5 and 6 show how the proposed system was compared using various ML methods reported in the literature. The comparison of the systems revealed that the optimized XGBoost produced a significant improvement over the other approaches.
The SVM and Logistic Regression techniques performed poorly when compared to the other systems. Random Forest shows most near approximates to the proposed method’s recall and accuracy values. The proposed XGB system is capable of handling large data dimensions while avoiding overtraining. Though, there are few limitations present in the study. Due to the unavailability of dataset for post-COVID-19 prior to the onset date, the analysis of high-risk trajectory prediction could not be included in this study. The available dataset only includes test reports obtained after 90 days of hospital discharge. This is surely a work to be included in future scope. The suggested model could also be utilized with a large-scale dataset including people from various geographical areas and age groups. The model has the potential to be a trustworthy tool for automatic analysis to aid in the diagnosis of pulmonary fibrosis.
Conclusion
As the comorbidities and complications due to COVID-19 have increased exponentially, many developing countries faced acute medical resource shortages. Hence, there is a need to identify every single complication at an early stage, which will reduce the burden on the medical society and healthcare system. The proposed XGBoost system to detect pulmonary fibrosis in COVID-19 patients could significantly help clinicians to examine patients with fibrotic complications by analyzing the electronic health reports or HRCT scans. This machine learning model achieved an accuracy of 99% and gave the best performance in terms of other evaluation metrics when compared with Decision Tree (97%), SVM (94%), Random Forest (90%), Logistic Regression (83%) and Naïve Bayes (63%). The precision, recall, and accuracy of the suggested system in this paper are higher than those of other approaches. This ensures its accuracy when it comes to the automatic classification of the pathology in this investigation. Finally, it is critical to note that the XGB approach has attained great performance, implying that this system will aid physicians in their decision-making.
Author Contributions
MJ was the principal author of this article. The literature search, data analysis, and drafting were conducted by MJ. Dr. RG and Prof. RS critically revised and edited the work.
Funding
The authors declare that no funds, grants, or other support were received during the preparation of this manuscript.
Availability of Data and Material
The dataset analyzed during the current study is publicly available and mentioned in the manuscript.
Code Availability
The source code is available from the corresponding author on reasonable request.
Declarations
Conflict of Interest
The authors have no relevant financial or non-financial interests to disclose.
Ethical Approval
Not applicable.
Consent to Participate
Not applicable.
Consent for Publication
Not applicable.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Kannan S Shaik Syed Ali P Sheeza A COVID-19 (novel coronavirus 2019)—recent trends SARS Eur Rev Med Pharmacol Sci 2020 24 4 2006 2011 10.26355/eurrev_202002_20378 32141569
2. West CP Montori VM Sampathkumar P COVID-19 testing: the threat of false-negative results Mayo Clin Proc 2020 95 6 1127 1129 10.1016/j.mayocp.2020.04.004 32376102
3. Oyebode O Ndulue C Mulchandani D COVID-19 pandemic: identifying key issues using social media and natural language processing J Healthc Inform Res 2022 10.1007/s41666-021-00111-w
4. Centers for Disease Control and Prevention. Novel coronavirus (2019-nCoV), Wuhan, China. 2019. https://www.cdc.gov/coronavirus/2019-nCoV/summary.html.
5. Wesner JS Van Peursem D Flores JD Forecasting hospitalizations due to COVID-19 in South Dakota, USA J Healthc Inform Res 2021 5 218 229 10.1007/s41666-021-00094-8 33969258
6. Jha M, Gupta R, Saxena R. A review on non-invasive biosensors for early detection of lung cancer. In 6th International Conference on Signal Processing and Communication (ICSC). 2020. 10.1109/ICSC48311.2020.9182775.
7. Song JW Do KH Kim MY Jang SJ Colby TV Kim DS Pathologic and radiologic differences between idiopathic and collagen vascular disease-related usual interstitial pneumonia Chest 2009 136 1 23 30 10.1378/chest.08-2572 19255290
8. Liu GY Ventura IB Achtar-Zadeh N Elicker BM Jones KD Wolters PJ Collard HR Adegunsoye A Strek ME Ley B Prevalence and clinical significance of antineutrophil cytoplasmic antibodies in North American patients with idiopathic pulmonary fibrosis Chest 2019 156 4 715 723 10.1016/j.chest.2019.05.014 31181198
9. Grasselli G Pesenti A Cecconi M Critical care utilization for the COVID-19 outbreak in Lombardy, Italy J Am Med Assoc. 2020 323 16 1545 1546 10.1001/jama.2020.4031
10. Gentile F Aimo A Forfori F Catapano G Clemente A Cademartiri F Emdin M Giannoni A COVID-19 and risk of pulmonary fibrosis: the importance of planning ahead Eur J Prev Cardiol. 2020 27 13 1442 1446 10.1177/2047487320932695 32551971
11. Da BL Im GY Schiano TD COVID-19 hangover: a rising tide of alcohol use disorder and alcohol-associated liver disease Hepatology 2020 10.1002/hep.31307
12. Zhang P Li J Liu H Long-term bone and lung consequences associated with hospital-acquired severe acute respiratory syndrome: a 15-year follow-up from a prospective cohort study Bone Res 2020 10.1038/s41413-020-0084-5
13. Li Y Li H Liu S Pan P Su X Tan H Wu D Zhang L Song C Dai M Li Q Mao Z Long Y Hu Y Hu C Pirfenidone ameliorates lipopolysaccharide-induced pulmonary inflammation and fibrosis by blocking NLRP3 inflammasome activation Mol Immunol. 2018 99 134 144 10.1016/j.molimm.2018.05.003 29783158
14. Saha A Vaidya PJ Chavhan VB Achlerkar A Leuppi JD Chhajed PN Combined pirfenidone, azithromycin and prednisolone in post-H1N1 ARDS pulmonary fibrosis Sarcoidosis Vasc Diffuse Lung Dis. 2018 35 1 85 90 32476885
15. Collins BF Raghu G Antifibrotic therapy for fibrotic lung disease beyond idiopathic pulmonary fibrosis Eur Respir Rev. 2019 28 153 190022 10.1183/16000617.0022-2019 31578210
16. Bruns D Kraguljac N Bruns T COVID-19: facts, cultural considerations, and risk of stigmatization J Transcult Nurs 2020 31 4 1 7 10.1177/1043659620917724
17. Carfì A Bernabei R Landi F For the Gemelli against COVID-19 post-acute care study group. Persistent symptoms in patients after acute COVID-19 J Am Med Assoc 2018 324 6 603 605 10.1001/jama.2020.12603
18. Zhao YM Shang YM Song WB Follow-up study of the pulmonary function and related physiological characteristics of COVID-19 survivors three months after recovery E Clin Med 2020 25 100463 10.1016/j.eclinm.2020.100463
19. Liu C Ye L Xia R Zheng X Yuan C Wang Z Lin R Shi D Gao Y Yao J Sun Q Wang X Jin M Chest Computed Tomography and Clinical Follow-Up of Discharged Patients with COVID-19 in Wenzhou City, Zhejiang China. Ann Am Thorac Soc. 2020 17 10 1231 1237 10.1513/AnnalsATS.202004-324OC 32692945
20. Yasin R Gomaa AAK Ghazy T Predicting lung fibrosis in post-COVID-19 patients after discharge with follow-up chest CT findings Egypt J Radiol Nucl Med 2021 10.1186/s43055-021-00495-0
21. Lichter Y Topilsky Y Taieb P Lung ultrasound predicts clinical course and outcomes in COVID-19 patients Intensive Care Med 2020 46 1873 1883 10.1007/s00134-020-06212-1 32860069
22. Li M Chest CT features and their role in COVID-19 Radiol Infect Dis. 2020 7 2 51 54 10.1016/j.jrid.2020.04.001 32309528
23. Spagnolo P Balestro E Aliberti S Pulmonary fibrosis secondary to COVID-19: a call to arms? Lancet Respir Med 2020 10.1016/S2213-2600(20)30222-8
24. Ali RMM Ghonimy MBI Post-COVID-19 pneumonia lung fibrosis: a worrisome sequelae in surviving patients Egypt J Radiol Nucl Med 2020 10.1186/s43055-021-00484-3
25. Chen R Gao Y Chen M Jian W Lei C Zheng J Li S Impaired pulmonary function in discharged patients with COVID-19: more work ahead Eur Respir J 2020 56 1 2002194 10.1183/13993003.02194-2020 32586883
26. Das KM Lee EY Singh R Follow-up chest radiographic findings in patients with MERS-CoV after recovery Indian J Radiol Imag 2020 27 3 342 349
27. Yu M Liu Y Xu D Zhang R Lan L Xu H Prediction of the development of pulmonary fibrosis using serial thin-section CT and clinical features in patients discharged after treatment for COVID-19 pneumonia Korean J Radiol 2020 21 6 746 755 10.3348/kjr.2020.0215 32410413
28. Diagnosis and Treatment of COVID-19. In: National Health Commission of China. 2020. http://www.nhc.gov.cn/. Accessed 15 Sep 2021.
29. Zou Y Wang BR Sun L Xu S Kong YG Shen LJ The issue of recurrently positive patients who recovered from COVID-19 according to the current discharge criteria: investigation of patients from multiple medical institutions in Wuhan, China J Infect Dis 2020 222 11 1784 1788 10.1093/infdis/jiaa301 32491178
30. George P Patterson C Ak R Lung transplantation for idiopathic pulmonary fibrosis Lancet Respir Med 2020 7 3 271 282 10.1016/S2213-2600(18)30502-2
31. Jaiswal AK Tiwari P Kumar S Identifying pneumonia in chest X-rays: a deep learning approach Measurement. 2020 145 511 518 10.1016/j.measurement.2019.05.076
32. Stony Brook University COVID-19 Positive Cases. In: The Cancer Imaging Archive (TCIA) Public Access. Available via https://wiki.cancerimagingarchive.net/pages/viewpage.action?pageId=89096912. 2020. Accessed 15 Sep 2021.
33. Friedman J Hastie T Tibshirani R The elements of statistical learning 2001 New York Springer series in statistics
34. Friedman JH Stochastic gradient boosting Comput Stat Data Anal 2002 38 4 367 378 10.1016/S0167-9473(01)00065-2
35. Matin S Fouladi N Pahlevan Y The sufficient vitamin D and albumin level have a protective effect on COVID-19 infection Arch Microbiol 2021 10.1007/s00203-021-02482-5
36. Chen T, Guestrin C. Xgboost: a scalable tree boosting system. In: Proceedings of the 22nd acm sigkdd international conference on knowledge discovery and data mining. ACM. 785–794. 2016.
37. Chen X Huang L Xie D EGBMMDA: extreme gradient boosting machine for miRNA-disease association prediction Cell Death Dis 2018 9 1 3 10.1038/s41419-017-0003-x 29305594
38. XGBoost. In: Wikipedia. Available via https://en.wikipedia.org/wiki/XGBoost. 2016. Accessed 15 Sep 2021.
39. Jha M Gupta R Saxena R A framework for in-vivo human brain tumor detection using image augmentation and hybrid features Health Inf Sci Syst 2022 10.1007/s13755-022-00193-9
40. Karadeniz T Tokdemir G Maraş HH Ensemble methods for heart disease prediction New Gener Comput 2021 39 569 581 10.1007/s00354-021-00124-4
41. Liu L Yu Y Fei Z An interpretable boosting model to predict side effects of analgesics for osteoarthritis BMC Syst Biol 2018 12 105 10.1186/s12918-018-0624-4 30463545
42. Miller A Subset selection in regression 2002 Boca Raton CRC Press
43. Cunningham P, Delany SJ. k-Nearest neighbour classifiers, Technical Report UCD-CSI-2002. 2007;4:1–17.
44. Evgeniou T, Pontil M. Support vector machines: theory and applications, ACAI 1999: Machine Learning and Its Applications. 2001;249–57.
45. Breiman L Random forests Mach Learn 2007 45 5 15 10.1023/A:1010933404324
| 0 | PMC9746584 | NO-CC CODE | 2022-12-15 23:21:56 | no | SN Comput Sci. 2023 Dec 13; 4(1):89 | utf-8 | SN Comput Sci | 2,022 | 10.1007/s42979-022-01526-x | oa_other |
==== Front
Soc Indic Res
Soc Indic Res
Social Indicators Research
0303-8300
1573-0921
Springer Netherlands Dordrecht
3047
10.1007/s11205-022-03047-9
Original Research
How to Enhance Citizens’ Sense of Gain in Smart Cities? A SWOT-AHP-TOWS Approach
Li Dezhi [email protected]
1
Wang Wentao [email protected]
1
Huang Guanying [email protected]
1
Zhou Shenghua [email protected]
1
Zhu Shiyao [email protected]
2
Feng Haibo [email protected]
3
1 grid.263826.b 0000 0004 1761 0489 Department of Construction and Real Estate, School of Civil Engineering, Southeast University, Nanjing, 210018 China
2 grid.260483.b 0000 0000 9530 8833 School of Transportation and Civil Engineering, Nantong University, Nantong, 226007 China
3 grid.42629.3b 0000000121965555 Department of Mechanical and Construction Engineering, Northumbria University, Newcastle Upon Tyne, UK
13 12 2022
134
30 11 2022
© The Author(s), under exclusive licence to Springer Nature B.V. 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
The previous technology-centric development of smart cities mainly focuses on the numbers, diversities, and types of applied intelligent technologies, while the citizen-centric smart city has become an important paradigm for improving the sustainability of cities around the world. The citizens’ sense of gain (CSG), which considers both material acquisition and spiritual feelings of smart city services, is thus proposed and regarded as one of the core orientations in the smart cities’ transformation development process from the centric of advanced technology applied to the centric of citizen subjective perception. To shift smart cities from being technology-centric to citizen-centric, it is critical to identify the factors influencing CSG and develop appropriate strategies to enhance CSG in smart cities. Hence, this work identifies 17 key CSG influencing factors based on the dimensions dissected from the definition of CSG and it further formulates 15 strategies for enhancing CSG by adopting the SWOT-AHP-TOWS method based on data collected from Nanjing citizens. The results indicate that the most important criteria for enhancing CSG in smart cities are the external opportunities, which are originated from citizens’ attitudes and behaviors, and the top-ranked strategy is “dividing smart infrastructure into different categories according to the hierarchy needs of citizens and promoting the synergy development of smart infrastructure within and among different categories”. Finally, four implications are proposed, including (i) strengthening publicity and encouraging citizen participation, (ii) clarifying the responsibilities of local governments, (iii) prioritizing citizens’ needs, and (iv) promoting age-friendly, vulnerable-friendly, and environmental-friendly development.
Keywords
Citizen-centric
Smart city
Citizens’ sense of gain
Influencing factor
SWOT-AHP-TOWS
China
http://dx.doi.org/10.13039/501100012456 National Social Science Fund of China 19BGL281 Li Dezhi
==== Body
pmcIntroduction
With the continuous growth of population and development of society, a series of problems have arisen in cities around the world, such as environmental pollution, data security, food safety, traffic chaos, and energy waste (Feng et al., 2020; Hu, 2019). The “smart city” has been reckoned as a promising solution for these issues to make cities more sustainable and habitable (Sharif & Pokharel, 2022). Smart city integrates information and communications technology (ICT) with conventional infrastructure and coordinates new digital technologies (Freudendal-Pedersen et al., 2019). As of 2021, over 1300 cities worldwide have proposed plans, acts, and initiatives concerning smart city development (Jang & Gim, 2022). Therefore, it can be seen that the development of smart cities has become an essential paradigm for the sustainability of cities all over the world.
The majority of smart cities’ developments at the early stages are technology-centric, (Liu et al., 2021), which focus on the deployments, applications, and innovations of ICT, such as Internet of Things (IoT) in e-commerce, cloud computing in transportation infrastructure, public service platform in government services, and big data in smart healthcare (Li et al., 2015; Long & Thill, 2015; Zhang et al., 2016). The concept of smart cities tended to be technocratic. The number, diversity, and scale of applied intelligent technology became the trademark of smart cities’ developments, while citizens’ feelings were left behind (Marsal-Llacuna, 2016). Such technology-centric smart city development may bring a series of demerits (e.g., incomplete consideration of citizens’ needs, invasion of privacy, exacerbated unfairness among different age groups, and low citizen satisfaction) (Ji et al., 2021). In response to the aforementioned problems brought by technology-centric smart cities, the development of smart cities shift from being technology-centric to citizen-centric for enhancing the spiritual well-being of citizens, which emphasizes citizen perceptions in addition to the technology applications in smart cities. The citizen-centric smart city was regarded as a new paradigm for smart city sustainable development (Krivy, 2018; Yigitcanlar et al., 2019). Many countries have started to explore the development of citizen-centric smart cities. For example, the citizen-involved governance structure in the U.S (Hu & Zheng, 2021), a series of citizen-centric policies to monitor technological innovation in the UK (Cowley et al., 2018), and citizen-centric smart city development plans in Japan and Singapore (Asher et al., 2015; Chatfield & Reddick, 2016).
As the largest developing country in the world, China has been advocating the development of smart cities since 2010 (Wang et al., 2020a, 2020b). Since the Ministry of Housing and Urban–Rural announced the first batch of 90 pilot smart cities in 2013, three-quarters of the cities at the prefecture or above level related to smart construction and digital services have been carried out in China (Song et al., 2022; Wang et al., 2021; Zhu et al., 2019). Beyond the applications of advanced technologies in smart cities, the Chinese government has begun to actively set up strategies to develop citizen-centric smart cities to achieve a shift from technology-centric (technicalism) to citizen-centric (humanism) (Zhu et al., 2019). In 2016, the “National Informatization Development Outline” promulgated by the State Council (The State Council of the PRC, 2016), for the first time included a new type of smart city in the policy while emphasizing “prioritizing public’s sense of gain and breaking through technology-centric theory” as the core goal. A range of national policies, such as The report of the 19th National Congress of the Communist Party of China and The 14th Five-Year Plan have been proposed to break the technocentric development mode and emphasize the improvement of the citizen sense of gain (CSG). CSG is defined as the sense of obtaining based on the satisfaction of material benefits and spiritual benefits (Feng & Zhong, 2021; Gu et al., 2020), and it holds the potential to help measure the effectiveness of citizen-centric development efforts in China (Wan & Guo, 2021). According to its definition, CSG can be divided into citizens’ sense of material gain and citizens’ sense of spiritual gain. Citizens’ sense of material gain refers to citizens’ material benefits obtained from the higher material level of living brought by smart city services, while citizens’ sense of spiritual gain refers to citizens’ mental well-being towards social and living in the background of smart city development (D’Acci, 2021; Gu et al., 2020; Huang et al., 2022; Ruan et al., 2022; Wang et al., 2022). In essence, citizens-centric smart cities are seeking to reposition advanced technologies in a way that improve what citizens subjectively feel and objectively acquire (König, 2021). The incorporation of CSG into the development of smart cities can facilitate sustainable development (Macke et al., 2018). Consequently, the issue that how to enhance citizens’ sense of gain in smart cities is an urgent need to be addressed when developing citizen-centric smart cities (Huang et al., 2022). However, despite the CSG concept being proposed, the influencing factors of CSG in smart city development are still under-investigated, and there is also a lack of strategies for enhancing CSG formulated from the perspective of citizens.
To fill these gaps, this paper aims to classify the influencing factors of CSG in smart cities and then use the SWOT-AHP-TOWS methods to analyze the strategies for enhancing CSG in smart cities. This work consists of three major steps, including (i) identifying influencing factors of CSG in smart cities, (ii)figuring out the strengths, weaknesses, opportunities, and threats of smart cities development from the perspective of CSG, and (iii) formulating strategies and providing suggestions for enhancing CSG in smart cities. The remainder of this paper is organized as follows. Section 2 reviews the literature on citizen-centered smart city, CSG, and the strategies for enhancing CSG of different domains. Section 3 presents the methodology step by step. Section 4 shows a case study of the sample city and the results of the SWOT-AHP-TOWS analysis. Section 5 discusses the results and policy implications.
Literature Review
Citizen-Centric Smart City
Smart cities’ developments are criticized as being excessively technocratic which may not bring tangible and expected benefits to citizens (Cardullo & Kitchin, 2019; Kitchin, 2019). The topics of “citizens’ benefits”, “citizens’ feelings”, and “humanistic” concerning smart cities have received increasing attention (Georgiadis et al., 2021; Zandbergen & Uitermark, 2020), and the citizen-centric smart city has gradually become a consensus among the public, scholars, and decision-makers (König, 2021). The material acquisition and spiritual feelings of citizens are both important criteria to reflect the citizen-centric level of a smart city (Ju et al., 2018). However, most studies have only analyzed the situation of citizens from a material or spiritual perspective. For example, the citizens’ material acquisition in smart cities has been analyzed from the smart environment, smart people, smart livelihood, smart economy, and smart governance aspects considering the citizens’ quality of life (Chen & Chan, 2022; Macke et al., 2018). From the perspective of citizens’ life in the smart city, the aspects of transportation, healthcare, safety, education, and environment were considered to assess the level of citizens’ material acquisition (Shami et al., 2022). Meanwhile, citizens’ spiritual feelings in smart cities have been evaluated from smart public service, smart public administration, and local culture integration aspects regarding the citizens’ satisfaction (Xu & Zhu, 2021; Yu et al., 2020). Some scholars have considered the consciousness of citizens from the income level and price level in smart cities (Lin et al., 2019). There were also many scholars who have adopted citizen participation (i.e., degree of citizen participation, approach of citizen participation, and feelings of citizen participation) as an important dimension to measure the spiritual feelings of citizens in smart cities (Feng, 2019; Guimaraes et al., 2020; Shami et al., 2022). Overall, although the existing researches cover many aspects (including citizens’ daily lives, citizen participation, citizen quality, and citizen satisfaction) in smart cities, there is still a lack of analysis of the effectiveness of smart cities development with the consideration of both citizens’ material acquisition and spiritual feelings.
Citizens’ Sense of Gain (CSG)
The Citizen’s sense of gain (CSG) is an emerging concept, and it has been applied to various domain-specific matters. There is a consensus that CSG is the combination of “sense of spiritual gain” and “sense of material gain” (Feng & Zhong, 2021; Gu et al., 2020). The sense of spiritual gain refers to one’s overall feelings regarding the advantages of economic and social growth, such as the right to enjoy fairness and justice, realizing self-worth and social value, and a rise in economic and social status. The sense of material gain refers to the feelings brought to people by objective material conditions, such as education, transportation, housing, medical care, and social security (Wan & Guo, 2021; Xie et al., 2020). Jia et al. (2022) established a fuzzy comprehensive evaluation model for farmers’ sense of gain in providing rural infrastructure and verified the validity of the model through investigation. Su and Li (2022) stated that subjective socioeconomic status has a positive statistical correlation with sense of gain in health-care. Despite an increasing number of CSG studies in various domain topics, there is currently a lack of attention and efforts on the influencing factors of CSG in smart cities.
Strategies for Enhancing CSG
To achieve citizen-centric, more and more studies focus on the CSG enhancement strategies. Feng and Zhong (2021) used a structural equation model to analyze the relationship between college students’ sense of gain, sense of security, and happiness, and proposed strategies (participate in social activities and enhance communication with others) to enhance college students’ sense of gain. Gu et al. (2020) developed a comprehensive framework for the concept of Employee Sense of Gain and proposed strategies to enhance environmental, social, and corporate governance ESG. Focusing on the link between capability deprivation and the subjective sense of gain of rural families, Huo et al. (2022) used the factor mixture model to analyze the group categories of capability deprivation and ordered probit regression to estimate the associations between the categories of ability deprivation and sense of gain. Wan & Gu (2021) analyzed the current situation of the demand for sense of gain, and proposed strategies (construct high-quality courses, cultivate people’s responsibilities, and implement “soft elimination” of training links) for enhancing the sense of gain of food science students. Xie et al., (2020) analyzed the impact mechanism of the digital business penetration rate of traditional villages in western China on farmers’ sense of economic gain through a combination of qualitative and quantitative study, and proposed strategies (improve farmers’ entrepreneurial intention, and enhance farmers’ attitude toward digitization) for improving farmers’ sense of gain. Although scholars have studied the strategies for enhancing the sense of gain across various groups of people for specific matters, there is limited research on the strategies for enhancing CSG in smart cities.
Progress and Gaps
In general, the existing research regarding either CSG or citizen-centric smart cities provides a firm foundation for our work, but few of them paid attention and efforts to the influencing factors of CSG in smart cities, as well as the strategies for enhancing CSG in smart city development. To fill such gaps, this paper proposes to identify the influencing factors of CSG in smart cities by extensively reviewing policies and publications and integrate SWOT-AHP-TOWS methods to analyze the CSG promotion strategies oriented to smart cities.
Methodology
A synthetic approach integrating qualitative and quantitative methods is proposed for critical influencing factors identification of CSG and the CSG enhancement strategy formulation in smart cities (Fig. 1). The devised approach consists of three major steps, (i) Identify the influence factors of CSG in smart cities, (ii) conduct a two-stage questionnaire survey to identify the SWOT criteria, and (iii) adopt the AHP-TOWS method to analyze the strategies for enhancing CSG in smart cities.Fig. 1 The flow chart of the methodology
Identify the Influencing Factors of CSG in Smart Cities
According to the related works on CSG and citizen-centric smart cities, this paper identifies the influencing factors of CSG in smart cities from two aspects: the sense of spiritual gain and the sense of material gain. The factors influencing citizens’ sense of material gain mainly refer to the smart city services and living conditions which affect their material acquisition. The factors influencing citizens’ sense of spiritual gain refer to the citizens’ perceptions and feelings toward smart cities (D’Acci, 2021; Gu et al., 2020; Huang et al., 2022; Ruan et al., 2022; Wang et al., 2022). The 17 influencing factors are identified from 6 dimensions, including CSG on public services, economic conditions, government affairs, safety, self-perception, and belonging feelings (Table 1).Table 1 The influencing factors of CSG in smart cities
Aspects of CSG Dimensions Influencing factors Explanation References
Citizens’ sense of material gain CSG on public services Public education Influence of smart education (e.g., MOOCs. A large number of students study by MOOCs during the COVID-2020.) on citizens’ acquisition for public education services Kranjac et al. (2021); Hudson et al. (2019); Williamson, (2017)
Healthcare Influence of smart healthcare (e.g., e-doctor, remote health monitoring) on citizens’ acquisition for public healthcare services Yu et al. (2020); Zhang et al. (2021); Trencher and Karvonen, (2019)
Transportation Influence of smart transportation (e.g., electronic navigation, smart parking, online car-hailing) on citizens’ acquisition for transportation services Jan et al. (2020); Peng et al. (2017); Carter et al. (2020)
Environmental governance Influence of smart environmental governance (e.g., air quality monitoring, water pollution monitoring) on citizens’ demand for natural environment Lin et al. (2019); Shami et al. (2022); Chen and Chan, (2022); Nikolic and Yang, (2020)
Social guarantee services Influence of smart social guarantee services (e.g., housing, unemployment, insurance, and vulnerable groups) on citizens’ acquisition for social guarantee services Khatoun and Zeadally, (2017); Alsamhi et al. (2019); Shami et al. (2022); Xie et al. (2019); Sookhak et al. (2019)
Aging services Influence of smart aging services (e.g., ageing-friendly facilities) on citizens’ acquisition for retirement services Ziganshina et al. (2020); Ivan et al. (2020); Li and Woolrych, (2021)
CSG on economic conditions Income level Influence of citizens’ income level in smart cities on citizens’ smart city living conditions Chanak and Banerjee, (2021); Pieroni et al. (2021)
Price level Influence of price level in smart cities on citizens’ smart city living conditions Pieroni et al. (2021); Chen and Chan, (2022)
CSG on government affairs Government online services Influence of smart government online services on citizens’ life convenience Valencia-Arias et al. (2021); Yeh, (2017); Vidiasova and Cronemberger, (2020); Dameri and Benevolo, (2016)
Political participation Influence of smart political participation on citizens’ demand for participation Feng, (2019); De Guimaraes et al. (2020); Shami et al. (2022); Simonofski et al. (2021); Szarek-Iwaniuk and Senetra, (2020)
CSG on safety Social public safety Influence of smart social public safety (e.g., electronic police, monitoring system) on citizens’ demand for public safety Alsamhi et al. (2019); Piro et al. (2014); Wereda et al. (2022)
Food hygiene and safety Influence of smart food hygiene and safety on citizens’ demand for food hygiene and safety Nagarajan et al. (2022); Ebenso et al. (2022)
Internet and data safety Influence of smart internet and data safety on citizens’ demand for internet and data safety Sookhak et al. (2019); Mohamed et al. (2020); Alsamhi et al. (2019); Braun et al. (2018)
Citizens’ sense of spiritual gain CSG on self- perception Self-worth Influence of citizens’ self-worth realization in smart cities on citizens’ spiritual feelings Feng and Zhong, (2021); Gu et al. (2020)
Socioeconomic status Influence of citizens’ socioeconomic status in smart cities on citizens’ spiritual feelings Xie et al. (2020); Wang et al. (2020a, 2020b)
CSG on belonging feelings Regional cultural integration Influence of regional cultural integration in smart cities on citizens’ spiritual feelings Xie and Yin, (2022); Yang and Ma, (2021)
Social fairness and justice Influence of right to enjoy social fairness and justice in smart cities on citizens’ spiritual feelings Bennati et al. (2018); Masucci et al. (2020)
Conduct a Two-Stage Questionnaire Survey and Determine the SWOT Sub-criteria
SWOT analysis is to enumerate various major internal strengths, weaknesses, and external opportunities and threats closely related to the research object through investigation, arrange them in the form of a matrix, and then uses the idea of systematic analysis to match various factors with each other analysis. It helps derive a series of corresponding conclusions for decision-making (Casebeer, 1993; Sharma & Bhatia, 1996). This approach is a well-known tool that many firms utilize to make better decisions and assess their strategic position (Dyson, 2004; Rizzo & Kim, 2005). In order to formulate strategies for enhancing CSG in smart cities, the characteristics of smart cities were regarded as the internal environment, while citizens’ attitudes and behaviors in smart cities were regarded as the external environment.
Based on the CSG influencing factors identified in the last step (Table 1), a questionnaire survey analyzing the effectiveness of smart city development from the perspective of CSG was distributed in a sample city.
The survey was divided into two stages. In the first stage, we used a questionnaire survey to ask citizens whether these smart city development supply indicators and CSG influencing factors can correspondingly have a positive impact on their CSG. The smart cities’ developments supply indicators were the integration of smart infrastructure and smart systems with all aspects of life (e.g., education, healthcare, transportation, and environment), according to “the plan to facilitate the development of the digital economy in the 14th Five-Year Plan period (2021–2025)” (The State Council of the PRC, 2021). A 5-level scale (−2,−1,0,1,2) was used, where −2 is the very strongly disagree, −1 is disagree, 0 means neutral, 1 is agree and 2 is strongly agree. The internal strengths and external opportunities of smart city development are summarized from the results with a score greater than the average, and the internal weaknesses and external threats are summarized from the results with a score less than the average (Shen et al., 2018).
Determine the Criteria/Sub-criteria Weights and Formulate Strategies
Based on the criteria and sub-criteria identified in the second step, the AHP method and TOWS method were used to determine the weights of criteria and sub-criteria and formulate strategies for enhancing CSG in smart cities.
The AHP analysis is used to complement the SWOT analysis. This is because it measures each aspect depending on its importance to the respective organization (Kim et al., 2017; Saaty, 1986). The four aspects of strength, weakness, opportunity, and threat are regarded as criteria in the AHP analysis, and the relative weights (RW) of each aspect are computed. Then, the sub-criteria (S1–S5, W1–W5, O1–O5, and T1–T5 in Table 2) for each criterion are compared in pairs in their own criterion group (Strengths, Opportunities, Weaknesses, and Threats in Table 2) to get the relative priority (RP in Table 4). Since RP is the priority of the sub-criteria within the criterion group, the total prioritization (TP in Table 4) requires multiplying RP by RW (Asadpourian et al., 2020). The TP of each sub-criteria is calculated in Eq. 1. The sub-criteria were ranked within the criterion group based on RP, while their total ranking among all sub-criteria is based on TP (Savari & Shokati Amghani, 2022).Table 2 The sub-criteria of smart city effectiveness from the perspective of GSG
Strengths Sources
(first stage) Approval rating
(second stage) (%)
S1: Improve citizens’ material quality of life Q1, Q2, Q3 89
S2: Provide a complete guarantee for citizens Q5 87
S3: Provide public safety protection for citizens Q12, Q14 85
S4: Provide a comfortable natural environment for citizens Q4 80
S5: Improve the convenience of citizens’ life Q16 83
Opportunities Sources Approval rating (%)
O1: Citizens’ increasing consumption level Q9, Q21 88
O2: Citizens’ high governmental institutional trust Q10, Q20, Q21 69
O3: Citizens’ ever-growing needs for a better life Q19, Q21 90
O4: Citizens’ high acceptance of the local government’s development planning Q18 71
O5: Citizens’ positive response to the national policy of benefiting the people Q20 85
Weaknesses Sources Approval rating (%)
W1: Low data synergy efficiency Q17 73
W2: Non-comprehensive legal system Q6 91
W3: Low urban resilience Q13 92
W4: Insufficient consideration of citizens’ needs Q24 93
W5: A regional imbalance in development Q23 97
Threats Sources Approval rating (%)
T1: Citizens’ low awareness of the connotation of the smart city Q8 78
T2: Citizens’ low willingness to participate in the development process Q11 62
T3: Citizens’ low sense of belonging Q22,Q23 70
T4: A high threshold for vulnerable groups to use public service Q7 85
T5: Citizens’ personal information data at risk Q15 73
The techniques for applying the AHP approach to derive weighting values between indicators are based on a judgment matrix. In this study, Experts are asked to rate each indicator on a paired basis using the Saaty numerical scale of 1 to 9 (Saaty et al., 2007). The indications with a higher number are more essential in the comparison judgment. Furthermore, the consistency ratio (CR in Table 4) is utilized to assess the judgment matrix’s sensitivity and consistency. According to Hummel et al. (2014), if CR > 0.1, the judgment matrix is irrational and must be re-determined.
An expertise committee from smart city development and administration, university faculty members, government officials, and business people are formed to decide the weighting and analyze the rationality of each indicator (step 3 in Fig. 1).
The TOWS analysis is conducted to derive CSG enhancement strategies. Hence, by using the TOWS matrix, strategies may be designed based on the identified strengths, weaknesses, opportunities, and threats (Gottfried et al., 2018; Seker et al., 2012). These strategies are developed by utilizing the strengths and possibilities of the various stakeholders while minimizing their weaknesses and risks. By combining each strength, weakness, opportunity, and threat, enhancement strategies are determined in four modes: SO, ST, WO, and WT. Each of the strategies is developed using a mix of sub-criteria at this step and the sub-criteria that make up each strategy come from different groups of criteria, thus to calculate the total weight (TW in Table 6) of each strategy, the sub-criteria weights between different groups must be multiplied (Asadpourian et al., 2020; Gottfried et al., 2018; Savari & Shokati Amghani, 2022). The priority of each strategy is ranked by TW (Savari & Shokati Amghani, 2022). The TWs of the four kinds (SO, ST, WO, and WT in Table 5) of strategy are respectively calculated in Eqs. 2–5, where TW refers to the total weight of a strategy, TP refers to the total prioritization of sub-criteria, n and m imply to the number of sub-criteria of a criterion group included in the strategy, and k means to the serial number of a certain type of strategy.
Here Eq. 2 is taken as the example to express the calculation process in detail, TWSOk (Eq. 2) refers to the total weight of the kth strategy in the SO strategy set, TPS1 means the total prioritization of S1, TPSi implies the total prioritization of the ith strength(S), TPO1 refers to the total prioritization of O1, TPOi means to the total prioritization of the ith opportunity(O), n refers the number of opportunities (sub-criteria in O criterion group) contained in the kth strategy in the SO strategy set, and m refers the number of strengths (sub-criteria in S criterion group) contained in the kth strategy in the SO strategy set.1 TP=RW×RP
2 TWSOk=fTPS1,TPS2,…,TPSm/TPO1,TPO2,…,TPOn=∑i=1m∑j=1nTPSi×TPOj
3 TWWOk=fTPW1,TPW2,…,TPWm/TPO1,TPO2,…,TPOn=∑i=1m∑j=1nTPWi×TPOj
4 TWSTk=fTPS1,TPS2,…,TPSm/TPT1,TPT2,…,TPTn=∑i=1m∑j=1nTPSi×TPTj
5 TWWTk=fTPW1,TPW2,…,TPWm/TPT1,TPT2,…,TPTn=∑i=1m∑j=1nTPWi×TPTj
Case Study
Case Area
Smart Nanjing City’s (SNC) development began early in China as one of the first pilot smart cities. It has excelled in smart transportation, smart education, smart aging services, smart governmance, and other smart developments. At the same time, Nanjing’s policies and measures for enhancing CSG (e.g., “Smart Nanjing” app to improve city services) are representative of China’s smart cities’ developments. In short, Nanjing’s experience will not only assist other cities of a similar scale (e.g., Suzhou, Zhengzhou, Qingdao.) in developing smart cities, but also give important lessons and references for other smart cities in China and elsewhere (Yuan et al., 2020).
SWOT Sub-criteria Identification Results
Before the surveys were conducted, a group of 25 experts, including smart city services providers, university researchers, and government authorities in the field of smart city, examined and refined the questionnaire to verify its comprehensiveness and validity. And then, the questionnaires were sent to a focus group of 94 citizens for a pilot investigation. These citizens, who were from different communities and had extensive experience using smart city services, were interviewed in-depth to ask for their input on changes to the questionnaires. After several revisions and feedback, the clarity of the questionnaires was improved.
SNC has an urban population of over 8.5 million, so in order to reach a 95% confidence level with a 5% confidence interval, at least 385 complete responses would need to be collected (Gu et al., 2019). To ensure the comprehensiveness of the sample, respondents from different types of communities in each district of Nanjing were selected randomly and contacted through the community residential committee. Respondents were asked to fill in online questionnaires based on their subject feelings toward smart city services. And the people who were disabled or too old to complete the questionnaire were interviewed patiently onsite and were helped to complete the paper questionnaire. The first stage questionnaire survey was carried out from April 1, 2022, to May 5, 2022, to preliminarily determine the SWOT criteria. A total of 751 questionnaires were collected, after removing short and incomplete questionnaires, 633 of which were valid, with an effective rate of 84.3%. And the second stage questionnaire survey was carried out from May 10, 2022, to June 5, 2022. A total of 503 questionnaires were collected, after removing short and incomplete questionnaires, 485 of which were valid, with an effective rate of 96.4%. The demographic distribution of the selected sample is shown in Table 7. Finally, based on the survey results, strengths, weaknesses, opportunities, and threats of smart cities’ developments from the perspective of CSG were identified.
Preliminarily Determine the SWOT Sub-criteria
Table 8 in Appendix showed the questions and survey results in the first stage. According to the average score of all the questions (Q1–Q24) and their relationship with the total average score, strengths, opportunities, weaknesses, and threats of smart cities’ developments from the perspective of CSG were determined (Fig. 2).Fig. 2 sub-criteria determination by Q1–Q24
To determine the sub-criteria, questions with an average score greater than the total average score were used to analyze the internal strengths and external opportunities. Similarly, questions with an average score smaller than the total average score were used to analyze the internal weaknesses and external threats (Fig. 2). The following examples (One each for S, W, O, T) illustrated the analysis process.
Internal strength (S): The average score of Q1, Q2, and Q3 were greater than the total average score, which showed that smart city development has improved citizens’ material quality of life.
Internal Weakness (W): The average score of Q17 was smaller than the total average score, indicating that low data synergy efficiency existed in smart city development.
External opportunities (O): According to Q9 and Q21, it could be speculated citizens’ consumption level will be increasing, which was regarded as beneficial to the development of smart cities (Rana et al., 2019).
External threats (T): The average score of Q22 and Q23 (smaller than the total average score) revealed that smart city development has made citizens’ low sense of belonging.
Based on the analysis like the above examples, the SWOT sub-criteria is preliminarily determined. Table 2 shows the strengths, weaknesses, opportunities,and threats of smart city development from the perspective of CSG, as well as their sources.
Finally Determine the SWOT Sub-criteria
In the second stage of the survey, a questionnaire was designed to investigate the citizens’ support rate for the results of the SWOT sub-criteria identified in the first stage. Each criterion had more than a 50% approval rating (last column of Table 2), so it can be considered effective.
The Relative Importance of Criteria and Sub-criteria for Enhancing CSG in Smart Cities Development
To determine the weights of the SWOT criteria and sub-criteria, thirteen experts made up the expertise committee in this study, and each expert has been working on the development of smart cities for more than five years. Table 3 reports detailed profiles of the experts. The final results are shown in Fig. 3 and Table 4.Table 3 Profile of the experts
Experts Organization Role
A Smart City Development Department in Government Director
B District government Deputy director
C Construction bureau in government Deputy director
D University Professor
E University Professor
F University Professor
G University Professor
H A large company related to smart city information technology Manager
I A large company related to smart city information technology Deputy manager
J A large construction company Manager
K A large construction company Manager
L A civil engineering association Director
M A civil engineering association Deputy director
Fig. 3 The weights of the SWOT criteria
Table 4 Relative priority and total priority of sub-criteria
Criteria RW
(Relative weight) Sub-criteria
(S1–S5, W1–W5, O1–O5, and T1–T5 in Table 2) RP
(Relative priority) TP
(Total prioritization) CR
(corresponding to the analysis of RP) CR
(corresponding to the analysis of RW)
Strengths 0.341 S1 0.296 0.101 0.04 0.03
S2 0.121 0.041
S3 0.180 0.061
S4 0.107 0.036
S5 0.296 0.101
Weaknesses 0.138 W1 0.187 0.026 0.03
W2 0.241 0.033
W3 0.103 0.014
W4 0.395 0.055
W5 0.074 0.010
Opportunities 0.432 O1 0.118 0.051 0.06
O2 0.097 0.042
O3 0.295 0.127
O4 0.218 0.094
O5 0.272 0.118
Threats 0.089 T1 0.108 0.010 0.07
T2 0.342 0.030
T3 0.077 0.007
T4 0.207 0.018
T5 0.266 0.024
As shown in Fig. 3, the RW of the SWOT criteria (S, W, O, T) corresponded to a CR of 0.03, which is lower than 0.1, proving that the analysis of RW was efficient. The opportunities (with weights of 0.4320) had the greatest influence on enhancing CSG in the development of smart cities. Other criteria in terms of weight include strengths (0.3409), weaknesses (0.1381), and threats (0.089). In this comparison, the inconsistency ratio is similarly smaller than 0.1. These findings suggest that, despite the strengths of smart city development (which are most closely connected to the characteristics of smart cities), opportunities (which originated from citizens’ characteristics and habits in smart cities) are the most essential factor for enhancing CSG. It indicates there are many opportunities to enhance CSG in the development of smart cities, which must be planned to turn into strengths.
After determining the prioritization of the SWOT criteria, the relative weights of each sub-criterion, relative prioritization, and total prioritization of enhancing CSG in smart city development were estimated (Table 4). In the last column of Table 4, the CR for each of the comparisons was shown, which was less than 0.1 in all four. Thus, none of the experts’ opinions contradicted those of the others.
Figure 4 depicts the relative weight and priority of each of these sub-criteria. The evaluation of internal items (strengths and weaknesses) revealed that “Improve the citizens’ material quality of life” and “Improve the convenience of citizens’ life” were ranked highest in strengths, while “Insufficient consideration of citizens’ needs” was placed at the top in weaknesses. Smart cities should develop corresponding intelligent systems along with the construction of infrastructure to enhance CSG, which can integrate various components of the city and make citizens’ life easier. At the same time, in the process of smart city development, it is also necessary to consider the needs of the citizens, so as to determine the supply factors of smart city development that match the needs of the citizens from the bottom-up, and ultimately achieve the goal of enhancing CSG.Fig. 4 The weights and ranks of the SWOT sub-criteria
In terms of the external parameters, “Citizens’ ever-growth needs for a better life” was the most important opportunity, while “Citizens’ low willingness to participate in the development process” was the main threat to the smart cities’ developments. These findings implied that although citizens had a continuous quest for a better living environment and city services, the government doesn’t provide appropriate avenues for citizens to participate, which resulted in a low willingness of citizens to participate. In order to enhance the CSG, the government should respond to the citizens’ pursuit of a better life and formulate relevant policies and participation avenues to motivate the citizens to participate in the development of smart cities. In the end, the citizens can realize their own interests in the process of participation, which will create a positive cycle to promote the citizens’ participation in the development of smart cities.
Develop and Rank Strategies for Enhancing CSG in Smart Cities Using the TOWS Matrix
At this phase, strategies for enhancing CSG in smart city development were developed by the strategic TOWS matrix. Different strategies were defined in four groups SO, ST, WO, and WT depending on the strategic space of the subject. After the formation of the Strategic Matrix SWOT, strategies SO, ST, WO, and WT were determined. According to the results in Table 5, a total of 14 strategies were developed to enhance CSG in SNC development.Table 5 The TOWS matrix for determining strategies to enhance CSG in smart cities
SO strategies ST strategies
SO1: Divide smart infrastructure into different categories according to the hierarchy needs of citizens and promote the synergy development of smart infrastructure within and among different categories ST1: Strengthen the publicity of smart city development and establish citizen participation paradigms that meet citizens’ participation interests and behaviors based on the functions of different departments
SO2: Create a convenient and safe consumption environment and promote citizens’ online e-commerce and offline smart services consumption ST2: Promote the age-friendly construction of smart cities
SO3: Clarify the role of local governments and departments in enhancing the CSG process for smart cities in terms of bottom-up analysis of local citizens' needs and top-down implementation of national policies ST3: Promote the integration of smart city development and regional cultural characteristics
SO4: Apply Internet and Internet of Things technologies to natural environment monitoring and promote the development of environmental-friendly smart cities ST4: Optimization of personal data protection for citizens in the smart systems and using blockchain technology to establish a multi-channel password lock mechanism for citizen information access
WO strategies WT strategies
WO1: Improve urban resilience, and improve the synergy of smart systems and smart infrastructure in all phases of disasters to protect the lives and property of citizens WT1: Improve the legal system for the protection of citizens’ personal information data
WO2: Conduct surveys of citizens’ needs, analyze the priority needs of various groups of citizens (e.g., different age, different careers, different incomes, and different gender), and formulate smart city development policies based on the priority of citizens’ needs for a better life WT2: Establish a feedback mechanism for citizens on the benefits of smart cities and make citizens share the dividends of development
WO3: Improve the supervision of smart city development and establish a multi-sectoral citizen feedback mechanism that allows citizens to participate in the supervision of smart city development WT3: Reduce the difficulty of using smart city public service
WO4: Improve the data synergy mechanism of smart city development and promote balanced development
Table 6 shows the pairwise comparisons as well as the final weights allocated to each element at four strategic dimensions. It also stated which sub-criteria were employed in each strategy. According to the results, “Divide smart infrastructure into different categories according to the hierarchy needs of citizens and promote the synergy development of smart infrastructure within and among different categories” and “Conduct surveys of citizens’ needs, analyze the priority needs of various groups of citizens (e.g., different ages, different careers, different incomes, and different gender), and formulate smart city development policies based on the priority of citizens’ needs for a better life” were found to be the most important strategies. While “Promote the integration of smart city development and regional cultural characteristics” and “Improve the legal system for the protection of citizens’ personal information data” were the weakest strategies. The overall weight of the SO strategies group is the largest, so the SO strategies group should be emphasized in enhancing CSG in smart cities (Fig. 5).Table 6 The strategies for enhancing citizens’ sense of gain in smart cities
Strengths Sub-criteria used for each strategy TW Rank
SO1 S1, S5, O2, O3 0.03414 1
SO2 S3, S5, O1 0.01636 4
SO3 S2, S3, O2, O3, O4 0.02018 3
SO4 S4, O2, O4 0.00490 9
SO – 0.06995 1
ST1 S1, S2, S5, T1, T2 0.00972 5
ST2 S1, S2, S5, T4 0.00437 10
ST3 S1, S4, T3 0.00096 15
ST4 S3, S5, T5 0.00389 11
ST – 0.01894 3
WO1 W3, O3, O4, O5 0.00533 8
WO2 W4, O2, O3, O4, O5 0.02096 2
WO3 W2, O4, O5 0.00700 7
WO4 W1, W5, O3, O4 0.00796 6
WO – 0.04124 2
WT1 W1, W2, T5 0.00142 14
WT2 W4, T1, T2, T3 0.00259 12
WT3 W3, W4, W5, T4 0.00142 13
WT – 0.00542 4
Fig. 5 The analysis of the strategic environment
Discussions and Implications
Discussions
The factors influencing CSG in smart cities (results in Table 1), SWOT criteria (results in Table 2) and the strategies for enhancing CSG in smart cities (results in Table 5) are discussed respectively combined with the analysis results.
Influencing Factors of CSG in Smart Cities
The development of smart cities is technocratic in the past (Cardullo & Kitchin, 2019; Kitchin, 2019), and citizen-centric smart cities are an emerging concept for achieving a shift from technocracy to humanism (Krivy, 2018; Yigitcanlar et al., 2019). This work integrates the concept of CSG into smart cities, as the material acquisition and spiritual feelings of citizens are both important criteria to reflect the development level of a smart city. The 17 critical CSG influencing factors (Table 1) are identified from the citizens’ sense of material and spiritual gain in aspects of various smart city services (i.e., smart education, smart healthcare, smart environment, smart transportation, smart governance, and smart aging), which can help evaluate smart city performance from the perspective of CSG (Ahvenniemi et al., 2017; D’Acci, 2021; Garau & Pavan, 2018; Huovila et al., 2019; Sharifi, 2020). Different from existing studies of smart city performance evaluations, this CSG-based work considers citizens’ needs in bottom-up manner, which emphasizes whether citizens’ actual needs are met by smart city services and advanced technologies. In comparison with the purely subjective indicators (e.g., satisfaction) (Lebrument et al., 2021; Xu & Zhu, 2021), CSG can reflect both the material acquisition and spiritual feelings of citizens in smart cities.
Smart cities are complex social-ecological systems that connect various parts of the city through advanced ICTs, and the incorporation of CSG can help in understanding the interplay between space and place in smart cities. “Urban Space” refers to the objective three-dimensional urban space in which things exist (i.e., the cartesian notion of coordinates), whereas “Urban Place” refers to human perception and experience therein (D’Acci, 2021; Lau et al., 2021; Szaszák & Kecskés, 2020). Based on the urban quality space–place conceptual framework developed by Cabrera-Barona & Merschdorf, 2018, smart city meant a human-centered, livable city in constant transformation, which corroborated the opinion of this paper. As an important mediator dimension in smart cities, advanced ICTs accelerate the space-place interaction process. However, the role and content of the citizen dimension are underappreciated. The two aspects (material acquisition and spiritual feelings) contained in CSG reposition the mediation role of advanced ICTs in facilitating the place-to-space improvement of smart city and emphasize the mediation intermediary role of the citizen dimension. The factors influencing citizens’ sense of spiritual gain included citizens’ sense of belonging and citizens’ sense of perception, which proved that citizens’ empowerment and participation were key drivers of the smart city. Meanwhile, factors influencing citizens’ sense of material gain represented the importance of handling advanced technologies to enhance citizens’ sense of “place”. In short, CSG provided a new theoretical framework and perspective for further research on space-place implications in smart cities.
Strengths, Weaknesses, Opportunities and Threats for Enhancing CSG in Smart Cities
Internal strengths for enhancing CSG in smart cities.
Strengths were proved to be the second most important factor for enhancing CSG in smart cities, indicating the development of smart cities has brought many conveniences to citizens. The results analyzed by AHP indicated that S1 and S5 were the most important strengths. According to the observations of research and the views of experts (Berglund et al., 2020; Rahouti et al., 2021; Rice & Martin, 2020; Singh et al., 2020; Voegler et al., 2017), a city’s infrastructure and information systems both play a vital role in the quality of life and perception of its citizens. In contrast, the weights of S2, S3, and S4 were smaller. One of the reasons is the development of citizen-centric smart cities in China is still in the initial stage of exploration, and technology is still the main factor in the deep perception of citizens.
External opportunities for enhancing CSG in smart cities.
The findings in Table 4 revealed that the most significant criterion to enhance CSG is the opportunities available in the progress of smart city development. When the opportunity criterion is the greatest of the four SWOT criteria, it may be deduced that many of the smart cities in question’s latent potentials have not yet been exploited to enhance CSG. There are three main reasons for this result. One is that citizens’ production activities are closely related to their consumption, and citizens’ income and consumption levels will increase with the development of the city. The second is that citizens know that a better life is beneficial to their interests, and governments can help them build citizen-centric smart cities to meet their needs (Bouzguenda et al., 2019; Nicolas et al., 2021). The third is that the contradiction between the people’s growing needs for a better life and the unbalanced and insufficient development is the main contradiction, which leads to citizens’ trust in governments.
Internal weaknesses for enhancing CSG in smart cities.
“ W4: Insufficient consideration of citizens’ needs” is the main weakness of smart cities. Some researchers (Ruhlandt, 2018) thought that the satisfaction of citizens’ needs was the final target for smart city sustainability development. And some experts (Vu & Hartley, 2018) also had the opinion that the emphasis on the “hardware” (e.g. technology) and “software” (e.g. citizens’ needs) should be balanced. The weight of W2 ranks second in weaknesses, indicating that a sound legal system is very important to the feelings of citizens. The possible reasons for W1, W3, and W5 are that citizens are dissatisfied with smart city data usage patterns, and the ability to deal with disasters and imbalanced development patterns.
External threats for enhancing CSG in smart cities.
T2 was identified to be the most important threat, showing that Citizen-Participation incentive policies for smart city development still need to be optimized. T5 was identified to be the second most important threat, indicating that numerous data breaches put citizens’ personal data at risk. According to T4, we can get citizens dissatisfied with the consideration of vulnerable groups in smart city development. According to T1, we can get publicity for smart cities development still needs to be optimized. T3 shows that smart city development need take into account regional cultural and humanistic needs.
Strategies for Enhancing CSG in Smart Cities
SO strategies for enhancing CSG in smart cities.
The SO strategies group is ranked the first overall weight (Fig. 5), indicating that smart cities should prioritize relying on internal technological strengths and leveraging external citizen demand to enhance CSG. Based on SO1, SO2, SO3, and SO4 (Tables 5 and 6), it can be found that as the core drivers of smart cities, the degree of synergistic development of smart infrastructure and smart systems is the basic guarantee of CSG enhancement. Besides, smart cities always gather a large number of high-quality education, transportation, healthcare, and business resources that attract citizens from surrounding towns, and cities land in them and make consumption. The capital brought by the migrant and local population promotes the speed of smart city development. So it’s important for smart cities to create a convenient and safe consumption environment and promote citizens’ online e-commerce and offline smart services consumption. What’s more, to enhance CSG in smart cities, the local governments are the only role who has the responsibility to analyze local citizens’ needs from the bottom-up. Local governments should be well aware of this, divide departmental responsibilities, and launch smart city development plans according to this principle. And it is also necessary to promote environmental-friendly construction thus providing a comfortable natural environment to satisfy citizens’ material needs.
ST strategies for enhancing CSG in smart cities.
The ST strategies group is ranked the third overall weight (Fig. 5). According to Tables 5 and 6, ST1 ranked 5 among all strategies, indicating that one of the most urgent tasks facing smart city CSG enhancement today is to make all citizens aware of the importance of smart city development and to engage them proactively in all aspects of city services. In order to deal with this problem, citizen participation paradigms that meets citizens’ participation interests and behaviors based on the functions of different departments can be established to attract citizens’ participation. The results of ST2 and ST3 show that with the aging of society and comprehensive social development, it’s of great importance for smart cities to promote age-friendly, vulnerable-friendly, and regional-cultural-integrated construction. The result of ST4 reveals that citizens are afraid of the numerous information leaks that have occurred in the past, so blockchain technology can be used in smart systems and a multi-channel password lock mechanism can be built to protect citizens’ data.
WO strategies for enhancing CSG in smart cities.
The WO strategies group is ranked the second overall weight (Fig. 5), indicating that smart cities should prioritize the use of external opportunities to compensate for internal weaknesses for transformational development thus enhancing CSG when adopting the WO strategies group. According to Table 6, WO2 ranked 2 among all strategies, confirming that the satisfaction of citizens’ material and spiritual needs is the core cause for enhancing CSG in smart cities. Thus it is suggested that the needs of various groups of citizens should be clarified and classified and policies related to smart cities’ developments should be formulated based on the priority needs of different groups of citizens. In response to W2, it’s suggested in WO4 that a multi-sectoral citizen feedback mechanism can be established to allow citizens to participate in the supervision of smart city development. Natural disasters such as the COVID-19 virus and urban flooding can pose a great threat to citizens’ lives and property thus decreasing CSG, so in order to improve urban resilience, we propose in WO1 and WO3 that the data synergy mechanism among different smart service systems should be improved, the synergy of smart systems and smart infrastructure in all phases of disasters should be promoted, and development of old community, new community, old district, and new district should be balanced.
WT strategies for enhancing CSG in smart cities.
Although the WT strategies group has the lowest overall weight ranking, three of its strategies need also to be taken seriously. As a complement to the ST4, WT1 proposes a strategy to enhance the security of citizens’ data from another perspective, which is “improve the legal system for the protection of citizens’ personal information data”. Considering citizens’ insufficient benefits acquisition in smart cities, it’s suggested that a feedback mechanism for citizens on the benefits of smart cities should be established to make citizens share the dividends of development. Besides, since it is often difficult for people who do not understand the operational process of smart city services to use them, we propose in WT3 that “reduce the difficulty of using smart city public service” so that every citizen can use enjoy smart city services.
Implications
Using the 15 strategies (Table 5) derived from the analysis result, four policy implications for enhancing CSG in smart cities are provided.
Strengthen Publicity and Encourage Citizen Participation
The government should develop incentive policies for participation in line with citizens’ interests and increase the publicity of smart cities to fully mobilize citizens’ enthusiasm. Besides, it’s also of significant importance for the government to motivate citizens to participate in smart cities through appropriate educational measures. This is to make them aware of the importance of smart cities to their future lives. At the same time, it’s suggested that the sectors that have positive and negative impacts on citizen participation should be analyzed. Specifically, this involves reforming each specific sector of implementation and optimizing the citizen participation system in light of this.
Clarify the Responsibilities of Local Governments
China’s current smart city development plans and strategies for enhancing CSG are guided by the central government. And then, local governments formulate policies and implement them based on the central government’s guidance. However, local governments should play a more substantial role than this in forming a connecting link between the preceding and the following. The central government’s policies are top-down guidance. Local governments should conduct citizen surveys and then analyze the demand structure of CSG based on survey results, so as to formulate bottom-up strategies. In the end, the local government should form the final smart city development strategies based on the guidance of the central government and the needs of the local CSG.
Prioritize Citizens’ Needs
It is necessary for the government to formulate policies related to the development of smart cities from the perspective of citizens’ needs for a comfortable urban life. In addition, it is critical to prioritize citizens' needs so that they can be met step by step. The suggested integrated services, which include smart infrastructure and smart systems, should be built to find the prioritized needs of citizens in a timely and accurate manner, with the assistance of high technologies. The interplay of technologies and citizens’ prioritized needs can give a citizen needs realization framework for developing smart city from the bottom-up perspective, while still needing governance from the top-down perspective to govern the city.
Promote Age-Friendly, Vulnerable-Friendly, and Environment-Friendly Development
In addition to the local population, the population of smart cities is also composed of a large number of floating populations. A larger proportion of these populations are elderly and low-income, but their contribution to the development of smart cities cannot be ignored. The elderly take care of their grandchildren, enabling their children to work, and the labor imported by the low-income populations brings a steady stream of power to smart city development. In order for every citizen to share the benefits of smart cities and enhance their CSG, the trend of age-friendly and vulnerable-friendly smart city is unstoppable. Besides, as an important element for human survival, the natural environment should not be destroyed by the development of smart cities. The concept of “Green mountains and clear water are equal to mountains of gold and silver” should be practiced, and environment-friendly smart cities should be developed.
Conclusions
This paper identified influencing factors of CSG in smart cities and proposed strategies to enhance CSG in smart cities. Firstly, based on the policies promulgated, the meaning of CSG, and existing literature, 17 influencing of CSG and 6 dimensions in smart cities were identified from the perspective of material and spiritual. Secondly, a two-stage questionnaire survey was conducted at SNC and 10 external sub-criteria, as well as 10 internal sub-criteria, were analyzed. Then, by analyzing the AHP, the sub-criteria were compared separately and “citizens’ ever-growth needs for a better life” ranked top. Finally, using TWOS analysis, 15 strategies for enhancing CSG were determined and ranked. The top three important strategies are (i) divide smart infrastructure into different categories according to the hierarchy needs citizens, and promote the synergy development of smart infrastructure within and among different categories, (ii) conduct surveys of citizens’ needs, analyze the priority needs of various groups of citizens (e.g., different ages, different careers, different incomes, and different gender), and formulate smart city development policies based on the priority of citizens’ needs for a better life, and (iii) clarify the role of local governments and departments in enhancing the CSG process for smart cities in terms of bottom-up analysis of local citizens' needs and top-down implementation of national policies.
The findings of this study also have to be seen in light of some limitations, such as the insufficient scope of the survey, and lacking to analyze the structure of citizens’ needs. In future research, citizen surveys can be carried out within the scope of smart cities across the whole nation to determine the material and spiritual demand hierarchy of citizens, and then formulate strategies to enhance CSG, which will be beneficial to help the development of citizen-centric smart cities in China.
Appendix
See Tables 7 and 8.Table 7 The demographic distribution of the sample
Item Category Amount
(the first-stage survey results) Amount
(the second-stage survey results)
Gender Male 327 251
Female 306 234
Age Under 18 83 63
18–30 296 227
31–40 161 124
41–50 63 48
Over 50 30 23
Educational level High school or less 130 100
College 407 312
Graduate school or higher 96 73
Income level Under 2500 CNY per month 126 96
2501–5000 CNY per month 261 201
5001–7500 CNY per month 136 104
7501–10,000 CNY per month 72 55
Over 10,001 CNY per month 38 29
Table 8 The questions and survey results of the first stage survey
Serial number Question Average score Greater/smaller than total average score
Q1 Do you agree that smart education construction has enhanced your sense of material gain in education? 1.63 Greater
Q2 Do you agree that smart healthcare construction has enhanced your sense of material gain in healthcare? 1.34 Greater
Q3 Do you agree that smart transportation construction has enhanced your sense of material gain in transportation? 1.78 Greater
Q4 Do you agree that smart environment construction has enhanced your sense of material gain in environment? 1.54 Greater
Q5 Do you agree that smart social security construction has enhanced your sense of material gain in social security? 1.22 Greater
Q6 Do you agree that smart social security construction are beneficial for you to defend your interest? −0.37 Smaller
Q7 Do you agree that smart aging construction has enhanced your sense of material gain in aging? 0.15 Smaller
Q8 Do you agree that the people around you have high awareness of the connotation of smart city? −0.76 Smaller
Q9 Do you agree that smart economy construction has enhanced your sense of material gain in economy? 1.68 Greater
Q10 Do you agree that smart government construction has enhanced your sense of material gain in daily government affairs? 1.12 Greater
Q11 Do you agree that smart government construction has enhanced your sense of material gain in participation? −1.03 Smaller
Q12 Do you agree that smart safety construction has enhanced your sense of material gain in social public safety? 1.23 Greater
Q13 Do you agree that smart emergency construction has enhanced your sense of material gain in social public safety? −0.45 Smaller
Q14 Do you agree that smart safety construction has enhanced your sense of material gain in food hygiene and safety? 1.34 Greater
Q15 Do you agree that smart safety construction has enhanced your sense of material gain in internet and data security? −0.98 Smaller
Q16 Do you agree that smart infrastructure construction are useful and convenience to you? 1.35 Greater
Q17 Do you agree that smart systems construction are efficiency and convenience to you? 0.41 Smaller
Q18 Do you agree that local government's smart city development plan can enhanced your sense of material gain? 0.99 Greater
Q19 Do you agree that smart cities development are helpful to realize self-worth? 0.93 Greater
Q20 Do you believe that the government’s policy of benefiting the people can enhance your sense of gain? 1.27 Greater
Q21 Do you agree that smart cities development are helpful to upgrade your socioeconomic status? 1.39 Greater
Q22 Do you agree that smart cities development are able to integrate into the regional cultural characteristics? −0.45 Smaller
Q23 Do you agree that smart cities development are able to maintain social fairness and justice? −0.21 Smaller
Q24 Do you agree that smart cities development supply are matched well with your material and spiritual needs? −1.48 Smaller
Total average score 0.59
Acknowledgements
This study was supported by the National Social Science Fund of China (No.19BGL281).
Declarations
Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Ahvenniemi H Huovila A Pinto-Seppä I Airaksinen M What are the differences between sustainable and smart cities? Cities 2017 60 234 245 10.1016/j.cities.2016.09.009
Al Sharif R Pokharel S Smart city dimensions and associated risks: Review of literature Sustainable Cities and Society 2022 10.1016/j.scs.2021.103542
Alsamhi SH Ma O Ansari MS Almalki FA Survey on collaborative smart drones and internet of things for improving smartness of smart cities Ieee Access 2019 7 128125 128152 10.1109/access.2019.2934998
Asadpourian Z Rahimian M Gholamrezai S SWOT-AHP-TOWS analysis for sustainable ecotourism development in the best area in Lorestan Province, Iran Social Indicators Research 2020 152 1 289 315 10.1007/s11205-020-02438-0
Asher MG Bali AS Kwan CY Public financial management in singapore: Key characteristics and prospects Singapore Economic Review 2015 10.1142/s0217590815500320
Bennati S Dusparic I Shinde R Jonker CM Volunteers in the smart city: Comparison of contribution strategies on human-centered measures Sensors 2018 10.3390/s18113707
Berglund EZ Monroe JG Ahmed I Noghabaei M Do J Pesantez JE Smart infrastructure: A vision for the role of the civil engineering profession in smart cities Journal of Infrastructure Systems 2020 10.1061/(asce)is.1943-555x.0000549
Bouzguenda I Alalouch C Fava N Towards smart sustainable cities: A review of the role digital citizen participation could play in advancing social sustainability Sustainable Cities and Society 2019 10.1016/j.scs.2019.101627
Braun T Fung BCM Iqbal F Shah B Security and privacy challenges in smart cities Sustainable Cities and Society 2018 39 499 507 10.1016/j.scs.2018.02.039
Cabrera-Barona P Merschdorf H A conceptual urban quality space-place framework: Linking geo-information and quality of life Urban Science 2018 2 3 73 10.3390/urbansci2030073
Cardullo P Kitchin R Being a ‘citizen’ in the smart city: Up and down the scaffold of smart citizen participation in Dublin, Ireland GeoJournal 2019 84 1 1 13 10.1007/s10708-018-9845-8
Carter E Adam P Tsakis D Shaw S Watson R Ryan P Enhancing pedestrian mobility in smart cities using big data Journal of Management Analytics 2020 7 2 173 188 10.1080/23270012.2020.1741039
Casebeer A Application of swot analysis British Journal of Hospital Medicine 1993 49 6 430 431 8472105
Chanak P Banerjee I Internet-of-things-enabled smartvillages: An overview IEEE Consumer Electronics Magazine 2021 10 3 12 18 10.1109/mce.2020.3013244
Chatfield AT Reddick CG Smart city implementation through shared vision of social innovation for environmental sustainability: A case study of Kitakyushu, Japan Social Science Computer Review 2016 34 6 757 773 10.1177/0894439315611085
Chen ZY Chan ICC Smart cities and quality of life: A quantitative analysis of citizens’ support for smart city development Information Technology & People 2022 10.1108/itp-07-2021-0577
Cowley R Joss S Dayot Y The smart city and its publics: Insights from across six UK cities Urban Research & Practice 2018 11 1 53 77 10.1080/17535069.2017.1293150
D’Acci LS Preferring or Needing Cities? (Evolutionary) psychology, utility and life satisfaction of urban living City, Culture and Society 2021 24 100375 10.1016/j.ccs.2021.100375
Dameri RP Benevolo C Governing smart cities: An empirical analysis Social Science Computer Review 2016 34 6 693 707 10.1177/0894439315611093
De Guimaraes JCF Severo EA Felix LA Da Costa W Salmoria FT Governance and quality of life in smart cities: Towards sustainable development goals Journal of Cleaner Production 2020 10.1016/j.jclepro.2019.119926
Dyson RG Strategic development and SWOT analysis at the University of Warwick European Journal of Operational Research 2004 152 3 631 640 10.1016/s0377-2217(03)00062-6
Ebenso B Otu A Giusti A Cousin P Adetimirin V Razafindralambo H Nature-based one health approaches to urban agriculture can deliver food and nutrition security Frontiers in Nutrition 2022 10.3389/fnut.2022.773746
Feng LL Zhong H Interrelationships and methods for improving University Students’ sense of gain, sense of security, and happiness Frontiers in Psychology 2021 10.3389/fpsyg.2021.729400
Feng ML Human-oriented smart city planning and management based on time-space behaviOR Open House International 2019 44 3 80 83 10.1108/OHI-03-2019-B0021
Feng XH Xiu CL Bai LM Zhong YX Wei Y Comprehensive evaluation of urban resilience based on the perspective of landscape pattern: A case study of Shenyang city Cities 2020 10.1016/j.cities.2020.102722
Freudendal-Pedersen M Kesselring S Servou E What is smart for the future city? Mobilities and automation Sustainability 2019 11 1 21 10.3390/su11010221
Garau C Pavan VM Evaluating urban quality: Indicators and assessment tools for smart sustainable cities Sustainability 2018 10.3390/su10030575
Georgiadis A Christodoulou P Zinonos Z Citizens’ perception of smart cities: A case study Applied Sciences 2021 11 6 2517 10.3390/app11062517
Gottfried O De Clercq D Blair E Weng X Wang C SWOT-AHP-TOWS analysis of private investment behavior in the Chinese biogas sector Journal of Cleaner Production 2018 184 632 647 10.1016/j.jclepro.2018.02.173
Gu T Li D Zhu S Wang Y Does sponge-style old community renewal lead to a satisfying life for residents? An investigation in Zhenjiang, China Habitat International 2019 90 102004 10.1016/j.habitatint.2019.102004
Gu YH Yang YY Wang J Research on employee sense of gain: The development of scale and influence mechanism Frontiers in Psychology 2020 11 16 10.3389/fpsyg.2020.568609 32140118
Hu Q Zheng YP Smart city initiatives: A comparative study of American and Chinese cities Journal of Urban Affairs 2021 43 4 504 525 10.1080/07352166.2019.1694413
Hu R The state of smart cities in China: The case of Shenzhen Energies 2019 12 22 18 10.3390/en12224375
Huang G Li D Yu L Yang D Wang Y Factors affecting sustainability of smart city services in China: From the perspective of citizens’ sense of gain Habitat International 2022 128 102645 10.1016/j.habitatint.2022.102645
Hudson L Wolff A Gooch D van der Linden J Kortuem G Petre M Supporting urban change: Using a MOOC to facilitate attitudinal learning and participation in smart cities Computers & Education 2019 129 37 47 10.1016/j.compedu.2018.10.012
Hummel JM Bridges JFP Ijzerman MJ Group decision making with the analytic hierarchy process in benefit-risk assessment: A tutorial Patient-Patient Centered Outcomes Research 2014 7 2 129 140 10.1007/s40271-014-0050-7 24623191
Huo ZH Zhang M Han JH Heterogeneity of capability deprivation and subjective sense of gain: Analysis of factor mixture models based on 892 rural households in six provinces International Journal of Environmental Research and Public Health 2022 10.3390/ijerph19074294
Huovila A Bosch P Airaksinen M Comparative analysis of standardized indicators for Smart sustainable cities: What indicators and standards to use and when? Cities 2019 89 141 153 10.1016/j.cities.2019.01.029
Ivan L Beu D van Hoof J Smart and age-friendly cities in romania: An overview of public policy and practice International Journal of Environmental Research and Public Health 2020 17 14 5202 10.3390/ijerph17145202 32708488
Jan T Azami P Iranmanesh S Sianaki OA Hajiebrahimi S Determining the optimal restricted driving zone using genetic algorithm in a smart city Sensors 2020 10.3390/s20082276
Jang SG Gim THT Considerations for encouraging citizen participation by information-disadvantaged groups in smart cities Sustainable Cities and Society 2022 76 8 10.1016/j.scs.2021.103437
Ji TT Chen JH Wei HH Su YC Towards people-centric smart city development: Investigating the citizens’ preferences and perceptions about smart-city services in Taiwan Sustainable Cities and Society 2021 10.1016/j.scs.2020.102691
Jia HT Zhu L Du J Fuzzy comprehensive evaluation model of the farmers’ sense of gain in the provision of rural infrastructures: The case of tourism-oriented rural areas of China Sustainability 2022 10.3390/su14105831
Ju J Liu L Feng Y Citizen-centered big data analysis-driven governance intelligence framework for smart cities Telecommunications Policy 2018 42 10 881 896 10.1016/J.TELPOL.2018.01.003
Khatoun R Zeadally S Cybersecurity and privacy solutions in smart cities Ieee Communications Magazine 2017 55 3 51 59 10.1109/mcom.2017.1600297cm
Kim N Park J Choi J-J Perceptual differences in core competencies between tourism industry practitioners and students using analytic hierarchy process (AHP) Journal of Hospitality Leisure Sport & Tourism Education 2017 20 76 86 10.1016/j.jhlste.2017.04.003
Kitchin R The timescape of smart cities Annals of the American Association of Geographers 2019 109 3 775 790 10.1080/24694452.2018.1497475
König PD Citizen-centered data governance in the smart city: From ethics to accountability Sustainable Cities and Society 2021 75 103308 10.1016/J.SCS.2021.103308
Kranjac M Tomic S Salom J Ilic-Kosanovic T Ilic D Bulajic S New methodology of teaching a smart city concept: A case study International Journal of Engineering Education 2021 37 1 204 214
Krivý M Towards a critique of cybernetic urbanism: The smart city and the society of control Planning Theory 2018 17 1 8 30 10.1177/1473095216645631
Lau BPL Ng BKK Yuen C Tuncer B Chong KH The study of urban residential’s public space activeness using space-centric approach IEEE Internet of Things Journal 2021 8 14 11503 11513 10.1109/JIOT.2021.3051343
Lebrument N Zumbo-Lebrument C Rochette C Roulet TJ Triggering participation in smart cities: Political efficacy, public administration satisfaction and sense of belonging as drivers of citizens’ intention Technological Forecasting and Social Change 2021 171 14 10.1016/j.techfore.2021.120938
Li DR Cao JJ Yao Y Big data in smart cities Science China-Information Sciences 2015 10.1007/s11432-015-5396-5
Li ML Woolrych R Experiences of older people and social inclusion in relation to smart “age-friendly” cities: A case study of Chongqing, China Frontiers in Public Health 2021 10.3389/fpubh.2021.779913
Lin CP Zhao GX Yu CP Wu YJ Smart city development and residents’ well-being Sustainability 2019 10.3390/su11030676
Liu KL Wang MJ Li JC Huang JJ Huang XH Chen SH Cheng BQ Developing a framework for spatial effects of smart cities based on spatial econometrics Complexity 2021 2021 8 10.1155/2021/9322112
Long Y Thill JC Combining smart card data and household travel survey to analyze jobs-housing relationships in Beijing Computers Environment and Urban Systems 2015 53 19 35 10.1016/j.compenvurbsys.2015.02.005
Macke J Casagrande RM Sarate JAR Silva KA Smart city and quality of life: Citizens’ perception in a Brazilian case study Journal of Cleaner Production 2018 182 717 726 10.1016/j.jclepro.2018.02.078
Marsal-Llacuna ML City indicators on social sustainability as standardization technologies for smarter (citizen-centered) governance of cities Social Indicators Research 2016 128 3 1193 1216 10.1007/s11205-015-1075-6
Masucci M Pearsall H Wiig A The smart city conundrum for social justice: Youth perspectives on digital technologies and urban transformations Annals of the American Association of Geographers 2020 110 2 476 484 10.1080/24694452.2019.1617101
Mohamed N Al-Jaroodi J Jawhar I Kesserwan N Data-driven security for smart city systems: Carving a trail IEEE Access 2020 8 147211 147230 10.1109/access.2020.3015510
Nagarajan SM Deverajan GG Chatterjee P Alnumay W Muthukumaran V Integration of IoT based routing process for food supply chain management in sustainable smart cities Sustainable Cities and Society 2022 10.1016/j.scs.2021.103448
Nicolas C Kim J Chi S Natural language processing-based characterization of top-down communication in smart cities for enhancing citizen alignment Sustainable Cities and Society 2021 10.1016/j.scs.2020.102674
Nikolic PK Yang H Designing playful cities: Audio-visual metaphors for new urban environment experience Mobile Networks & Applications 2020 25 3 846 852 10.1007/s11036-020-01514-6
Peng GCA Nunes MB Zheng LQ Impacts of low citizen awareness and usage in smart city services: The case of London’s smart parking system Information Systems and E-Business Management 2017 15 4 845 876 10.1007/s10257-016-0333-8
Pieroni C Giannotti M Alves BB Arbex R Big data for big issues: Revealing travel patterns of low-income population based on smart card data mining in a global south unequal city Journal of Transport Geography 2021 10.1016/j.jtrangeo.2021.103203
Piro G Cianci I Grieco LA Boggia G Camarda P Information centric services in Smart Cities Journal of Systems and Software 2014 88 169 188 10.1016/j.jss.2013.10.029
Rahouti M Xiong K Xin Y Secure software-defined networking communication systems for smart cities: Current status, challenges, and trends IEEE Access 2021 9 12083 12113 10.1109/access.2020.3047996
Rana NP Luthra S Mangla SK Islam R Roderick S Dwivedi YK Barriers to the development of smart cities in Indian context Information Systems Frontiers 2019 21 3 503 525 10.1007/s10796-018-9873-4
Rice J Martin N Smart infrastructure technologies: Crowdsourcing future development and benefits for Australian communities Technological Forecasting and Social Change 2020 10.1016/j.techfore.2018.03.027
Rizzo A Kim GJ A SWOT analysis of the field of virtual reality rehabilitation and therapy Presence-Teleoperators and Virtual Environments 2005 14 2 119 146 10.1162/1054746053967094
Ruan WQ Yang TT Zhang SN Restoration path of small tourism enterprise managers’ confidence in the COVID-19 period Journal of Travel and Tourism Marketing 2022 39 2 137 151 10.1080/10548408.2022.2061675
Ruhlandt RWS The governance of smart cities: A systematic literature review Cities 2018 81 1 23 10.1016/j.cities.2018.02.014
Saaty TL axiomatic foundation of the analytic hierarchy process Management Science 1986 32 7 841 855 10.1287/mnsc.32.7.841
Saaty TL Peniwati K Shang JS The analytic hierarchy process and human resource allocation: Half the story Mathematical and Computer Modelling 2007 46 7–8 1041 1053 10.1016/j.mcm.2007.03.010
Savari M Shokati Amghani M SWOT-FAHP-TOWS analysis for adaptation strategies development among small-scale farmers in drought conditions International Journal of Disaster Risk Reduction 2022 67 102695 10.1016/j.ijdrr.2021.102695
Seker, S., Ozgurler, M., Beykent Univ Istanbul; Gebze Inst Technol; Emerald Publishing, G., & Elsevier Ltd Publishing, G. (2012). Analysis of the Turkish consumer electronics firm using SWOT-AHP method. In 8th international strategic management conference (Vol. 58, pp. 1544–1554). Barcelona, SPAIN. 10.1016/j.sbspro.2012.09.1141
Shami MR Rad VB Moinifar M The structural model of indicators for evaluating the quality of urban smart living Technological Forecasting and Social Change 2022 10.1016/j.techfore.2021.121427
Sharifi A A typology of smart city assessment tools and indicator sets Sustainable Cities and Society 2020 53 101936 10.1016/j.scs.2019.101936
Sharma M Bhatia G The voluntary community health movement in India: A strengths, weaknesses, opportunities, and threats (SWOT) analysis Journal of Community Health 1996 21 6 453 464 10.1007/bf01702605 8912121
Shen LY Huang ZH Wong SW Liao SJ Lou YL A holistic evaluation of smart city performance in the context of China Journal of Cleaner Production 2018 200 667 679 10.1016/j.jclepro.2018.07.281
Simonofski A Valle T Serral E Wautelet Y Investigating context factors in citizen participation strategies: A comparative analysis of Swedish and Belgian smart cities International Journal of Information Management 2021 10.1016/j.ijinfomgt.2019.09.007
Singh SK Jeong Y-S Park JH A deep learning-based IoT-oriented infrastructure for secure smart City Sustainable Cities and Society 2020 10.1016/j.scs.2020.102252
Song ML Tan KH Wang JL Shen ZY Modeling and evaluating economic and ecological operation efficiency of smart city pilots Cities 2022 124 16 10.1016/j.cities.2022.103575
Sookhak M Tang H He Y Yu FR Security and privacy of smart cities: A survey, research issues and challenges IEEE Communications Surveys and Tutorials 2019 21 2 1718 1743 10.1109/comst.2018.2867288
Su Y Li YP Relationship between subjective socioeconomic status and sense of gain of health-care reform and the mediating role of self-rated health: A cross-sectional study in China BMC Public Health 2022 10.1186/s12889-022-13106-y
Szarek-Iwaniuk P Senetra A Access to ICT in Poland and the co-creation of urban space in the process of modern social participation in a smart city-a case study Sustainability 2020 10.3390/su12052136
Szaszák G Kecskés T Universal open space design to inform digital technologies for a disability-inclusive place-making on the example of Hungary Smart Cities 2020 3 4 1293 1333 10.3390/smartcities3040063
Tan Yigitcanlar M Kamruzzaman MF Sabatini-Marques J da Costa E Ioppolo G Can cities become smart without being sustainable? A systematic review of the literature Sustainable Cities and Society 2019 45 348 365 10.1016/j.scs.2018.11.033
The State Council of the PRC (2016). National Informatization Development Outline. Retrieved from http://www.gov.cn/gongbao/content/2016/content_5100032.htm
The State Council of the PRC (2021). The plan to facilitate the development of the digital economy in the 14th Five-Year Plan period (2021–2025). Retrieved from http://www.gov.cn/zhengce/zhengceku/2022-01/12/content_5667817.htm
Trencher G Karvonen A Stretching “smart”: Advancing health and well-being through the smart city agenda Local Environment 2019 24 7 610 627 10.1080/13549839.2017.1360264
Valencia-Arias A Urrego-Marin ML Bran-Piedrahita L A Methodological model to evaluate smart city sustainability Sustainability 2021 10.3390/su132011214
Vidiasova L Cronemberger F Discrepancies in perceptions of smart city initiatives in Saint Petersburg, Russia Sustainable Cities and Society 2020 10.1016/j.scs.2020.102158
Voegler M Schleicher JM Inzinger C Dustdar S Ahab: A cloud-based distributed big data analytics framework for the Internet of Things Software-Practice & Experience 2017 47 3 443 454 10.1002/spe.2424
Vu K Hartley K Promoting smart cities in developing countries: Policy insights from Vietnam Telecommunications Policy 2018 42 10 845 859 10.1016/j.telpol.2017.10.005
Wan YL Guo ZM Improving the sense of gain of graduate students in food science Journal of Food Quality 2021 2021 7 10.1155/2021/3525667
Wang B Loo BPY Huang GZ Becoming smarter through smart city pilot projects: Experiences and lessons from China since 2013 Journal of Urban Technology 2021 10.1080/10630732.2021.1962695
Wang J Gu Y Luo Y Huang Y Liao L The mechanism of the influence of coaching leadership behavior on subordinate’s sense of gain at work Leadership and Organization Development Journal 2022 43 4 638 652 10.1108/LODJ-07-2021-0342
Wang MM Zhou T Wang D Tracking the evolution processes of smart cities in China by assessing performance and efficiency Technology in Society 2020 10.1016/j.techsoc.2020.101353
Wang YL Yang C Hu XY Chen H The mediating effect of community identity between socioeconomic status and sense of gain in Chinese adults International Journal of Environmental Research and Public Health 2020 10.3390/ijerph17051553
Wereda W Moch N Wachulak A The importance of stakeholders in managing a safe city Sustainability 2022 10.3390/su14010244
Williamson B Computing brains: Learning algorithms and neurocomputation in the smart city Information Communication & Society 2017 20 1 81 99 10.1080/1369118x.2016.1181194
Xie CF Zhu YQ Zhao Q How digital business penetration influences farmers’ sense of economic gain: The role of farmers’ entrepreneurial orientation and market responsiveness IEEE Access 2020 8 187744 187753 10.1109/access.2020.3031110
Xie D Yin C Research on the communication strategy of history and culture in Shaanxi based on BP neural network model Computational Intelligence and Neuroscience 2022 10.1155/2022/8965622
Xie JF Tang HE Huang T Yu FR Xie RC Liu J Liu YJ A survey of blockchain technology applied to smart cities: Research issues and challenges IEEE Communications Surveys and Tutorials 2019 21 3 2794 2830 10.1109/comst.2019.2899617
Xu HL Zhu WL Evaluating the impact mechanism of citizen participation on citizen satisfaction in a smart city Environment and Planning B-Urban Analytics and City Science 2021 48 8 2466 2480 10.1177/2399808320980746
Yang XF Ma LN AHP-based analysis factors influencing the construction of a smart city with three-dimensional regional color Journal of Sensors 2021 10.1155/2021/1778399
Yeh HP The effects of successful ICT-based smart city services: From citizens’ perspectives Government Information Quarterly 2017 34 3 556 565 10.1016/j.giq.2017.05.001
Yu CP Ye BS Lin CP Wu YJ Can smart city development promote residents’ emotional well-being? Evidence from China IEEE Access 2020 8 116024 116040 10.1109/access.2020.3004367
Yuan J Xie H Yang D Xiahou X Skibniewski MJ Huang W Strategy formulation for the sustainable development of smart citiES: A case study of Nanjing, China International Journal of Strategic Property Management 2020 24 6 379 399 10.3846/ijspm.2020.13345
Zandbergen D Uitermark J In search of the smart citizen: Republican and cybernetic citizenship in the smart city Urban Studies 2020 57 8 1733 1748 10.1177/0042098019847410
Zhang F Zhao JJ Tian C Xu CZ Liu X Rao L Spatiotemporal segmentation of metro trips using smart card data Ieee Transactions on Vehicular Technology 2016 65 3 1137 1149 10.1109/tvt.2015.2409815
Zhang WZ Cao J He J Chen L City health examination in China: A methodology and empirical study Chinese Geographical Science 2021 31 6 951 965 10.1007/s11769-021-1239-z
Zhu SY Li DZ Feng HB Is smart city resilient? Evidence from China Sustainable Cities and Society 2019 50 14 10.1016/j.scs.2019.101636
Ziganshina LE Yudina EV Talipova LI Sharafutdinova GN Khairullin RN Smart and age-friendly cities in Russia: An exploratory study of attitudes, perceptions, quality of life and health information needs International Journal of Environmental Research and Public Health 2020 17 24 9212 10.3390/ijerph17249212 33317150
| 0 | PMC9746588 | NO-CC CODE | 2022-12-15 23:21:56 | no | Soc Indic Res. 2022 Dec 13;:1-34 | utf-8 | Soc Indic Res | 2,022 | 10.1007/s11205-022-03047-9 | oa_other |
==== Front
Stoch Environ Res Risk Assess
Stoch Environ Res Risk Assess
Stochastic Environmental Research and Risk Assessment
1436-3240
1436-3259
Springer Berlin Heidelberg Berlin/Heidelberg
2354
10.1007/s00477-022-02354-4
Original Paper
Spatio-temporal modeling of infectious diseases by integrating compartment and point process models
http://orcid.org/0000-0003-3748-6801
Amaral André Victor Ribeiro [email protected]
González Jonatan A.
Moraga Paula
grid.45672.32 0000 0001 1926 5090 CEMSE Division, Statistics Program, King Abdullah University of Science and Technology (KAUST), Thuwal, Saudi Arabia
13 12 2022
115
30 11 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Infectious disease modeling plays an important role in understanding disease spreading dynamics and can be used for prevention and control. The well-known SIR (Susceptible, Infected, and Recovered) compartment model and spatial and spatio-temporal statistical models are common choices for studying problems of this kind. This paper proposes a spatio-temporal modeling framework to characterize infectious disease dynamics by integrating the SIR compartment and log-Gaussian Cox process (LGCP) models. The method’s performance is assessed via simulation using a combination of real and synthetic data for a region in São Paulo, Brazil. We also apply our modeling approach to analyze COVID-19 dynamics in Cali, Colombia. The results show that our modified LGCP model, which takes advantage of information obtained from the previous SIR modeling step, leads to a better forecasting performance than equivalent models that do not do that. Finally, the proposed method also allows the incorporation of age-stratified contact information, which provides valuable decision-making insights.
Supplementary Information
The online version contains supplementary material available at 10.1007/s00477-022-02354-4.
Keywords
Compartment SIR model
Infectious diseases
Log-Gaussian Cox process
Spatial point process
Spatio-temporal modeling
==== Body
pmcIntroduction
The spread of infectious diseases such as COVID-19 may overload healthcare systems and have devastating health, social and economic impacts at different levels (Kaye et al. 2021; Pak et al. 2020; Hossain et al. 2020). Disease modeling is essential to understand spreading dynamics and may provide valuable insights into disease prevention and control.
In this regard, one common approach to describe epidemic dynamics consists of splitting the population into compartments and modeling the rates that describe how individuals move from one compartment to another. Kermack and McKendrick (1927) initially proposed a model of this kind, assigned individuals to three compartments, namely Susceptible (S), Infected (I), and Recovered (R), and modeled the transition events S→I and I→R. This model aimed to describe how the number of individuals in each compartment evolves in time by solving a system of differential equations. Several extensions with additional compartments and transition events have been proposed. For instance, we may allow infected individuals to become susceptible again, or we can even include a new compartment—named Exposed (E)—between the susceptible and infected blocks in such a way that it addresses the latent period an individual may go through after being exposed to a disease but before being effectively infected. The last extension is known as SEIR, and both (and others) are described in Keeling and Rohani (2011).
Other extensions can arise from the usual model assumptions. As discussed in Britton et al. (2019), one can assume that individuals are homogeneously and uniformly mixed or primarily in contact with households and close friends. For sexually transmitted diseases, contact may only be defined among one’s sexual partners. Also, we may assume that all individuals have similar responses when exposed to a certain disease; however, people may vary, for example, regarding their immune systems and whether or not they should be included in the susceptible group. Another variable is related to the overall population size; we might assume it is constant over time (or at least approximately constant), but especially for groups observed over longer periods, migration is likely to play a role in the epidemic dynamics.
One could also consider a population divided into strata (e.g., based on gender, age and job type) to account for the different contact frequencies among individuals from these groups. As described by Mossong et al. (2008), elderly individuals have much less contact with other people when compared to younger classes. Age also plays an essential role in the disease severity—for example, there is evidence that the risk of death increases with age for COVID-19 (Wu et al. 2020; Moraga et al. 2020). Moreover, since children were already shown to be more susceptible to influenza than adults (Viboud et al. 2004), age might also impact the infectious rate for some specific diseases. Motivated by that, one can extend the SIR model by including such contact information among age groups to better characterize the interaction in the overall population; that is the approach we take throughout this paper.
Spatial and spatio-temporal statistical models have also been developed to understand disease geographic and temporal patterns, especially over the recent years, due to the increased availability of georeferenced health information (Moraga 2019). However, many of these works that focus on spatio-temporal modeling for infectious disease-related problems analyze areal data from aggregated point-level information or geostatistical data from surveys. For instance, Giuliani et al. (2020) and Lawson and Kim (2021) modeled COVID-19 areal data in space and time in Italy and South Carolina (USA), respectively. On the other hand, methods for dealing with point pattern data for the locations of infected individuals seem to be less explored—partly due to the challenge of obtaining the exact locations of infected individuals. An important example of point pattern data being used in this context comes from Diggle et al. (2005). In such a work, they introduced a log-Gaussian Cox point process model for spatio-temporal disease modeling and applied it to online spatio-temporal surveillance of non-specific gastrointestinal diseases in the United Kingdom. Similarly, there have also been some recent advances in including spatial components in the aforementioned compartment models; e.g., Geng et al. (2021) extended the SIR model to accommodate for the spatial spreading of the studied disease across the considered region by using a spatial kernel, and Lau et al. (2017) proposed a framework based on the SEIR model for describing a disease transmission pattern in space and time.
In this paper, we propose a spatio-temporal modeling framework that describes infectious disease dynamics and allows us to make accurate predictions for the number of infected cases in the future across space. To do so, we integrate the SIR compartment and log-Gaussian Cox process models and use the age-stratified contact information and point pattern data corresponding to the infected individuals’ locations in space-time to fit the model. Although the modeling scheme proposed by Diggle et al. (2005) already decomposes the intensity process into purely spatial, purely temporal and spatio-temporal dependent terms, our approach extends it by using the SIR compartment model output as the temporal component, which has a significant impact on the obtained results, especially when making predictions. More specifically, we propose to model the epidemic spatio-temporal dynamics in two steps: (1) fitting a temporal compartment SIR model taking into account contact patterns of different age groups, and (2) fitting a log-Gaussian Cox process for the point pattern data that represents the locations of infected individuals in the studied region and time interval, such that the mean of such a process depends on the information obtained from the previous step. Also, our fitting procedure for step (2) relies upon a faster computational technique, namely integrated nested Laplace approximation (INLA) (Rue et al. 2009)—as opposed to a Markov chain Monte Carlo (MCMC) algorithm—making it more appealing for practitioners.
The remainder of this paper is as follows. Section 2 presents the prerequisites for SIR modeling and introduces the key concepts in point pattern data analysis. In Sect. 3, we give the details of our proposed framework by describing the integration of the modeling steps in time and space-time. In Sect. 4, we conduct a simulation study to assess the performance of our approach and compare it with a null model that does not incorporate information from the SIR model. In this section, we simulate an epidemic scenario in São Paulo, Brazil, discuss the fitting procedures, and explore the obtained results. Section 5 shows a case study of the number of COVID-19-infected individuals in Cali, Colombia. In the last section, we briefly discuss our methodology and results; we also detail its limitations and possible extensions.
SIR and spatial point process models
Here, we briefly review the basics of infectious disease modeling through compartment models; in particular, we consider a SIR compartment model in time that accounts for age classes. We also present the relevant concepts of spatial point processes and describe one possible approach, named “log-Gaussian Cox process,” to model this data.
SIR model
Suppose we are interested in studying how an infectious disease may spread among individuals living in a given region. To model the evolution of such a disease in time, we can start by dividing the population into three groups, namely Susceptible (S), Infected (I), and Recovered (R), such that individuals are transferred from one compartment to another in the following way: S→I→R. Then, under the assumptions of (1) a homogeneous population with uniform mixing (meaning that individuals meet each other uniformly at random, (2) constant infectious rate and recovery (or death) rate over time, and (3) preserved population mass (that is, the total number of individuals in the population is constant over the total time), we can state the so-called deterministic base-SIR model.
For t∈T⊆R+={x∈R;x≥0}, let S(t), I(t), and R(t) denote the number of susceptible, infected, and recovered individuals at time t. Also, let N(t)=S(t)+I(t)+R(t) denote the total number of individuals in the population, such that N(t)=N, ∀t. Moreover, let β>0 and γ>0 be the infectious and recovery (or death) rates, respectively. In this case, the deterministic base-SIR model describes the evolution of the epidemic through the solution of the following system of Ordinary Differential Equations (ODEs)dS(t)dt=-βS(t)I(t)NdI(t)dt=+βS(t)I(t)N-γI(t)dR(t)dt=+γI(t).
Under initial conditions (S(0),I(0),R(0)), such that R(0) is usually assumed to be equal to 0, this system can be numerically solved—for instance, one can use the Euler’s Method (Atkinson 2008). Also, β and γ are disease-specific parameters and can be estimated as described in Sect. 3.1.1.
For this so-called base-SIR model, another quantity of interest, R0 (Base Reproductive Number), is essential when studying an infectious disease outbreak. Roughly speaking, R0 represents the expected number of secondary cases arising from a primary case in a completely susceptible population (Blackwood and Childs 2018). In that way, for the above simplistic model, it can be computed as R0=β/γ. Here, notice that R0 depends on a “single” infectious case and ignores the individual variability. As we will see next, the base reproductive number is computed differently when considering structured populations.
Often, we cannot assume a homogeneous population with uniform mixing since the number of contacts a person from a particular age group has highly depends on the other individuals’ ages (Mossong et al. 2008). Consequently, we might want to incorporate this contact information for a structured population in our base-SIR model. To do so, we rely on a contact matrix Cij that represents the average number of contacts made by an individual of age group i with an individual of age group j and on the proportion of individuals in each age group, namely fi. In this regard, notice that (1) Cij is not symmetric, and (2) the total number of contacts of group i with group j should be equal to the total number of contacts of group j with group i, i.e., Cij·fi=Cji·fj, for all pairs (i, j). Although, as discussed in Fitzgerald et al. (2020), (2) might not hold for all real data sets—especially when one tries to aggregate data from different sources, as we will do.
Based on Cij and fi, the age-groups-SIR model will describe the evolution of the epidemic for the different population classes as the solution of the following system of differential equations1 dSi(t)dt=-βSi(t)∑jCij·Ij(t)NjdIi(t)dt=+βSi(t)∑jCij·Ij(t)Nj-γIi(t)dRi(t)dt=+γIi(t),
such that Si(t), Ii(t), and Ri(t) correspond to the number of susceptible, infected and recovered individuals in group i and at time t, and Ni(t)=Ni, ∀t, is the total number of individuals in group i. The other quantities are defined as before.
From Model (1), notice that the infectious and recovery (or death) rates are still common for all age groups i. Letting β and γ vary with the population class is one possible extension (among many others) of such a model. Also, although there are different methods to determine R0 in this structured-population scenario (Li et al. 2011), the base reproductive number can be numerically computed as the largest eigenvalue of (β/γ·Mij), such that Mij is a matrix with elements Cij·fi/fj for all pairs (i, j) (Blackwood and Childs 2018; Towers and Feng 2012). This method is best known as the “next-generation matrix” (Diekmann et al. 1990, 2010) and has been used in different works (Li et al. 2020; Davies et al. 2020).
Finally, as mentioned before, an analytical solution for systems of differential equations like in Model (1) might not be possible to determine. Instead, we will approach such a problem numerically. To do so, we define a partition of T=[0,T], given by {tk;k=0,1,…,n}, such that 0=t0<t1<⋯<tn=T. For such a discretization, and under initial conditions (Si(0),Ii(0),Ri(0)), ∀i, we can use a solver for initial value problems for ODEs to determine an approximate solution at all {tk}k. In particular, we will use a solver named lsoda (Hindmarsh 1983; Petzold 1983) when implementing the methodology described in Sect. 3.
Spatial point process models
This section describes how disease spread can be modeled in space and space-time. In particular, we see how to model the observed space-time locations of infectious individuals in the studied population. To do so, we will briefly discuss the key ideas about point processes and introduce the notation used in the paper.
For each discretized time point tk from the previous subsection, we assume that the locations of infectious individuals arise as a realization of a point process in space. For u∈U⊆R2 and tk defined as before, a spatial point process ξ(tk) will be defined as a locally finite random subset of U; that is, #(ξ(tk)∩A) is finite for all bounded subsets A⊆U, such that #(x) denotes the cardinality of x. From this definition, we will let ξi(tk) denote a point process in U and at t=tk for the locations of infected individuals from a class i in our studied area at a given instant in time. One can refer to Moller and Waagepetersen (2003) for a rigorous introduction to (spatial) point processes.
For spatial point processes, we may define an intensity function λ:U→[0,+∞), such that ∫Aλ(u;tk)du<+∞, for all bounded A⊆U, in the following wayE[N(A;tk)]=∫Aλ(u;tk)du,A⊆U,
where N(A;tk) counts the number of points in A at t=tk. Thus, λi(u;tk) will correspond to the intensity function of the point process ξi(tk) for all classes i.
Spatial point processes can be modeled using a Poisson point process. A point process ξ(tk) on U is a Poisson point process with intensity function λ(u;tk) if the following properties are satisfied For any bounded A⊆U, N(A;tk)∼Poisson∫Aλ(u;tk)du; and
For any bounded A⊆U and n∈N, conditional on N(A;tk)=n, points in ξ(tk)∩A are independent and identically distributed with density proportional to λ(u;tk).
Therefore, getting back to the disease spread modeling problem, ξi(tk) would be the process that describes the locations, at tk, of the infected individuals in age group i. Also, if ξi(tk) is a Poisson point process, its average will depend on the corresponding intensity function λi(u;tk).
Log-Gaussian Cox process
Due to its simplicity, the Poisson point process may not be valid for describing more complex scenarios. However, such a process can be extended to a more general class of models named Cox process (Cox 1955). A Cox process can be seen as a doubly stochastic process since its intensity function is a random process itself. More specifically, ξ(tk) is a Cox process driven by Λ(u;tk) if {Λ(u;tk);u∈U} is a non-negative valued stochastic process, and
Conditional on {Λ(u;tk)=λ(u;tk);u∈U}, ξ(tk) is a Poisson process with intensity function λ(u;tk).
A particular case of a Cox process named log-Gaussian Cox process can be constructed by setting log{Λ(u;tk)}=μ⋆(u;tk)+ζ(u;tk), such that μ(u;tk)=exp{μ⋆(u;tk)} is possibly interpreted as the mean structure of Λ(u;tk), and ζ(u;tk) is a stationary Gaussian process, such that E(ζ(u;tk))=-σ2/2, ∀k and u, and Cov(ζ(u1;tk),ζ(u2;tk))=ϕ(h;tk)=σ2ρ(h;tk), where h=||u1-u2|| and σ2 is the variance of ζ(u;tk). In this case, E(Λ(u;tk))=μ(u;tk), and Cov(Λ(u1;tk),Λ(u2;tk))=ψ(u1,u2;tk)=τ2θ(u1,u2;tk)=μ(u1;tk)μ(u2;tk)[exp{ϕ(||u1-u2||;tk)}-1], where τ2=2μ(u;tk)[exp{σ2}-1] is the variance of Λ(u;tk).
We set ξi(tk), ∀i, as log-Gaussian Cox processes for the events of observing infected individuals in each age class and at tk. In particular, the corresponding Gaussian processes ζi(u;tk) will describe the spatio-temporal dependence structure for the underlying processes Λi(u;tk). In Sect. 3.2, we see different ways to model ζi(u;tk).
Methodology
We propose a model to describe how infectious diseases evolve in space and time and how to make predictions for the future number of cases across the study region. We do this by integrating the SIR compartment modeling in time and a point process modeling approach in space-time. By combining these methods, we incorporate knowledge about the mechanistic approach that drives the temporal dynamics into the model that accounts for spatial dependence.
Our approach is divided into two steps. First, we aggregate the data in the whole region for each discretised time window and estimate the epidemic dynamics in time using a SIR model. This gives us the corresponding curves for the number of infected individuals for all age groups i and at all time points tk. Then, for each population class, we estimate the intensity of the spatial-temporal point process that originated each observed point pattern of infected individuals. We do this by incorporating the estimates of the total number of infectious derived from the SIR model into the mean component of the point process and adding a spatio-temporal random structure to take variation into account. This approach allows us to describe past spatio-temporal patterns and predict the spatio-temporal evolution of the disease in future times. Figure 1 shows a diagram that illustrates such a procedure.Fig. 1 Diagram for the spatio-temporal modeling approach of infected individuals in all age groups i. From left to right, we have the infected individuals’ locations (“∘” denotes Group 1 and “×” denotes Group 2), the observed and estimated Ii(tk) curves (here, notice that we collected data up to t50 and made predictions up to t99), the base functions that represent the population at risk (see Sect. 3.2), and the estimated intensity functions for all time windows
Throughout this paper, we will use the following notation. The observed data are the locations and times of infected individuals for each age group i—we will denote them by {ξi(tk)}k for all tk. That is, for t∈T=[0,T], such that T is partitioned into {tk;k=0,1,…,n} and u=(u1,u2)∈U⊆R2, we have that {ξi(tk)}k, for all i, are sequences of spatial point patterns in U observed at time t=tk, such that the corresponding point process describes the locations of the infected individuals in each class and time window.
Temporal modeling
Our first step concerns modeling and estimating the temporal structures of the disease spread process in the population of interest. In this regard, the counting processes for the number of susceptible, infected and recovered individuals in all age classes, namely Si(t), Ii(t), and Ri(t), respectively, will be modeled according to Model (1).
Based on data given by a set of locations for each observed infected individual, denoted by {ξi(tk)}k, such that k∈{0,1,…,n}, we can start by aggregating all location observations over space in such a way that ii(tk)=#(ξi(tk)), ∀i,k. Therefore, ii(t) will correspond to the observed counting process for the number of infected individuals in each group at t=t0,t1,…,tn. We can estimate the parameters based on Model (1) from such data and solve for Si(tk), Ii(tk), and Ri(tk), ∀i,k, for inference and prediction.
Inference and prediction in time
A solution for Model (1) can be approximately computed using a numerical solver, as described in Sect. 2.1. In this way, for a set of initial values (Si(0),Ii(0),Ri(0)), ∀i, and initial guesses for β and γ, we can solve the system of ODEs for Si(tk), Ii(tk), and Ri(tk) by employing such a method. For later reference, we will name these solutions SiODE(tk), IiODE(tk), and RiODE(tk), respectively.
Suppose we have obtained ii(tk), ∀i,k, i.e., the number of infected individuals in all groups and at all time points. One way to model such data is by assuming they come from a specific probability distribution with the mean given by the ODE solution IiODE(tk), ∀i,k. In particular, as we are dealing with counting data and aiming to account for the possible overdispersion when fitting a Poisson model, we will assume a Negative Binomial distribution for the observed number of infected individuals, that is,2 Ii(tk)∼Negative Binomial(IiODE(tk),φ),
such that φ is the overdispersion parameter; this implies that E(Ii(tk))=IiODE(tk) and Var(Ii(tk))=IiODE(tk)(1+(1/φ)·IiODE(tk)).
For such an approach, notice that this is an iterative procedure. (I) First, we set initial guesses for β, γ, and φ. (II) Then, given the parameter values, we solve Model (1) for Si(tk), Ii(tk), and Ri(tk). (III) After that, we plug the IiODE(tk) curve into the mean component of Model (2) and evaluate the corresponding likelihood function. (IV) Next, regardless of the estimation framework, and after updating the values of β, γ, and φ, we get back to step (II) and repeat this sequential procedure until convergence.
The model from Eqs. (1) and (2) can be fitted in different ways. Here, we adopt a Bayesian framework and use RStan (Stan Development Team 2021) to estimate the posterior distribution of θ=(β,γ,φ)⊤, always assuming reasonable prior distributions for β, γ, and φ and the Negative Binomial likelihood for the observed counting. Also, when making predictions, assuming we can generate quantities from the fitted model, we need to solve Eq. (1) for Si(tk), Ii(tk), and Ri(tk) for any i and any k beyond the observation range.
Spatio-temporal modeling
Once we have estimated the model parameters and, therefore, the Ii(tk) curves, we can start modeling the intensity functions for the spatial processes for all age groups i at all times tk. In particular, we assume that the observed point patterns are originated from log-Gaussian Cox processes evaluated at the same partition of T=[0,T] defined before, namely {tk;k=0,1,…n}. Thus, the counting number of infected individuals in each age group i and at tk, namely Ni(tk), will be modeled as followsNi(tk)|Λi(u;tk)=λi(u;tk)∼Poisson∫Uλi(u;tk)du,
and the corresponding intensity functions will be described by3 Λi(u;tk)=μi(u;tk)·exp{ζi(u;tk)}
such that all quantities are defined as in Sect. 2.2.1. Here, μi(u;tk) represents the large-scale component of the model, and exp{ζi(u;tk)} represents the random variation around it. This model formulation for Λi(u;tk) is similar to the one proposed by Diggle et al. (2005). Notice that although ζi(u;tk) is a stationary process, if μi(u;tk) is a non-constant function for a fixed tk, then Λi(u;tk) has spatially varying mean and covariance functions depending on the points’ locations. In this case, the resulting Cox process with such intensity is called an intensity-reweighed stationary point process (Baddeley et al. 2000), similar to a real-valued process with varying mean and stationary residuals (Diggle et al. 2013).
Concerning Eq. (3), we have to define the mean structure μi(u;tk) and the correlation function ρi(h;tk). Starting with μi(u;tk), we will set it as a function of the previously estimated Ii(tk) curve, ∀i,k. We aim to have the expected number of infected individuals in the studied region somehow similar to what we estimated using the compartment model. This strategy will be beneficial when making predictions, as discussed in Sect. 4. In particular, we set the mean structure to4 μi(u;tk)=λ0,i(u;tk)·Ii(tk)·exp{ω1,ix1,i(u;tk)+⋯+ωp,ixp,i(u;tk)},
where λ0,i(u;tk) is a non-negative real-valued function, such that ∫Uλ0,i(u;tk)=1. In practice, λ0,i(u;tk) will be set as proportional to the population density function for class i. Here, notice that λ0,i(u;tk) depends on tk and might evolve over time if the population structure also changes. By setting it that way, we are using the total population distribution as a proxy for the locations of the individuals at risk of infection. Also, for each group, we can include a vector of p spatio-temporal covariates representing risk factors (x1,i(u;tk),…xp,i(u;tk)) with associated coefficients (ω1,i,…,ωp,i)⊤.
Now, to define the covariance structure of ζi(u;tk), first, notice that it can be written asζi(u;tk)=β0,i+ϑi(u;tk),
where β0,i is the mean component of ζi(u;tk), and ϑi(u;tk) is a zero-mean spatially dependent Gaussian process with covariance function ϕi(h;tk)=σi2ρi(h;tk). In particular, we assume a Matérn model (Matérn 1960) for the correlation function, a flexible correlation model appearing in many fields. Thus,5 ρi(h;tk)=12νi-1Γ(νi)(κi·h)νiKνi(κi·h),
such that νi,k=νi and κi,k=κi, ∀k are unknown parameters, and Kνi(·) is a modified Bessel function of 2nd order for age class i. For later reference, we will name this approach REF.
Aiming for more flexible models, we can still add other structures to the REF model, for instance, independent and identically distributed (i.i.d.) and autoregressive components. More specifically, we will also define the IID and AR1 model extensions. The IID model is defined asζi(u;tk)=β0,i+ϑi(u;tk)+εi(u;tk),∀i,
where εi(u;tk) is a zero-mean independent Gaussian process with variance σi,ε2. Here, εi(u;tk) acts like an unstructured exchangeable component modeling the uncorrelated noise in space and time. Also, the AR1 model is defined asζi(u;tk)=β0,i+ϑi(u;tk)+εi(u;tk)+υi(tk),∀i,
such thatυi(tk)=αiυi(tk-1)+ϱi(tk),fork=1,…,n,
where |αi|<1, ∀i, and ϱi(tk) is a zero-mean temporally independent Gaussian process with variance σi,ϱ2. In that case, υi(tk)∼Normal(0,σi,ϱ2/(1-αi2)), ∀k, such that Cov(νi(tk),νi(tk+m))=αi|m|·σi,ϱ2/(1-αi2). Also, notice that υi(tk) does not depend on the location, and therefore, the corresponding Cox process can account for spatial clustering but not temporal clustering. In that way, we are not modeling the local dependence of ζi(u;tk) over time; instead, the temporal dependence is commonly included for all u∈U. To overcome this issue, one could define ϱi, ∀i, as a zero-mean temporally independent but spatially dependent Gaussian process for each tk with, for example, a Matérn covariance model. However, this would result in a more computationally expensive inference procedure that typically requires larger data sets.
As a final remark, notice that another natural extension would be including a random effect for the interaction between space and time [e.g., as in Knorr-Held (2000)]. However, due to the typical number of data points observed in our infectious-disease-modeling problems, it may become computationally prohibitive fitting such models.
Inference and prediction in space and time
The final model for the counting number of infected individuals in all age groups i at tk is specified as follows6 Ni(tk)|Λi(u;tk)=λi(u;tk)∼Poisson∫Uλi(u;tk)du,∀i,kΛi(u;tk)=μi(u;tk)·exp{ζi(u;tk)}μi(u;tk)=λ0,i(u;tk)·Ii(tk)ζi(u;tk|ηi)∼Gaussian Process(β0,i,ϕi(h;tk|ηi))ηi∼priors,
where ϕi(h;tk|ηi) is a covariance function that depends on the selected model from Sect. 3.2, and ηi is a vector of parameters and hyperparameters.
We fit this model, using R-INLA (Rue et al. 2009), by employing the gridding approach described by Moraga (2020). Specifically, we create a regular grid over the studied region and model the number of occurring events in each grid cell cj as Ni,j(tk)∼Poisson(θi,j,k), such that θi,j,k=∫cjλi(u;tk)du. In that way, for sufficiently small cells, we can approximate such an integral by |cj|·λi(u;tk), for any u∈cj, where |·| denotes area. Also, in R-INLA, ζi(u;tk) is defined as a zero-mean process; therefore, to accommodate this change in the mean of the intensity process, we must include an intercept in the linear predictor for the Poisson regression formula when fitting the model.
As a final comment, notice that if we have “well predicted” values for Ii(tk) for some k in the future, the mean structure of the spatio-temporal modeling, namely μi(u;tk), can greatly help us to explain λi(u;tk) for non-observed time points, as we will see in Sect. 4.
Simulation study
In this section, we perform a simulation study to assess the performance of our model and compare it with a null model in which information from the estimated Ii(tk) curves is not used. Since the exact locations of infected individuals over time are still difficult to obtain (Hernández-Orallo et al. 2020), we use a combination of real and synthetic data. Specifically, we consider many scenarios and simulate point patterns for the locations of infected individuals in all age groups i observed over space and time. Then, we fit a null and our proposed model to assess and compare their prediction performances. The code to reproduce such analyses is available on https://github.com/avramaral/PP_SIR.
Data simulation
We consider as a study region an area of approximately 3 km2 in São Paulo, Brazil (Figure SF1, Supplementary Information). In such a region, we use the estimated population size (WorldPop 2020) defined in each of the (approximately) 100 × 100 m cells (with 39,040 individuals in total) as a base function that mimics the real intensity that describes how infectious individuals are distributed over space. Then, we simulate daily observations for 100 days.
We divide the population into three age groups, namely 0–19, 20–59, and 60+, so that the proportion of individuals in Brazil that fall into each cathegory is given by 0.33, 0.60, and 0.07, respectively (Nations 2019). The contact matrix Cij is also defined for the same three age groups and determined by the estimates from Prem et al. (2017) (Table ST1, Supplementary Information). Population age distribution and contact matrix data were retrieved using the COVOID package (Fitzgerald et al. 2020).
Regarding the simulation of the infected curves, we will do this in two ways. First, we will assume the data comes from the model described by Eqs. (1) and (2). Second, to assess how well our approach behaves for an incorrectly specified model for the temporal component, we will also generate data from a different model, i.e., we will violate the SIR assumptions and conduct a sensitivity analysis under these conditions. In particular, we will simulate from the model described in Section SS2 (Supplementary Information), based on Chen et al. (2016) and Held et al. (2005). For reference, name it Autoregressive Conditional Negative Binomial (ACNB) model.
Given β, γ and φ, for the SIR scenarios, we can simulate the epidemic dynamics from Eqs. (1) and (2) and sample it at tk, ∀k. In particular, when considering the SIR temporal modeling, notice that our goal is estimating the Si(tk), Ii(tk), and Ri(tk) curves for all age groups i. To do this, we will consider two scenarios, namely “early peak” (EP) and “flat curve” (FC) for the infected individuals. The corresponding parameters β, γ, and φ, will be set as 0.04, 0.2, and 100 for scenario EP and 0.0175, 0.1, and 100 for scenario FC. The chosen scenarios aim to cover the situations in which (1) we observe a high peak for the infectious in the very beginning and, therefore, it is easy to make predictions for future times under the SIR model assumptions, and (2) the epidemic is still half-way through and, therefore, it is not trivial to predict what will happen next. Finally, the simulated number of infected individuals for the SIR model in each group and under each scenario (EP and FC) can be seen in Figure SF2 (Supplementary Information). Similarly, the data simulated from the ACNB model can be seen in Figure SF3 (Supplementary Information).
Once we have simulated the Ii(tk) curves, ∀i,k, the true generated intensity functions will be sampled from the following modelΛi(u;tk)=λ0,i(u;tk)·Ii(tk)·exp{ζi(u;tk)},for eachiandk=0,1,…,99,
where λ0,i(u;tk) is the normalized populational grid, and ζi(u;tk) is defined according to the IID or AR1 models, as in Sect. 3.2, with parameters specified in Table ST3 (Supplementary Information). Then, the number of infected individuals and their respective locations will be generated from a Poisson process with the corresponding intensity function λi(u;tk).
Fitted models
Based on the data described in Sect. 4.1, we will fit a null model (M0) and our proposed alternative model (M1). The difference is that information from the estimated Ii(tk) curves will not be used for the null model. In particular, for the counting number of infected individuals in all age groups i and at tk modeled as followsNi(tk)|Λi(u;tk)=λi(u;tk)∼Poisson∫Uλi(u;tk)du,
the null model (M0) will be given by7 Λi(u;tk)=exp{ω0,i+ζi(u;tk)},
such that ω0,i is an unknown intercept, and the alternative (M1) model will be given by8 Λi(u;tk)=μi(u;tk)·exp{ζi(u;tk)},
such that μi(u;tk) is defined as in Eq. (4); that is, the alternative model will depend on both the base function λ0,i(u;tk) and the estimated Ii(tk) curve. Also, ζi(u;tk) will be defined in all scenarios based on one of the structures introduced in Sect. 3.2, namely IID and AR1.
Implementation and results
For the simulated data sets introduced in Sect. 4.1, we can fit the models (both null and alternative) that we have just described. In particular, the temporal component will be modeled as in Sect. 3.1 with the SIR approach, and the spatial modeling for all age groups i and all tk will be performed as detailed in Sect. 3.2; that is, we will consider the point pattern of infected individuals in each of the discretized time windows to estimate and predict the corresponding intensity functions. Here, we specify the remaining terms and analyze the obtained results.
Starting with the temporal modeling, the number of individuals in each compartment, the Si(tk), Ii(tk), and Ri(tk) curves, ∀i,k, will be described by Eq. (1)—with an approximated solution obtained as mentioned in Sect. 2.1. Also, the stochasticity from the sampling procedure will be given by Eq. (2). In that case, the prior distributions were specified as Normal(0.5,1) for β and γ and Normal(1,100) for φ, all truncated at 0. This choice for the priors is vaguely enough for our problem, and the chains for the posterior sampled values were well mixed in all cases (Figures SF4 and SF5 for the data generated from the SIR model (scenarios EP and FC, respectively), and Figure SF6 for the data generated from ACNB model—Supplementary Information). In particular, using RStan, we set the number of chains, the number of iterations and the burn-in size as 4, 4, 000, and 2, 000, respectively.
Furthermore, as we also want to make predictions, all models throughout this section will be fitted with data up to t49. The estimated parameters for all scenarios, namely SIR-EP, SIR-FC, and ACNB, can be seen in Table 1. Additionally, the estimated curves for the three scenarios (SIR-EP, SIR-FC, and ACNB) and age group 20–59 can be seen in Fig. 2. The other two sets of fitted curves for the age groups 0–19 and 60+ can be seen in Figure SF7 (Supplementary Information).Table 1 Estimated parameters (and 95% equal-tail credible interval) for the two SIR scenarios, namely EP and FC, and for the model fitted with data from ACNB. Models were fitted with data up to tk=49
Scenario True value Posterior mean 95% equal-tail credible interval
β SIR-EP 0.04 0.0400 (0.0399, 0.0402)
γ SIR-EP 0.2 0.1998 (0.1972, 0.2024)
1/φ SIR-EP 0.01 0.0095 (0.0070, 0.0132)
β SIR-FC 0.0175 0.0173 (0.0170, 0.0176)
γ SIR-FC 0.1 0.0964 (0.0912, 0.1013)
1/φ SIR-FC 0.01 0.0091 (0.0067, 0.0131)
β ACNB – 0.0272 (0.0227, 0.0315)
γ ACNB – 0.2972 (0.2184, 0.3682)
1/φ ACNB – 0.9843 (0.7894, 1.2548)
Fig. 2 Estimated Ii(tk) curves (black curves) for the age group 20–59 in the two SIR scenarios, namely EP (left panel) and FC (middle panel), and the ACNB scenario (right panel). Models were fitted with data up to t49 (vertical dashed line). The red curves correspond to the observed number of infected individuals over time
From Table 1, we can see that the model parameters for the SIR scenarios were well estimated, and their true values were always within the 95% equal-tail estimated credible interval. Also, from Figs. 2 and SF7 (Supplementary Information), the estimated curves (given by the solution of Model (1) averaged over all sampled β, γ, and φ parameters) approximate the observed count data well. However, for the data that violates the SIR assumptions, i.e., data from the ACNB model, the corresponding fitted model presents higher variability (expected since the fitted SIR model is not that flexible). Moreover, based on Figs. 2 and SF7 (Supplementary Information) for the ACNB scenario, we can observe that the estimated curves under this setting do not approximate well the true ones. Nonetheless, as we will confirm next, even in that case, our alternative model still performs better than the null model when using such a poorly fitted mean component.
Provided that we have fitted the temporal model and estimated the Ii(tk) curves, ∀i,k, we can now employ our modeling approach in space-time, as in Sect. 3.2. Here, we will use the same regular grid over U as in Figure SF1 (Supplementary Information) for the inference and prediction steps. Also, as discussed in Sect. 3, we will fit the spatio-temporal model under different scenarios (IID and AR1); in particular, we will compare a null model (M0), as in Eq. (7), with our proposed alternative (M1) model, as in Eq. (8). So that we can assess whether bringing information (even if it is not good quality) from the SIR compartment modeling approach, as in Eq. (4), helps to describe the intensity function for the observed point processes and make predictions. Table 2 lists all possible combinations for data generation and model fitting.Table 2 Different space-time scenarios (indexed by the “ST Sce.” columns) under which we will generate data and fit the spatio-temporal models. “T Sce.” refers to the two possible SIR scenarios, namely EP and FC, and the ACNB one—all for temporal data generation. “Data Gen.” refers to the spatio-temporal models, as in Sect. 3.2, when generating ζi(u;tk). Furthermore, “Model Fit.” refers to the possible fitted models (M0 and M1) with the different proposed spatio-temporal structures (IID and AR1)
ST Sce T Sce Data Gen Model Fit ST Sce T Sce Data Gen Model Fit ST Sce T Sce Data Gen Model Fit
01 SIR-EP IID M0 IID 09 SIR-FC IID M0 IID 17 ACNB IID M0 IID
02 SIR-EP IID M1 IID 10 SIR-FC IID M1 IID 18 ACNB IID M1 IID
03 SIR-EP IID M0 AR1 11 SIR-FC IID M0 AR1 19 ACNB IID M0 AR1
04 SIR-EP IID M1 AR1 12 SIR-FC IID M1 AR1 20 ACNB IID M1 AR1
05 SIR-EP AR1 M0 IID 13 SIR-FC AR1 M0 IID 21 ACNB AR1 M0 IID
06 SIR-EP AR1 M1 IID 14 SIR-FC AR1 M1 IID 22 ACNB AR1 M1 IID
07 SIR-EP AR1 M0 AR1 15 SIR-FC AR1 M0 AR1 23 ACNB AR1 M0 AR1
08 SIR-EP AR1 M1 AR1 16 SIR-FC AR1 M1 AR1 24 ACNB AR1 M1 AR1
To fit Model (6), we used R-INLA and set the priors for the parameters and hyperparameters as the default distributions, ω0,i∼Normal(0,(1/ι0,i)), such that ι0,i=0, log(8νi/κi)∼LogGamma(1,5×10-5), log(1/σi2)∼LogGamma(1,0.01), log(1/σi,ε2)∼LogGamma(1,5×10-5), log((1+αi)/(1-αi))∼Normal(0,(1/0.15)), and log((1/σi,ϱ2)·(1-αi))∼LogGamma(1,5×10-5). Besides, the Gaussian processes ζi(u;tk), ∀i were defined as following a Matérn model for the correlation structure, as in Eq. (5), with smoothing parameter νi=ν, for all age groups i. Then, all possible combinations for the “data generation procedure” and “model fitting” from Table 2 were fitted for data observed up to t49. In that way, based on the samples drawn from the posterior distributions of the model parameters, we could directly estimate λi(u;tk) for all i and tk, such that k≤49, and predict all remaining k. Figure SF8 (Supplementary Information) shows the estimated number of infected individuals at t49, age group 20–59, and space-time scenario 16, as per Table 2. As a remark, under this setting, the true number of infected individuals in the entire region was 7,435, while the estimated value was, rounding it to zero decimal places, 7,444 (with a 95% equal-tail credible interval given by 6,781—8,160).
Following this procedure, we could fit our model and estimate (and predict) the intensity functions for all age groups—which is the same as estimating (and predicting) the number of infected individuals in each cell. As described in Sect. 3.2.1, assuming we have “reasonably well predicted” values for Ii(tk), we expect better quality prediction when describing the underlying process that generated the observed point pattern. This happens since Ii(tk) is plugged-in as the mean of the spatio-temporal process, which improves the overall model performance. The following section assesses the null and alternative models regarding their error for the estimated or predicted intensity functions compared to the true values for all scenarios listed in Table 2.
Model assessment
Aiming to compare the null model (M0) with our proposed alternative model (M1), we compute a measure of error for all scenarios from Table 2 for the difference between the estimated (or predicted) values and the true values for a quantity proportional to the intensity function. In particular, we analyze the difference between the estimated and true values for the number of infected individuals per cell—more specifically, we are interested in such errors when making predictions. However, we might want to use a scale-independent error measure to compare the data sets with different scales, such as data from different age groups. The Absolute Percentage Error (APE) has been widely employed in this regard (Bowerman et al. 2005), and it can be defined as follows for each group i and at tk,APEi,j,k=(fi,j,k-f^i,j,k)fi,j,k,
where fi,j,k and f^i,j,k correspond to the true and predicted number of infectious in group i, cell cj, and tk, respectively. However, since it produces infinite values if fi,j,k=0 for any j, it may not be suitable for our problem. Instead, we will use a modified version of APE, namely Arctangent Absolute Percentage Error (AAPE) (Kim and Kim 2016), which can be defined as follows9 AAPEi,j,k=arctan(fi,j,k-f^i,j,k)fi,j,k.
From Eq. (9), we can compute the error in predicting Ii(tk). In that way, considering the different scenarios presented in Table 2, our goal will be to compare the output obtained from the null and alternative models; that is, we will plot the errors for scenarios 01 and 02, 03 and 04, etc. Figure 3 shows the AAPEs comparing scenarios 15 and 16. Also, Figures SF9, SF10, SF11, SF12, SF13, SF14, SF15, SF16, SF17, SF18, and SF19 (Supplementary Information) show the other plots for the remaining pairs of corresponding scenarios.Fig. 3 Computed AAPEs for groups 1, 2, and 3 (0–19, 20–59, and 60+, respectively), all cells, and for all k. Scenarios 15 (upper row) and 16 (lower row). Models were fitted with data up to t49 (vertical solid line). The upper row corresponds to the errors for the fitted null model (M0), and the lower for the fitted alternative model (M1)
From Fig. 3 (and all others in the Supplementary Information), we can see that, although the estimated intensity functions (and therefore, the estimated number of infected individuals per cell) for the null and alternative models approximate equally well the true process for k≤49, when making predictions. That is, for k>49, the null model (M0) tends to mispredict the process values—if compared to the alternative model (M1) under the same setting. Also, when the epidemic ends, as we can observe at later times in the EP scenarios, the null model overestimates the number of infectious individuals, while for the alternative model, the estimated Ii(tk) curve greatly contributes to the correct predictions.
Case study
In this section, we model the initial number of COVID-19 cases in Cali, the third-biggest city in Colombia and one of the most populated. The data were provided by the Municipal Public Health Secretary of the city1 and recorded the confirmed COVID-19 cases at an individual-level from 21/01/2020 to 18/06/2020 (Fig. 4, top-left panel). In total, 4,518 unique individuals were observed. The data were collected so that we have access to the initial locations of the infected individuals and their first symptoms dates. We make two assumptions to obtain the infectious curves: (1) all infected individuals will not move while being infected, and (2) the recovery time is assumed to be five days—as suggested by He et al. (2020). Also, as no covariates (e.g., age) were made available, we do not divide the population into different age groups; instead, the Cij will only have one element (representing the average number of contacts a person of any age has). As in Sect. 4, an estimate for Cij was retrieved from the COVOID package (Fitzgerald et al. 2020).Fig. 4 Top-left panel: Map of Cali, Colombia, and 4,518 unique infected (COVID-19) individual locations (red dots) from 21/01/2020 to 18/06/2020. The map also shows the grid used to fit the model. Top-right panel: Estimated Ii(tk) curve. The model was fitted with data up to t129 (vertical dashed line). Bottom-left panel: Computed AAPEs based on the fitted Null Model (M0) for all tk. The model was fitted with data up to t129 (vertical solid line). Bottom-right panel: Computed AAPEs based on the fitted Alternative Model (M1) for all tk. The model was fitted with data up to t129 (vertical solid line)
Based on the approach described in Sects. 3 and 4.2, we fit the null (M0) and alternative (M1) models to the data. In particular, we rely on Eqs. (7) and (8), respectively, for describing the intensity processes. In this regard, λ0,i(u;tk) is set as proportional to the population density in Cali (WorldPop 2020), Ii(tk) is obtained from the temporal modeling approach described in Sect. 3.1 (Fig. 4, top-right panel), and ζi(u;tk) is defined according to the IID structure introduced in Sect. 3.2. Finally, we use data from 21/01/2020 to 29/05/2020 (130 days) to fit the model and make predictions for the remaining period (20 days). The error in predicting the future number of infected individuals in all 131 cells is computed based on Eq. (9).
Figure 4 (bottom-left and bottom-right panels) shows the computed errors for both fitted null and alternative models. From these plots, we can see that, for all tk, such that k≤139, the two models have similar performance; in fact, the Null Model seems to perform better than the Alternative Model. This happens because the SIR model is too rigid and fails to capture the shape of the true infectious curve for the observed time points (Fig. 4, top-right panel). The estimated SIR parameters are presented in Table ST4 (Supplementary Information). However, when forecasting, the information provided by the SIR step improves the quality (for the AAPE) of the Alternative Model, making it better than the Null Model for almost all cells.
Discussion
Infectious disease modeling is essential to understand epidemic dynamics in the past and predict its future. By doing this, researchers can identify highly infectious areas, and decision-makers can use such information to choose where and when to focus their resources. In this paper, we have introduced a new infectious disease modeling approach for spatio-temporal point pattern data. In particular, we proposed a two-step framework for modeling data on the infectious locations at each time for different age groups. Firstly, we used a compartment SIR model with contact matrix information to characterize the epidemic dynamics in time. Secondly, we incorporated this information into the mean component of a Cox process that models the epidemic dynamics in space-time for all groups. By doing this, provided that the temporal model in the first step is “reasonably well specified”, the spatio-temporal modeling is also guaranteed to produce reasonable estimates in space for time points also in the future.
Regarding implementation and always under a Bayesian framework, for the temporal and spatio-temporal models, we used RStan and R-INLA, respectively. Using these tools, we can be very flexible in the model specification without considerably changing the implemented procedure. For instance, in the temporal-modeling step, many different models for counting data could have been employed to describe the number of infected individuals over time. We believe that extending our approach in that direction may result in more flexible and, therefore, more valuable methods. Also, for the spatio-temporal modeling step, notice that the log-Gaussian Cox process (LGCP) is easily implemented in R-INLA, so we can take advantage of its speed when fitting the corresponding models. Alternatively, other models, e.g., self-exciting point processes (Reinhart 2018), may replace the LGCP in such a framework.
Aiming to assess our model, we first analyzed a combination of real and simulated data. From that, we have seen that when we compare the null model with our two-step modeling framework for space-time, our approach provides similar results for past values but performs much better for time points in the future (for the AAPE). In other words, by describing the mean component of our spatial Cox process through the estimated Ii(tk) curves from the compartment modeling step, we gain information about the epidemic dynamics and make better predictions. Second, we analyzed a COVID-19 data set. In particular, by modeling the number of infectious cases in space and time according to our approach, we observed that, even though the temporal component was not very well fitted to the observed count data, the spatio-temporal modeling still benefited from such estimates when making predictions for almost all cells.
However, our approach also has limitations. For instance, recall that the proposed framework assumes a SIR model with age structure to model the number of infected individuals over time and a log-Gaussian Cox process to model spatio-temporal patterns. These assumptions may not be appropriate for all scenarios, and other structures can be included depending on the application. For instance, although we focused on the SIR model with age groups throughout this work, other temporal models could have been used as the first step of our modeling approach. Even non-compartment models can be implemented—however, the main advantage of the SIR model in predicting future times well (assuming that the assumptions hold) might be lost. Also, note that the presented model uses infected individuals’ exact locations and times. These data may be challenging to obtain or present underreporting issues. However, if good quality data is available, our model can help us to understand infectious diseases spreading and contribute to health policies.
Thus, in future work, we might be interested in extending such an approach for a broader scenario, relaxing some of the assumptions made in Sect. 2.1, or proposing alternative ways to deal with the challenge of obtaining point pattern data for the infected individuals. Finally, to increase our model performance, we could add covariates that are known to affect infection transmission and structures that consider reporting delays and underreporting. This might help us to describe the underlying intensity processes better and obtain smaller errors when forecasting.
Supplementary Information
Below is the link to the electronic supplementary material.Supplementary file 1 (pdf 6332 KB)
Acknowledgements
The authors acknowledge the support of Professor Francisco J. Rodríguez-Cortés for his valuable help in obtaining the agreements for using the COVID-19 dataset and the Municipal Public Health Secretary of Cali, Colombia, for providing it.
Funding
The authors have not disclosed any funding.
Declarations
Conflict of interest
The authors declare that they have no conflict of interest.
1 https://www.cali.gov.co/salud/
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Atkinson KE An introduction to numerical analysis 2008 New York Wiley
Baddeley AJ Møller J Waagepetersen R Non-and semi-parametric estimation of interaction in inhomogeneous point patterns Stat Neerl 2000 54 329 350 10.1111/1467-9574.00144
Blackwood JC Childs LM An introduction to compartmental modeling for the budding infectious disease modeler 2018 London Taylor & Francis
Bowerman BL O’Connell RT Koehler AB Forecasting, time series, and regression: an applied approach 2005 Nashville South-Western Pub
Britton T Pardoux E Ball F Laredo C Sirl D Tran VC Stochastic epidemic models with inference 2019 Berlin Springer
Chen CW So MK Li JC Sriboonchitta S Autoregressive conditional negative binomial model applied to over-dispersed time series of counts Stat Methodol 2016 31 73 90 10.1016/j.stamet.2016.02.001
Cox DR Some statistical methods connected with series of events J R Stat Soc Ser B (Methodol) 1955 17 129 157
Davies NG Klepac P Liu Y Prem K Jit M Eggo RM Age-dependent effects in the transmission and control of COVID-19 epidemics Nat Med 2020 26 1205 1211 10.1038/s41591-020-0962-9 32546824
Diekmann O Heesterbeek JAP Metz JA On the definition and the computation of the basic reproduction ratio R0 in models for infectious diseases in heterogeneous populations J Math Biol 1990 28 365 382 10.1007/BF00178324 2117040
Diekmann O Heesterbeek J Roberts MG The construction of next-generation matrices for compartmental epidemic models J R Soc Interface 2010 7 873 885 10.1098/rsif.2009.0386 19892718
Diggle PJ Rowlingson B Su T-L Point process methodology for on-line spatio-temporal disease surveillance Environmetrics 2005 16 423 434 10.1002/env.712
Diggle PJ Moraga P Rowlingson B Taylor BM Spatial and spatio-temporal log-Gaussian Cox processes: extending the geostatistical paradigm Stat Sci 2013 28 542 563 10.1214/13-STS441
Fitzgerald O, Hanly M, Churches T (2020) COVOID: age-structured epidemic models in COVOID
Geng X Katul GG Gerges F Bou-Zeid E Nassif H Boufadel MC A kernel-modulated SIR model for Covid-19 contagious spread from county to continent Proc Natl Acad Sci 2021 118 e2023321118 10.1073/pnas.2023321118 33958443
Giuliani D Dickson MM Espa G Santi F Modelling and predicting the spatio-temporal spread of COVID-19 in Italy BMC Infect Dis 2020 20 1 10 10.1186/s12879-020-05415-7
He X Lau EH Wu P Deng X Wang J Hao X Lau YC Wong JY Guan Y Tan X Temporal dynamics in viral shedding and transmissibility of COVID-19 Nat Med 2020 26 672 675 10.1038/s41591-020-0869-5 32296168
Held L Höhle M Hofmann M A statistical framework for the analysis of multivariate infectious disease surveillance counts Stat Model 2005 5 187 199 10.1191/1471082X05st098oa
Hernández-Orallo E Manzoni P Calafate CT Cano J-C Evaluating how smartphone contact tracing technology can reduce the spread of infectious diseases: the case of COVID-19 IEEE Access 2020 8 99083 99097 10.1109/ACCESS.2020.2998042 34192101
Hindmarsh AC ODEPACK, a systematized collection of ODE solvers Sci Comput 1983 1 55 64
Hossain MM, Tasnim S, Sultana A, Faizah F, Mazumder H, Zou L, McKyer ELJ, Ahmed HU, Ma P (2020) Epidemiology of mental health problems in COVID-19: a review. F1000Research 9
Kaye AD Okeagu CN Pham AD Silva RA Hurley JJ Arron BL Sarfraz N Lee HN Ghali GE Gamble JW Economic impact of COVID-19 pandemic on healthcare facilities and systems: international perspectives Best Pract Res Clin Anaesthesiol 2021 35 293 306 10.1016/j.bpa.2020.11.009 34511220
Keeling MJ Rohani P Modeling infectious diseases in humans and animals 2011 Princeton Princeton University Press
Kermack WO McKendrick AG Containing papers of a mathematical and physical character Proc R Soc Lond Ser A 1927 115 700 721
Kim S Kim H A new metric of absolute percentage error for intermittent demand forecasts Int J Forecast 2016 32 669 679 10.1016/j.ijforecast.2015.12.003
Knorr-Held L Bayesian modelling of inseparable space-time variation in disease risk Stat Med 2000 19 2555 2567 10.1002/1097-0258(20000915/30)19:17/18<2555::AID-SIM587>3.0.CO;2-# 10960871
Lau MS Gibson GJ Adrakey H McClelland A Riley S Zelner J Streftaris G Funk S Metcalf J Dalziel BD A mechanistic spatio-temporal framework for modelling individual-to-individual transmission-With an application to the 2014–2015 West Africa Ebola outbreak PLoS Comput Biol 2017 13 e1005798 10.1371/journal.pcbi.1005798 29084216
Lawson AB Kim J Space-time covid-19 Bayesian SIR modeling in South Carolina PLoS ONE 2021 16 e0242777 10.1371/journal.pone.0242777 33730035
Li R Pei S Chen B Song Y Zhang T Yang W Shaman J Substantial undocumented infection facilitates the rapid dissemination of novel coronavirus (SARS-CoV-2) Science 2020 368 489 493 10.1126/science.abb3221 32179701
Li J, Blakeley D et al (2011) The failure of R0. Computational and mathematical methods in medicine 2011
Matérn B Spatial variation 1960 Berlin Springer
Moller J Waagepetersen RP Statistical inference and simulation for spatial point processes 2003 Boca Raton CRC Press
Moraga P Geospatial health data: modeling and visualization with R-INLA and shiny 2019 Boca Raton Chapman and Hall/CRC
Moraga P Species distribution modeling using spatial point processes: a case study of sloth occurrence in Costa Rica RJ 2020 12 1 10 10.32614/RJ-2021-017
Moraga P, Ketcheson DI, Ombao HC, Duarte CM (2020) Assessing the age- and gender-dependence of the severity and case fatality rates of COVID-19 disease in Spain. Wellcome Open Res 5
Mossong J Hens N Jit M Beutels P Auranen K Mikolajczyk R Massari M Salmaso S Tomba GS Wallinga J Social contacts and mixing patterns relevant to the spread of infectious diseases PLoS Med 2008 5 e74 10.1371/journal.pmed.0050074 18366252
Nations U (2019) World Population Prospects 2019: Department of Economic and Social Affairs
Pak A Adegboye OA Adekunle AI Rahman KM McBryde ES Eisen DP Economic consequences of the COVID-19 outbreak: the need for epidemic preparedness Front Public Health 2020 8 241 10.3389/fpubh.2020.00241 32574307
Petzold L Automatic selection of methods for solving stiff and nonstiff systems of ordinary differential equations SIAM J Sci Stat Comput 1983 4 136 148 10.1137/0904010
Prem K Cook AR Jit M Projecting social contact matrices in 152 countries using contact surveys and demographic data PLoS Comput Biol 2017 13 e1005697 10.1371/journal.pcbi.1005697 28898249
Reinhart A A review of self-exciting spatio-temporal point processes and their applications Stat Sci 2018 33 299 318
Rue H Martino S Chopin N Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations J R Stat Soc Ser B (Stat Methodol) 2009 71 319 392 10.1111/j.1467-9868.2008.00700.x
Stan Development Team (2021) R package version 2.21.3
Towers S Feng Z Social contact patterns and control strategies for influenza in the elderly Math Biosci 2012 240 241 249 10.1016/j.mbs.2012.07.007 22877728
Viboud C Boëlle P-Y Cauchemez S Lavenu A Valleron A-J Flahault A Carrat F Risk factors of influenza transmission in households Br J Gen Pract 2004 54 684 689 15353055
WorldPop (2020) The spatial distribution of population in 2020, Brazil
Wu JT Leung K Bushman M Kishore N Niehus R de Salazar PM Cowling BJ Lipsitch M Leung GM Estimating clinical severity of COVID-19 from the transmission dynamics in Wuhan, China Nat Med 2020 26 506 510 10.1038/s41591-020-0822-7 32284616
| 0 | PMC9746591 | NO-CC CODE | 2022-12-15 23:21:56 | no | Stoch Environ Res Risk Assess. 2022 Dec 13;:1-15 | utf-8 | Stoch Environ Res Risk Assess | 2,022 | 10.1007/s00477-022-02354-4 | oa_other |
==== Front
Biomass Convers Biorefin
Biomass Convers Biorefin
Biomass Conversion and Biorefinery
2190-6815
2190-6823
Springer Berlin Heidelberg Berlin/Heidelberg
3645
10.1007/s13399-022-03645-0
Original Article
Multiobjective response and chemometric approaches to enhance the phytochemicals and biological activities of beetroot leaves: an unexploited organic waste
Chaari Moufida 1
Elhadef Khaoula 1
Akermi Sarra 1
Hlima Hajer Ben 2
Fourati Mariam 1
Chakchouk Mtibaa Ahlem 1
Sarkar Tanmay 3
Shariati Mohammed Ali 45
Rebezov Maksim 456
D’Amore Teresa 7
Mellouli Lotfi 1
http://orcid.org/0000-0002-8839-7377
Smaoui Slim [email protected]
1
1 grid.412124.0 0000 0001 2323 5644 Laboratory of Microbial Biotechnology and Engineering Enzymes (LMBEE), Center of Biotechnology of Sfax (CBS), University of Sfax, Road of Sidi Mansour Km 6, P.O. Box 1177, 3018 Sfax, Tunisia
2 grid.412124.0 0000 0001 2323 5644 Laboratory of Enzymatic Engineering and Microbiology, Algae Biotechnology Unit, Biological Engineering Department, National School of Engineers of Sfax, University of Sfax, 3038 Sfax, Tunisia
3 grid.505920.b Department of Food Technology, Malda Polytechnic, West Bengal State Council of Technical Education, Govt. of West Bengal, Malda, 732102 West Bengal India
4 grid.446163.2 0000 0000 9194 3477 Department of Scientific Research, Russian State Agrarian University-Moscow Timiryazev Agricultural Academy, Moscow, 127550 Russia
5 grid.496798.d Department of Scientific Research, K.G. Razumovsky Moscow State University of Technologies and Management (The First Cossack University), 109004 Moscow, Russia
6 grid.465377.4 0000 0004 5940 5280 Department of Scientific Research, V. M. Gorbatov Federal Research Center for Food Systems, Moscow, 109316 Russia
7 grid.508082.7 0000 0004 1755 4106 Chemistry Department, Istituto Zooprofilattico Sperimentale Della Puglia E Della, Foggia, Italy
13 12 2022
115
6 10 2022
2 12 2022
5 12 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Research on medicinal plants is developing each day due to inborn phytochemicals, which can encourage the progress of novel drugs. Most plant-based phytochemicals have valuable effects on well-being. Among them, beetroot leaves (BL) are known for their therapeutic properties. Here, three solvents, namely, acetonitrile, ethanol, and water, and their combinations were developed for BL extraction and simultaneous assessment of phytochemical compounds and antioxidant and antifoodborne pathogen bacteria activities. By using the augmented simplex-centroid mixture design, 40.40% acetonitrile diluted in water at 38.74% and ethanol at 20.86% favored the recovery of 49.28 mg GAE/mL (total phenolic content (TPC)) and 0.314 mg QE/mL (total flavonoid content (TFC)), respectively. Acetonitrile diluted in water at 50% guarantees the best antioxidant activity, whereas the optimal predicted mixture for the highest antibacterial activity matches 24.58, 50.17, and 25.25% of acetonitrile, ethanol, and water, respectively. These extraction conditions ensured inhibition of Staphylococcus aureus, Salmonella enterica, and Escherichia coli, respectively, at 0.402, 0.497, and 0.207 mg/mL. Under optimized conditions, at three concentrations of BL, minimal inhibitory concentration (MIC), 2 × MIC, and 4 × MIC, a linear model was employed to investigate the inhibition behavior against the three tested bacteria. The early logarithmic growth phase of these bacteria illustrated the bactericidal effect of optimized extracted BL with a logarithmic growth phase inferior to 6 h. Therefore, BL extract at 4 × MIC, which corresponds to 1.608, 1.988, and 0.828 mg/mL, was more efficient against S. aureus, S. enterica, and E. coli.
Keywords
Beetroot leaves
Mixture design
Phytochemical contents
Antioxidant and antibacterial activities
Foodborne bacterial pathogen inactivation
Chemometric approaches
==== Body
pmcIntroduction
The disastrous effects of the COVID-19 pandemic conjointly with additional challenges in health, nutrition, and demography incite researchers and expert practitioners to investigate sustainable and preventive remedies. For instance, medicinal plants and derived products, with eminent bioactive compounds and biological potential [1–3], have been recommended to strengthen the immune system during the COVID-19 pandemic [4, 5].
Among these medicinal plant-derived natural products, beetroot (Beta vulgaris L.) has been announced for numerous bioactivities. For instance, this plant has long been used in traditional medicine to treat a wide variety of diseases [6, 7]. The claimed therapeutic use of beetroot includes antianxiety, anti-inflammatory, antihypertensive, anticancer, antiobesity, neuroprotective, cardioprotective, and antioxidant effects [7–10].
In contrast to the considerable amount of research conducted on beetroot extract and juice, there are comparatively few studies on other beetroot by-products, such as leaves, which are discarded as waste during food processing. Nevertheless, the literature displayed that this unexploited biomass has been reviewed as a good source of bioactive compounds including proteins [11], minerals [12], and multiple phytochemicals [13] with biomedical applications [14–16]. Among beetroot leaf (BL) phytochemicals, the flavonoid glycosides acquired from apigenin, termed vitexin (such as vitexin-2-O-rhamnoside and vitexin-2-O-xyloside), which is analyzed with liquid chromatography-mass spectrometry [17], could be reliable for the inhibition of α-glucosidase activity and therefore arbitrate this hypoglycemic effect [18]. Lorizola et al. [18] confirmed that BL rich in vitexin derivatives can protect the liver from damage induced by a high-fat diet [19]. Likewise, β-aldehydes and phenolic compounds found in beetroot have been connected to enhance low-density lipoprotein (LDL) resistance to oxidation [18]. These compounds present a preventive capacity against cardiovascular diseases by lowering the oxidative effects of free radicals on lipids [10, 20] and decreasing the ratio between LDL cholesterol (LDL-C) and serum cholesterol (TC) levels [20, 21]. There are certain papers on the compositions of BL [18–20], but none of them have revealed the simultaneous exploration of the bioactive compounds of leaves’ phenolic constituents, antioxidants, and biological activities against foodborne bacterial pathogens, such as Staphylococcus aureus, Salmonella enterica, and Escherichia coli. When these pathogens come into contact with food, they cause infections and food poisoning, which is a major public health concern [22–24].
On the other hand, considering the recovery of bioactive compounds’ polarity may notably vary with numerous conditions [25], it is complicated to evolve a perfect method for the extraction of all phenolic compounds, and extraction systems can engender inefficient results. In this way, variables like solvent type, solvent-sample ratio, solvent–solvent ratio, extraction time, temperature, and pH can control the efficiency of extraction of total phenolic compounds and consequently the antioxidant/antimicrobial activities [26, 27]. In this respect, the blend of solvents, which can vary from binary, ternary, and even multicomponent mixtures at different polarities, has been recommended for the effective extraction of phytochemicals [28, 29]. Thereby, a hypothesis is that solvent mixtures produce efficient extraction of phenolic compounds. Statistical methods such as the mixture design have been widely used to investigate the effects of process variables on specific responses in studies in food technology and science [28]. This method can evaluate the interaction effect of multiple factors in different ranges employing a three-dimensional graph. In addition, this approach’s results allow identifying the synergetic effects of mixtures and predicting models that provide answers such as high phenolic recovery with maximal biological activities [29, 30]. The advantage is that multiple parameters can be calculated over the duration with fewer experiments and provide quantitative results [31]. Smaoui et al. [30] used an augmented simplex-centroid mixture design to examine the synergistic effect of solvents (acetone, ethanol, and water) at different polarities on the simultaneous production of phytochemical contents and antibacterial activity of Phoenix dactylifera L. (date) seeds. El Ksibi et al. [29] determined the best mixture of solvents (water, ethanol, and acetone) to optimize the recovery of colored phenolics from red pepper (Capsicum annum L.) by-products as a potential source of natural dye and assessment of its antimicrobial activity. These studies provide a reference for here to choose the mixing design model and predict the proportion of multicomponent mixed solvents.
The present paper aims to identify the most adequate mixture for the extraction of phenolic content from Tunisian BL in such a way as to produce the dry extract with high antioxidant and antibacterial activities. In addition, findings from this study provide an understanding of optimized extract’s efficacy versus common foodborne pathogenic microorganisms: Staphylococcus aureus, Salmonella enterica, and E. coli.
Materials and methods
Beetroot leaf extract preparation
Beetroot leaves (BL) were purchased from a local market based in Sfax, Tunisia. Leaves were washed with distilled water, dried at about 40 °C until it attains constant weight, and ground in a laboratory heavy-duty grinder (IKA, Germany) [32]. Finally, BL were stored at 4 °C in the dark until used for analysis.
By maceration with a constant stirring rate (150 rpm for 2 h), BL extracts were obtained in a proportion sample/solvent of 1/30 at 45 °C. Later, each mixture was centrifuged at 10.000 × g for 10 min [33]. The obtained supernatant was concentrated in a rotary evaporator for further evaluation of phytochemical content and antioxidant and antibacterial activity.
Phytochemical study of BL extracts
Total phenolic content quantification (TPC)
TPC of BL extract was evaluated using the reagent of Folin–Ciocalteu according to the method of López-Froilán et al. [34]. Briefly, 100 μL of the Folin–Ciocalteu (0.2 N) reagent was mixed with 100 μL of BL extract. Subsequently, 400 μL of Na2CO3 (7% (w/v)) and 250 μL of distilled water were added. All the mixtures were preceded to incubation at room temperature for 90 min. Gallic acid (GA) was used as standard (0–200 μg/L), and the absorbance was measured at 750 nm. TPCs were determined in mg GA equivalents/g (mg GAE/g).
Total phenolic content quantification (TFC)
TFC was assessed using the aluminum chloride (AlCl3) colorimetric assay described by Saharan et al. [35]. 500 μL of 2% AlCl3 was added to 500 μL of BL extract and incubated at room temperature for 1 h. Then, the absorbance was read at 420 nm. Quercetin was used as standard (0–50 μg/mL). TFCs were determined in milligrams of quercetin equivalents mg QE/mL.
Antibacterial activity: determination of minimum inhibitory concentration (MIC)
The target bacterial strains including Gram-positive bacteria Staphylococcus aureus ATCC 6538 and two Gram-negative bacteria Salmonella enterica ATCC 14,028 and Escherichia coli ATCC 25,922 were selected for the antibacterial activity assays according to the method proposed by Fourati et al. [36]. MIC values are the lowest concentration of BL, which could be detected when the microorganisms do not manifest any visible growth after incubation. MIC test was performed in sterile 96-well microplates with a final volume of 100 μL per well [37]. The corresponding extract of BL was transferred to each successive well in a view to perform a two-fold serial dilution from the original sample. Then, 10 μL of the cell suspension was added to each test well with a final inoculum concentration of 106 CFU/mL of each bacterium. Plates were incubated at 37 °C for 24 h. Later, 25 μL of MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) solution (0.5 mg/mL) was added to the wells as an indicator of microorganism growth and the plate was incubated again at 37 °C for 30 min.
DPPH assay
The DPPH (2-diphenyl-1-picrylhydrazyl) antioxidant activity of the BL extracts was determined according to Xu et al. [27]. Briefly, 1 mL aliquots of each sample were added to 1 mL of DPPH solution. After 30 min in the dark, the absorbance was measured at 517 nm. The concentration that provides 50% of the action of scavenging radicals (IC50), expressed in mg/mL, was determined from the plot of serial dilutions vs. present inhibition.
Mixture design
The simplex-centroid design was performed to obtain the optimum composition between solvents for maximum phytochemical content and antioxidant and antibacterial activities [38]. The factors represent the fraction of each solvent in the mixture, which ranges from 0 to 1. As a result, the experiment’s number for this design was equal to ten with various water, ethanol, and acetonitrile mixtures. The mixture representation is as follows: three pure products corresponding to X1: water, X2: ethanol, X3: acetonitrile, mixtures of two pure solvents (1/2:1/2): corresponding to midpoints (4, 5, 6), central point (1/3:1/3:1/3) (7), and three augmented points (8, 9, 10) ascribed to ternary combinations.
Linear, quadratic, and special cubic regression models were used for variations of all the effects of the interactions between the proportions within the same response. The following equation represents these models:Y=∑βi×Xi+∑βij×Xi×Xj+∑βijk×Xi×Xj×Xk+ε.
Y is the predicted response, while βi, βij, and βijk correspond to the regression coefficients for linear, binary, and ternary interaction effect terms, respectively. Xi, Xj, and Xk are the variables and ε is the random error. Minitab 16 software was employed for experimental design and data analysis.
Mode of action of BL extract
The effect of BL extract on S. aureus, S. enterica, and E. coli inhibition was determined in Luria Broth (LB) medium. The bacteria’s growth initially reached the start of the exponential phase (∼104 CFU/mL). Different concentrations of BL extract, obtained under optimized conditions, at 1 × MIC, 2 × MIC, and 4 × MIC, were added individually on an LB medium with S. aureus (0.402, 0.804, and 1.608 mg/mL), S. enterica (0.497, 0.994, and 1.988 mg/mL), and E. coli (0.207, 0.414, and 0.828 mg/mL), respectively, and incubated for 26 h at 37 °C. The number of CFU/mL was measured by plating the samples on LB medium at various times throughout the incubation duration. Samples were compared with the growth of the control sample (without BL extract), which were made under the same experimental circumstances. The assay was carried out in triplicate, and the means were expressed as log10 CFU/mL.
Statistical analysis
Two chemometric techniques consisting of principal component analysis (PCA) and hierarchical cluster analysis (HCA) were carried out. These techniques were conducted to identify the possible links between phytochemical contents (TPC and TFC) of different extract mixtures and their corresponding biological activities (antioxidant and antibacterial activities). The Ward technique and the squared Euclidean distance matrix were performed to define each cluster, resulting in a dendrogram. These approaches were achieved using XLSTAT software for Windows (v.2014.1.08, Addinsoft, New York, USA). The PCA type was Pearson (n), the plot type was correlation biplot, and the coefficient was automatic. HCA methodology was also applied to evaluate the relationships between trials/responses.
On the other hand, all tests were assayed in triplicate and expressed as the mean ± standard deviation of the measurements. The statistical program Statistical Package for the Social Sciences (SPSS) version 21.00 for Windows (SPSS Inc., Chicago, IL, USA) was utilized to analyze data. The variance was analyzed by one-way analysis of variance (ANOVA), and Tukey test was applied to compare each parameter at p < 0.05.
Results and discussion
Measured and predicted responses: TPC, TFC, and MICs of the tested bacteria (S. aureus, S. enterica, and E. coli) and DPPH (IC50) values are shown in Table 1. As a function of solvent characteristics using ANOVA of regression models (Tables 2, 3, and 4), the corresponding responses were traced in Figs. 1, 2, and 3. The resulted contour plots are composed of vertices, which reflect the response value, and the triangle edges represent the concentration, which depicts the individual components in their binary mixing. Gathered data from the trials were utilized to analyze the coefficients using regression analysis. The adequacy and fitness of the models were investigated using ANOVA, and p values were utilized to determine the significance of the coefficients. In addition, the coefficient of determination (R2) was used to assess the efficacy of models in approaching the desired output of each response. In fact, R2 values of 84.17, 70.20, 89.16, 71.11, 79.54, and 74.34% were obtained for TPC, TFC, MIC S. aureus, MIC S. enterica, MIC E. coli, and DPPH respectively, indicating that these fitted models properly present the variability of all the response. Additionally, it enables checking the quality of fit between predicted and experimental findings. It is a typical criterion for evaluating statistical performance [30].Table 1 Experimental variables and responses used in the mixture design
Variables TPC (mg GAE/mL) TFC (mg QE/mL) MIC S. aureus (mg/mL) MIC S. enterica (mg/mL) MIC E. coli (mg/mL) DPPH (IC 50) (mg/mL)
Runs Water
X1 Ethanol
X2 Acetonitrile
X3 Exp Pred Exp Pred Exp Pred Exp Pred Exp Pred Exp Pred
1 1.000 0.000 0.000 44.84±2.24b 44.915 0.222±0.011 b 0.244 2.500±0.125 c 2.352 1.250±0.063 c 1.150 0.625±0.031 c 0.586 1.225±0.061 d 1.134
2 0.000 1.000 0.000 42.32±2.12 a 41.242 0.289±0.014 d 0.275 0.625±0.031 a 0.704 0.312±0.016 a 0.430 0.312±0.016 b 0.303 0.420±0.021 b 0.482
3 0.000 0.000 1.000 41.00±2.05 a 41.580 0.291±0.015 d 0.287 0.625±0.031 a 0.648 0.625±0.031 b 0.580 0.312±0.016 b 0.317 1.500±0.075 e 1.424
4 0.500 0.500 0.000 44.12±2.21 b 43.460 0.221±0.011 b 0.290 0.625±0.031 a 0.556 0.625±0.031 b 0.743 0.312±0.016 b 0.266 0.630±0.032 bc 0.601
5 0.500 0.000 0.500 49.40±2.47 c 50.055 0.221±0.011 b 0.240 1.250±0.063 b 1.125 1.250±0.063 c 1.108 0.156±0.008 a 0.126 0.510±0.026 bc 0.343
6 0.000 0.500 0.500 40.00±2.00 a 39.500 0.251±0.013 c 0.235 0.625±0.031 a 0.727 0.625±0.031 b 0.798 0.156±0.008 a 0.156 1.120±0.056 d 1.106
7 0.333 0.333 0.333 49.28±2.46 c 48.012 0.231±0.012 b 0.274 0.625±0.031 a 0.830 0.312±0.016 a 0.533 0.312±0.016 b 0.200 0.750±0.038 c 0.434
8 0.6667 0.1667 0.1667 48.00±2.40 c 48.197 0.310±0.016 e 0.265 0.625±0.031 a 0.780 0.625±0.031 b 0.747 0.156±0.008 a 0.298 0.135±0.007 a 0.513
9 0.1667 0.6667 0.1667 39.80±1.99 a 44.265 0.213±0.011 a 0.274 0.625±0.031 a 0.604 1.250±0.063 c 1.037 0.156±0.008 a 0.214 0.630±0.032 bc 0.550
10 0.1667 0.1667 0.6667 47.20±2.36 c 45.880 0.250±0.012 c 0.261 0.625±0.031 a 0.775 0.625±0.031 b 0.675 0.156±0.008 a 0.172 0.443±0.022 b 0.778
For each run, responses with different letters are significantly different (one-way analysis of variance, p < 0.05; Tukey’s test), ± standard deviation (SD) of three measurements
Exp experimental, Pred predicted
Table 2 ANOVA results of regression models from mixture design for solvent optimization of TPC
Source DF Seq SS Adj SS Adj MS F value p value
Regression 6 105.975 105.9746 17.6624 4.66 0.031
Linear 2 42.896 9.0649 4.5325 2.68 0.042
Quadratic 3 55.646 38.2037 12.734 2.92 0.046
X1 × X2 1 1.458 0.0010 0.0010 2.00 0.041
X1 × X3 1 53.935 31.3152 31.3152 8.71 0.048
X2 × X3 1 0.253 2.4616 2.4616 2.37 0.050
Special cubic 1 7.432 7.4322 7.4322 3.12 0.074
X1 × X2 × X3 1 7.432 7.4322 7.4322 3.12 0.074
Residual error 3 19.934 19.9340 6.6447
Total 9 125.909
Table 3 ANOVA results of regression models from mixture design for solvent optimization of TFC
Source DF Seq SS Adj SS Adj MS F value p value
Regression 6 0.0030 0.0030 0.0005 2.19 0.042
Linear 2 0.0009 0.0011 0.0005 2.21 0.031
Quadratic 3 0.0016 0.0019 0.0006 2.24 0.028
X1 × X2 1 0.0003 0.0006 0.0006 2.23 0.024
X1 × X3 1 0.0002 0.0004 0.0004 2.17 0.042
X2 × X3 1 0.0011 0.0015 0.0015 2.55 0.056
Special cubic 1 0.0004 0.0004 0.0004 2.15 0.063
X1 × X2 × X3 1 0.0004 0.0004 0.0004 2.15 0.063
Residual error 3 0.0080 0.0080 0.0027
Total 9 0.0110
Table 4 ANOVA results of regression models from mixture design for solvent optimization of DPPH
Source DF Seq SS Adj SS Adj MS F value p value
Regression 6 1.182 1.182 0.197 2.45 0.010
Linear 2 0.305 0.511 0.256 2.88 0.039
Quadratic 3 0.863 0.666 0.222 3.63 0.079
X1 × X2 1 0.054 0.028 0.028 2.21 0.056
X1 × X3 1 0.801 0.592 0.592 4.35 0.068
X2 × X3 1 0.007 0.015 0.015 1.12 0.066
Special cubic 1 0.013 0.013 0.013 2.10 0.042
X1 × X2 × X3 1 0.013 0.013 0.013 2.10 0.042
Residual error 3 0.410 0.408 0.136
Total 9 1.590
Fig. 1 2D contour plot for the effect of different mixtures of BL extract on a TPC (mg GAE/mL) and b TFC (mg QE/mL)
Fig. 2 2D contour plot for the effect of different mixtures of BL extract on DPPH (expressed by IC50 (mg/mL))
Fig. 3 2D contour plot for the effect of different mixtures of BL extract on a S. aureus, b S. enterica, and c E. coli (expressed by MIC (mg/mL))
The effect of the system solvents on TPC
The most efficient mixture for TPC extraction was acetonitrile diluted in water at 50% (49.40 ± 2.47 mg GAE/mL), as mentioned in Table 1. Biondo et al. [19] detected a TPC of 15 mg/mL in BL methanolic extract which is a lower value compared to those found in this study. However, using just water or acetonitrile led to a TPC equal to 44.84 mg GAE/mL and 41.00 mg GAE/mL, respectively. According to [39], the TP content of Beta vulgaris extract using only water was equal to 27.72 mg GAE/mL. In fact, the binary combination exhibits a more effective interaction than when a solvent is separated, indicating that both solvents work in synergy. In this case, the solvent combination may raise polarity and expand plant components by allowing the solvent to easily penetrate the solid matrix [40]. This agrees with the study of Ameer et al.[41], who elucidate that the addition of water to organic solvents increases the solubility of polyphenols by modulating the polarity of the organic solvent and this increase may be due to the weakening of hydrogen bonds in aqueous solutions [30, 41]. It could also be due to the increase of basicity and ionization of polyphenols in these solvents. Additionally, the extractability of bioactive chemicals from plant materials could also be affected by solvent viscosity [42]. For instance, the low solvent viscosity of acetonitrile and water allows better diffusion into the pores of plant matrices and subsequently enhances the extraction of the compounds [43]. As a result, the polarity and viscosity of the solvents utilized influence polyphenol extractability. For TPC, p values of linear (p = 0.042) and quadratic (p = 0.046) were significant (Table 2). The generalized polynomial equation (Eq. 1) considered is as follows:1 TPCmgGAE/mL=44.915×X1+41.242×X2+41.58×X3+0.154×X1×X2+27.231×X1×X3-7.635×X2×X3+87.459×X1×X2×X3
The contour plots enable us to confirm the interaction influence between the mixtures. The dark green regions in all response surface plots denote the places where the maximum was identified (Fig. 1). Indeed, the binary mixture of water and acetonitrile at 55.56 and 44.44% extracted the highest level of TPC, as depicted in Fig. 1a. This agrees with the study of Alcântara et al. [28] who confirm the necessity of the use of various solvents to optimally extract all the polyphenols from plants. Likewise, the chemical complexity of the investigated plant, as well as the type and the polarity of the solvent utilized for extraction, may have considerable effects on the extraction of the phenolics.
The effect of the system solvents on TFC
The results of the TFC extracted from BL ranged from 0.213 ± 0.011 to 0.310 ± 0.016 mg QE/mL. The highest flavonoid content (0.310 mg QE/mL) was obtained at 66.667, 16.667, and 16.667% of water, ethanol, and acetonitrile, respectively. In fact, flavonoid glycosides are more soluble in water, a solvent proton donor, so they must be used with polar solvents [44]. For TFC, p values of linear and quadratic (X1 × X2 and X1 × X3) were significant (p < 0.05) (Table 3). The mathematical model for representing the response is expressed as follows:2 TFCmgQE/mL=0.2438×X1+0.2754×X2+0.2874×X3+0.1197×X1×X2-0.1043×X1×X3-0.1867×X2×X3+0.6472×X1×X2×X3
Figure 1b reveals that the ternary mixture (water–ethanol-acetonitrile) allows for high flavonoid extraction. The extraction efficiency of TFC in the extracts decreased as the water content in the ethanol decreased; this was in agreement with Vural et al. [45] study. Hence, the solubility of flavonoids in hydrophilic solvent systems is due to the presence of hydroxyl groups and glycosylic moieties in their molecular structure [46]. Therefore, the polarity of the solvents and the solubility of each chemical in the solvent employed for the extraction procedure might have different effects on flavonoids [47]. As a result, the extraction solvent had a significant impact on the obtained concentration of flavonoids present.
The effect of the system solvents on the antioxidant activity
Oxidative stress has been detected as a principal cause of the development of various diseases [38]. As a result, finding new effective antioxidant agents is critical. In the present investigation, the antioxidant activity of the different mixture BL extracts was determined using the DPPH assay. The BL extract has revealed an antioxidant activity ranging between 1.5 ± 0.075 and 0.135 ± 0.007 mg/mL. High antioxidant activity with an IC50 of 0.135 mg/mL was envisaged by a combination of 16.667, 66.667, and 16.667% of water, ethanol, and acetonitrile, respectively (Table 1). Antioxidant activity was correlated with the presence of bioactive substances. These chemicals, which dissolve differentially due to varying polarity of the employed solvents, might explain the variance in DPPH scavenging activities seen in the BL extract [48]. Among these compounds, the phenolic group is the most abundant and extensively distributed found in plants, and they are extremely powerful antioxidants [43]. Thus, due to its high antioxidant content, Silva et al. [15] highlighted the potential use of BL as a functional food for the adjuvant treatment of dyslipidemias.
The polynomial equation (Eq. 3) executed for the antioxidant activity is as follows with p values of linear and quadratic < 0.05 (Table 4).3 IC50mg/mL=1.134×X1+0.482×X2+1.424×X3-0.828×X1×X2-3.745×X1×X3+0.611×X2×X3-3.759×X1×X2×X3
As indicated in Fig. 2, a binary mixture of water (53.54%) and acetonitrile (46.46%) contributed to the highest antioxidant activity. These findings corroborate a previous study, which found that acetonitrile and water extracts of Kandelia candel bark provide the highest activity [49]. In fact, antioxidants such as phytochemicals neutralize free oxygen radicals and preserve the genetic material of cells from harm [50]. Consequently, it may effectively treat diseases caused by oxidative stress.
The antioxidant activity of beet extracts has been reported in several studies [51–53]. However, the level of antioxidant activity and polyphenol content fluctuates. The variation in the reported data could be attributed to various variables, including postharvest nutritional quality losses increased by physical damage, longer storage times, high temperatures, and low relative humidity.
The effect of the system solvents on the antibacterial activity
The microdilution method was employed to find the MIC and the lowest concentration of an antibacterial agent required to inhibit a bacterium. The MICs of the chosen bacteria are represented in Table 1. These findings reveal variability in the sensitivity of the bacteria to the different mixtures. Indeed, Gram-negative bacteria, E. coli and S. enterica, have a low MIC of 0.312 ± 0.016 and 0.156 ± 0.008 mg/mL, respectively, with all the binary and the ternary mixtures. Ethanolic beetroot extract exhibits a MIC of E. coli equal to 1.5 mg/mL [54]. On the other hand, the BL extract showed a MIC equal to 0.625 mg/mL against S. aureus. These findings were better than those investigated by Čanadanović-Brunet et al. [54]. These authors used ethanol as an extraction solvent and reported a concentration of 0.750 mg/mL to inhibit S. aureus. For antibacterial activities, p values of linear and cubic were significant (p < 0.05), and the resulting regression models (Table 5) are shown in the following equations:Table 5 ANOVA results of regression models from mixture design for solvent optimization of MICs of S. aureus, S. enterica, and E. coli
Source DF Seq SS Adj SS Adj MS F value p value
S. aureus
Regression 6 2.925 2.9256 0.581 8.11 0.047
Linear 2 1.866 2.062 1.031 16.70 0.016
Quadratic 3 1.038 0.0712 0.346 4.00 0.071
X1 × X2 1 0.876 0.638 0.881 10.38 0.053
X1 × X3 1 0.16147 0.095 0.161 2.80 0.066
X2 × X3 1 0.00034 0.00176 0.00034 2.01 0.051
Special cubic 1 0.02085 0.02085 0.02085 4.18 0.043
X1 × X2 × X3 1 0.02085 0.02085 0.02085 4.18 0.043
Residual error 3 0.35564 0.35564 0.11855
Total 9 3.281
S. enterica
Regression 6 0.479 0.479 0.08 2.33 0.042
Linear 2 0.304 0.263 0.131 2.54 0.031
Quadratic 3 0.066 0.088 0.030 2.12 0.072
X1 × X2 1 0.056 0.006 0.006 4.03 0.058
X1 × X3 1 0.005 0.039 0.039 4.16 0.061
X2 × X3 1 0.005 0.039 0.039 5.16 0.084
Special cubic 1 0.109 0.109 0.109 6.45 0.042
X1 × X2 × X3 1 0.109 0.109 0.109 6.45 0.042
Residual error 3 0.732 0.732 0.244
Total 9 1.212
E. coli
Regression 6 0.156 0.156 0.026 2.94 0.013
Linear 2 0.057 0.057 0.028 4.13 0.045
Quadratic 3 0.093 0.087 0.029 4.16 0.079
X1 × X2 1 0.015 0.021 0.021 3.61 0.055
X1 × X3 1 0.066 0.071 0.072 12.34 0.078
X2 × X3 1 0.011 0.016 0.016 2.20 0.094
Special cubic 1 0.006 0.006 0.006 2.44 0.035
X1 × X2 × X3 1 0.006 0.006 0.006 2.44 0.035
Residual error 3 0.040 0.04 0.013
Total 9 0.197
4 MICS.aureusmg/mL=2.352×X1+0.704×X2+0.648×X3-3.887×X1×X2-1.501×X1×X3+0.204×X2×X3+4.632×X1×X2×X3
5 MICS.entericamg/mL=1.15×X1+0.53×X2+0.58×X3-0.39×X1×X2+0.97×X1×X3+0.97×X2×X3-10.60×X1×X2×X3
6 MICE.colimg/mL=0.586×X1+0.303×X2+0.317×X3-0.716×X1×X2-1.304×X1×X3-0.617×X2×X3+2.461×X1×X2×X3
As mentioned in the contour plot (Fig. 3a, b), the binary combination mixture (water–ethanol) yielded the best results of MIC of S. aureus and S. enterica. Meanwhile, the ternary mixture extracts of BL were more efficient against E. coli than the mono and binary mixture (Fig. 3c). Beta vulgaris extracts were shown to have antiquorum sensing and antibiofilm (E. coli) activities [55]. Therefore, BL extract’s antibacterial action might be attributed to the presence of a high concentration of phytochemical contents, which have been shown in earlier studies to have effective antimicrobial properties. Flavonoids, such as vitexin, were investigated to modulate cell surface hydrophobicity in order to combat S. aureus biofilm [56]. Additionally, polyphenols like galloyl catechins can intercalate in membranes and generate significant biophysical alterations [57]. These alterations can inhibit biofilm formation and limit the dissemination of the biosynthetic machinery of the cell wall, which is linked to beta-lactam antibiotic resistance. Indeed, if natural bioactives can inhibit pathogenic bacteria from communicating and proliferating, they can also inhibit them from being virulent and resistant. As a result, this vegetable is suggested as a bioactive alternative to standard antibacterial and antipathogenic agents.
Mixture optimization
To find the most suitable response, a numerical optimization approach of the different mixtures was used. Optimum levels of solvents were obtained to maximize all the responses. This method looks for a set of factor values that fits the design’s requirements with each response at the same time. The optimal mixture for all the responses is the ternary mixture with different percentages. For TPC and TFC, the % of solvents were 38.74, 20.86, and 40.40% for water, ethanol, and acetonitrile, respectively (Fig. 4a). For the antibacterial activities, the % of solvents were 25.25, 50.17, and 24.58% of water, ethanol, and acetonitrile, respectively, with an overall desirability value of 0.780 (Fig. 4b). For the antioxidant activity, the percentages of solvents were 53.53, 1.02, and 45.45% of water, ethanol, and acetonitrile, respectively, with an overall desirability value of 0.831 (Fig. 4c). Accordingly, the predicted models with optimum values of TPC, TFC, MIC S. aureus, MIC S. enterica, MIC E. coli, and antioxidant (IC50) activities were 49.287 mg GAE/mL, 0.314 mg QE/mL, 0.402 mg/mL, 0.497 mg/mL, 0.207 mg/mL, and 0.337 mg/mL, respectively (Fig. 4a–c).Fig. 4 Response optimizer at the optimum conditions for the maximum response of the tested a phytochemical contents, b MICs of S. aureus, S. enterica, and E. coli, and c the antioxidant (DPPH) activity
Chemometric analysis
With the purpose of revealing hidden patterns and similarities between the different samples of BL extract mixtures, phytochemical contents and their bioactivities were investigated. The dataset obtained was studied using chemometric analysis, which included PCA and HCA.
PCA generates score plots of the distribution of the different samples following the major components. In this study, the obtained score plots (Fig. 5a, b) presented the differences between the sample mixtures, phytochemical contents, and their bioactivities. The major principal component is F1, which contributed to 42.63% of the total variances. In fact, F1 has the maximum eigenvalue of 2.56, while F2, F3, F4, F5, and F6 have eigenvalues of 1.5, 1.07, 0.52, 0.30, and 0.04, and elucidate 24.88, 17.94, 8.74, 5.10, and 0.71% of the variance in the data, respectively. Consequently and based on the loading plot of PCA, F1 and F2, which contributed to 67.51% of the total variances, were utilized to supply a relevant visual guide for mentioning data differences (Fig. 5a, b). Indeed, the F1 axis is associated with the majority of variables (TFC, MIC of S. aureus, MIC of S. enterica, and MIC of E. coli) with high values, whereas the F2 axis is linked to TPC and DPPH.Fig. 5 Principal component analysis (PCA) plots of a phytochemical contents (TPC and TFC), biological (antioxidant and antibacterial) activities, and b all samples (1–10) extracted with different solvent mixtures
In fact, the score plot depicts a clear difference between the 10 samples (Fig. 5b). Overall, the first sample was associated with MIC of S. aureus, S. enterica, and E. coli, while samples 2 and 8 were correlated with TFC. Additionally, samples 5 and 9 were correlated with the TPC (Fig. 5a, b). In addition, the loading plot showed that the variables, which contributed positively to F1, are DPPH, MIC of S. aureus, MIC of S. enterica, and MIC E. coli while TPC and TFC contributed negatively to the same axis. On the other hand, F2 was positively correlated with TFC, DPPH, and MIC of E. coli and negatively correlated with MIC of S. aureus, MIC of S. enterica, and TPC. The results revealed that BL extracts with high antioxidant activity could have a better antibacterial activity and the MIC values of the chosen bacteria were related to the aqueous extract of beetroot. According to a previous report, phenolic compounds have been demonstrated to present better antibacterial activity [58].
The result acquired by PCA plots was corroborated by developing a hierarchical cluster analysis (HCA). HCA is a clustering technique that investigates the way of sample organization. Furthermore, by illustrating hierarchy, it enables the identification of similarities and differences within and across groups [59]. This method was exhibited using the matrix of squared Euclidean distances to get further information about a probable classification of the sample mixtures according to their contents and activities. The findings of HCA have usually illustrated in a dendrogram a plot that displays the arrangement of samples and their relationships in a tree shape. Indeed, the generated dendrogram (Fig. 6) indicated the presence of five clusters (1, (5–9), (3–6), (7–4-10), and (2–8)), which agree with PCA. In investigations involving bioactive chemicals and functional characteristics, PCA and HCA are frequently utilized together [60–62]. For instance, Sicak et al. [63] used PCA to determine the connections between the chemical components and the biological activities (antioxidant, antiproliferative, and enzyme inhibition) of pine and thyme honey extracts.Fig. 6 HCA dendrogram showing clustering of BL extracts of all samples (1–10) extracted with three different solvents
Mode of action of BL extract
Using a linear model (ANOVA), the effect of the concentrations of BL extract on the growth of S. aureus (0.402, 0.804, and 1.608 mg/mL), S. enterica (0.497, 0.994, and 1.988 mg/mL), and E. coli (0.207, 0.414, and 0.828 mg/mL) was assessed. To investigate this effect, the bacterial growth was measured for 26 h and compared to the control culture. A notable decrease trend in the viable cells was detected, after 6 h, when the different concentrations of BL extracts (1 × MIC, 2 × MIC, and 4 × MIC) were applied, as depicted in Fig. 7a–c. For 1 × MIC and 2 × MIC, a significant (p < 0.05) decrease in the tested bacteria’s growth appeared after 8 h compared to the control. At the same time, BL extract at 1.608 and 0.828 mg/mL (4 × MIC) showed a significant (p < 0.05) decrease in cell numbers of S. aureus (2.7 log 10 CFU/mL) and E. coli (2.8 log 10 CFU/mL), respectively (Fig. 7a–c). Meanwhile, a highly significant (p < 0.05) inhibition of S. enterica (2.05 log 10 CFU/mL) was displayed by 1.988 mg/mL (4 × MIC) of BL extract after 20 h incubation (Fig. 7b). Thus, BL extract exhibited a remarkable antimicrobial dose-dependent effect against the tested bacteria. Our results were in good agreement with Lee et al. [64], who found that passion fruit peel extract (PPE) could effectively reduce the growth of E. coli and L. monocytogenes within 3 h.Fig. 7 Influence of the dose of BL extract on the growth of a S. aureus, b S. enterica, and c E. coli (log 10 CFU/mL) during 26 h
Therefore, BL extract at 4 × MIC, which corresponds to 1.608, 1.988, and 0.828 mg/mL, was more efficient against S. aureus, S. enterica, and E. coli, respectively. Accordingly, BL extract revealed a bactericidal activity against the tested bacteria, after 26 h. Thereby, the high antibacterial effectiveness of BL extract may be accorded to the plant’s metabolites [65].
Conclusion
The phytochemical contents and biological activities of BL were enhanced by using an augmented simplex-centroid mixture design approach. Three solvents were mixed at different concentrations, and the response optimizer permits to maximize the phytochemical contents and antioxidant and antibacterial activity. Solvent combination leads to better extraction of bioactive compounds and antibacterial activity than using each solvent separately. In addition, BL extract at optimized conditions was effective against the tested foodborne pathogens (S. aureus, S. enterica, and E. coli). The mode of action confirms these results and indicates that BL extract at optimized conditions exerts a dose-dependent bactericidal effect against all tested bacteria. By employing chemometric techniques, PCA and HCA, all the data enable to segregate all of the mixtures and connect phytochemical contents (TPC and TFC) to antioxidant and antibacterial activities using correlation models.
The findings of this study suggested that Beta vulgaris might be an alternative source of biomolecules, which implies its potential commercial usage in nutraceuticals and medications that are helpful to human health.
Author contribution
Moufda Chaari, Khaoula Elhadef, Sarra Akermi, and Lotf Mellouli were involved in methodology, data curation, and writing—original draft preparation; Hajer Ben Hlima and Mariam Fourati contributed to resources and formal analysis; Ahlem Chakchouk-Mtibaa, Tanmay Sarkar, Mohammed Ali Shariati Maksim Rebezov, and Teresa D'Amore were involved in software and validation; and Slim Smaoui was involved in supervision and project administration.
Funding
This research was funded by the Tunisian government PEJC project (2019) no. 19PEJC07-03.
Data and materials availability
The data and materials used and/or analyzed during the present study are available. The authors will provide additional details if required.
Declarations
Ethics approval
All authors were governed by ethics and professionalism.
Conflict of interest
The authors declare no competing interests.
Publisher's note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Beegum PPS, Pandiselvam R, Ramesh SV, et al (2022) A critical appraisal on the antimicrobial, oral protective, and anti-diabetic functions of coconut and its derivatives. Qual Assur Saf Crops Foods 14:86–100. 10.15586/qas.v14i2.1040
2. Wani SA, Naik H, Wagay J, et al (2022) Mentha: A review on its bioactive compounds and potential health benefits. Qual Assur Saf Crops Foods 14:154–168. 10.15586/qas.v14i4.1129
3. Xie F, Wang H, Cao Q, et al (2022) The effects of Oldenlandia diffusa water extract on glucose metabolism and inflammation level in rats with streptozotocin-induced gestational diabetes mellitus. Qual Assur Saf Crops Foods 14:24–30. 10.15586/qas.v14i1.970
4. Castro-Muñoz R Díaz-Montes E Gontarek-Castro E A comprehensive review on current and emerging technologies toward the valorization of bio-based wastes and by products from foods Compr Rev Food Sci Food Saf 2022 21 46 105 10.1111/1541-4337.12894 34957673
5. Calatayud-Sáez FM (2022) The traditional Mediterranean diet is effective in the prevention and treatment of acute and recurrent inflammatory diseases of childhood. aei 50:93–95. 10.15586/aei.v50i2.565
6. Prabhu S Rajasekar M Identification of colour pigments in root vegetables - carrot (Daucus carota) and beetroot (Beta vulgaris) by foldscope Int j sci eng res 2019 10 4
7. Nobari H Saedmocheshi S Johnson K A brief overview of the effects of exercise and red beets on the immune system in patients with prostate cancer Sustainability 2022 14 6492 10.3390/su14116492
8. Amirpoor A Zavar R Amerizadeh A Effect of beetroot consumption on serum lipid profile: a systematic review and meta-analysis Curr Probl Cardiol 2022 47 100887 10.1016/j.cpcardiol.2021.100887 34154819
9. Nade VS Kawale LA Zambre SS Kapure AB Neuroprotective potential of Beta vulgaris L. in Parkinson’s disease Indian J Pharmacol 2015 47 403 408 10.4103/0253-7613.161263 26288473
10. dos S. Baião D, da Silva DVT, Paschoalin VMF, Beetroot, a remarkable vegetable: its nitrate and phytochemical contents can be adjusted in novel formulations to benefit health and support cardiovascular disease therapies Antioxidants 2020 9 960 10.3390/antiox9100960 33049969
11. Akyüz A, Ersus S (2021) Optimization of enzyme assisted extraction of protein from the sugar beet (Beta vulgaris L.) leaves for alternative plant protein concentrate production. Food Chemistry 335:127673. 10.1016/j.foodchem.2020.127673
12. Asadi SZ Khan MA The effect of beetroot (Beta vulgaris L.) leaves powder on nutritional, textural, sensorial and antioxidant properties of cookies J Culin Sci Technol 2021 19 424 438 10.1080/15428052.2020.1787285
13. Goyeneche R Di Scala K Ramirez CL Fanovich MA Recovery of bioactive compounds from beetroot leaves by supercritical CO2 extraction as a promising bioresource J Supercrit Fluids 2020 155 104658 10.1016/j.supflu.2019.104658
14. Gawlik-Dziki U Dziki L Anisiewicz J Leaves of white beetroot as a new source of antioxidant and anti-inflammatory compounds Plants 2020 9 944 10.3390/plants9080944 32722637
15. Silva LGS Morelli AP Pavan ICB Protective effects of beet ( Beta vulgaris ) leaves extract against oxidative stress in endothelial cells in vitro Phytother Res 2020 34 1385 1396 10.1002/ptr.6612 31989717
16. Abd El-Ghffar EA Hegazi NM Saad HH HPLC-ESI- MS/MS analysis of beet (Beta vulgaris) leaves and its beneficial properties in type 1 diabetic rats Biomed Pharmacother 2019 120 109541 10.1016/j.biopha.2019.109541 31629949
17. Gholamalipour Alamdari E, Taleghani A (2022) New bioactive compounds characterized by liquid chromatography–mass spectrometry and gas chromatography–mass spectrometry in hydro-methanol and petroleum ether extracts of Prosopis farcta (Banks & Sol.) J. F. Macbr weed. J Mass Spectrom 57:. 10.1002/jms.4884
18. Lorizola IM, Furlan CPB, Portovedo M, et al (2018) Beet stalks and leaves (Beta vulgaris L.) protect against high-fat diet-induced oxidative damage in the liver in mice. Nutrients 10:872. 10.3390/nu10070872
19. Biondo PBF Boeing JS Barizão ÉO Evaluation of beetroot (Beta vulgaris L.) leaves during its developmental stages: a chemical composition study Food Sci Technol 2014 34 94 101 10.1590/S0101-20612014005000007
20. de Oliveira SPA do Nascimento HMA, Sampaio KB, de Souza EL, A review on bioactive compounds of beet ( Beta vulgaris L. subsp. vulgaris ) with special emphasis on their beneficial effects on gut microbiota and gastrointestinal health Crit Rev Food Sci Nutr 2021 61 2022 2033 10.1080/10408398.2020.1768510 32449379
21. Hikmawanti NPE Dwita LP Wisnunanda DW Farista F The effect of different extracts of beetroots as antioxidant and anti-anaemia on phenylhydrazine-induced rats Eur Pharm J 2021 68 1 7 10.2478/afpuc-2020-0014
22. Grispoldi L Karama M Armani A Staphylococcus aureus enterotoxin in food of animal origin and staphylococcal food poisoning risk assessment from farm to table Ital J Anim Sci 2021 20 677 690 10.1080/1828051X.2020.1871428
23. Monte DF Lincopan N Berman H Genomic features of high-priority Salmonella enterica serovars circulating in the food production chain, Brazil, 2000–2016 Sci Rep 2019 9 11058 10.1038/s41598-019-45838-0 31363103
24. Punia Bangar S, Sharma N, Bhardwaj A, Phimolsiripol Y (2022) Lactic acid bacteria: a bio-green preservative against mycotoxins for food safety and shelf-life extension. Qual Assur Saf Crops Foods 14:13–31. 10.15586/qas.v14i2.1014
25. Adnan Md, Oh KK, Azad MOK, et al (2020) Kenaf (Hibiscus cannabinus L.) leaves and seed as a potential source of the bioactive compounds: effects of various extraction solvents on biological properties. Life 10:223. 10.3390/life10100223
26. Aboulghazi A Bakour M Fadil M Lyoussi B Simultaneous optimization of extraction yield, phenolic compounds and antioxidant activity of Moroccan propolis extracts: improvement of ultrasound-assisted technique using response surface methodology Processes 2022 10 297 10.3390/pr10020297
27. Xu H-R, Zhang Y-Q, Wang S, et al (2022) Optimization of functional compounds extraction from Ginkgo biloba seeds using response surface methodology. Qual Assur Saf Crops Foods 14:102–112. 10.15586/qas.v14i1.1033
28. Alcântara MA de Lima Brito Polari I, de Albuquerque Meireles BRL, Effect of the solvent composition on the profile of phenolic compounds extracted from chia seeds Food Chem 2019 275 489 496 10.1016/j.foodchem.2018.09.133 30724224
29. Ksibi IE Slama RB Faidi K Mixture approach for optimizing the recovery of colored phenolics from red pepper (Capsicum annum L.) by-products as potential source of natural dye and assessment of its antimicrobial activity Ind Crops Prod 2015 70 34 40 10.1016/j.indcrop.2015.03.017
30. Smaoui S, Ben Hlima H, Fourati M, et al (2020) Multiobjective optimization of Phoenix dactylifera L. seeds extraction: mixture design methodology for phytochemical contents and antibacterial activity. J Food Process Preserv 44:. 10.1111/jfpp.14822
31. Gasemloo S Khosravi M Sohrabi MR Response surface methodology (RSM) modeling to improve removal of Cr (VI) ions from tannery wastewater using sulfated carboxymethyl cellulose nanofilter J Clean Prod 2019 208 736 742 10.1016/j.jclepro.2018.10.177
32. Chaari M, Akermi S, Elhadef K, et al (2022) From modeling and optimizing extraction of peels beetroot (Beta vulgaris L.) betalains to in silico probing of their antibacterial multitarget mechanisms. Biomass Conv Bioref. 10.1007/s13399-022-03140-6
33. Lezoul NEH Belkadi M Habibi F Guillén F Extraction processes with several solvents on total bioactive compounds in different organs of three medicinal plants Molecules 2020 25 4672 10.3390/molecules25204672 33066273
34. López-Froilán R Hernández-Ledesma B Cámara M Pérez-Rodríguez ML Evaluation of the antioxidant potential of mixed fruit-based beverages: a new insight on the Folin-Ciocalteu method Food Anal Methods 2018 11 2897 2906 10.1007/s12161-018-1259-1
35. Saharan P Sadh PK Duhan S Duhan JS Bio-enrichment of phenolic, flavonoids content and antioxidant activity of commonly used pulses by solid-state fermentation Food Measure 2020 14 1497 1510 10.1007/s11694-020-00399-z
36. Fourati M Smaoui S Ennouri K Multiresponse optimization of pomegranate peel extraction by statistical versus artificial intelligence: predictive approach for foodborne bacterial pathogen inactivation Evid Based Complementary Altern Med 2019 2019 1 18 10.1155/2019/1542615
37. Elhadef K Akermi S Ben Hlima H Tunisian pistachio hull extracts: phytochemical content, antioxidant activity, and foodborne pathogen inhibition J Food Qual 2021 2021 1 18 10.1155/2021/9953545
38. Zeouk I Ouedrhiri W Jiménez IA Intra-combined antioxidant activity and chemical characterization of three fractions from Rhamnus alaternus extract: mixture design Ind Crops Prod 2020 144 112054 10.1016/j.indcrop.2019.112054
39. Aykın-Dinçer E Güngör KK Çağlar E Erbaş M The use of beetroot extract and extract powder in sausages as natural food colorant Int J Food Eng 2021 17 75 82 10.1515/ijfe-2019-0052
40. Park BI Kim J Lee K Flavonoids in common and tartary buckwheat hull extracts and antioxidant activity of the extracts against lipids in mayonnaise J Food Sci Technol 2019 56 2712 2720 10.1007/s13197-019-03761-2 31168153
41. Ameer K Shahbaz HM Kwon J-H Green extraction methods for polyphenols from plant matrices and their byproducts: a review: polyphenols extraction by green methods Compr Rev Food Sci Food Saf 2017 16 295 315 10.1111/1541-4337.12253 33371540
42. Ivanović M Islamčević Razboršek M Kolar M Innovative extraction techniques using deep eutectic solvents and analytical methods for the isolation and characterization of natural bioactive compounds from plant material Plants 2020 9 1428 10.3390/plants9111428 33114332
43. Wijekoon MMJO Bhat R Karim AA Effect of extraction solvents on the phenolic compounds and antioxidant activities of bunga kantan (Etlingera elatior Jack.) inflorescence J Food Compost Anal 2011 24 615 619 10.1016/j.jfca.2010.09.018
44. Galanakis CM Goulas V Tsakona S A knowledge base for the recovery of natural phenols with different solvents Int J Food Prop 2013 16 382 396 10.1080/10942912.2010.522750
45. Vural N Algan Cavuldak Ö Akay MA Anlı RE Determination of the various extraction solvent effects on polyphenolic profile and antioxidant activities of selected tea samples by chemometric approach Food Measure 2020 14 1286 1305 10.1007/s11694-020-00376-6
46. Handa CL de Lima FS Guelfi MFG Multi-response optimisation of the extraction solvent system for phenolics and antioxidant activities from fermented soy flour using a simplex-centroid design Food Chem 2016 197 175 184 10.1016/j.foodchem.2015.10.124 26616938
47. dos Santos C, Mizobucchi AL, Escaramboni B, et al (2020) Optimization of Eugenia punicifolia (Kunth) D. C. leaf extraction using a simplex centroid design focused on extracting phenolics with antioxidant and antiproliferative activities. BMC Chemistry 14:34. 10.1186/s13065-020-00686-2
48. Ng ZX, Samsuri SN, Yong PH (2020) The antioxidant index and chemometric analysis of tannin, flavonoid, and total phenolic extracted from medicinal plant foods with the solvents of different polarities. J Food Process Preserv 44:.10.1111/jfpp.14680
49. Jasna T K, Khaleel K M (2022) Evaluation of antioxidant activities and quantitative estimation of phytochemicals in Kandelia candel (L.) Druce. Int J Res Pharm Sci 13:166–172. 10.26452/ijrps.v13i2.138
50. Naseem Z Hanif MA Zahid M Ultrasound-assisted deep eutectic solvent–based extraction of phytochemicals from Mentha arvensis: optimization using Box-Behnken design Biomass Conv Bioref 2021 10.1007/s13399-021-01617-4
51. Chhikara N Kushwaha K Jaglan S Nutritional, physicochemical, and functional quality of beetroot ( Beta vulgaris L.) incorporated Asian noodles Cereal Chem 2019 96 154 161 10.1002/cche.10126
52. Edziri H Jaziri R Haddad O Phytochemical analysis, antioxidant, anticoagulant and in vitro toxicity and genotoxicity testing of methanolic and juice extracts of Beta vulgaris L S Afr J Bot 2019 126 170 175 10.1016/j.sajb.2019.01.017
53. Righi Pessoa da Silva H, da Silva C, Bolanho BC (2018) Ultrasonic-assisted extraction of betalains from red beet ( Beta vulgaris L.). J Food Process Eng 41:e12833. 10.1111/jfpe.12833
54. Čanadanović-Brunet JM, Savatović SS, Ćetković GS, et al (2011) Antioxidant and antimicrobial activities of beet root pomace extracts. Czech J Food Sci 29:575–585. 10.17221/210/2010-CJFS
55. Pellegrini MC Ponce AG Beet (Beta vulgaris) and leek (Allium porrum) leaves as a source of bioactive compounds with anti-quorum sensing and anti-biofilm activity Waste Biomass Valor 2020 11 4305 4313 10.1007/s12649-019-00775-x
56. Das M, Das A, Samaddar S, et al (2018) Vitexin altersStaphylococcus aureussurface hydrophobicity to interfere with biofilm formatio
57. Martinez RM Hohmann MS Longhi-Balbinot DT Analgesic activity and mechanism of action of a Beta vulgaris dye enriched in betalains in inflammatory models in mice Inflammopharmacol 2020 28 1663 1675 10.1007/s10787-020-00689-4
58. Tayebeh B, Soraya K, Khaneghah AM (2021) Antioxidant and antibacterial activity of ethanolic extract of safflower with contrasting seed coat colors. Qual Assur Saf Crops Foods 13:94–100. 10.15586/qas.v13i2.866
59. Miladinović DL Dimitrijević MV Mihajilov-Krstev TM The significance of minor components on the antibacterial activity of essential oil via chemometrics LWT 2021 136 110305 10.1016/j.lwt.2020.110305
60. Ferreira DS, Rocha JCB, Arellano DB, Pallone JAL (2022) Discrimination of South American grains based on fatty acid. Qual Assur Saf Crops Foods 14:30–42. 10.15586/qas.v14i3.1064
61. Cerqueira da Silva VH da Silva S junior AL, Pinheiro Lôbo I, Chemometric tools in the optimization of a microwave-assisted digestion procedure for guarana-based drink samples and data analysis from elemental, caffeine, and epicatechin contents Food Chem 2021 365 130468 10.1016/j.foodchem.2021.130468 34392233
62. Tan J, Li M-F (2021) Rapid and nondestructive identification of Belgian and Netherlandish Trappist beers by front-face synchronous fluorescence spectroscopy coupled with multiple statistical analysis. Qual Assur Saf Crops Foods 13:83–92. 10.15586/qas.v13i1.839
63. Sıcak Y Şahin-Yağlıoğlu A Öztürk M Bioactivities and phenolic constituents relationship of Muğla thyme and pine honey of Turkey with the chemometric approach Food Measure 2021 15 3694 3707 10.1007/s11694-021-00940-8
64. Lee C-H Woo H-J Kang J-H Song KB Electrostatic spraying of passion fruit (Passiflora edulis L.) peel extract for inactivation of Escherichia coli O157:H7 and Listeria monocytogenes on fresh-cut lollo rossa and beetroot leaves Food Bioprocess Technol 2021 14 898 908 10.1007/s11947-021-02608-z
65. Bouslamti M, El Barnossi A, Kara M, et al (2022) Total polyphenols content, antioxidant and antimicrobial activities of leaves of Solanum elaeagnifolium Cav. from Morocco. Molecules 27:4322. 10.3390/molecules27134322
| 0 | PMC9746593 | NO-CC CODE | 2022-12-15 23:21:56 | no | Biomass Convers Biorefin. 2022 Dec 13;:1-15 | utf-8 | Biomass Convers Biorefin | 2,022 | 10.1007/s13399-022-03645-0 | oa_other |
==== Front
J Comput Soc Sci
J Comput Soc Sci
Journal of Computational Social Science
2432-2717
2432-2725
Springer Nature Singapore Singapore
193
10.1007/s42001-022-00193-5
Research Article
Varieties of corona news: a cross-national study on the foundations of online misinformation production during the COVID-19 pandemic
Caliskan Cantay [email protected]
1
Kilicaslan Alaz 2
1 grid.16416.34 0000 0004 1936 9174 Goergen Institute for Data Science, University of Rochester, Rochester, USA
2 grid.267484.b 0000 0001 0087 1429 Department of Sociology, Criminology and Anthropology, University of Wisconsin-Whitewater, Whitewater, USA
13 12 2022
153
30 4 2022
11 11 2022
© The Author(s), under exclusive licence to Springer Nature Singapore Pte Ltd. 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Misinformation in the media is produced by hard-to-gauge thought mechanisms employed by individuals or collectivities. In this paper, we shed light on what the country-specific factors of falsehood production in the context of COVID-19 Pandemic might be. Collecting our evidence from the largest misinformation dataset used in the COVID-19 misinformation literature with close to 11,000 pieces of falsehood, we explore patterns of misinformation production by employing a variety of methodological tools including algorithms for text similarity, clustering, network distances, and other statistical tools. Covering news produced in a span of more than 14 months, our paper also differentiates itself by its use of carefully controlled hand-labeling of topics of falsehood. Findings suggest that country-level factors do not provide the strongest support for predicting outcomes of falsehood, except for one phenomenon: in countries with serious press freedom problems and low human development, the mostly unknown authors of misinformation tend to focus on similar content. In addition, the intensity of discussion on animals, predictions and symptoms as part of fake news is the biggest differentiator between nations; whereas news on conspiracies, medical equipment and risk factors offer the least explanation to differentiate. Based on those findings, we discuss some distinct public health and communication strategies to dispel misinformation in countries with particular characteristics. We also emphasize that a global action plan against misinformation is needed given the highly globalized nature of the online media environment.
Supplementary Information
The online version contains supplementary material available at 10.1007/s42001-022-00193-5.
Keywords
Misinformation
Fake news
Social media
Computational social science
Cross-country analysis
Natural language processing
==== Body
pmcIntroduction
“Anybody that wants a test can get a test. That’s what the bottom line is.” [1] was a sentence uttered by the president of the United States at a time when the daily testing capacity was about 75,000 nationally. Misinformation ranging from harmless rumors to extremely complex and dangerous conspiracy theories has been one of the most defining characteristics of social media use in recent years. The intensification of the use of social media as an information-sharing tool, coupled with a directly and indirectly forced lockdown during the COVID-19 pandemic, has multiplied the production of misinformation.
As provided in the example above, people of different walks of life including world leaders, Instagram celebrities, troll factories, and many others have been intentionally or unintentionally spreading misinformation. Assuming that these entities act rationally, people and institutions that use their own resources to produce those falsehoods must have a purpose. Inspired from the national-level analysis and empirical approach of the book Varieties of Capitalism: The Institutional Foundations of Comparative Advantage [2] and other articles and books that use ‘Varieties of…’ in their titles,1 this article studies the relationship between country-specific variables, and the topics and content of misinformation. The study specifically looks at the variation in topic ratios, and the creativity in misinformation content over time by providing extensive data analysis and using the most extensive datasets on misleading and false news created during the global pandemic.2
Findings indicate that the ‘Varieties of…” literature can be applied only to certain cases in this context. Specifically, countries struggling in a range of areas including press freedom and human development tend to produce news content similar to each other. In addition, news on animals, predictions and symptoms are the three biggest differentiators between countries, whereas news on conspiracies, medical equipment and risk factors offer the least explanation to differentiate. Based on those findings, we discuss some distinct public health and communication strategies to dispel misinformation in countries with particular characteristics, such as efforts to repair Western governments and pharmaceutical companies’ tarnished reputation in low human development countries. We also emphasize that a global action plan against misinformation is needed given the highly globalized nature of the online media environment and the ubiquitousness of the major conspiracy theories around COVID-19. The dataset used in this study is the largest dataset of online misinformation about COVID-19 that can be found in the literature as it comprises over 10 thousand falsehoods from 129 countries. Our study is also unique in the number of variables used: the countries are compared based on 14 variables aiming to measure economic, informational, political, and socio-cultural environments in each, such as their level of democracy, trust in science, income inequality, and healthcare strength.
Misinformation during a pandemic
First, a clarification on the terms we use in this article. In parallel to the rise of social media and other online platforms over the past 15 years, there has been a proliferation of studies looking at the emergence, spread, consumption, and effects of misleading and false information by analyzing the phenomenon in various terms: how rumors spread on Twitter [9], real-world impacts of hoaxes at Wikipedia [10], how individuals consumed fake news prior to 2016 US presidential election [11], or how regular people, and not just state-supported media, actively participate in generating disinformation in Russia [12]. In this paper, we chose to use the term misinformation over the aforementioned alternatives. As defined in Merriam-Webster’s dictionary [13], misinformation refers to “incorrect or misleading information”. According to this definition, any piece of information that is partly or fully false can be labeled as misinformation, irrespective of an intent to deceive on the part of those who produce or diffuse the information. Thus, misinformation is a broader term, which comprises disinformation, i.e., false information that is “deliberately” spread “in order to influence public opinion or obscure the truth” [13], as well as those shared inadvertently or unintentionally. As discussed in the “Data” section below, this term befits the dataset we use. In line with some other scholars [14, 15], we also avoid the popular term fake news both because it is polarizing and politically charged and also it is rather limited to forms of misinformation that are deliberately designed to mimic news from established and mainstream news organizations [16].
Misinformation during the COVID-19 pandemic has reached such high levels that World Health Organization (WHO) and other United Nations bodies recognized the need to fight against false and misleading information as a critical part of the global pandemic strategy [17]. Many of the falsehoods regarding COVID-19 have been inspired by and interacting with conspiracy theories that predated the pandemic, including those involving “Big Pharma”, GMOs, Bill Gates, “deep state”, or a ring of satanic pedophiles [18–20]. The outcomes have been tangible: misinformation has significantly contributed to the spread of the illness and preventable deaths by promoting ineffective and harmful treatments and discouraging people from basic prevention such as wearing a mask or maintaining social distance. For instance, a study shows that over just a few months in the first of half 2020, approximately 800 people died and many more were hospitalized after drinking methanol as a cure for coronavirus [21]. More recently, vaccine-related misinformation has curtailed the vaccination efforts of many countries around the world, including the US where as of September 2021 the rate of vaccinated individuals fell short of the Biden administration’s original ambitions [22]. This was particularly worrying as hospitalization and death rates were much higher among the unvaccinated [23].
Unsurprisingly, there is a growing literature around misinformation regarding COVID-19, which can be arguably grouped into three main streams of research. A first stream strives to understand different types of misinformation in terms of their sources, spread patterns, and influence as well as analyzing how misinformation differs from more accurate and factual information in those respects. The main findings indicate that most news shared online are accurate; however, they are less likely to be shared than inaccurate ones [24, 25], which are produced and spread by denser and more organized communities [26, 27]. In addition, we learn that more misinformation circulates on social media platforms than on traditional news media [28], and misinformation coming from public figures such as celebrities or politicians, although making up a relatively small part of online misinformation about COVID-19, generates higher levels of engagement and support than other types of falsehoods [29, 30].
A second group of research looks at what specific factors push individuals to believe in or share misinformation. Accordingly, they show that certain psychological predispositions, political leanings, and daily habits such as tendency to reject expert information [31, 32], political conservatism [33], and right-leaning media consumption [34] are positively correlated with expressing and propagating misinformed views about COVID-19; whereas others such as greater science knowledge and cognitive reflection [35, 36], and trust in science and scientists [37] are negatively associated with those. Furthermore, it is interesting that an individual’s worry for personal health doesn’t seem to have an effect on their propensity to share COVID-19 related misinformation [38].
A third category of research articles focuses on how exposure to misinformation affects health behavior during the COVID-19 pandemic. The findings are clear as follows: consuming and/or believing in misinformation has a significant negative effect on willingness to take preventive measures such as wearing a face mask [39], maintaining physical distancing [40, 41], or getting vaccinated [42, 43]. In addition, belief in COVID-19 related falsehoods, including conspiracies, predicts greater use of pseudoscientific practices such as consuming garlic [44] or hydroxychloroquine [45]. Taken together with the research streams discussed previously, these studies shed light on the level of threat that misinformation poses on public health.
In parallel with these three streams, a relatively small but emerging group of studies analyze how economic, political, and socio-cultural differences among countries impact misinformation during the pandemic. This body of work illustrates the significant effect of a number of county-level variables on sources and types of as well as exposure and susceptibility to misinformation. Hence, countries’ levels of uncertainty avoidance [46]; political and media freedom, and mobile connectivity [47]; human development [48]; political conservatism and political control over media [49]; Gross Domestic Product [50]; media fragmentation and partisanship [51] have been found to shape the misinformation environment beyond the individual-level factors.
This study contributes to the existing literature by using the largest dataset of online misinformation to our knowledge: over 10 thousand falsehoods produced and shared across 129 countries.3 It is worth noting that, unlike most other studies focusing on online misinformation [14, 25, 52, 53], our study does not solely rely on Twitter, but uses false and misleading information from various sources, including social media platforms such as Facebook, WhatsApp, Instagram, and Telegram. In fact, Facebook leads the list by making up 4347 of those pieces of misinformation. We find this important, because Facebook is not only the biggest social media platform globally with its 2.8 billion users [54], it is also the most popular one for COVID-19-related misinformation [24]. For comparison, Twitter had slightly less than 400 million active users as of July 2021 [54]. The prominent role of Facebook in the spread of Covid-19-related misinformation has been documented in a number of settings including the USA (2020), and Egypt (2021) [55–57].
Another unique aspect of this study is the wide range of variables it uses to compare COVID-19 misinformation across countries. As the “Data” section below discusses, the countries are compared based on 14 variables aiming to measure political, economic, informational, and socio-cultural environments in each, such as the levels of corruption perception, human development, press freedom, and healthcare strength. The Health Belief Model posits that when facing a health threat like the COVID-19 pandemic, individuals’ likelihood of engaging in preventive health behaviors depend on the perceived level of threat and perceptions regarding the potential benefits of and barriers of participating in such behaviors [58]. The literature laid out above shows how misinformation affects those perceptions by downplaying the threat, distorting the facts around the origin and spread of the pandemic, and offering ineffective or harmful treatments. Therefore, we believe that our study can help public health authorities better leverage their country-level resources while fighting against online misinformation such as by strengthening media freedom or improving scientific literacy.
Data
The data for this study include 10,131 falsehoods about the COVID-19 Pandemic in varying intensities of misinformation.4 Observations have been collected from the Poynter CoronaVirus Facts/DatosCoronaVirus Alliance Database provided by the Poynter Institute [1]. Poynter Institute is a non-profit NGO focused on journalism and research and based in St. Petersburg, Florida (https://www.poynter.org/). The dataset is updated daily and provides a comprehensive understanding on the evolution of misinformation during the progression of the pandemic. The dataset covers the period between January 2020 and February 2021. For each observation, Poynter Institute provides the fact-checker that has provided information on the intensity of misinformation to the institute, the date on which the story was published, origin country/countries/continents, the intensity of misinformation, a brief summary of the story, the original text, and the link to the story. Most of the misinformation has initially been published on social media outlets, such as Facebook, Instagram or YouTube. In most cases, it is hard to know the true origin of the misinformation, since they are posted on different social media outlets on the same day. We should note that our study is not the only one that uses Poynter Institute’s dataset. A few other articles [14, 29, 47] also rely on the Poynter CoronaVirus Facts/DatosCoronaVirus Alliance Database. However, our study uses the highest number of falsehoods—over 10 000—not just among the studies that use this particular dataset, but among all published research articles that focus on misinformation around COVID-19.
This study stands apart from other work in that it uses a carefully controlled hand-labeling of topics of falsehood. More specifically, to extract more information from the dataset, we manually labeled 28 different topic categories that are mentioned by the Poynter Institute in association with COVID-19. (The topics that have been identified by the Poynter Institute are aid, animals, conspiracies, crime, cures, detection, food, governments, hospitals, individuals, insurance, laws, lockdown, medical equipment, medicine, origins, other diseases, predictions, prevention, religion, risk factors, spread, symptoms, travel, vaccines, videos, technology, and NGOs.) Manual labeling is a common practice in NLP-research and has been used in other studies, as well.5 In contrast to the suggestion by the Poynter Institute that each news belongs to exactly one topic, using excellent research assistance, we identified the news belonging to more than one topic and marked those accordingly. We used stratified sampling to check for the quality and consistency of the data collection process. The original dataset has later been enriched by using NLP by removing the stop words, using contraction mapping, removing links, emojis and hashtags, POS-tagging the words, and lemmatizing the news content. Brief and long summaries have then been merged to provide more information. The descriptive table below provides a few interesting facts about the dataset (Table 1).Table 1 Descriptive table for the falsehood dataset
Description Minimum Average Maximum
News count Vanuatu (1) 85.7 per country India (1892)
Brief summary length 3 words 17.1 words 113 words
Long summary length 2 words 27.3 words 198 words
Brief summary length (lemmatized) 1 word 10 words 63 words
Long summary length (lemmatized) 1 word 14.1 words 112 words
Merged summary length (lemmatized) 4 words 24.1 words 120 words
1st 2nd 3rd
Social media source Facebook (4347) Twitter (1088) WhatsApp (811)
Topic count Individuals (4769) Governments (2003) Videos (1308)
Productive dates 03/19/20 (170) 03/23/2020 (154) 03/17/2020 (148)
The map below shows the aggregate count of misinforming and misleading stories coming from each country. Most of the falsehoods were produced in India, the USA, Spain, Brazil and a few other European countries. As indicated by the map in Fig. 1, there is some correlation between the intensity of the pandemic and the amount of misinformation production. For example, as of September 2021, the USA, India, and Brazil were the top three countries in the world in terms of both the number of confirmed COVID-19 cases and deaths [65]. Also, as shown below, the period between March 2020 and May 2020 was the peak of misinformation production worldwide, and during that time, Spain had a higher number of cases than other European nations such as Italy, France, or Germany, despite having a significantly smaller population size than them [17].Fig. 1 Geographic distribution of the falsehood counts
The stacked bar graph below shows the distribution of topics for each day, and the red line plot shows the number of stories published per day. As expected, there is a strong correlation between the two; nevertheless, there are some news that could not be assigned to any topic and some others with more than one topic. As one can observe below, the dataset shows signs of seasonality and trend. Specifically, the number of falsehoods reached its peak in the period March 2020-May 2020, in a time when there were significant uncertainties about the definition and implications of the Sars-Cov2 virus. It is worth noting that starting from June 2020, significantly fewer falsehoods have been published worldwide, arguably thanks to clarifications and new information regarding the origin and spread of the virus as well as methods of treatment and prevention. Three key moments in the pandemic may have played important roles in reducing the amount of misinformation regarding COVID-19:On May 27, 2020, Dr. Anthony Fauci, the director of the National Institute of Allergy and Infectious Diseases, announced that a vaccine would be ready by December 2020 [66], which offered a promise of an end to the pandemic and lockdowns.
On June 17, 2020, the WHO announced that it was stopping its trial of the hyped anti-malaria drug hydroxychloroquine after new data suggested that the drug was not effective for COVID-19. This helped dispel the myths regarding its benefits that were spread by individuals and organizations, including the influential French physician Didier Raoult, Donald Trump, and Russian state-owned media [67].
On July 7, 2020, the WHO announced that COVID-19 may be an airborne disease mainly transmitted through respiratory droplets [68]. This not only reinforced the message around the importance of mask-wearing, but also helped limit the spread of misinformation concerning transmission through food or packaging (Figs. 2, 3).
Fig. 2 News count per day
Fig. 3 Topics and co-occurrences
As noted above, a significant contribution of this study is the assignment of the falsehoods to topics. Similarly, the exploration of co-occurrences between topics is of great importance to understand how groups of countries with varying characteristics produce misinformation. The graph below shows the aggregated results for topic co-occurrences. In the graph, the diagonal values show the counts for topics in the dataset, and the non-diagonal values indicate the number of co-occurrences between topics. Most falsehoods refer to an individual creating a story, and therefore have been classified as ‘Individuals’. Other topics with a significant amount of co-occurrence are ‘Governments’, ‘Conspiracies’, ‘Cures’, ‘Food’, ‘Prevention’, ‘Spread’, and ‘Medicine’. We explain these co-occurrences as follows: a great percentage of falsehoods were created between January 2020 and May 2020, a period in which there were still lots of unknowns about the origin, spread, cure, and prevention of the disease. This created a fertile ground for fully or partially misleading stories (e.g., benefits of eating a particular food as a protective measure) as well as outright conspiracy theories (e.g., Sars-Cov2 virus being a biological weapon). Starting from June 2020, the aforementioned announcements by leading figures and organizations seem to have helped reduce the misinformation around those subjects. However, despite a significant reduction in its amount, misinformation continued to be produced, and started to focus on other issues such as vaccines, politicians, and the impact of COVID-19 on healthcare systems. The count values in the graph have been colorized using a logarithmic scale, and there are a few instances where no co-occurrence has occurred (those observations have been colored as gray). On average, each falsehood has 2.165 co-occurring topics, and the maximum number of topics found in a single observation is 8.
During the collection and processing of the dataset, we were pleasantly surprised to see news with a varying degree of credibility and fantasy. There were stories with complete divergence from reality, nevertheless, they sounded credible; others gave the impression that they were produced in distant corners of a world of fantasy. Overall, we can divide the stories into two main categories. On the one hand, there were those that had some ground in reality but still distorted the facts and misled the reader, whether intentionally or not. One example was a story published in Brazil that claimed the FDA had warned the public that COVID-19 vaccines cause stroke. In reality, the FDA had prepared a table listing all possible side effects that must be looked after. That did not mean those side effects were known to happen after immunization with vaccines. On the other hand, some stories did not have any basis in facts and deserved to be labeled conspiracy theories. For instance, a story published in Georgia used a fake quote attributed to an American virologist and claimed that 5G high-frequency towers were installed to control humans implanted with microchips through vaccines, and the Rockefeller and Mason families were the financiers of this project. A range of examples showing this variation and the diversity of topics are in the Table 2.Table 2 Examples of falsehoods
Brief description of the story Origin Publication medium Topics assigned Fact checking Nature of the misinformation
The WHO stated that wearing a mask while playing sports increases the levels of carbon dioxide we absorb Spain Various media outlets Medical equipment The WHO recommends not wearing a mask when carrying out intense sports activity, but it never said it is due to an increase in carbon dioxide Distortion of reality
Studies show the flu shot increases a person’s susceptibility to COVID-19 Australia Facebook Vaccines, risk factors, other diseases In fact, the study referred to in the post shows that the flu shot does not make people more susceptible to any other respiratory infection Complete fabrication
There is a sign outside the Bill & Melinda Gates Foundation that reads as “Center for Global Human Population Reduction.” United States Facebook Individuals, conspiracies There is no such center run by the Gates, nor a sign as described Complete fabrication
A diet rich in alkaline foods can eliminate the coronavirus Brazil Facebook Cures, food There is no food capable of preventing or curing someone from Covid-19 Complete fabrication
Bing Liu, a researcher who was about to discover a vaccine for COVID-19, was murdered in the US Brazil Instagram Crime, individuals, vaccines Although a researcher named Bing Liu was killed in the US on May, he was not about to discover a vaccine for COVID-19 Distortion of reality
President Donald Trump attended a recital of Quran to fight coronavirus India Facebook Governments, individuals, religion The video is real, but it is from 2017, when Trump attended the inaugural prayer service after taking the oath as the 45th president of the United States Distortion of reality
As mentioned above, we compared the 129 countries in the dataset based on 14 variables aiming to measure economic (e.g. income inequality), political (e.g. trust in government), informational (e.g. press freedom), and socio-cultural (e.g. trust in science) environments in each. While selecting those variables, our goal was to draw a thorough picture of countries to understand what factors impact the production of COVID-19-related misinformation in different settings. A detailed discussion on the social, economic, and political variables we have used can be found in “Part I” of the Appendix.
Research questions
The rich dataset on misinformation provides opportunities to make statistical comparisons between countries (e.g., their social and political characteristics), topics of the news, and the content of their text. In addition, since the evolution of the pandemic was a socially dynamic phenomenon, the fourth aspect is time. The examination of the dataset shows that there is considerably high variation between individual observations over time, and this study aims to find if these micro-variations can lead to meaningful macro-level comparisons. The richness of detail and the opportunities offered by the variables guided us to construct a methodological framework to find the larger patterns in the data.
By attempting to cover a large ground, this study aims to conceptually and empirically contribute to the literature by grouping the countries according to the predominant types of falsehoods they produce. In order to cover the dynamic evolution of misinformation over the course of 13 months during the pandemic, the paper looks at four pillars of analysis that can be grouped under topic analysis and content analysis. With this background in mind, the paper aims to be one of the pioneering contributions to literature. Despite the early optimism stemming from vaccine rollouts, as of September 2021, COVID-19 was still a major health threat around the world due to factors including new mutations and many countries missing their vaccination targets. Particularly, we can predict that a sizable number of low- and middle-income countries will be fighting against it for a long time due to unacceptably low vaccination rates—e.g., only around 2% received at least one dose of a vaccine in low-income countries [83]. In addition, socially and biologically, the field is still characterized by known unknowns and unknown unknowns; therefore, conceptual contributions may provide helpful guidance to researchers and policymakers. Theoretically, the paper also aims to give comparative politics and sociology scholars opportunities to look deeper into the reasons why different countries produce different types of falsehoods and to analyze which socio-cultural, economic, and political variables affect the misinformation environment more than others. Methodologically, the paper takes advantage of a variety of statistical techniques, including a selection of network similarity algorithms. The use of network similarity algorithms to compare texts has largely been neglected in the computational social science literature.
It is also important to mention that before conducting this study, we considered a different strategy as well. In fact, initially, we extracted hundreds of millions of tweets from more than ten countries around the world and calculated the similarity between those tweets and the dataset on falsehoods. However, this approach resulted in no findings, and using different text similarity algorithms, we were not able to identify any matches. This led us to believe that a targeted approach to analyze misinformation could be more effective than trying to discover patterns in large but random samples.
A closer elaboration of the research questions has been provided below.
Topic analysis
The first two sets of research questions analyze the causal mechanism of topic selection by different groups of countries. The countries have been grouped by using economic, informational, political, and socio-cultural variables that we have introduced in the Data section. The set of questions below help to understand the macro patterns in the dataset by minimizing the errors associated with labeling through manual classification.
RQ1) Divisive and connective topics
RQ1a) In terms of topic creation, what are the topics that two groups of countries utilize in the most comparable amount vis-a-vis each other? In other words, what are the topics that are the most connective?
RQ1b) What are the topics that are the most divisive?
RQ2) Topic co-occurrences
RQ2a) What are the topics that co-occur the most?
RQ2b) Are some groups of countries statistically significantly different from others in terms of topic co-occurrence?
RQ2c) Is there a time frame in the evolution of COVID-19 in which topics were more similar to each other?
Content analysis
The second group of questions look more deeply at the content of the news by calculating the similarity between and across news associated with different topics and also analyze the news from the perspective of ‘unusualness’. (This will be explained in greater detail in the Methods section.) By doing this, we aim to understand if there is any association between topic correlation or content similarity across different groups of countries. The goal is to find out if countries have been inspired from each other in terms of content creation and how this relates to variables collected.
RQ3) Content similarity
RQ3a) Are there groups of countries that produce news that are significantly more similar to each other?
RQ3b) How does the similarity between news change over time?
RQ4) Misinformation unusualness/creativity
RQ4a) Are there groups of countries that are more creative than others in content formation?
RQ4b) How does creativity evolve over time?
As the research questions suggest, the paper aims to offer a descriptive perspective into the creation of a framework on misinformation production. The statistical tools used in the paper are elaborated closely in the next section.
Methods
The methodological tools used in the paper have been chosen to find similarities and differences between individual and groups of observations. To answer the four sets of research questions indicated above, we employed a variety of tools, including t-test, calculation of entropy and GINI index as a measure of information gain, k-means++ clustering, network similarity algorithms, and content comparison algorithms (NLP). (For the analysis, Python programming language and associated libraries were used.)
The data on falsehoods were collected from the Poynter Institute using webscraping techniques. (Poynter Institute allows the use of their data for research purposes.) The data was later manually processed to associate each observation with at least one topic from a collection of 28 different topics (topics were identified through the examples provided on Poynter Institute’s website). Around 500 observations could not be associated with any topic and therefore discarded. For the classification of topics, a few unsupervised clustering options have been tested, such as latent Dirichlet allocation [84] and non-negative matrix factorization [85]; nevertheless, the most coherent results have been obtained through manual labeling.
As previously mentioned, the starting point of this paper is the assumption that the nature of misinformation production is highly dependent on the personalities of countries that can be associated with certain socio-cultural, political, informational, and economic characteristics. In that sense, our study follows previous work such as Sauvy’s “Three Worlds, One Planet” (1952) [86] hat coined the term Third World; Huntington’s “The Clash of Civilizations?” (2000) [87]; Hall and Soskice’s Varieties of Capitalism: The Institutional Foundations of Comparative Advantage (2001) [2]; and Wallerstein’s The Capitalist World Economy (1979) [88] with its core versus periphery distinction. However, as different from them, we do not prioritize a specific dimension (e.g., economic systems or “culture”) as the primary distinguishing variable; instead, we aim at drawing a more comprehensive picture of countries by using 14 variables ranging from income inequality to trust in science and scientists to colonization history. In order to simplify and automate the classification of countries, a generally accepted and useful clustering algorithm, k-Means++ [89] was used. To pre-process the data for clustering, categorical variables have been converted into a 5-point Likert scale, and 0–1 normalization has been applied to all variables. The optimal number of clusters was determined using the cluster variation (SSE) and the ‘elbow method’. Two clusters came out as the optimal number, and six as the second optimal choice; to better represent the variation among countries, we chose six.
To handle the missing data in the datasets, different imputation methods were used. For the missing social, economic and political observations, a technique called “multivariate feature imputation” was implemented [90]. This technique uses a two-dimensional matrix as the input and models each feature with missing values as a function of other features using an iterated round-robin fashion. This is suitable for our case, since the variables at hand possibly have causal connections. To fill in the missing values in the time series datasets (for similarity and unusualness), K-nearest neighbors (KNN) [91–94] algorithm was used. The assumption behind this algorithm is that missing observations can be approximated by the values of the closest points, most frequently by taking the average of ‘k’ many points around the missing observation. As argued in a multitude of works using KNN for imputation (This approach is believed to work well in time series data with missing observations for which the best predictor of the missing points are the values temporally closest to them.
In addition, the same set of social, political, cultural, and economic variables were used to reduce dimensionality using principal components analysis. For the PCA, two dimensions came out to be the optimal choice based on the scree plot. Two other dimension reduction techniques, namely t-SNE [95] and spectral embedding [96], were considered; however, PCA was preferred as a traditional method to obtain two variables that are not correlated. The correlation map between the variables used and the reduced dimensions can be seen in the plot below (significant correlation values are marked with a black box). As Fig. 4 shows, most of the social, cultural, and political variation can be explained by four variables: corruption perceptions index and health coverage (PCA—Dimension 1), GINI Index, and trust in government (PCA—Dimension 2). Figure 5 is a representation of the variables after dimension reduction (PCA). As evidenced by it, countries in the dataset can be successfully grouped into six clusters (by using k-means++) with the help of the variables listed below. Findings have been consistent and robust after running the clustering algorithm ten times.Fig. 4 Correlations between variables
Fig. 5 Countries in different social, economic and political clusters
In order to compare the use of topics across groups of countries, an idea employed by decision trees was used. When decision trees are applied for classification goals, entropy and GINI index are the two most frequently used cost functions to calculate information gain/purity of the classes obtained by splitting the data. Thus, we wanted to find, given two groups of countries that are different by a single feature (for example, two groups of countries with different levels of democracy), which topic (in terms of its frequency) is the most different and which topic is the most similar among the two. Across the social, cultural, economic, and political variables, hundreds of comparisons between groups of countries were made, and the count values for most divisive (most different) and the most connective (most similar) topics were identified. Finally, these count values were inversely weighted by the count of the associated topic in the dataset to obtain a ranking for the most divisive and connective topics.
The paper assumes that the diversity of word usage in news reflects creativity; thus, more ‘unusually’ worded news are more creative. To measure the unusualness of the observations, 3-g and 4-g for the cleaned and lemmatized news were found. These n-grams were then used to calculate the TF-IDF score of each observation, which corresponds to the sum of TF-IDF scores for each n-gram associated with that observation. Observations with higher TF-IDF scores are believed to be more important and more creative; those with lower scores are considered as less unusual. This information was then used to compute how the unusualness changes over time and across groups of countries.
To calculate the document similarity between different observations, several considerations and attempts have been made, ranging from more generally accepted and earlier algorithms to more advanced techniques. Specifically, Word Mover Distance [97], Universal Sentence Encoder provided by Google [98], BERT embeddings [99], and Knowledge-based Measures [100] have been explored in the earlier phases of the analysis. All these models have turned out to be computationally too expensive to find the cross-similarities between over 10,000 documents. To solve this problem, TF-IDF (term frequency-inverse document frequency) scores have been calculated for all documents following the cleaning and lemmatization process [101]. Eventually, cosine similarities have been found in over 50 million cross-comparisons. These similarities have then been aggregated to make cross-country comparisons using t tests.
Finally, network similarity algorithms were applied to compare adjacency matrices composed of bi-weekly aggregated topic correlations between documents. Topic similarities can be represented as graph data since one document can only have a limited number of topics, and more than one topic presented in a single document can change the impact of the misinformation dramatically (holistic assumption). To calculate the similarities between topic correlation matrices, the following two advanced graph similarity algorithms were used: Frobenius distance [102] and quantum-JSD distance [103]. The aggregated relative similarity matrix between topic ratios has been provided as an example in Fig. 6. Topics with yellow-to-red colored cross-similarities are more closely related, and topics with yellow-to-blue colored cross-similarities are rarely mentioned together. A closer elaboration on this relationship will be provided in the “Results” section. For more information on how network similarities have been calculated, please refer to the “Appendix” section. In a similar approach, PERMANOVA [104] and Anosim [105] techniques that allow the comparison of n × n-dimensional topic correlation matrices were put to test; however, ultimately, the tests were not reported because of the impact of data size on the results.Fig. 6 Relative similarities between falsehoods of all topics
Empirical results
Topic analysis
The topic analysis focuses on two questions as previously mentioned: divisive and connective topics (i) and topic co-occurrences (ii). The results obtained for the first case indicate significant variance in the power of topics to differentiate groups of countries from each other. Thus, clusters of countries can be strongly associated with topics and vice versa. Topics were used to separate countries into clusters and these clusters were compared with the groups generated through the use of social, economic, cultural, and political variables. The results show that some topics lead to a much greater amount of cluster purity when used for generating groups. The analysis to calculate purity has been repeated by using two cost functions, entropy, and GINI Index, and the results are the same. The table below provides a ranking for the most connective and divisive topics. In each comparison the name for the most connective and divisive topic has been obtained and the number of times a topic appears as the most connective or most divisive has been recorded. Finally, the count values have been inversely weighted by the falsehood count associated with that particular topic. The table below shows the ranking for most connective and divisive topics and these values. The inversely weighted values explain how strongly connective or divisive a topic is compared to others (Table 3).Table 3 Ranking of most connective and most divisive topics
Minimum entropy/GINI index (most connective) Maximum entropy/GINI index (most divisive)
1 Conspiracies (0.071) Animals (0.387)
2 Medical equipment (0.058) Predictions (0.135)
3 Risk factors (0.053) Symptoms (0.067)
4 Other diseases (0.048) Laws (0.063)
5 Vaccines (0.047) Travel (0.046)
6 Travel (0.046) Risk factors (0.046)
7 Aid (0.029) Origins (0.030)
8 Religion (0.029) Prevention (0.029)
9 Animals (0.028) Spread (0.017)
10 Symptoms (0.026) Hospitals (0.015)
11 Laws (0.024) Aid (0.012)
12 Origins (0.023) Conspiracies (0.012)
13 Predictions (0.021) Other diseases (0.010)
14 Medicine (0.017) Medicine (0.009)
15 Hospitals (0.015) Detection (0.008)
16 Crime (0.014) Medical equipment (0.008)
17 Cures (0.012) Religion (0.006)
18 Detection (0.009) Food (0.005)
19 Videos (0.008) Crime (0.004)
20 Food (0.005) Videos (0.002)
21 Lockdown (0.004) Vaccines (0.001)
22 Governments (0.003) Governments (0.000)
23 Spread (0.003) Cures (0.000)
24 Prevention (0.002) Individuals (0.000)
25 Individuals (0.002) Lockdown (0.000)
As seen above, “conspiracies” is the most shared topic category across groups of countries. We explain this finding as follows: the major conspiracy theories, including those pointing to Bill Gates-led plots to implant digital microchips to control people, marking the virus as a biological weapon created by Chinese or American scientists, and those demonizing pharmaceutical companies as agents that worsen the pandemic and conceal the effective treatments, are produced by a small number of individuals and organizations with political and financial goals. Then, these are shared globally in the form of news stories occasionally through media outlets, but primarily via social media posts. In fact, a recent investigation conducted by the Associated Press and the Atlantic Council’s Digital Forensic Research Lab found that a few “superspreaders”—people and organizations such as Kevin Barrett, an anti-Semitic former lecturer on Islam, and the Montreal based “Centre for Research on Globalization”—were responsible for a great percentage of conspiracies on the origin of COVID-19 circulating online [106]. Similarly, a study published right before the COVID-19 pandemic found that 54% of all anti-vaccine ads on Facebook were funded by two organizations, even though most of the ads appeared to be grass-root discussions by concerned parents and neighborhood groups [107]. Thus, in addition to raising concerns around the use of Facebook and similar platforms to spread misinformation, this finding indicates that conspiracy theories regarding COVID-19 have a global appeal cutting across socio-cultural, economic, informational, and political variables that divide the countries.
Secondly, we looked at the co-occurrence dynamics of the topics in the topic analysis section. A one-to-one match between each topic gives close to 400 possibilities for topic pairs. Among those, co-occurring topics with an aggregated relative similarity of more than 0.1 have been selected and their number of co-occurrences have been inversely weighted by the total count of both topics (comparable to Jaccard similarity) in periods of two weeks. In other words, a matrix similar to the one in Fig. 5 has been produced for every two weeks topic pairs with high relative similarity have been observed. This gave us Fig. 6 below. The high similarity co-occurring topics are food-cures, individuals-governments, lockdown-governments, lockdown-individuals, medicine-cures, origins-conspiracies, other diseases-medical equipment, prevention-cures, prevention-food, spread-detection, spread-individuals, videos-individuals, and videos-religion. The figure below suggests that there is a pattern in the co-occurrence of the topics and the time series dataset can be clustered into the following two groups: before April 2020 and after. Higher values correspond to greater weighted topic co-occurrence and lower values indicate that co-occurrence has become weaker (Fig. 7).Fig. 7 Relative similarity between high-correlation topics over time
Lastly, bi-weekly relative similarity matrices have been treated as networks and compared to each other using network similarity algorithms. This provided a systematic way to compare the dynamics of topic ratios in the misinformation dataset over time. To validate the results, two different algorithms (Frobenius distance and quantum-JSD distance) have been used and the results have been evaluated against each other. The results indicate that in the first few months of the pandemic, topic ratios have been comparable to each other; specifically, starting from February 2020 until the end of June 2020, results suggest that there has not been much variation. This finding is also reinforced by the results provided in Fig. 6 relatively more conservatively: highly correlated topics formed a pattern until the end of May 2020. This suggests intense cross-country exchanges and learning from each other in the first few months of the pandemic. The graphs below show the similarities (or distances) between the bi-weekly relative similarity graphs. The rectangles in the intersection of two time points show the distance between two topic ratio graphs. Red values are associated with greater similarity and blue values correspond to lower similarity scores. In addition, to have complete data for the bi-weekly periods, the first and the last time series observations have been truncated (Fig. 8).Fig. 8 Frobenius and quantum-JSD distances between the bi-weekly relative similarity graphs of topic ratios
We interpret those results in line with the discussion in the “Data” section above. As we also mentioned there, the period until May/June 2020—the first few months of the pandemic— was characterized by uncertainties about the definition and implications of the Sars-Cov2 virus and the highest intensity of misinformation production. More specifically, there were still lots of unknowns about the origin, spread, cure, and prevention of the disease; each among the topics with a significant amount of co-occurrence. Starting from June 2020, with key announcements by leading figures and organizations regarding the origin and spread of the virus as well as methods of treatment and prevention—e.g. the WHO’ announcement that COVID-19 may be an airborne disease mainly transmitted through respiratory droplets—we saw a decline in the number of falsehoods related to those popular and highly co-occurring topics, while misinformation started to focus on other issues such as vaccines, politicians, and the impact of COVID-19 on healthcare systems.
Content analysis
Content similarity
In this section, we tried to understand if countries of similar social, economic, and political backgrounds produce news with similar content, or if there is a statistically significant difference between countries with different social, economic and political endowment in terms of content creation. The assumption was that the behavior of people is strongly related to national variables [108, 109] and this ultimately translates into writing. In fact, there is an extensive literature showing that individual’s everyday behaviors such as financial decisions [110], consumption habits [111], and health behaviors [112] are associated with national variables, including cultural values, human development levels, or business systems. Similarly, scholars point to how national-level factors such as political systems, economic indicators, or press freedom have determining impacts on the “journalistic cultures” [113], which, in turn, shape how different topics, including climate change [114] and international migration [115] are covered.
We broke the countries down into groups and compared the aggregated mean of pairwise similarity between the news for bi-weekly periods and for the whole dataset. The instances in which a comparison results in statistically significantly higher similarity results than the other sets of comparisons have been identified. These instances can be observed from the graphs below. The remainder of the comparisons can be found in the “Appendix”. On the whole, countries in the fourth cluster, countries in West and South-Asia, socialist/Arab-oil-based/advanced city economies, countries with low HDI (human development index), countries with very serious press freedom problems produce news that are more similar to each other (the remainder of the comparisons have been reported in the “Appendix”) (Figs. 9, 10, 11, 12, 13).Fig. 9 Similarity between news over time, cluster comparison
Fig. 10 Similarity between news over time, HDI comparison
Fig. 11 Similarity between news over time, culture comparison
Fig. 12 Similarity between news over time, business systems comparison
Fig. 13 Similarity between news over time, press freedom comparison
We believe some of those findings are particularly worth discussing here. First, it should be noted that an overwhelming percentage of countries classified as having low HDI are located in sub-Saharan Africa. A number of studies indicate that the COVID-19-related misinformation in the African continent has a few distinct characteristics, and we speculate that this might explain why the news are more similar to each other. Specifically, falsehoods related to unproven local remedies [116] and those stemming from religious beliefs [117] are found to be particularly common in Africa. In addition, distrust towards international bodies [118] and the history of unethical Western medical practices in the continent [119] are some of the other factors fuelling misinformation. In fact, our dataset offers some interesting examples. In Ivory Coast, stories claiming that neem leaf works against COVID-19 were posted thousands of times on social media despite no evidence. Similarly, falsehoods claiming that the Rwandan president Paul Kagame censured the WHO for rejecting a herbal tonic were widely shared on Facebook and Twitter across African nations, including Nigeria.
Countries classified as having “very serious” press freedom issues by the reporters without borders also produced more similar news. In line with our discussion on the “journalistic cultures” above, we believe that this might reflect the effects of government control and censorship of the media, which largely shape both the content and tone of the coverage with regard to the pandemic. In fact, in this group of countries— including Egypt, Iran, Saudi Arabia, China, and Vietnam, among others—not just the traditional media sources, but also the social media networks are subject to heavy government control [72]. For instance, a report by the social media exchange—an NGO working to advance digital rights in the Arabic-speaking region—shows how Egyptian authorities prosecuted a number of journalists, doctors, and activists who circulated news on the social media on the COVID-19 outbreak—e.g., the number of infections or deaths—that did not match the official discourse and numbers [120].
A third interesting finding is that the countries in the fourth cluster—the light pink colored cluster in Fig. 3 above—generated more similar news in terms of the content. Some of the countries in this group are Iraq, the Democratic Republic of Congo, Venezuela, Honduras, Kenya, Bolivia, Uganda, and Yemen. A few of the members of this cluster have also low HDI and/or very serious press freedom issues; therefore, the explanations above can partially apply to those countries. However, four distinct characteristics identify the countries in this cluster: a high perception of corruption of the public sector, a high degree of mistrust towards the government, a high level of economic inequality, and largely ineffective health service provision. Taken together, these factors point to an environment of weak state capacity and a low level of trust in public institutions. In fact, studies show that there are remarkable correlations among those variables. For example, while a study finds a strong relationship between high levels of economic inequality and low levels of trust in national institutions across the EU member countries [121], another one conducted in post-Soviet nations shows that there is a negative association between perception of corruption and trust in public institutions such as police, national and regional governments, and courts [122]. Given that the state capacity and trust in public institutions are integral to an effective pandemic strategy—affecting people’s compliance with restrictions and willingness to get vaccinated, governments’ success in enforcing lockdowns and other isolation practices, etc.—it is not surprising that those countries produced more similar news. Our dataset includes several fascinating falsehoods particularly common to the countries in this cluster. Reflecting the mistrust towards the government and its capacity to supervise the pandemic efforts, a news story in Zimbabwe alleged that a medical laboratory conducted clinical trials for a possible vaccine and led to the death of 68 out of 80 volunteers in total. Similarly, mistrust towards the government and a high level of political polarization undoubtedly fostered misinforming news such as the one in Bolivia that was published in July 2020 and claimed that President Maduro was extending the full lockdown until January 2021.
Misinformation unusualness/creativity
In the last part of the paper, groups of countries have been compared against each other in terms of the creative word usage in the news they published. The average value of unusualness of one group was compared with the average unusualness extracted from the other group. Since the difference is taken into consideration, statistically significant results should be much higher than zero. We provided the average aggregated unusualness score in Fig. 14. The figure shows that an initial lack of creativity in the first few months of the pandemic was followed by an increase and relative stability throughout 2020.Fig. 14 Aggregated average unusualness scores over time
Breaking down the countries into groups and comparing the levels of creativity between them did not provide any results that are statistically significantly different from zero. Thus, our expectation that countries of different backgrounds would choose word-groups according to their taste did not come true. A comparison between clusters of countries created using the social, cultural, economic, and political variables in the dataset has been provided below in Fig. 14. We believe that this lack of meaningful difference across clusters of countries in terms of creative word usage can be explained by three main factors. First, it points to a highly globalized media environment in the sense that media outlets across nations share vocabulary and discourses to a great extent. The digital media, and the Internet more broadly, have created “a new global language” [123] with specific neologisms and novel syntactic, orthographic, and lexical commonalities among world languages, such as heavy use of emojis and emoticons, abbreviations, and acronyms [124]. Second, research shows significant differences between truthful news and falsehoods regarding their linguistic characteristics as the latter use more words related to anxiety, more superlatives, sensationalistic writing, and overly emotional language [125, 126]. Third, as previously mentioned, a great percentage of falsehoods in our dataset are from Facebook and a few other social media outlets. Given the studies showing that misinformation spreads really fast on social media platforms [24, 127] and that most COVID-19-related falsehoods were produced by a very small number of individuals [128], the lack of statistically significant difference among the groups of countries is not too surprising (Fig. 15).Fig. 15 Comparison of unusualness scores between clusters of countries
Discussion and conclusion
The “Varieties of… “literature has influenced more than a generation of scholars and practitioners worldwide. There have been politicians to use the arguments first offered by Hall and Soskice to transform the institutional structure of their countries (such as the British Labour Party politician, Ed Miliband, when he was Leader of the Opposition) and many other scholars who expanded the typologies first proposed by Hall and Soskice. Academically, we hope that our paper will provide a strong comparative perspective to an emerging literature. We also agree that the “Varieties of…” conceptualization is deterministic in nature; however, as recent media-viewer experience suggests, local and global media, policymakers, and transnational institutions have also been looking at pandemic-related policy success and failure from a cross-national, and mostly deterministic perspective. Thus, many are wondering why some countries have been more successful than others in mitigating the human costs of the pandemic, while others have been less so in an environment where local political leaders are looking for the best non-local practices. From a practical sense, we believe that the arguments and facts laid out here may contribute to the public health efforts to fight against misinformation, which continues to take lives in a myriad of ways, such as by discouraging people from getting vaccinated or promoting fraudulent and dangerous products.
To conclude, we want to reiterate four of the key contributions that this paper provides to the literature, and particularly to the tools to be used by global public health circles. First, our study is truly unique in terms of its data and methodology—it comprises over 10 thousand falsehoods from 129 countries; its data come from a variety of sources, including the most widely used social media platform globally, i.e., Facebook; and it uses 14 different variables aiming to measure political, economic, informational, and socio-cultural environments in each country in order to compare COVID-19 misinformation across them. We believe that the resulting clustering of countries into groups offers avenues for developing distinct public health and communication strategies to dispel misinformation in countries with particular characteristics.
Second, and relatedly, the findings give clues on what those strategies should be. For instance, our analysis suggests that countries with low HDI (mainly located in sub-Saharan Africa) produce misinformation related to unproven local remedies and those stemming from certain religious beliefs as well as from distrust of international organizations and Western medical practices. This shows the importance of working with local religious leaders and healers and repairing Western governments and pharmaceutical companies’ tarnished reputation. Likewise, given that countries with severe press freedom issues (e.g., those implementing outright censorship of news and bans on social media platforms) generate similar news, global public health circles should design an anti-misinformation strategy specific to those nations, which necessitates going beyond using online platforms that are at risk of being censored.
Third, our study indicates that there have been successful anti-misinformation efforts throughout the pandemic, but significant challenges persist. More specifically, we found that the types of falsehoods that were particularly common in the first few months of the pandemic and were widely shared across countries (mainly those about the origin, spread, cure, and prevention of the disease) got effectively addressed by announcements coming from leading figures and organizations such as WHO or Anthony Fauci, resulting in a decline in the number of falsehoods related to those topics. However, the findings also reveal two worrying trends, among others: (1) conspiracy theories are common among all groups of countries, which can be explained by the fact that they are originated by a small number of individuals and organizations (aka misinformation “superspreaders”), but are effectively disseminated across the globe; (2) in countries with weak state capacity and a low level of trust in public institutions, misinformation creates a particularly dangerous vicious circle—distrust of government fosters the production of falsehoods, which in turn further weakens governments’ ability to supervise the pandemic efforts. Accordingly, we argue that international organizations and leading figures in global health should strengthen their efforts to reach out to those populations and develop effective strategies against the dissemination of the major conspiracies.
Fourth, we found that while the most prominent misinformation topics vary across groups of countries, the word-groups used in misinforming news stories are remarkably similar. In line with the “glocalization” literature [129, 130], we interpret this as the coexistence of globalizing and localizing processes—on the one hand, socio-economic, cultural, political, and informational characteristics of countries clearly affect the types of falsehoods Internet users are exposed to; but, one the other hand, the tone and structure of falsehoods do not show much variance, which points to a highly globalized online media environment. Given those findings, we argue that even though implementing country-specific strategies (e.g., improving scientific literacy in a country where it is currently weak) is crucial, a global action plan against misinformation is also very much needed.
Electronic supplementary material
Below is the link to the electronic supplementary material and the dataset.Supplementary file1 (Data Documentation) (DOCX 7 KB)
Supplementary file2 (COVID-19 Falsehoods and Topics Dataset) (XLSX 3757 KB)
Appendix
Part I: Social, economic, and political variables
See Table 4.
Table 4 Country information data
Name of the index Organization/individual that compiled the index Year of data collection Number of countries represented Indicators Assigned categories
Freedom in the World Freedom House [69, 70] 2019 129 Electoral process; political pluralism and participation; functioning of government; freedom of expression and belief; associational and organizational rights; rule of law; personal autonomy and individual rights Free; partly free; not free
Edelman Trust Barometer Edelman Data and Intelligence [71] 2018 26 Public trust in government through an online survey Scale (lowest = 21
highest 86)
Corruption Perceptions Index Transparency International [72] 2019–2020 127 How corrupt the public sector is perceived to be (based on expert opinion and surveys of business people) Scale (lowest = 12 highest = 88)
World Press Freedom Index Reporters Without Borders [73] 2019 127 Online questionnaire filled by lawyers, sociologists, and media professionals evaluating: pluralism, media independence, media environment and self-censorship, legislative framework, abuses, transparency, and the quality of the infrastructure that supports the production of news and information Good situation, satisfactory situation, problematic situation, difficult situation, very serious situation
Reuters Institute Digital News Report Reuters Institute for Study of Journalism [74] 2019 37 Online questionnaire filled by a representative sample of people evaluating overall trust in the news % of those who trust in the news (lowest = 22, highest = 59)
The Wellcome Global Monitor Wellcome Trust [75] 2018 115 Questionnaire filled by a representative sample of people face-to-face or via telephone evaluating people’s trust in scientists % of people who have a high level of trust in scientists (lowest = 2, highest = 54)
Index of Populist Rhetoric Luigi Curini (Universita degli Studi di Milano) (2019) [76] 2012–2018 37 The populist discourse of the most recent presidents and prime ministers who came to power since 2012 (using The Global Populism Database by the Guardian and Team Populism) Scale (1 = not populist, 8 = very populist)
World Cultural Map World Values Survey [77] 2017–2020 92 Face-to-face interview with a representative sample of people Confucian, Orthodox Europe, African-Islamic, Latin America, West and South Asia, Catholic Europe, English-speaking, Protestant Europe
Clusters of Business Systems Michael A. Witt et al. (2018) [78] 2013 50 Education, employment relations, finance, interfirm relations, internal dynamics, ownership and governance, social capital, state Socialist economies, emergent economies, Arab oil-based economies, advanced city economies, advanced emerging economies, European peripheral economies, liberal market economies, coordinated market economies, highly coordinated economies
Human Development Index United Nations Development Program [79] 2018–2019 126 Health, education, income Very high, high, medium, low
GINI Index World Bank [80] 2019 117 Distribution of income across a population GINI coefficient: 1 = perfect inequality, 0 = perfect equality (lowest = 0.25, highest = 0.63)
Effective coverage of health services index Global Burden of Diseases 2019 Universal Health Coverage Collaborators (2020) [81] 2019 127 Promotion, prevention, treatment Scale (lowest = 32 highest = 96)
Colonization history N/A N/A 129 History of being colonized by a Western power (excluding settler colonies such as Australia) Yes, no
United Nations Geoscheme United Nations Statistics Division [82] N/A 129 Geographic subregions Northern America, Central America, Caribbean, South America, Northern Europe, Western Europe, Eastern Europe, Southern Europe, Western Asia, Central Asia, Eastern Asia, Southern Asia, Southeastern Asia, Melanesia, Polynesia, Australia and New Zealand, Northern Africa, Western Africa, Central Africa, Southern Africa, Eastern Africa
Part II: An explanation on network similarity formulas
A closer elaboration of the formulas has been taken from a chapter in Financial Data Analytics: Theory and Applications [131]:
“Frobenius Distance [103] computes the similarity between two graphs by “locally” comparing the individual connections between pairs of nodes. This is considered a known node correspondence (KNC) method (an algorithm that needs information about which nodes should be compared to each other). If ai,j and bi,j represent the connections between two nodes i and j that belong to two different graphs G1 and G2 such that ai,j is in G1 and bi,j is in G2, Frobenius distance d(G1, G2) is the following:dG1,G2=∑i,j||aij-bij||.
Quantum-JSD distance [103] compares the spectral entropies of the density matrices. This is done by calculating the ‘Quantum’ Jensen–Shannon divergence between two graphs. The authors create a connection-based density matrix to calculate the von Neumann entropy of a network. The algorithm proposed uses the whole network, instead of a subset of network features. Most importantly, the algorithm allows the authors to quantify the distance between ‘complex’ networks. Classical algorithms attempt to quantify the amount of information about a probability distribution (entropy), and quantum JSD expands this definition by introducing divergences (also known as quantum relative entropy). The distance is calculated by using a generalized Jenson–Shannon divergence between two graphs:Jq(ρ||σ)=Sqρ+σ2-12[Sq(ρ)+Sq(σ)]
In the equation above, ρ and σ represent the density matrices and q represents the order parameter. The density matrix looks like the following:ρ=e-βLZ,
whereZ=∑i=1Ne-βλi(L)
and λi(L) represents an imaginary diffusion process i, over the network with time parameter β > 0.
Part III: Content similarity comparison graphs
A list of the visuals that demonstrate insignificant comparison results have been listed below (Figs. 16, 17, 18, 19, 20, 21, 22, 23, 24).Fig. 16 Similarity between news over time, Freedom House Index comparison
Fig. 17 Similarity between news over time, trust in government comparison
Fig. 18 Similarity between news over time, GINI Index comparison
Fig. 19 Similarity between news over time, trust in science comparison
Fig. 20 Similarity between news over time, Corruption Perceptions Index comparison
Fig. 21 Similarity between news over time, trust in news media comparison
Fig. 22 Similarity between news over time, colonization comparison
Fig. 23 Similarity between news over time, populism comparison
Fig. 24 Similarity between news over time, Health Services Index comparison
Part IV: Unusualness differences over time
A list of the visuals that demonstrate the comparison results that are insignificantly different from each other have been listed below (Figs. 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36)Fig. 25 Unusualness differences between varieties of countries over time, HDI comparison
Fig. 26 Comparison of unusualness scores between clusters of cultures
Fig. 27 Unusualness differences between varieties of countries over time, Freedom House Index comparison
Fig. 28 Unusualness differences between varieties of countries over time, business systems comparison
Fig. 29 Unusualness differences between varieties of countries over time, press freedom comparison
Fig. 30 Unusualness differences between varieties of countries over time, trust in government comparison
Fig. 31 Unusualness differences between varieties of countries over time, GINI Index comparison
Fig. 32 Unusualness differences between varieties of countries over time, trust in science comparison
Fig. 33 Unusualness differences between varieties of countries over time, Corruption Perceptions Index comparison
Fig. 34 Unusualness differences between varieties of countries over time, trust in news media comparison
Fig. 35 Unusualness differences between varieties of countries over time, populism comparison
Fig. 36 Unusualness differences between varieties of countries over time, Health Services Index comparison
Declarations
Conflict of interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Data availability statement
The authors confirm that all data generated or analyzed during this study will be made available as a supplement with the publication of this article.
1 Some well-known recent works from political science literature that use the ‘Varieties of…’ in their titles are [3–8].
2 On this note, we wholeheartedly thank our attentive research assistant Che Hoon Jeong from Denison University in Granville, Ohio for his extensive work on compiling the topics for the misinformation dataset.
3 It is worth indicating that geolocation and translation into English has been done by the sources providing data to Poynter Institute [1].
4 The types of misinformation and their count as provided in the original (raw) dataset is provided here. False: 8615, Misleading: 655, MISLEADING: 383, Partly false: 132, NO EVIDENCE: 128, Mostly false: 104, misleading: 64, No evidence: 58, No Evidence: 49, PARTLY FALSE: 46, Explanatory: 38, Mostly False: 31, partly false: 21, Partially false: 18, MOSTLY FALSE: 14, Partly False: 13, no evidence: 12, missing context: 9, mostly false: 9, MOSTLY TRUE: 8, MIsleading: 7, Missing context: 6, mainly false: 6, HALF TRUE: 6, false context: 4, Mostly True: 4, MISSING CONTEXT: 3, Partially False: 3, Partially true: 3, Two Pinocchios: 3, Fake: 3, Half True: 3, Inaccurate: 2, Partly FALSE: 2, mislEADING: 2, half true: 2, PARTLY TRUE: 2, Misleading/False: 2, Unproven: 2, "(Org. doesnt apply rating)": 2, Correct: 2, Missing Context: 1, partially false: 1, MISLEADING/FALSE: 1, EXPLANATORY: 1, mainly correct: 1, UNPROVEN: 1, True but: 1, Partly true: 1, Partially correct: 1, IN DISPUTE: 1, Mostly true: 1, false and misleading: 1, Mixed: 1, HALF TRUTH: 1, MiSLEADING: 1, Unlikely: 1, Misinformation / Conspiracy theory: 1, Fake news: 1, Unverified: 1.
5 Hand-labeling of textual data is a technique that has been implemented in other related studies, as well [59–64].
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Poynter Institute. IFCN Covid-19 Misinformation [Internet]. Poynter. [cited 2022 Feb 25]. https://www.poynter.org/ifcn-covid-19-misinformation/. Accessed 13 Nov 2022.
2. Hall PA Soskice D Varieties of capitalism: The institutional foundations of comparative advantage 2001 Oxford University Press
3. Benz A Broschek J Federal dynamics: Continuity, change, and the varieties of federalism 2013 OUP Oxford
4. Bochsler D Kriesi H Kriesi H Varieties of democracy Democracy in the age of globalization and mediatization 2013 Palgrave Macmillan
5. Devinney TM Hartwell CA Varieties of populism Global Strategy Journal 2020 10 1 32 66 10.1002/gsj.1373
6. Dorman SR The varieties of nationalism in Africa Current History 2015 114 772 189 193 10.1525/curh.2015.114.772.189
7. Rothstein H Demeritt D Paul R Varieties of risk regulation in Europe: Coordination, complementarity and occupational safety in capitalist welfare states Socio-Economic Review 2019 17 4 993 1020 10.1093/ser/mwx029
8. Saraceno C Varieties of familialism: Comparing four southern European and East Asian welfare regimes Journal of European Social Policy 2016 26 4 314 326 10.1177/0958928716657275
9. Zubiaga A Liakata M Procter R Hoi GWS Tolmie P Analysing how people orient to and spread rumours in social media by looking at conversational threads PLoS ONE 2016 11 3 e0150989 10.1371/journal.pone.0150989 26943909
10. Kumar, S., West, R., Leskovec, J. (2016). Disinformation on the web: Impact, characteristics, and detection of Wikipedia hoaxes. In: Proceedings of the 25th international conference on World Wide Web [Internet]. Republic and Canton of Geneva, CHE: International World Wide Web Conferences Steering Committee. [cited 2022 Apr 23]. p. 591–602. (WWW ’16). 10.1145/2872427.2883085.
11. Allcott H Gentzkow M Social media and fake news in the 2016 election Journal of Economic Perspectives. 2017 31 2 211 236 10.1257/jep.31.2.211
12. Mejias UA Vokuev NE Disinformation and the media: The case of Russia and Ukraine Media, Culture and Society. 2017 39 7 1027 1042 10.1177/0163443716686672
13. Merriam-Webster. The dictionary [Internet]. [cited 2022 Apr 23]. https://www.merriam-webster.com/. Accessed 13 Nov 2022.
14. Shahi GK Dirkson A Majchrzak TA An exploratory study of COVID-19 misinformation on Twitter Online Social Networks and Media. 2021 1 22 100104 10.1016/j.osnem.2020.100104
15. Wardle, C., Derakhshan, H. (2017). Information disorder: Toward an interdisciplinary framework for research and policy making. Council of Europe Report. [cited 2022 Apr 23]. https://rm.coe.int/information-disorder-toward-an-interdisciplinary-framework-for-researc/168076277c. Accessed 13 Nov 2022.
16. Guess, A. M., Lyons, B. A. (2020). Misinformation, Disinformation, and Online Propaganda. In Social Media and Democracy: The State of the Field, Prospects for Reform, edited by Joshua A. Tucker and Nathaniel Persily, 10–33. SSRC Anxieties of Democracy. Cambridge: Cambridge University Press. https://www.cambridge.org/core/books/social-media-anddemocracy/misinformation-disinformation-and-online-propaganda/D14406A631AA181839ED896916598500.
17. World Health Organization. (2020). Coronavirus disease (COVID-19): situation report, 162 [Internet]. World Health Organization. [cited 2022 Apr 23]. https://apps.who.int/iris/handle/10665/332970. Accessed 13 Nov 2022.
18. Lewis, T. (2020). Eight persistent COVID-19 myths and why people believe them [Internet]. Scientific American. [cited 2022 Apr 23]. https://www.scientificamerican.com/article/eight-persistent-covid-19-myths-and-why-people-believe-them/. Accessed 13 Nov 2022.
19. Lynas, M. (2020). COVID: Top 10 current conspiracy theories [Internet]. Alliance for Science. [cited 2022 Apr 23]. https://allianceforscience.cornell.edu/blog/2020/04/covid-top-10-current-conspiracy-theories/. Accessed 13 Nov 2022.
20. Spring, M., Wendling, M. (2020). How Covid-19 myths are merging with the QAnon conspiracy theory. BBC News [Internet]. [cited 2022 Apr 23]. https://www.bbc.com/news/blogs-trending-53997203. Accessed 13 Nov 2022.
21. Islam MS Sarkar T Khan SH Kamal A-HM Hasan SM Kabir A COVID-19-related infodemic and its impact on public health: A global social media analysis The American Journal of Tropical Medicine and Hygiene. 2020 103 4 1621 10.4269/ajtmh.20-0812 32783794
22. Murphy, J. (2021). Biden has new vaccination goals. See if the country is on pace to hit them. [Internet]. NBC News. [cited 2022 Apr 23]. https://www.nbcnews.com/politics/white-house/graphic-track-biden-fourth-july-vaccination-goals-n1268803. Accessed 13 Nov 2022.
23. Centers for Disease Control and Prevention. (2021). Delta variant: What we know about the science. 2021. https://www.cdc.gov/coronavirus/2019-ncov/variants/delta-variant.html. Accessed 13 Nov 2022.
24. Obiała J Obiała K Mańczak M Owoc J Olszewski R COVID-19 misinformation: Accuracy of articles about coronavirus prevention mostly shared on social media Health Policy and Technology. 2020 10 182 186 10.1016/j.hlpt.2020.10.007 33163352
25. Singh, L., Bansal, S., Bode, L., Budak, C., Chi, G., Kawintiranon, K., et al. (2020). A first look at COVID-19 information and misinformation sharing on Twitter. arXiv:2003.13907v1.
26. Chen E Chang H Rao A Lerman K Cowan G Ferrara E COVID-19 misinformation and the 2020 US presidential election 2021 The Harvard Kennedy School Misinformation Review
27. Memon, S. A., Carley, K. M. (2020). Characterizing COVID-19 misinformation communities using a novel Twitter dataset. arXiv:200800791 [cs] [Internet]. [cited 2022 Apr 23]. http://arxiv.org/abs/2008.00791. Accessed 13 Nov 2022.
28. Bridgman, A., Merkley, E., Loewen, P. J., Owen, T., Ruths, D., Teichmann, L., et al. (2020). The causes and consequences of COVID-19 misperceptions: Understanding the role of news and social media. Harvard Kennedy School Misinformation Review, 1(3).
29. Brennen, J. S., Simon, F. M., Howard, P. N., Nielsen, R. K. (2020). Types, sources, and claims of COVID-19 misinformation. Ph.D. thesis. University of Oxford.
30. Enders, A. M., Uscinski, J. E., Klofstad, C., Stoler, J. (2020). The different forms of COVID-19 misinformation and their consequences. The Harvard Kennedy School Misinformation Review.
31. Stecula DA Pickup M How populism and conservative media fuel conspiracy beliefs about COVID-19 and what it means for COVID-19 behaviors Research and Politics. 2021 8 1 2053168021993979 10.1177/2053168021993979
32. Uscinski, J. E., Enders, A. M., Klofstad, C., Seelig, M., Funchion, J., Everett, C., et al. (2020). Why do people believe COVID-19 conspiracy theories? Harvard Kennedy School Misinformation Review [Internet]. [cited 2022 Apr 23];1(3).
33. Lobato, E. J. C., Powell, M., Padilla, L. M. K., Holbrook, C. (2020). Factors predicting willingness to share COVID-19 misinformation. Front. Psychol. [Internet]. [cited 2022 Apr 23];11. 10.3389/fpsyg.2020.566108. Accessed 13 Nov 2022.
34. Motta M Stecula D Farhart C How right-leaning media coverage of COVID-19 facilitated the spread of misinformation in the early stages of the pandemic in the US Canadian Journal of Political Science. 2020 53 2 335 342 10.1017/S0008423920000396
35. Pennycook G McPhetres J Zhang Y Lu JG Rand DG Fighting COVID-19 misinformation on social media: experimental evidence for a scalable accuracy-nudge intervention Psychological Science 2020 31 7 770 780 10.1177/0956797620939054 32603243
36. Sallam M Dababseh D Yaseen A Al-Haidar A Taim D Eid H COVID-19 misinformation: Mere harmless delusions or much more? A knowledge and attitude cross-sectional study among the general public residing in Jordan PLoS ONE 2020 15 12 e0243264 10.1371/journal.pone.0243264 33270783
37. Agley J Xiao Y Misinformation about COVID-19: Evidence for differential latent profiles and a strong association with trust in science BMC Public Health 2021 21 1 1 12 10.1186/s12889-020-10103-x 33388037
38. Laato S Islam AKMN Islam MN Whelan E Why do people share misinformation during the COVID-19 pandemic? European Journal of Information Systems. 2020 29 3 288 305 10.1080/0960085X.2020.1770632
39. Roozenbeek J Schneider CR Dryhurst S Kerr J Freeman ALJ Recchia G Susceptibility to misinformation about COVID-19 around the world Royal Society Open Science 2020 7 10 201199 10.1098/rsos.201199 33204475
40. Hornik R Kikut A Jesch E Woko C Siegel L Kim K Association of COVID-19 misinformation with face mask wearing and social distancing in a nationally representative US sample Health Communication. 2021 36 1 6 14 10.1080/10410236.2020.1847437 33225745
41. Imhoff R Lamberty P A bioweapon or a hoax? The link between distinct conspiracy beliefs about the Coronavirus disease (COVID-19) outbreak and pandemic behavior Social Psychological and Personality Science. 2020 11 8 1110 1118 10.1177/1948550620934692
42. Loomba S de Figueiredo A Piatek SJ de Graaf K Larson HJ Measuring the impact of COVID-19 vaccine misinformation on vaccination intent in the UK and USA Nature Human Behaviour 2021 5 3 337 348 10.1038/s41562-021-01056-1
43. Romer D Jamieson KH Conspiracy theories as barriers to controlling the spread of COVID-19 in the US Social Science and Medicine. 2020 263 113356 10.1016/j.socscimed.2020.113356 32967786
44. Teovanović P Irrational beliefs differentially predict adherence to guidelines and pseudoscientific practices during the COVID-19 pandemic Applied Cognitive Psychology 2021 35 486 496 10.1002/acp.3770 33362344
45. Bertin P Conspiracy beliefs, rejection of vaccination, and support for hydroxychloroquine: A conceptual replication-extension in the COVID-19 pandemic context Frontiers in Psychology 2020 11 565128 10.3389/fpsyg.2020.565128 33071892
46. Chen Y Impact of national culture on the severity of the COVID-19 pandemic Current Psychology 2022 10.1007/s12144-022-02906-5
47. Shirish A Srivastava SC Chandra S Impact of mobile connectivity and freedom on fake news propensity during the COVID-19 pandemic: A cross-country empirical examination European Journal of Information Systems. 2021 30 3 322 341 10.1080/0960085X.2021.1886614
48. Hammes LS Rossi AP Pedrotti LG Pitrez PM Mutlaq MP Rosa RG Is the press properly presenting the epidemiological data on COVID-19? An analysis of newspapers from 25 countries Journal of Public Health Policy 2021 42 3 359 372 10.1057/s41271-021-00298-7 34341478
49. Al-Zaman, M. S. (2021). Prevalence and source analysis of COVID-19 misinformation of 138 countries [Internet]. medRxiv. [cited 2022 Feb 25]. p. 2021.05.08.21256879. 10.1101/2021.05.08.21256879v1.
50. Cha M Cha C Singh K Prevalence of misinformation and factchecks on the COVID-19 pandemic in 35 countries: observational infodemiology study JMIR Human Factors. 2016 8 1 e23279 10.2196/23279
51. De Coninck, D., Frissen, T., Matthijs, K., d’Haenens, L., Lits, G., Champagne-Poirier, O., et al. (2021). Beliefs in conspiracy theories and misinformation about COVID-19: Comparative perspectives on the role of anxiety, depression and exposure to and trust in information sources. Frontiers in Psychology [Internet]. [cited 2022 Feb 26];12. 10.3389/fpsyg.2021.646394.
52. Sharma, K., Seo, S., Meng, C., Rambhatla, S., Liu, Y. (2020). COVID-19 on social media: Analyzing misinformation in Twitter conversations. arXiv:200312309 [cs] [Internet]. [cited 2022 Apr 23]. http://arxiv.org/abs/2003.12309.
53. Xaudiera, S., Cardenal, A. S. (2020). Ibuprofen narratives in five European countries during the COVID-19 pandemic. Harvard Kennedy School Misinformation Review [Internet]. [cited 2022 Apr 23];1(3).
54. Statista. Most popular social networks worldwide as of October 2021, ranked by number of active users. [Internet] [cited 2022 Apr 23]. https://lb-aps-frontend.statista.com/statistics/272014/global-social-networks-ranked-by-number-of-users/.
55. Apuke OD Omar B Fake news and COVID-19: Modelling the predictors of fake news sharing among social media users Telematics and Informatics 2020 56 101475 10.1016/j.tele.2020.101475 34887612
56. Shehata A Eldakar M An exploration of Egyptian Facebook users’ perceptions and behavior of COVID-19 misinformation Science and Technology Libraries 2021 40 4 390 415 10.1080/0194262X.2021.1925203
57. Champion VL Glanz K The health belief model Health behavior and health education: Theory, research, and practice 2008 San Francisco Jossey-Bass 45 65
58. Pranesh, R. R., Farokhnejad, M., Shekhar, A., Vargas-Solar, G. (2021). Looking for COVID-19 misinformation in multilingual social media texts arXiv:2105.03313 [Internet]. [cited 2022 Apr 23]; https://arxiv.org/abs/2105.03313.
59. Caliskan, C. (2021). How does “A Bit of Everything American” state feel about COVID-19? A quantitative Twitter analysis of the pandemic in Ohio. Journal of Computational Social Science [Internet]. [cited 2022 Apr 23]; 10.1007/s42001-021-00111-1.
60. Himelboim I McCreery S Smith M Birds of a feather tweet together: Integrating network and content analyses to examine cross-ideology exposure on Twitter Journal of Computer-Mediated Communication 2013 18 2 154 174 10.1111/jcc4.12001
61. Kušen E Strembeck M Politics, sentiments, and misinformation: An analysis of the Twitter discussion on the 2016 Austrian presidential elections Online Social Networks and Media 2018 5 37 50 10.1016/j.osnem.2017.12.002
62. Lai, M., Bosco, C., Patti, V., Virone, D. (2015) Debate on political reforms in Twitter: a hashtag-driven analysis of political polarization. In 2015 Ieee international conference on data science and advanced analytics (Dsaa) (pp. 1–9). IEEE.
63. Lansdall-Welfare, T., Dzogang, F., Cristianini, N. (2016). Change-point analysis of the public mood in UK Twitter during the Brexit Referendum. In 2016 IEEE 16th international conference on data mining workshops (ICDMW) (pp. 434–439). IEEE.
64. Sharma S Black Twitter? Racial hashtags, networks and contagion New Formations 2013 78 78 46 64 10.3898/NewF.78.02.2013
65. World Health Organization. WHO Coronavirus (COVID-19) Dashboard [Internet]. [cited 2022 Apr 23]. https://covid19.who.int. Accessed 13 Nov 2022.
66. CNN BMM Hayes, M., Alfonso III, F., Rocha, V. (2020). US coronavirus news [Internet]. CNN. [cited 2022 Apr 23]. https://www.cnn.com/us/live-news/us-coronavirus-update-05-27-20/index.html. Accessed 13 Nov 2022.
67. Furlong, H. (2020). WHO ends hydroxychloroquine study. Politico. [Internet]. [cited 2022 Apr 23]. https://www.politico.com/news/2020/06/17/who-ends-hydroxychloroquine-study-326238. Accessed 13 Nov 2022.
68. WHO officials are reviewing new evidence of airborne transmission, importance of ventilation in fighting coronavirus [Internet]. CNBC. 2020 [cited 2022 Apr 23]. https://www.cnbc.com/2020/07/07/who-officials-are-reviewing-new-evidence-of-airborne-transmission-importance-of-ventilation-in-fighting-coronavirus.html. Accessed 13 Nov 2022.
69. Freedom House. Freedom in the World: 2019 Scores [Internet]. [cited 2022 Apr 23]. https://freedomhouse.org/report/freedom-world/2019/scores. Accessed 13 Nov 2022.
70. Freedom House. Freedom on the Net [Internet]. [cited 2022 Apr 23]. https://freedomhouse.org/report/freedom-net. Accessed 13 Nov 2022.
71. Edelman. 2019 Edelman Trust Barometer [Internet]. [cited 2022 Apr 23]. https://www.edelman.com/trust/2019-trust-barometer. Accessed 13 Nov 2022.
72. Transparency International. 2020 Corruption Perceptions Index [Internet]. [cited 2022 Apr 23]. https://www.transparency.org/en/cpi/2020. Accessed 13 Nov 2022.
73. Reporters Without Borders. (2020). 2020 World Press Freedom Index: “Entering a decisive decade for journalism, exacerbated by coronavirus” [Internet]. [cited 2022 Apr 23]. https://rsf.org/en/2020-world-press-freedom-index-entering-decisive-decade-journalism-exacerbated-coronavirus. Accessed 13 Nov 2022.
74. Newman, N., Fletcher, R., Kalogeropoulos, A., Nielsen, R. K. (2019). Reuters Institute Digital News Report 2019 [Internet]. Rochester, NY: Social Science Research Network. [cited 2022 Apr 23]. Report No.: ID 3414941. https://papers.ssrn.com/abstract=3414941. Accessed 13 Nov 2022.
75. Wellcome Trust. Trust in science and health professionals | Wellcome Global Monitor 2018 [cited 2022 Apr 23] [Internet]. Wellcome. https://wellcome.org/reports/wellcome-global-monitor/2018/chapter-3-trust-science-and-health-professionals. Accessed 13 Nov 2022.
76. Curini, L. (2019). Can the market stop populism? [Internet]. IREF Europe EN. [cited 2022 Apr 23]. https://en.irefeurope.org/publications/online-articles/article/can-the-market-stop-populism/. Accessed 13 Nov 2022.
77. Haerpfer, C., Inglehart, R., Moreno, A., Welzel, C., Kizilova, K., Diez-Medrano, J., et al. (2021). World values survey time-series (1981–2020) Cross-National Data-Set [Internet]. World Values Survey Association. [cited 2022 Apr 23]. https://www.worldvaluessurvey.org/WVSEVStrend.jsp. Accessed 13 Nov 2022.
78. Witt M Mapping the business systems of 61 major economies: a taxonomy and implications for varieties of capitalism and business systems research Socio-Economic Review 2018 16 5 38 10.1093/ser/mwx012
79. United Nations Development Programme. Latest Human Development Index Ranking. [cited 2022 Apr 23]. https://hdr.undp.org/en/content/latest-human-development-index-ranking. Accessed 13 Nov 2022.
80. World Bank. Gini index (World Bank estimate) | Data. [cited 2022 Apr 23]. https://data.worldbank.org/indicator/SI.POV.GINI. Accessed 13 Nov 2022.
81. Lozano R Fullman N Mumford JE Knight M Barthelemy CM Abbafati C Measuring universal health coverage based on an index of effective coverage of health services in 204 countries and territories, 1990–2019: A systematic analysis for the global burden of disease study 2019 The Lancet. 2020 396 10258 1250 1284 10.1016/S0140-6736(20)30750-9
82. United Nations Statistics Division. Methodology: Standard country or area codes for statistical use. [Internet] [cited 2022 Apr 23]. https://unstats.un.org/unsd/methodology/m49/. Accessed 13 Nov 2022.
83. Ritchie, H., Mathieu, E., Rodés-Guirao, L., Appel, C., Giattino, C., Ortiz-Ospina, E., et al. (2020). Coronavirus pandemic (COVID-19). Our world in data [Internet]. [cited 2022 Apr 23]. https://ourworldindata.org/coronavirus. Accessed 13 Nov 2022.
84. Blei DM Ng AY Jordan MI Latent dirichlet allocation Journal of Machine Learning Research. 2003 3 993 1022
85. Shahnaz F Berry MW Pauca VP Plemmons RJ Document clustering using nonnegative matrix factorization Information Processing and Management 2006 42 2 373 386 10.1016/j.ipm.2004.11.005
86. Sauvy A Trois Mondes, Une Planète L’Observateur. 1952 118 14
87. Huntington, S.P. (2000) The clash of civilizations? In Culture and politics (pp. 99–118). Springer.
88. Wallerstein I The capitalist world-economy 1979 Cambridge University Press
89. Vassilvitskii, S., & David, A. (2006). K-Means++: The Advantages of Careful Seeding. In Proceedings of the Eighteenth Annual ACM-SIAM Symposium on Discrete Algorithms (pp. 1027–1035).
90. van Buuren S Groothuis-Oudshoorn K Mice: Multivariate imputation by chained equations in R Journal of Statistical Software. 2011 12 45 1 67
91 Ahn H Sun K Kim K Comparison of missing data imputation methods in time series forecasting Computers, Materials and Continua 2021 70 1 767 779 10.3260/cmc.2022.019369
92. Daberdaku S Tavazzi E Di Camillo B A Combined interpolation and weighted K-nearest neighbours approach for the imputation of longitudinal ICU laboratory data Journal of Healthcare Informatics Research 2020 4 2 174 188 10.1007/s41666-020-00069-1 35415441
93. Sun, B., Ma, L., Cheng, W., Wen, W., Goswami, P., Bai, G. 2017. An improved K-nearest neighbours method for traffic time series imputation. In 2017 Chinese Automation Congress (CAC) (pp. 7346–7351). 10.1109/CAC.2017.8244105.
94. Troyanskaya O Cantor M Sherlock G Brown P Hastie T Tibshirani R Botstein D Altman RB Missing value estimation methods for DNA microarrays Bioinformatics 2001 17 6 520 525 10.1093/bioinformatics/17.6.520 11395428
95. Van der Maaten L Hinton G Visualizing data using t-SNE Journal of Machine Learning Research 2008 9 2579 2605
96. Belkin M Niyogi P Laplacian eigenmaps for dimensionality reduction and data representation Neural Computation. 2003 15 6 1373 1396 10.1162/089976603321780317
97. Kusner, M. J., Sun, Y., Kolkin, N. I., Weinberger, K. Q. (2015). From word embeddings to document distances. In Proceedings of the 32nd international conference on international conference on machine learning—volume 37 (ICML’15) (pp. 957–966). JMLR.org.
98. Cer, D., Yang, Y., Kong, S., Hua, N., Limtiaco, N., John, R. S., et al. (2018). Universal sentence encoder for English. In Proceedings of the 2018 conference on empirical methods in natural language processing: system demonstrations (pp. 169–174).
99. Devlin, J., Chang, M.-W., Lee, K., Toutanova, K. (2018). Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:181004805.
100. Mihalcea, R., Corley, C., Strapparava, C. (2006). Corpus-based and knowledge-based measures of text semantic similarity. In Proceedings of the 21st national conference on artificial intelligence—volume 1 (AAAI’06) (pp. 775–780). AAAI Press.
101. Bun KK Ishizuka M Emerging topic tracking system in WWW Knowledge-Based Systems. 2006 19 3 164 171 10.1016/j.knosys.2005.11.008
102. Sankowski P Improved distance queries and cycle counting by frobenius normal form Theory of Computing Systems 2019 63 1049 1067 10.1007/s00224-018-9894-x 31396014
103. Majtey, A. P., Borras, A., Casas, M., Lamberti, P. W., Plastino, A. (2008). Jensen Shannon divergence as a measure of the degree of entanglement. arXiv:08043662 [quant-ph] [Internet]. [cited 2022 Apr 23]. http://arxiv.org/abs/0804.3662.
104. Anderson MJ A new method for non-parametric multivariate analysis of variance Austral Ecology. 2001 26 1 32 46
105. Clarke KR Non-parametric multivariate analyses of changes in community structure Australian Journal of Ecology. 1993 18 1 117 143 10.1111/j.1442-9993.1993.tb00438.x
106. Bandeira L Aleksejeva N Knight T Le Roux J Weaponized: How rumors about Covid-19’s origins led to a narrative arms race 2021 Atlantic Council
107. Jamison AM Broniatowski DA Dredze M Wood-Doughty Z Khan D Quinn SC Vaccine-related advertising in the facebook ad archive Vaccine. 2020 38 3 512 520 10.1016/j.vaccine.2019.10.066 31732327
108. Cooper CL Culture’s consequences: international differences in work related values, Geert Hofstede, Sage Publications, London and Beverly Hils, 1980. No. of Pages: 475. Price £18.75 Journal of Organizational Behavior 1982 3 202 204 10.1002/job.4030030208
109. Schwartz, S. (1994). Beyond individualism/collectivism: New cultural dimensions of values. In: Cross-cultural research and methodology (pp. 85–119).
110. Mättö M Niskanen M Religion, national culture and cross-country differences in the use of trade credit: Evidence from European SMEs International Journal of Managerial Finance. 2019 15 3 350 370 10.1108/IJMF-06-2018-0172
111. Jung HJ Oh KW Kim HM Country differences in determinants of behavioral intention towards sustainable apparel products Sustainability. 2021 13 2 558 10.3390/su13020558
112. Hong W Liu R-D Ding Y Hwang J Wang J Yang Y Cross-country differences in stay-at-home behaviors during peaks in the covid-19 pandemic in China and the United States: The roles of health beliefs and behavioral intention International Journal of Environmental Research and Public Health. 2021 18 4 2104 10.3390/ijerph18042104 33670078
113. Nguyen A Tran M Science journalism for development in the global south: A systematic literature review of issues and challenges Public Understanding of Science 2019 28 8 973 990 10.1177/0963662519875447 31533528
114. Hase V Mahl D Schäfer MS Keller TR Climate change in news media across the globe: An automated analysis of issue attention and themes in climate change coverage in 10 countries (2006–2018) Global Environmental Change. 2021 70 102353 10.1016/j.gloenvcha.2021.102353
115. McNeil, R., Karstens, E. (2018). Comparative report on cross-country media practices, migration, and mobility [Internet]. The Reminder Project. [cited 2022 Apr 23]. https://www.reminder-project.eu/publications/reports/comparative-report-on-cross-country-media-practices-migration-and-mobility/. Accessed 13 Nov 2022.
116. World Health Organization Regional Office for Africa. (2021). On the frontlines in the fight against dangerous misinformation [Internet]. [cited 2022 Apr 23]. https://www.afro.who.int/news/frontlines-fight-against-dangerous-misinformation. Accessed 13 Nov 2022.
117. Okereke M Ukor NA Ngaruiya LM Mwansa C Alhaj SM Ogunkola IO COVID-19 misinformation and infodemic in Rural Africa American Journal of Tropical Medicine and Hygiene 2021 104 2 453 456 10.4269/ajtmh.20-1488
118. Larson, H., Tajudeen, R. (2022). Vaccinating Africa against COVID-19: riding a roller coaster of poor information [Internet]. The conversation. [cited 2022 Apr 23]. http://theconversation.com/vaccinating-africa-against-covid-19-riding-a-roller-coaster-of-poor-information-159716. Accessed 13 Nov 2022.
119. Menezes, N. P., Simuzingili, M., Debebe, Z. Y., Pivodic, F., Massiah, E. (2021). What is driving COVID-19 vaccine hesitancy in Sub-Saharan Africa?. World Bank Blogs. [Internet]. [cited 2022 Apr 23]. https://blogs.worldbank.org/africacan/what-driving-covid-19-vaccine-hesitancy-sub-saharan-africa. Accessed 13 Nov 2022.
120. Farahat, M. (2021). Coronavirus trials in Egypt: Blurring the lines between fake news and freedom of expression [Internet]. SMEX. [cited 2022 Apr 23]. https://smex.org/coronavirus-trials-in-egypt-blurring-the-lines-between-fake-news-and-freedom-of-expression/. Accessed 13 Nov 2022.
121. Lipps J Schraff D Regional inequality and institutional trust in Europe European Journal of Political Research. 2021 60 4 892 913 10.1111/1475-6765.12430
122. Habibov N Afandi E Cheung A Sand or grease? Corruption-institutional trust nexus in post-Soviet countries Journal of Eurasian Studies. 2017 8 2 172 184 10.1016/j.euras.2017.05.001
123. Rodríguez AB Social networks: A source of lexical innovation and creativity in contemporary peninsular Spanish Languages. 2021 6 3 138 10.3390/languages6030138
124. Herring, S. C. (2012). Grammar and electronic communication. The encyclopedia of applied linguistics. (pp. 1–9).
125. Asr, F. T. (n.d). The language gives it away: How an algorithm can help us detect fake news. The conversation. http://theconversation.com/the-language-gives-it-away-how-an-algorithm-can-help-us-detect-fake-news-120199. Accessed 2 Dec 2022.
126. Rashkin, H., Choi, E., Jang, J., Volkova, S., Choi, Y. (2017). Truth of varying shades: Analyzing language in fake news and political fact-checking. (pp. 2931–2937).
127. Vosoughi S Roy D Aral S The spread of true and false news online Science 2018 8 1 1146 1151 10.1126/science.aap9559
128. Center for Countering Digital Hate. The Disinformation Dozen [Internet]. [cited 2022 Apr 23]. https://www.counterhate.com/disinformationdozen. Accessed 13 Nov 2022.
129. Robertson R European glocalization in global context 2014 Palgrave Macmillan
130. Roudometof V Glocalization: A critical introduction 2016 Routledge & CRC Press
131. Caliskan C Derindere-Koseoglu S Network modeling: Historical perspectives, agent-based modeling, correlation networks, and network similarities Financial data analytics: Theory and application 2022 Springer
| 0 | PMC9746594 | NO-CC CODE | 2022-12-15 23:21:57 | no | J Comput Soc Sci. 2022 Dec 13;:1-53 | utf-8 | J Comput Soc Sci | 2,022 | 10.1007/s42001-022-00193-5 | oa_other |
==== Front
Z Gesundh Wiss
Z Gesundh Wiss
Zeitschrift Fur Gesundheitswissenschaften
2198-1833
1613-2238
Springer Berlin Heidelberg Berlin/Heidelberg
1791
10.1007/s10389-022-01791-3
Original Article
Determining attitudes toward e-learning: what are the attitudes of health professional students?
https://orcid.org/0000-0003-1241-9370
Güllü Ayla [email protected]
1
https://orcid.org/0000-0002-1655-7132
Kara Mustafa [email protected]
2
https://orcid.org/0000-0002-6604-4343
Akgün Şenay [email protected]
3
1 grid.14352.31 0000 0001 0680 7823 Faculty of Health Sciences, Department of Nursing, Hatay Mustafa Kemal University, Hatay, Turkey
2 grid.411741.6 0000 0004 0574 2441 Afşin Health School Department of Nursing, Kahramanmaraş Sütçü İmam Unıversıty, Kahramanmaraş, Turkey
3 Faculty of Health Sciences, Department of Nursing, Alanya Alaaddin Keykubat University, Antalya, Turkey
13 12 2022
18
15 9 2022
30 11 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
Aim
The research was carried out to determine the attitudes of nursing undergraduate students toward e-learning implemented during the COVID-19 pandemic.
Subject and methods
The study sample consisted of 320 undergraduate students studying at the Faculty of Health Sciences, Nursing Department of a state university. Personal Information Form and the Test of e-Learning Related Attitudes were administered to the research participants.
Results
Of the students, 68.8% were female and 51.6% were between the ages of 21–24. The attitude of 55.3% (177) toward e-learning is negative. Attitude scores of students aged 25–29 were significantly higher compared to those aged 17–20 and 21–24 years old (p = 0.002). The attitude scores of the students who have a computer are significantly higher than those who do not (p = 0.001). Most students did not agree with the statement “E-learning will provide me with better learning opportunities than conventional learning methods.” (disagree n = 121; 37.8%, strongly disagree n = 110; 34.4%).
Conclusion
It is noticed that e-learning is not sufficient in subjects such as conducting clinical and laboratory practices in health sciences fields with practical training like nursing, and students’ attitudes are generally negative. For undergraduate health sciences education, face-to-face and online education for support purposes can be used together in theoretical courses. In addition, it is recommended to use effective online communication techniques in online courses.
Keywords
Attitude
E-learning
Students
Technology
==== Body
pmcIntroduction
Coronavirus disease (COVID-19) was first reported in December 2019 in the city of Wuhan, Hubei Province, China. Later, on March 11, 2020, the World Health Organization declared the disease a pandemic (WHO 2020). The fact that this virus spread rapidly and became a threat to the whole world and impacted the global economy and countries brought about different measures and practices. Following the health sector, the education sector has been one of the sectors most affected by this situation (Telli and Altun 2020). The COVID-19 pandemic has impacted the education system worldwide and brought radical organizational changes to conventional teaching techniques. Many countries have resorted to e-courses and exams for medical/nursing students, and e-learning has been used as the primary method of teaching the curriculum (Singh et al. 2021).
The term e-learning is a relatively new term that emerged with the development and advancement of information and communication technologies. E-learning is considered a broad concept that describes the asynchronous or simultaneous transfer of knowledge to learners through electronic systems. The historical origin of the term e-learning is not exactly known. However, it started to develop with the use of the internet and personal computers (Sweileh 2021; Wang et al. 2021).
The success of e-learning depends on various factors such as accessibility, use of appropriate methods, course content, and assessment criteria. Like any teaching method, e-learning has pros and cons for both students and teachers. Problems related to internet access, lack of digital skills, social isolation, and lack of student–teacher interaction make up its limitations (Abbasi et al. 2020a, b; Bączek et al. 2021). In general terms, the reasons for its acceptability include its advantages such as applicability, ease of use, and flexibility (Abbasi et al. 2020a, b).
Health science education is a key element in the sustainability of healthcare services in any country. There is a growing interest in e-learning among health educators (Ruiz et al. 2006). The American Association of Critical-Care Nurses (AACN) reported that the courses and programs delivered through distance learning in nursing education are increasing gradually and it is indispensable to set some standards to achieve higher quality nursing education (AACN 2003). In addition, even if all the necessary criteria are met in the development of distance learning environments, the learner’s attitudes and approaches toward these environments are monumental (Usta et al. 2016). The data obtained from this research will enable the assessment of nursing students’ attitudes toward e-learning and will guide the training to be planned.
Methods
Sample selection
As in the whole world, Turkey had to apply distance education in higher education due to the COVID-19 pandemic. On March 23, 2020, education started in the digital environment in higher education institutions in Turkey. At the university where the research was carried out during the pandemic period, the courses were processed synchronously or asynchronously in the form of taking videos with the Camtasia program and uploading them to the system or defining online courses. Face-to-face education started in the 2021–2022 academic year.
The research was designed as a cross-sectional study. The population of the research consisted of 549 students studying at the Nursing Department of the Faculty of Health Sciences in the 2020–2021 academic year at a state university located in the Mediterranean region of Turkey. No sample selection was made, so it was aimed to reach all nursing students who agreed to participate in the study. Data were collected through an online survey between May–August, 2021. A Google Form containing the study questionnaire was sent to WhatsApp groups consisting of nursing students through class representatives to ensure proper selection of study participants. In the research, 1st, 2nd, 3rd, and 4th grade representatives were contacted and these students were provided the questionnaire to deliver to other participants. A friendly reminder was sent to the participants so that the survey could be answered by eligible participants. The research was completed with 320 participants.
Data collection tools
Two types of forms were used in this study. These are “Personal Information Form” and “Test of e-Learning Related Attitudes”.
Personal information form
The personal information form consisted of eight questions to identify some characteristics of individuals such as age, gender, income level, internet access, and the presence of their computer.
Test of e-learning related attitudes
A Test of e-Learning Related Attitudes (TeLRA) scale was developed by Kisanga (2016) to identify attitudes toward e-learning; the scale is a four-point Likert type and is rated as 1-strongly disagree, 2-disagree, 3-agree and 4-strongly agree. The adaptation of the scale to Turkish culture and language was made by Biçer (2019). The reliability of this scale, which includes 23 items and four factors, adapted to identify attitudes toward e-learning, is α = 0.789. Within this regard, the factors have been termed as “tendency to use technology,” “satisfaction,” “motivation,” and “usefulness”. The scale includes some reverse worded items (Biçer and Korucu 2020; Kisanga 2016). In this study, the Cronbach’s Alpha value was 0.896 (Table 1).Table 1 Descriptive statistics for TeLRA and subscales
Variable Mean Std.
Deviation Minimum–maximum Skewness Kurtosis Cronbach’s Alpha
TeLRA total score (23 items) 56.74 12.52 28–92 .047 −.111 0.896
TeLRA mean score 2.45 0.5 1–4 .047 −.111
Tendency to use technology (6 items) 15.50 4.1 6–24 −.100 −.373 0.768
Satisfaction (5 items) 13.24 3.5 5–20 −.199 −.466 0.743
Motivation (6 items) 13.70 4.3 6–24 .238 −.292 0.847
Usefulness (6 items) 13.30 4.1 6–24 −.020 −.643 0.777
Scoring: 1–4-point Likert-type scale from strongly disagree to strongly agree
Statistical analysis of data
SPSS (IBM SPSS for Windows, ver.26) statistical package software was used for the analysis. Measurement data were expressed as numbers, percent (%), and mean ± standard deviation. As the data showed normal distribution (Table 1), the t-test was used for the analysis of two variables, and one way ANOVA was used for the analysis of more than two variables. Post-hoc Bonferroni test was conducted to determine between-groups-variance for three or more groups. The results were evaluated at the 95% confidence level, at the p < 0.05 level.
Ethical aspect of the study
Ethics Committee approval was obtained from the Non-Invasive Clinical Research Ethics Committee for the study. (Date: October 30th, 2019 Session: 2019/20 Decision No: 04). Consent of the participants was obtained through Google Forms, after reading the explanation text of the research before starting the survey and clicking the “Yes” button to the question that they participated in the study voluntarily.
Results
Of the nursing students participating in the study, 68.8% were female and 51.6% were between the ages of 21–24; 34.7% of the students were in the 3rd grade and 6.9% of them live in the dormitory; 52.2% of the students have a computer and 86.9% have internet access; 78.8% of the students stated their income level as middle income; 70.3% of nursing students considered themselves adequate in terms of technological knowledge (Table 2).Table 2 The analysis results of some demographical variances of nursing students and their attitude toward E-learning
Characteristics N (%) Mean t/F p value Bonferroni test
Age (n = 320)
17-20 years (1) 140 (43.8) 55.25 ± 12.4 3 > 1
21-24 years (2) 165 (51.6) 57.10 ± 12.1 3 > 2
25-29 years (3) 15 (4.7) 67.28 ± 13.8 6.216 a0.002
Gender (n = 320)
Female 220 (68.8) 56.76 ± 12.8
Male 100 (31.3) 56.67 ± 11.8 0.065 b0.948
The class of study (n = 320)
1st 105 (32.8) 56.25 ± 13.1
2nd 48 (15.0) 53.64 ± 15.7
3rd 111 (34.7) 58.19 ± 10.3
4th 56 (17.5) 57.39 ± 12.1 1.589 a0.192
With whom does he/she live (n = 320)
Family 283 (88.4) 56.57 ± 12.5
Dormitory 22 (6.9) 55.63 ± 10.3
With friends at a student house 15 (4.7) 61.33 ± 15.3 1.118 a0.328
Ownership of a computer (n = 320)
Yes 167 (52.2) 59.22 ± 12.4
No 153 (47.8) 54.02 ± 12.1 3.782 b0.001
Do you have an internet connection? (n = 320)
Yes 278 (86.9) 57.13 ± 12.7
No 42 (13.1) 54.11 ± 10.9 1.456 b0.146
How do you define your economic condition?(n = 320)
Bad 44 (13.8) 54.70 ± 14.3
Medium 252 (78.8) 56.62 ± 11.8
Good 24 (7.5) 61.66 ± 15.4 2.470 a0.086
How do you define your technological know-how? (n = 320)
Sufficient 225 (70.3) 57.57 ± 11.7
Insufficient 54 (16.9) 53.24 ± 12.5
Very insufficient 41 (12.8) 56.75 ± 15.7 2.631 a0.074
a ANOVA, b t test 1,2,3, post hoc bonferroni. The percentages are calculated over N
When the analysis results of some demographic variables of nursing students and their general attitude scores toward E-learning are examined, it was found that students aged 25–29 obtained higher scores for E-learning compared to those aged 17–20 and 21–24 years old (F = 6.216, p = 0.002). The general attitude scores of the students who have a computer are significantly higher compared to those who do not (t = 3.782, p = 0.001). However, no significant correlation was found between gender (t = 0.065, p = 0.948), grade (F = 1.589, p = 0.192), persons the student lives with (F = 1.118, p = 0.328), economic status (F = 2.470, p = 0.086), availability of internet connection (t = 1.456, p = 0.146), evaluation of technological knowledge (F = 2.631, p = 0.074), and general attitude scores (Table 2).
The distribution of students’ attitudes toward e-learning is presented in Fig. 1. The analysis revealed that 55.3% (n = 177) of the students had a negative attitude toward e-learning, and 44.7% (n = 143) had a positive attitude toward e-learning (Fig. 1).Fig. 1 Attitudes toward e-learning
The results of the analysis of the sub-dimension scores of the scale according to descriptive information are given in Table 3. Students aged 25–29 scored significantly higher in the sub-dimensions of usefulness (p = 0.000) and motivation (p = 0.007) compared to those aged 17–20. Students with a computer scored significantly higher than students without a computer in all of the sub-dimensions of using technology (p = 0.001), satisfaction (p = 0.034), motivation (p = 0.002), and usefulness (p = 0.047). The tendency to use the technology sub-dimension scale score of students with internet connections were found to be significantly higher than those without internet connections (p = 0.025). The students with sufficient technological knowledge had a higher tendency to use technology sub-dimension score than those with insufficient technological knowledge (p = 0.013) (Table 3).Table 3 Comparison of scale sub-dimension scores according to descriptive information
Variables Tendency to use technology Satısfactıon Motıvatıon Usefulness
Mean p value Mean p value Mean p value Mean p value
Gender
Female 15.46 13.21 13.61 14.48
Male 15.59 b0.794 13.29 b0.857 13.89 b0.596 13.90 b0.235
Age
17-20(1) 15.32 12.84 13.12 13.96
21-24(2) 15.45 13.50 13.92 14.22
25-29(3) 17.86 a0.091 14.07 a0.175 16.78 a0.007
3 > 1
18.57 a0.000
3 > 1,3 > 2
The class of study
1st 15.60 12.96 13.35 14.34
2nd 15.13 12.21 12.81 13.50
3rd 15.74 13.84 14.18 14.43
4th 15.16 a0.756 13.45 a0.041 14.13 a0.193 14.64 a0.497
With whom does he/she live
Family 15.50 13.19 13.68 14.21
Dormitory 14.95 13.23 13.13 14.32
With friends at a student house 16.27 a0.642 14.20 a0.554 14.86 a0.479 16.00 a0.250
Ownership of a computer
Yes 13.63 16.46 14.40 14.73
No 12.80 b0.001 14.46 b0.034 12.93 b0.002 13.83 b0.047
Do you have an internet connection?
Yes 15.70 13.32 13.72 14.38
No 14.17 b0.025 12.67 b0.259 13.52 b0.777 13.76 b0.357
How do you define your economic condition?
Bad 15.57 11.95 14.02 14.16
Medium 15.56 13.44 13.42 14.20
Good 16.58 a0.142 13.50 a0.032 15.95 a0.020 15.63 a0.251
How do you define your technological know-how?
Sufficient(1) 15.88 13.50 13.77 14.42
Insufficient(2) 14.09 12.94 12.74 13.46
Very insufficient(3) 15.27 a0.013
1 > 2
12.17 a0.065 14.56 a0.113 14.76 a0.223
a ANOVA, b t test 1,2,3, post hoc bonferroni
Table 4 shows students’ attitudes toward e-learning statements. One hundred twenty-one (37.8%) students disagreed and 110 (34.4%) strongly disagreed with the statement “E-learning will provide me with better learning opportunities than traditional learning methods” in item 15. This item received the lowest score compared to other items on the scale (Table 4).Table 4 Attitudes of study participants toward statement for e-learning
1(%) 2(%) 3(%) 4(%) mean ± SD
Tendency to use technology
1. I make errors frequently when using a Computer. 48(15.0) 68(21.3) 126(39.4) 78(24.4) 2.73 ± 0.9
2. It will be difficult for me to become skilful in the use of e-learning tools. 36(11.3) 63(19.7) 136(42.5) 85(26.6) 2.84 ± 0.9
3. Using a computer at home is very frustrating. 55(17.2) 68(21.3) 99(30.9) 98(30.6) 2.75 ± 1.0
4. I find computer online interaction unexciting. 90(28.1) 97(30.3) 80(25.0) 53(16.6) 2.30 ± 1.0
5. Communicating through electronic mails is annoying. 105(32.8) 90(28.1) 80(25.0) 45(14.1) 2.20 ± 1.0
6. E-learning infrastructure is very expensive for the government to afford. 41(12.8) 100(31.3) 102(31.9) 77(24.1) 2.67 ± 0.9
Satısfactıon
7. E-learning is very economical for educational institutions to adopt. 43(13.4) 95(29.7) 125(39.1) 57(17.8) 2.61 ± 0.9
8. I believe using e-learning will improve the quality of my work. 76(23.8) 104(32.5) 85(26.6) 55(17.2) 2.37 ± 1.0
9. Computers make work more interesting. 75(23.4) 104(32.5) 83(25.9) 58(18.1) 2.38 ± 1.0
10. It is easier to revise electronic educational materials than printed material. 43(13.4) 63(19.7) 104(32.5) 110(34.4) 2.88 ± 1.0
11. I prefer using a computer to prepare my lessons. 30(9.4) 57(17.8) 118(36.9) 115(35.9) 2.99 ± 0.9
Motıvatıon
12. Working with computers is exciting. 87(27.2) 114(35.6) 82(25.6) 37(11.6) 2.22 ± 0.9
13. My institution has enough teaching-learning resources to carry out e-learning. 107(33.4) 107(33.4) 68(21.3) 38(11.9) 2.12 ± 1.0
14. I like discussing about new e-learning innovations. 68(21.3) 112(35.0) 100(31.3) 40(12.5) 2.35 ± 0.9
15. E-learning will provide me with better learning opportunities than traditional means of learning. 110(34.4) 121(37.8) 62(19.4) 27(8.4) 2.02 ± 0.9
16. Using e-learning technologies will allow me to accomplish more work than would otherwise be possible. 38(11.9) 99(30.9) 136(42.5) 47(14.7) 2.60 ± 0.8
17. I enjoy teaching using computers. 67(20.9) 106(33.1) 99(30.9) 48(15.0) 2.40 ± 0.9
Usefulness
18. E-learning reduces quality of knowledge attained. 73(22.8) 81(25.3) 115(35.9) 51(15.9) 2.45 ± 1.0
19. E-learning requires expensive technical support. 82(25.6) 97(30.3) 102(31.9) 39(12.2) 2.31 ± 0.9
20. Delivering a lecture through electronic technologies is very difficult. 59(18.4) 83(25.9) 101(31.6) 77(24.1) 2.61 ± 1.0
21. Interacting with the computer system is often frustrating. 68(21.3) 111(34.7) 100(31.3) 41(12.8) 2.36 ± 0.9
22. Discussions on e-learning technologies are uninteresting. 65(20.3) 98(30.6) 118(36.9) 39(12.2) 2.41 ± 0.9
23. Teaching through e-learning is tiresome. 93(29.1) 111(34.7) 86(26.9) 30(9.4) 2.17 ± 0.9
The percentages are calculated over N, 1- strongly disagree, 2- disagree, 3- agree, 4- strongly agree, SD, standard deviation
Discussion
Our study demonstrated that 55.3% of the nursing students included in the study had negative attitudes toward e-learning. Likewise, in the study by Diab and Elgahsh (2020), 61.6% of nursing students were found to have a negative attitude toward e-learning (Diab and Elgahsh 2020). Studies by Bączek et al. (2021) and Olum et al. (2020) with students studying health science, including medical, dentistry, and nursing students, have revealed a generally negative attitude toward e-learning (Bączek et al. 2021; Olum et al. 2020). Our study supported the literature in this regard.
In this study, it was found that students aged 25–29 had significantly higher scores on e-learning compared to those aged 17–20 and 21–24 years old. Moreover, students aged 25–29 scored significantly higher on the usefulness and motivation scale sub-dimensions compared to those aged 17–20. Contrary to our study, a study argued that e-learning imposes more limitations on older students (Ramos-Morcillo et al. 2020). It was considered that the students in the older age group were most likely in the upper grades and had experience in hospital practice before the COVID-19 pandemic, which might have affected their TeLRA scores. Medical and health science students need physical contact with patients to learn necessary skills and personally experience patient care (Abbasi et al. 2020a, b; Rose 2020). On the other hand, it has been revealed in the literature that there is a lack of preparation, skills, and training during the pandemic (Farooq et al. 2020; Ramos-Morcillo et al. 2020).
In the study, the computer ownership status of the students impacted their attitude scores. The attitude score of the students who had a computer was found to be significantly higher than those who did not have a computer. In our study, no significant correlation was found between students’ considering themselves technologically competent and the scale total score, but the tendency to use technology sub-dimension scores of students who consider themselves sufficient in terms of technological knowledge were found to be significantly higher. Furthermore, students with internet access had a significantly higher “technology usage tendency” sub-scale score. It is well-known that e-learning is interrupted due to technical problems such as computer and connection problems, even if the student is technologically competent. In the study by Sharma et al. (2021) with medical and dentistry students, most of the students stated that they had difficulties in online classes due to internet and electricity problems (Sharma et al. 2021). In the study of Abbasi et al. (2020a, b) with health sciences students, 41% of the students stated that e-learning was hindered due to network problems (Abbasi et al. 2020a, b). In fact, in many studies, it has been suggested that situations such as telecommunications infrastructure, technological difficulties, limited technical skills, technical and management support, instructor’s qualifications, power cuts, internet costs, family distraction, communication difficulties, and lack of security are the major reasons that will increase dissatisfaction with online learning (Al-Balas et al. 2020; Diab and Elgahsh 2020; Dost et al. 2020; Fawaz and Samaha 2020; Mukasa et al. 2021; Olum et al. 2020; Singh et al. 2021).
In this study, the majority of the students did not agree with the statement “E-learning will provide me with better learning opportunities than traditional learning methods.” Similar to our study, in the study of Eltaybani et al. (2021), it was stated that 72.9% of nursing students believed that traditional learning was more effective than e-learning (Eltaybani et al. 2021). In the study by Singh et al. (2021) on e-learning methods in nursing and medical education, only 20.4% of the students stated that they felt that e-learning could replace traditional teaching (Singh et al. 2021). Studies have demonstrated that students studying in health sciences have low satisfaction with e-learning (Al-Balas et al. 2020; Olum et al. 2020; Sindiani et al. 2020).
Regarding the preferred learning methods in the literature, Dost et al. (2020)’s study with medical students showed that most students preferred face-to-face teaching (Dost et al. 2020). In the studies conducted by Bhattarai et al. (2021) and Olum et al., (2020) with health science students such as medicine, nursing, and dentistry, most students preferred to combine online learning with traditional learning (Bhattarai et al. 2021; Olum et al. 2020). Moreover, in the study by Elsalem et al. (2020) with nursing, dentistry, pharmacy, and applied medical sciences students, 32% of the students stated that they experience more stress in e-exams (Elsalem et al. 2020).
E-learning provides students with time savings, flexibility, and the ability to study at their own pace and their convenience. Hence, flexibility and convenience are the primary factors behind the demand for online education (Al-Balas et al. 2020; Dost et al. 2020). On the other hand, e-learning is not effective in acquiring clinical and technical skills (Olum et al. 2020). E-learning has proven to be equivalent to or superior to conventional learning, but when used to replace conventional programs, unique challenges arise in clinical teaching and learning experiences (Tashkandi 2021). Teaching health evaluation in nursing in the digital age is a challenging process. It involves cognitive, procedural, and psychomotor learning and is a subject that fails to provide “readiness for practice” (McDonald et al. 2018). According to Sheikhaboumasoudi et al. (2018), combining traditional learning methods with e-learning methods could be an effective support in improving the clinical skills of nursing students. The key factor in designing a training program is to make learning simple and effective (Sheikhaboumasoudi et al. 2018). E-learning content should be compatible with its design, teaching pedagogy, and learning outcomes (Prosen et al. 2022). To make e-learning more meaningful for evidence-based nursing practice, educators should plan, perform, and assess appropriate online activities (Song and Park 2021).
Limitations
The research results are based on the statements of the participants. The fact that the study was single-centered and the sample size was small limits the generalization of the study.
Conclusion
E-learning, which has been applied quickly and widely due to the COVID-19 pandemic, has increased the interest of researchers in assessing the students’ attitudes. This cross-sectional study was carried out to determine the attitudes of nursing students, who continued distance learning due to the COVID-19 pandemic, toward e-learning. In our study, it was revealed that more than half of the nursing students had a negative attitude toward e-learning. Having a computer, having an internet connection, age, and technological knowledge variables were found to be associated with the attitude and affected the scores attitude toward e-learning. Nevertheless, it is seen that e-learning is not adequate in subjects such as clinical and laboratory practices in health sciences with applied education such as nursing. For undergraduate health sciences education, face-to-face and online education for support purposes can be used together in theoretical courses. In addition, it is recommended to solve the existing technical problems in online courses and to use effective online communication techniques. Further studies can be conducted to compare students’ attitudes toward e-learning in theoretical and applied courses.
Authors’ contributions
All authors contributed to the study conception and design. Material preparation, data collection and analysis were performed by AG, MK, and ŞA. The first draft of the manuscript was written by AG and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.
Funding
This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.
Declarations
Ethics approval
Ethics Committee approval was obtained from the Non-Invasive Clinical Research Ethics Committee for the study. (Date: October 30th, 2019 Session: 2019/20 Decision No: 04).
Consent to participate
Each study participant provided written informed consent. Privacy and confidentiality of the respondents during data collection were maintained strictly.
Consent for publication
The corresponding author confrms that the manuscript has been read and approved for submission by all the named authors.
Conflict of interest
All authors declare no conflict of interest.
Publisher’s note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Abbasi MS, Ahmed N, Sajjad B et al (2020a) E-learning perception and satisfaction among health sciences students amid the COVID-19 pandemic. Work 1–8. 10.3233/wor-203308
Abbasi S Ayoob T Malik A Perceptions of students regarding E-learning during Covid-19 at a private medical college Pakistan J Med Sci 2020 36 COVID19-S4 S57 S61 10.12669/pjms.36.covid19-s4.2766
Al-Balas M Al-Balas HI Jaber HM Distance learning in clinical medical education amid COVID-19 pandemic in Jordan: current situation, challenges, and perspectives BMC Med Educ 2020 20 1 341 10.1186/s12909-020-02257-4 33008392
American Association of Colleges of Nursing (AACN), 2003 Position statements: alliance for nursing accreditation statement on distance education policies,.https://www.aacnnursing.org/News-Information/Position-Statements-White-Papers/Distance-Education. Accessed 01 Feb 2022
Bączek M Zagańczyk-Bączek M Szpringer M Students’ perception of online learning during the COVID-19 pandemic Medicine 2021 100 7 e24821 10.1097/md.0000000000024821 33607848
Bhattarai B Gupta S Dahal S Perception of online lectures among students of a medical College in Kathmandu: a descriptive cross-sectional study J Nepal Med Assoc 2021 59 235 234 238 10.31729/jnma.6276
Biçer H (2019) E-öğrenmeye yönelik tutum: Ölçek uyarlama çalışması (Yüksek lisans tezi). Necmettin Erbakan Üniversitesi Eğitim Bilimleri Enstitüsü, Konya. https://tez.yok.gov.tr/UlusalTezMerkezi/TezGoster?key=Mir2lXQK1dkmQ9Ige3PZbo10lgsYSfAHJl590YKzfzyshIqNjhnk9vnqCxSHz_Ke. Accessed 10 Oct 2019
Biçer H Korucu AT E-Öğrenmeye Yönelik Tutum Ölçeğinin Türkçeye Uyarlanmasi Eğitim Teknolojisi Kuram ve Uygulama 2020 10 1 237 256 10.17943/Etku.632178
Diab GMAEH Elgahsh NF E-learning during COVID-19 pandemic: obstacles faced nursing students and its effect on their attitudes while applying it. American journal of Nurs Sci 2020 9 4 295 309 10.11648/j.ajns.20200904.33
Dost S Hossain A Shehab M Perceptions of medical students towards online teaching during the COVID-19 pandemic: a national cross-sectional survey of 2721 UK medical students BMJ Open 2020 10 11 e042378 10.1136/bmjopen-2020-042378
Elsalem L, Al-Azzam N, Jum’ah AA et al. (2020) Stress and behavioral changes with remote E-exams during the Covid-19 pandemic: a cross-sectional study among undergraduates of medical sciences. Ann Med Surg 60:271–279. 10.1016/j.amsu.2020.10.058
Eltaybani S Abdelhalim GE Abdelgawad ME Nursing students’ and educators’ experience with e-learning during a pandemic: an online survey Nurs Forum 2021 56 4 878 888 10.1111/nuf.12634 34318491
Farooq F Rathore FA Mansoor SN Challenges of online medical education in Pakistan during COVID-19 pandemic J Coll Physicians Surg Pak 2020 30 6 67 69 10.29271/jcpsp.2020.Supp1.S67
Fawaz M Samaha A E-learning: depression, anxiety, and stress symptomatology among Lebanese university students during COVID-19 quarantine Nurs Forum 2020 56 1 52 57 10.1111/nuf.12521 33125744
Kisanga DH Determinants of teachers’ attitudes towards e-learning in Tanzanian higher learning institutions. The international review of research in open and distributed Learning 2016 17 5 109 125 10.19173/irrodl.v17i5.2720
McDonald EW Boulton JL Davis JL E-learning and nursing assessment skills and knowledge – an integrative review Nurse Educ Today 2018 66 166 174 10.1016/j.nedt.2018.03.011 29705504
Mukasa J Otim M Monaco B Nursing students’ perspectives and readiness to transition to E-learning during COVID-19 in the UAE: a cross-sectional study Adv Med Educ Pract 2021 12 1505 1512 10.2147/amep.s335578 35221744
Olum R Atulinda L Kigozi E Medical education and E-learning during COVID-19 pandemic: awareness, attitudes, preferences, and barriers among undergraduate medicine and nursing students at Makerere University, Uganda J Med Educ Curric Dev 2020 7 238212052097321 10.1177/2382120520973212
Prosen M Karnjuš I Ličen S Evaluation of E-learning experience among health and allied health professions students during the COVID-19 pandemic in Slovenia: an instrument development and validation study Int J Environ Res Public Health 2022 14 19(8) 4777 10.3390/ijerph19084777
Ramos-Morcillo AJ Leal-Costa C Moral-García JE Experiences of nursing students during the abrupt change from face-to-face to e-learning education during the first month of confinement due to COVID-19 in Spain Int J Environ Res Public Health 2020 17 15 5519 10.3390/ijerph17155519 32751660
Rose S Medical student education in the time of COVID-19 Jama 2020 323 21 2131 2132 10.1001/jama.2020.5227 32232420
Ruiz JG Mintzer MJ Leipzig RM The impact of E-learning in medical education Acad Med 2006 81 3 207 212 10.1097/00001888-200603000-00002 16501260
Sharma N Bhusal CK Subedi S Perception towards online classes during COVID-19 among MBBS and BDS students in a medical College of Nepal: a descriptive cross-sectional study J Nepal Med Assoc 2021 59 235 276 279 10.31729/jnma.5348
Sheikhaboumasoudi R Bagheri M Hosseini SA Improving nursing Students' learning outcomes in fundamentals of nursing course through combination of traditional and e-learning methods Iran J Nurs Midwifery Res 2018 23 3 217 221 10.4103/ijnmr.IJNMR_79_17 29861761
Sindiani AM Obeidat N Alshdaifat E Distance education during the COVID-19 outbreak: a cross-sectional study among medical students in north of Jordan Ann Med Surg 2020 59 186 194 10.1016/j.amsu.2020.09.036
Singh HK Joshi A Malepati RN A survey of E-learning methods in nursing and medical education during COVID-19 pandemic in India Nurse Educ Today 2021 99 104796 10.1016/j.nedt.2021.104796 33607513
Song CE Park H Active learning in E-learning programs for evidence-based nursing in academic settings: a scoping review J Continuing Educ Nurs 2021 52 9 407 412 10.3928/00220124-20210804-05
Sweileh WM Global research activity on E-learning in health sciences education: a bibliometric analysis Med Sci Educator 2021 31 2 765 775 10.1007/s40670-021-01254-6
Tashkandi E E-learning for undergraduate medical students Adv Med Educ Pract 2021 12 665 674 10.2147/amep.s314509 34163282
Telli SG Altun D Coronavirüs ve Çevrimiçi (Online) Eğitimin Önlenemeyen Yükselişi Üniversite Araştırmaları Dergisi 2020 3 1 25 34 10.32329/uad.711110
Usta İ Uysal Ö Okur MR Çevrimiçi öğrenme tutum ölçeği: Geliştirilmesi, geçerliği ve güvenirliği Uluslararası Sosyal Araştırmalar Dergisi 2016 9 43 2215 2222
Wang ZY Zhang LJ Liu YH Process evaluation of E-learning in continuing medical education: evidence from the China-Gates Foundation tuberculosis control program Infect Dis Poverty 2021 10 1 23 10.1186/s40249-021-00810-x 33750423
WHO (2020) Director-General’s opening remarks at the media briefing on COVID-19–11 March 2020. https://www.who.int/director-general/speeches/detail/who-director-general-s-opening-remarks-at-the-media-briefing-on-covid-19—11-march-2020. Accessed 18 Feb 2022
| 0 | PMC9746595 | NO-CC CODE | 2022-12-15 23:21:57 | no | Z Gesundh Wiss. 2022 Dec 13;:1-8 | utf-8 | Z Gesundh Wiss | 2,022 | 10.1007/s10389-022-01791-3 | oa_other |
==== Front
Qual Theory Dyn Syst
Qual Theory Dyn Syst
Qualitative Theory of Dynamical Systems
1575-5460
1662-3592
Springer International Publishing Cham
707
10.1007/s12346-022-00707-6
Article
Viral Infection Model with Diffusion and Distributed Delay: Finite-Dimensional Global Attractor
Rezounenko Alexander [email protected]
grid.18999.30 0000 0004 0517 6080 V.N.Karazin Kharkiv National University, Kharkiv, 61022 Ukraine
13 12 2022
2023
22 1 119 9 2022
26 11 2022
© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
We study a virus dynamics model with reaction-diffusion, logistic growth terms and a general non-linear infection rate functional response. The model has a distributed delay, including the case of state-selective delay. We construct a dynamical system in a Hilbert space and prove the existence of a finite-dimensional global attractor.
Keywords
Delay equations
Reaction-diffusion
Evolution equations
Attractor
Virus infection model
Mathematics Subject Classification
Primary 93C23
Secondary 34D45
35K57
issue-copyright-statement© Springer Nature Switzerland AG 2023
==== Body
pmcIntroduction
We are interested in qualitative properties of mathematical models of viral infections. Such models attract much attention during last years, especially after wide spread of viral diseases, including COVID-19, HIV, hepatitis B and C. Many viruses continue to be a major global public health issues.
World Health Organization (WHO) reports [41, p.3] “As of 20 April 2022, more than 504.4 million confirmed COVID-19 cases and over 6.2 million related deaths had been reported to WHO.” Moreover [41, p.2] “the long-term impact of infection on people’s health is not yet fully understood.” “A recent systematic review reported a prevalence of persistent symptoms in patients after mild COVID-19 infection ranged from 10% to 35% . Cognitive impairment, various neuropsychiatric symptoms, fatigue, headaches and other complaints are among the conditions reported four or more weeks after the initial infection. A detailed understanding of these long-lasting symptoms has not yet been achieved.”
According to WHO ’354 million people globally live with a hepatitis B or C infection’ (296 million with chronic hepatitis B and 58 million with chronic hepatitis C).1 More data (2017) are collected in [40].
In such a situation, any step toward understanding the qualitative (particularly, long-time asymptotic) behaviour of viral infection models is important.
The classical models [21, 22] contain ordinary differential equations (without delay) for three variables: susceptible host cells T, infected host cells T∗ and free virus particles V. The intracellular delay is an important property of the biological problem, so we start discussion with the delay problem (for a particular case of DeAngelis-Beddington functional response f see e.g. [9, 15])1.1 T˙(t)=λ-dT(t)-f(T(t),V(t)),T˙∗(t)=e-ωhf(T(t-h),V(t-h))-δT∗(t),V˙(t)=NδT∗(t)-cV(t).
In (1.1), susceptible cells T are produced at a rate λ, die at rate dT, and become infected at rate f(T, V). Properties and examples of incidence function f are discussed below. Infected cells T∗ die at rate δT∗, free virions V are produced by infected cells at rate NδT∗ and are removed at rate cV(t). In (1.1) h>0 denotes the delay between the time a virus particle contacts a target susceptible cell and the time the cell becomes actively infected (start to produce new virions). It is clear that the constancy of the discrete delay is an extra assumption which essentially simplifies the analysis, but has no biological background.
Since the precise value of the discrete delay h may be difficult to find, viral models with a distributed delay may provide another way to study the dynamics of a disease. For ODE cases and motivations see e.g. [12, 20, 37] and references therein. For motivations for a state-selective delay see Introduction in [23].
For discussion of different viral infection models we refer to [12–15, 19, 28, 29, 31, 36, 37]. PDE models which take into account spatial mobility of cells and virus, possibility of cell-to-cell transmission of the infection (see e.g. [3]) as well as natural time delay effects are discussed in many papers (see e.g. [18, 30, 38, 39] and references therein).
Let Ω⊂R3 be a connected bounded domain with a smooth boundary. Let T(t,x),T∗(t,x),V(t,x) represent the densities of uninfected cells, infected cells and free virions at position x∈Ω at time t.
In this note we are interested in the following PDEs system with delay1.2 T˙(t,x)=rT(t,x)1-T(t,x)TK-dT(t,x)-f(T(t,x),V(t,x))+d1ΔT(t,x),T˙∗(t,x)=e-ωh∫-h0f(T(t+θ),x),V(t+θ),x))ξ(θ,x,ut)dθ-δT∗(t,x)+d2ΔT∗(t,x),V˙(t,x)=NδT∗(t,x)-cV(t,x)+d3ΔV(t,x),x∈Ω.
Here the dot over a function denotes the partial derivative with respect to time i.e, T˙(t,x)=∂T(t,x)∂t, all the constants d,δ,N,c,r,N,ω including di,i=1,2,3 (diffusion coefficients) are positive. We consider a general functional response f(T, V) satisfying natural assumptions presented below. In earlier models (with constant or without delay) the study was started in case of bilinear f(T,V)=const·TV and then extended to more general classes of non-linearities. For more details and discussion see [18, 30].
As usual, for a delay system we denote ut=ut(θ)≡u(t+θ) for θ∈[-h,0],h>0. For general theory on delay equations see [8, 10, 16, 33, 42].
We need initial conditions for the delay problem (1.2):1.3 u(θ,x)=φ(θ,x)≡(T(θ,x),T∗(θ,x),V(θ,x)),θ∈[-h,0]
or shortly u0=φ.
In the papers cited above, the main goal was to study asymptotic stability of stationary solutions. In case of a globally asymptotically stable stationary solution, the long-time behaviour of the system looks very simple. We are interested in a more general case when the existence of a global attractor (for the corresponding dynamical system) may, in general, allow the co-existence of multiple stationary solutions, periodic orbits, invariant manifolds. In case when the existence of a global attractor is proved, the important question arises if the attractor is finite-dimensional (for general facts on attractors see e.g. [4, 34]). The existence of a finite-dimensional global attractor forms a theoretical basis for finite-dimensional approximations which reflect all the asymptotic behaviour of the original system.
Our main mathematical tool in studying of the asymptotic behaviour of solutions is the quasi-stability method developed by I.D.Chueshov (for more details and definitions see [6]). For applications of this metod to delay PDEs see [5] where a model with discrete state-dependent delay is studied. For connections between PDEs with a discrete state-dependent delay [25–27] and considered in the current paper PDEs with a distributed state-selective delay see [23, 24].
To the best of our knowledge, the existence of a finite-dimensional global attractor for a viral infection model has not been investigated before (except the simple case of a globally asymptotically stable stationary solution). It is important to mention that the problem under consideration is infinite-dimensional in both space coordinate (as PDE [4, 34]) and time coordinate (as delay problem [8, 10, 16]).
Main Results
We combine two lines of investigations, one is in a Hilbert space, while the other is in a Banach space.
Study in L2(Ω). Part 1
We start with the Hilbert space approach.
Define the following linear operator2.1 -A=diagd1Δ-d2,d2Δ-δ2,d3Δ-c2inH≡[L2(Ω)]3
with domain D(A)≡D(d1Δ)×D(d2Δ)×D(d3Δ). Here we set D(diΔ)≡{v∈L2(Ω):Δv∈L2(Ω),∂v(x)∂n|∂Ω=0}. We consider operator -Δ in L2(Ω) with the Neumann boundary conditions. This type of conditions is more adequate to biological nature of the problem.
Operator A is a positive self-adjoint operator in [L2(Ω)]3. It is a positive self-adjoint operator with discrete spectrum i.e. there exists an othonormal basis {ek}k=1∞ of H, where ek are eigenvectors of A : Aek=λkek,0<λ1≤λ2≤...,limk→∞λk=+∞ (see e.g., [6, Definition 4.1.1] ). Hence we can define spaces Hα≡D(Aα), H0=H.
Let Cα≡C([-h,0];Hα)⊂C0=C≡C([-h,0];H) for α∈[0,1).
We write, the system (1.2) in the following abstract form2.2 ddtu(t)+Au(t)=F(ut),t>0.
The non-linear mapping F:C≡C0→H is defined by2.3 F(φ)(x)=rφ1(0,x)1-φ1(0,x)TK-d2φ1(0,x)-f(φ1(0,x),φ3(0,x))B(φ,x)-δ2φ2(0,x)Nδφ2(0,x)-c2φ3(0,x).
Here φ=(φ1,φ2,φ3)∈C and the nonlinear distributed delay term has the following form2.4 B(φ,x)≡e-ωh∫-h0f(φ1(θ,x),φ3(θ,x))ξ(θ,x,φ)dθ,x∈Ω.
We assume the following
(H1) Let f:R2→R be continuous, ξ:C→L1(-h,0;L∞(Ω)) be continuous and bounded, i.e.∫-h0||ξ(θ,·,φ)-ξ(θ,·,ψ)||L∞(Ω)dθ→0,as||φ-ψ||C→0,∫-h0||ξ(θ,·,φ)||L∞(Ω)dθ≤Mξ,1,∀φ∈C.
Considering ||B(φ,·)-B(ψ,·)||L2(Ω) one can check that assumptions (H1) imply continuity B:C→L2(Ω). It implies continuity B:Cα→L2(Ω) for α∈[0,1).
Hence, the form (2.3) and the above continuity of B give that F is a nonlinear continuous mapping from Cα into H for all α∈[0,1) and is bounded on bounded sets in Cα.
We use the standard
Definition 2.1
We call a function u∈C([-h,T];Hα) a mild solution (Hα-mild) of the problem (2.2), (1.3) if u0=φ and2.5 u(t)=e-tAu(0)+∫0te-(t-τ)AF(uτ)dτ,t∈[0,T].
The local existence of a mild solution to (2.2), (1.3) is standard due to the continuity of F, its boundedness on bounded sets and Schauder’s fixed point theorem (see [35]).
Now we assume the nonlinear term B has the form (2.4) and
(H2) f is Lipschitz, ξ is Lipschitz and bounded in the following norms2.6 ∫-h0||ξ(θ,·,φ)-ξ(θ,·,ψ)||L∞(Ω)dθ≤Lξ,1||φ-ψ||C,
2.7 ∫-h0||ξ(θ,·,φ)||L∞(Ω)dθ≤Mξ,1,∀φ∈C.
Theorem 2.2
Let f, ξ satisfy (H2). Then, for any initial φ∈Cα with α∈[0,1) there exist T=Tφ>0 and an unique mild solution to (2.2), (1.3) on [-h,T]. The solution continuously depends on initial function φ i.e. for two mild solutions ||ui(t)||α≤R,i=1,2 one has2.8 ||u1(t)-u2(t)||α≤CT^,R||φ1-φ2||Cα,t∈[0,T^],T^≡min{Tφ1;Tφ2}.
Proof
To prove Theorem 2.2 we need only to show that F:Cα→H is locally Lipschitz. We start with the following
Lemma 2.3
Assume (H2) is satisfied. Then the nonlinear distributed delay term B (2.4) is locally Lipschitz continuous2.9 ||B(φ,·)-B(ψ,·)||L2(Ω)≤LB(R)||φ-ψ||C,∀||φ||C,||ψ||C≤R
with LB(R)=e-ωhLf2Mξ,1+Lξ,1·R.
Proof of Lemma
Consider the differenceB(φ,x)-B(ψ,x)=e-ωh∫-h0f(φ1(θ,x),φ3(θ,x))-f(ψ1(θ,x),ψ3(θ,x))ξ(θ,x,φ)dθ+e-ωh∫-h0f(ψ1(θ,x),ψ3(θ,x))ξ(θ,x,φ)-ξ(θ,x,ψ)dθ.
We have ||B(φ,·)-B(ψ,·)||L2(Ω)≤e-ωh∫-h0||(f(φ1(θ,·),φ3(θ,·))-f(ψ1(θ,·),ψ3(θ,·)))ξ(θ,·,φ)||L2(Ω)dθ+e-ωh∫-h0||f(ψ1(θ,·),ψ3(θ,·))ξ(θ,·,φ)-ξ(θ,·,ψ)||L2(Ω)dθ≡I.
To estimate the L2(Ω)-norm in the first term we consider (notice the square of the norm)∫Ω|f(φ1(θ,x),φ3(θ,x))-f(ψ1(θ,x),ψ3(θ,x))|2|ξ(θ,x,φ)|2dx≤|||f(φ1(θ,·),φ3(θ,·))-f(ψ1(θ,·),ψ3(θ,·))|2||L1(Ω)|||ξ(θ,·,φ)|2||L∞(Ω)≤2Lf2||φ-ψ||C2·||ξ(θ,·,φ)||L∞(Ω)2.
Here we used properties ∫Ω|a(x)b(x)|dx≤||a||L1(Ω)||b||L∞(Ω), |||b|2||L∞(Ω)=||b||L∞(Ω)2 and the Lipschitz property of f. Similar properties for the second term give∫Ω|f(ψ1(θ,·),ψ3(θ,·)|2|ξ(θ,·,φ)-ξ(θ,·,ψ)|2dx≤Lf2||ψ||C2·||ξ(θ,·,φ)-ξ(θ,·,ψ)||L∞(Ω)2.
These estimates and (H2) allow to continueI≤e-ωhLf2Mξ,1+Lξ,1·||ψ||C||φ-ψ||C≤LB(R)||φ-ψ||C,
∀||φ||C,||ψ||C≤R with LB(R)=e-ωhLf2Mξ,1+Lξ,1·R. It completes the proof of lemma.
Next, we notice that for α>0 and v∈Hα one has ||v||≤λ1-α||v||α. Hence (2.9) implies similar Lipschitz property in smaller space2.10 ||B(φ,·)-B(ψ,·)||L2(Ω)≤LB,α(R)||φ-ψ||Cα,∀||φ||Cα,||ψ||Cα≤R
with LB,α(R)=λ1-αe-ωhLf2Mξ,1+Lξ,1·R. Finally, (2.10) and (2.3) give the local Lipschitz property of F (since all the other terms are polynomials): for every R>0 there exists LF,α(R) such that2.11 ||F(φ)-F(ψ)||L2(Ω)≤LF,α(R)||φ-ψ||Cα,∀||φ||Cα,||ψ||Cα≤R.
The rest of the proof is standard (see e.g. [35, 6, theorem 6.1.6]). We do not repeat it here. □
Study in C(Ω¯)
We use the basic functional framework described in [17] and applied in [30].
Define the following linear operator -A0 =diag(d1Δ-d2,d2Δ-δ2,d3Δ-c2) in C(Ω¯;R3) with D(A0)≡D(d1Δ)×D(d2Δ)×D(d3Δ). Here, for di≠0 we set D(diΔ)≡{v∈C2(Ω¯):∂v(x)∂n|∂Ω=0} and D(djΔ)≡C(Ω¯) for dj=0. We omit the space coordinate x, for short, for unknown u(t)=(T(t),T∗(t),V(t))∈X≡[C(Ω¯)]3≡C(Ω¯;R3). It is well-known that the closure -A =-AC (in X) of the operator -A0 generates a C0-semigroup e-At on X which is analytic and nonexpansive [17, p.5]. We denote the space of continuous functions by CX≡C([-h,0];X) equipped with the sup-norm ||ψ||CX≡maxθ∈[-h,0]||ψ(θ)||X.
We can use the abstract form (2.2) and nonlinear map (2.3), changing linear operator (A=AC instead of A, see (2.1)) and corresponding spaces.
Definition 2.4
We call a function u∈C([-h,T];X) a mild solution (C(Ω¯)-mild) of the problem (2.2), (1.3) if u0=φ and (2.5) holds with A=AC instead of A.
We notice that Definitions 2.1 and 2.4 give different notions of mild solutions (belong to different spaces and use different semigroups e-tA on Hα and e-tAC on X).
Now we assume the nonlinear term B has the form (2.4) and (c.f. (H2))
(H3) f is Lipschitz, ξ≥0 is Lipschitz and bounded in the following norms2.12 ∫-h0||ξ(θ,·,φ)-ξ(θ,·,ψ)||C(Ω¯)dθ≤Lξ,C||φ-ψ||CX,
2.13 ∫-h0||ξ(θ,·,φ)||C(Ω¯)dθ≤Mξ,C,∀φ∈CX.
We need further assumptions on Lipschitz function f :2.14 (Hf1+)f(T,0)=f(0,V)=0,andf(T,V)>0for allT>0,V>0;fis strictly increasing in both coordinates for allT>0,V>0;there existsμ>0such that|f(T,V)|≤μ|T|for allT,V∈R.
Define the set2.15 Ωlog≡φ=(φ1,φ2,φ3)∈C=CX:0≤φ1(θ)≤M1≡r2dTK,0≤φ2(θ)≤M2≡e-ωhμrdδTKMξ,C,0≤φ3(θ)≤M3≡2e-ωhNμrdcTKMξ,C
where θ∈[-h,0], μ is defined in (Hf1+) and all the inequalities hold pointwise w.r.t. x∈Ω¯.
We have the following result
Theorem 2.5
Let non-linear Lipschitz function f satisfy (Hf1+) (see (2.14)), ξ satisfy (H3). Then Ωlog is invariant i.e. for any φ∈Ωlog the unique C(Ω¯)-mild solution to problem (2.2), (1.3) exists and satisfies ut∈Ωlog for all t≥0.
Proof
We start with the local Lipschitz property of B:CX→X. Assumptions (H3) give2.16 ||B(φ,·)-B(ψ,·)||C(Ω¯)≤LB,C(R)||φ-ψ||CX,∀||φ||CX,||ψ||CX≤R
with LB,C(R)=e-ωhLfMξ,C+Lξ,C·R.
One can check that F:CX→X is locally Lipschitz. The existence and uniqueness of a mild solution u∈C([-h,T];X) to the problem (2.2), (1.3) is standard. The proof of the invariance part follows the invariance result of [17] with the use of the Lipschitz property of nonlinearity F. The estimates (for the subtangential condition) are the same as for the constant delay case, see e.g. [18, Theorem 2.2].
Consider ρ≥0 and φ∈Ωlog.
φ(0,x)+ρF(φ,x) =φ1(0,x)+ρrφ1(0,x)1-φ1(0,x)TK-ρd2φ1(0,x)-ρf(φ1(0,x),φ3(0,x))φ2(0,x)+ρB(φ,x)-ρδ2φ2(0,x)φ3(0,x)+ρNδφ2(0,x)-ρc2φ3(0,x)
We use notation F=(F1,F2,F3)T and estimate separately each of three coordinates above.
(a) We notice that the logistic term (see the first equation in (1.2)) rT1-TTK has its maximum at point T=TK/2, so rT1-TTK≤14rTK for all T∈R.
Hence, for small enough ρ≥0 and φ1(0,x)∈[0,M1] (see (2.15)) we haveφ1(0,x)+ρF1(φ,x)≤φ1(0,x)+ρ14rTK-d2φ1(0,x)≤M1.
(b) For the second coordinate we use (see (2.13) and (2.14))B(φ,x)≤e-ωh∫-h0μ|φ1(θ,x)|ξ(θ,x,φ)dθ≤e-ωhμr2dTKMξ,C.
This estimate gives for small enough ρ≥0 and φ2(0,x)∈[0,M2] (see (2.15)):φ2(0,x)+ρF2(φ,x)≤φ2(0,x)+ρe-ωhμr2dTKMξ,C-ρδ2φ2(0,x)≤M2.
(c) For small enough ρ≥0 and φ3(0,x)∈[0,M3] (see (2.15)) one hasφ3(0,x)+ρF3(φ,x)≤φ3(0,x)+ρNδM2-ρc2φ3(0,x)≤NδM22c=M3.
Combining the estimates above, one can check that for small enough ρ≥0 and φ∈Ωlog000≤φ(0,x)+ρF(φ,x)≤M1M2M3≡M
or shortly φ(0,x)+ρF(φ,x)∈[0,M]⊂R3 for all x∈Ω.
The above implies limρ→0+dist{φ(0,·)+ρF(φ,·);[0,M]X}=0,∀φ∈Ωlog⊂CX.
It gives the subtangential condition and allows to apply the invariance result of [17, 32].
The proof of Theorem 2.5 is complete. □
Study in L2(Ω). Part 2
In this section we continue our study of Hα-mild solutions and use results of Theorem 2.5 obtained for C(Ω¯)-mild solutions. The key point here is the Sobolev imbedding theorem [1, p.85] which suggests values of α for which the imbedding Hα→C(Ω¯) holds.
Let us remind the part we need of the Sobolev imbedding theorem [1, p.85].
Let Ω be a domain in Rn. Let j≥0,m≥1 be integers and let 1≤p<∞. Suppose Ω satisfies the strong loc.Lipschitz condition. If mp>n>(m-1)p, then Wj+m,p(Ω)→Cj,λ(Ω¯) for 0<λ≤m-np.
In our case, n=3,p=2. We have (see condition mp>n>(m-1)p) that m∈(32,52). We are interested in m<2, so consider m∈(32,2).
In case A=-Δ, the condition m∈(32,2) corresponds to α∈(34,1).
We notice the importance of the restriction α∈(34,1) which guaranties u(t)∈X=[C(Ω¯)]3.
Combining this property with the uniqueness results for both Hα-mild solutions and C(Ω¯)-mild solutions (both for the same initial function φ) one has the following key property: for any initial φ∈Cα,α∈(34,1) the Hα-mild solution is C(Ω¯)-mild solution.
Our goal is to construct a dynamical system in phase space Ωαlog≡Cα∩Ωlog,α∈(34,1). On this space we define evolution operator Stφ=ut,t≥0, where u is the unique mild solution of problem (2.2), (1.3).
Now we need the following spaceYβ≡{v∈Cα:|v|Yβ<∞},
where β∈(α,1) and2.17 |v|Yβ≡maxθ∈[-h,0]||Aβv(θ)||+maxθ1,θ2∈[-h,0],θ1≠θ2||Aα(v(θ1)-v(θ2))|||θ1-θ2|β-α.
We remind the following result (formulated for an abstract equation of the form (2.2)).
Proposition 2.6
[6, p. 293] Let A be a linear positive self-adjoint operator with discrete spectrum on H. Let F:Cα→H be a locally Lipschitz mapping i.e. for every R>0 (2.11) holds. Assume that the problem (2.2), (1.3) generates a dynamical system (Cα,St). Let D be a forward invariant bounded set in Cα.
Then For every t>h the set StD is bounded in Yβ for arbitrary β∈(α,1). Moreover, for every δ>0 there exists Rδ such that 2.18 StD⊂Bβ={u∈Yβ:|u|Yβ≤Rδ}for allt≥δ+h.
In particular, this means that the dynamical system (Cα,St) is conditionally compact and thus asymptotically smooth.
The mapping St is Lipschitz from D into Yβ. Moreover, for every h<a<b<+∞ there exists a constant MD(a,b) such that 2.19 |Stφ-Stψ|Yβ≤MD(a,b)||φ-ψ||Cα,t∈[a,b],φ,ψ∈D.
In particular, this means that the dynamical system (Cα,St) is quasi-stable at any time t∈[a,b].
We remind (see, e.g., [4, 34])
Definition 2.7
A global attractor of the dynamical system (Cα,St) is defined as a bounded closed set U⊂Cα which is invariant (StU=U for all t>0) and uniformly attracts all bounded setslimt→+∞sup{distCα(Sty,U):y∈B}=0foranyboundedsetBinCα.
Our main result is the following
Theorem 2.8
Let α∈(34,1), non-linear Lipschitz function f satisfy (Hf1+) (see (2.14)), ξ satisfy (H2), (H3). Then the pair (St;Ωαlog) constitutes a dynamical system constructed by problem (2.2), (1.3). This dynamical system possesses a finite-dimensional global attractor.
Proof
The well-posedness of the problem (2.2), (1.3) (the existence, uniqueness and continuous dependence on initial function φ∈Cα) is given by Theorem 2.2.
First we remind an important estimate (and its derivation) which is a part of the property (2.18). For more details, see [6, p.294]. Let D be a forward invariant bounded set in Cα (for this part α≥0). Consider β∈(α,1) and a mild solution u(t)=Stφ, see (2.5). We use property ||Aαe-tA||≤αetα,t>0,α≥0 (with the rule 00=1) to get||u(t)||β≤β-αe(t-s)β-α||u(s)||α+∫stβe(t-τ)β||F(uτ)||dτ
for all t>s≥0. Since Stφ∈D for all t≥0 one has ||u(t)||α≤CD,∀t≥0. So||u(t)||β≤β-αe(t-s)β-αCD+KD(F)βeβ|t-s|1-β1-β
for all t>s≥0, where KD(F)=sup{||F(v)||:v∈D}. If we choose s=t-δ, then2.20 StD⊂{u∈Cβ:||u||β≤Rδ∗},for allt≥δ+h,
whereRδ∗≡β-αeδβ-αCD+KD(F)βeβδ1-β1-β.
This estimate is a part of the property (2.18), see (2.17).
Since parameter α is a smoothness parameter of the space Cα and phase space Ωαlog, we change notations in (2.20) to adopt it for the proof of the dissipativeness (the existence of a bounded absorbing set). More precisely, we consider a solution ||u(t)||γ≤CD,∀t≥0. Here γ≥0 instead of α. Now we estimate ||u(t)||α,α∈(γ,1) instead of ||u(t)||β. The estimate, similar to (2.20) gives2.21 StD⊂{u∈Cα:||u||α≤R^δ∗},for allt≥δ+h,
whereR^δ∗≡α-γeδα-γCD+KD(F)αeαδ1-α1-α.
Notice that for any v∈[C(Ω¯)]3⊂[L2(Ω)]3 one has ||v||0=||v||[L2(Ω)]3≤||v||[C(Ω¯)]3·|Ω| with |Ω|≡∫Ω1dx.
We apply the above property (2.21) for γ=0 and D=Ωαlog⊂Ωlog (bounded in CX). Hence for γ=0 the property ||u(t)||0≤CD,∀t≥0 holds. As a result, (2.21) implies (a) mild solutions are global (defined for all t≥-h) and (b) the dissipativeness of the dynamical system (St;Ωαlog) for each α∈(34,1).
Now by Proposition 2.6 [6, p.293] our dynamical system (St;Ωαlog) is quasi-stable.
We can apply [6, Theorem 6.1.12] to the dynamical system (St;Ωαlog) to get the main result - the existence of a finite-dimensional global attractor.
It completes the proof of Theorem 2.8. □
Examples of the Distributed Delay Term
Consider the nonlinear delay term B of the form (2.4). We present a simple example of function ξ:[-h,0]×Ω×C→R (c.f. [23, 24])ξ(θ,x,φ)=e-σ(-η(φ)-θ)2g(x),σ>0,
where (i) η:C→[0,h] is Lipschitz continuous and g∈L∞(Ω) to satisfy (H2) and
(ii) η:CX→[0,h] is Lipschitz continuous and g∈C(Ω¯) to satisfy (H3).
For motivations for such a state-selective delay see e.g. [23]. The profile function e-σ(c-θ)2 was chosen for simplicity to show that the delay term of the form ∫-h0e-σ(c-θ)2ϕ(θ)dθ has the maximal historical impact in a neighbourhood of the time moment c (the maximum of function e-σ(c-θ)2 at point θ=c). In our example this maximum point can be state-selective [23] (state-dependent) c=-η(φ)∈[-h,0].
We mention some well-known examples of non-linear functions f used for viral infection models. The first one is the DeAngelis-Bendington [2, 7] functional response f(T,V)=kTV1+k1T+k2V, with k,k1≥0,k2>0. We also mention that the functional response includes as a special case (k1=0) the saturated incidence rate f(T,V)=kTV1+k2V. Another example of the nonlinearity is the Crowley-Martin incidence rate f(T,V)=kTV(1+k1T)(1+k2V), with k≥0,k1,k2>0 and more general the Hattaf-Yousfi functional response of the form kTVk0+k1T+k2V+k3TV [11]. For more general class of functions f see, e.g. [11, 18, 29]. We notice that, in contrast to [11, 18], we do not assume here the differentiability of f.
We also mention that our assumptions on f are naturally less restrictive comparing to the ones in the mentioned above works where asymptotic stability of stationary solutions are discussed.
Conclusion
In this paper we study a virus dynamics model with reaction-diffusion, logistic growth terms and a general non-linear infection rate functional response. The model has a distributed delay, including the case of state-selective delay which is a distributed ‘analog’ to a discrete state-dependent delay.
Our main mathematical tool in studying of the asymptotic behaviour of solutions is the quasi-stability method developed by I.D.Chueshov [6]. We construct a dynamical system in a Hilbert space and prove the existence of a finite-dimensional global attractor. To prove the natural for a virus dynamics model dissipativness of the dynamical system we conduct a parallel study in a Banach space.
Acknowledgements
The author would like to thank five anonymous reviewers for their useful comments and suggestions. This paper is dedicated to the memory of my father, Vyacheslav O. Rezunenko (February 1941–August 2022), who passed away recently.
Funding
No funding was received to assist with the preparation of this manuscript.
Declarations
Conflict of interest
The authors have no conflicts of interest to declare that are relevant to the content of this article.
1 https://www.who.int/health-topics/hepatitis.
This work was completed with the support of our TeX-pert.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
1. Adams, R.A., Fournier, J.F.: Sobolev spaces. Elsevier, 2nd Edition, 320 p (2003)
2. Beddington JR Mutual interference between parasites or predators and its effect on searching efficiency J. Anim. Ecol. 1975 44 331 340 10.2307/3866
3. Carloni G Crema A Valli MB Ponzetto A Clementi M HCV infection by cell-to-cell transmission: choice or necessity? Curr. Mol. Med. 2012 12 83 95 10.2174/156652412798376152 22082483
4. Chueshov, I.D.: Introduction to the theory of infinite-dimensional dissipative systems, acta, Kharkov, 1999, english translation (2002). http://www.emis.de/monographs/Chueshov/
5. Chueshov ID Rezounenko AV Finite-dimensional global attractors for parabolic nonlinear equations with state-dependent delay Commun. Pure Appl. Anal. 2015 14 5 1685 1704 10.3934/cpaa.2015.14.1685
6. Chueshov I Dynamics of Quasi-Stable Dissipative Systems 2015 Cham Springer
7. DeAngelis DL Goldstein RA O’Neill RV A model for tropic interaction Ecology 1975 56 881 892 10.2307/1936298
8. Diekmann O van Gils S Verduyn Lunel S Walther H-O Delay Equations: Functional, Complex, and Nonlinear Analysis 1995 New York Springer-Verlag
9. Gourley SA Kuang Y Nagy JD Dynamics of a delay differential equation model of hepatitis B virus infection J. Biol. Dyn. 2008 2 140 153 10.1080/17513750701769873 22880697
10. Hale JK Theory of Functional Differential Equations 1977 Berlin- Heidelberg- New York Springer
11. Hattaf K Yousfi N A generalized HBV model with diffusion and two delays Comput. Math. Appl. 2015 69 31 40 10.1016/j.camwa.2014.11.010
12. Hattaf K Yousfi N A class of delayed viral infection models with general incidence rate and adaptive immune response Int. J. Dyn. Control 2016 4 254 265 10.1007/s40435-015-0158-1
13. Hattaf K On the stability and numerical scheme of fractional differential equations with application to biology Computation 2022 10 97 10.3390/computation10060097
14. Hews S Eikenberry S Nagy JD Rich dynamics of a hepatitis B viral infection model with logistic hepatocyte growth J. Math. Biol. 2010 60 4 573 590 10.1007/s00285-009-0278-3 19533136
15. Huang G Ma W Takeuchi Y Global analysis for delay virus dynamics model with Beddington-DeAngelis functional response Appl. Math. Lett. 2011 24 1199 1203 10.1016/j.aml.2011.02.007
16. Kuang Y Delay Differential Equations with Applications in Population Dynamics, Mathematics in Science and Engineering 1993 Boston, MA Academic Press Inc
17. Martin RH Jr Smith HL Abstract functional-differential equations and reaction-diffusion systems Trans. Amer. Math. Soc. 1990 321 1 44
18. McCluskey C Yang Yu Global stability of a diffusive virus dynamics model with general incidence function and time delay Nonlinear Anal. Real World Appl. 2015 25 64 78 10.1016/j.nonrwa.2015.03.002
19. Murray JM Kelleher AD Cooper DA Timing of the components of the HIV life cycle in productively infected CD4+ T cells in a population of HIV-Infected individuals J. Virol. 2011 85 20 10798 10805 10.1128/JVI.05095-11 21835801
20. Nakata Y Global dynamics of a cell mediated immunity in viral infection models with distributed delays J. Math. Anal. Appl. 2011 375 1 14 27 10.1016/j.jmaa.2010.08.025
21. Nowak M Bangham C Population dynamics of immune response to persistent viruses Science 1996 272 74 79 10.1126/science.272.5258.74 8600540
22. Perelson A Neumann A Markowitz M Leonard J Ho D HIV-1 dynamics in vivo: virion clearance rate, infected cell life-span, and viral generation time Science 1996 271 1582 1586 10.1126/science.271.5255.1582 8599114
23. Rezounenko A Wu J A non-local PDE model for population dynamics with state-selective delay: local theory and global attractors J. Comput. Appl. Math. 2006 190 1–2 99 113 10.1016/J.CAM.2005.01.047
24. Rezounenko AV Partial differential equations with discrete and distributed state-dependent delays J. Math. Anal. Appl. 2007 326 1031 1045 10.1016/j.jmaa.2006.03.049
25. Rezounenko AV Differential equations with discrete state-dependent delay: Uniqueness and well-posedness in the space of continuous functions Nonlinear Anal. Theory Methods Appl. 2009 70 3978 3986 10.1016/j.na.2008.08.006
26. Rezounenko AV Non-linear partial differential equations with discrete state-dependent delays in a metric space Nonlinear Anal. Theory Methods Appl. 2010 73 1707 1714 10.1016/j.na.2010.05.005
27. Rezounenko AV Zagalak P Non-local PDEs with discrete state-dependent delays: well-posedness in a metric space Discrete Contin. Dyn. Syst. Ser. A 2013 33 2 819 835 10.3934/dcds.2013.33.819
28. Rezounenko AV Stability of a viral infection model with state-dependent delay, CTL and antibody immune responses Discrete Contin. Dyn. Syst. Ser. B 2017 22 1547 1563 10.3934/dcdsb.2017074
29. Rezounenko AV Continuous solutions to a viral infection model with general incidence rate, discrete state-dependent delay, CTL and antibody immune responses Electron. J. Qual. Theory Differ. Equ. 2016 79 1 15 10.14232/ejqtde.2016.1.79
30. Rezounenko AV Viral infection model with diffusion and state-dependent delay: stability of classical solutions Discrete Contin. Dyn. Syst. Ser. B 2018 23 3 1091 1105 10.3934/dcdsb.2018143
31. Shudo E Ribeiro RM Talal AH Perelson AS A hepatitis C viral kinetic model that allows for time-varying drug effectiveness Antivir. Ther. 2008 13 919 926 10.1177/135965350801300711 19043926
32. Smith HL Monotone Dynamical Systems, An Introduction to the Theory of Competitive and Cooperative Systems, Mathematical Surveys and Monographs 1995 Providence, RI American Mathematical Society
33. Smith H Introduction to Delay Differential Equations with Sciences Applications to the Life, Texts in Applied Mathematics 2011 New York, Dordrecht, Heidelberg, London Springer
34. Temam R Infinite Dimensional Dynamical Systems in Mechanics and Physics 1988 Berlin-Heidelberg-New York Springer
35. Travis CC Webb GF Existence and stability for partial functional differential equations Transact. AMS 1974 200 395 418 10.1090/S0002-9947-1974-0382808-3
36. Wang X Liu S A class of delayed viral models with saturation infection rate and immune response Math. Methods Appl. Sci. 2013 36 2 125 142 10.1002/mma.2576
37. Wang J Huang G Takeuchi Y Global asymptotic stability for HIV-1 dynamics with two distributed delays Math. Med. Biol. 2012 29 283 300 10.1093/imammb/dqr009 21734220
38. Wang W Ren X Ma W Lai X New insights into pharmacologic inhibition of pyroptotic cell death by necrosulfonamide: a PDE model Nonlinear Anal. Real World Appl. 2020 56 103173 10.1016/j.nonrwa.2020.103173
39. Wang W Wang X Feng Z Time periodic reaction-diffusion equations for modeling 2-LTR dynamics in HIV-infected patients Nonlinear Anal. Real World Appl. 2021 57 103184 10.1016/j.nonrwa.2020.103184
40. World Health Organization, Global hepatitis report-2017, ISBN: 978-92-4-156545-5 http://apps.who.int/iris/bitstream/10665/255016/1/9789241565455-eng.pdf?ua=1
41. World Health Organization, World health statistics 2022: monitoring health for the SDGs, sustainable development goals, 2022, Global report, https://www.who.int/publications/i/item/9789240051157
42. Wu J Theory and Applications of Partial Functional Differential Equations 1996 New York Springer-Verlag
| 0 | PMC9746596 | NO-CC CODE | 2022-12-15 23:21:57 | no | Qual Theory Dyn Syst. 2023 Dec 13; 22(1):11 | utf-8 | Qual Theory Dyn Syst | 2,022 | 10.1007/s12346-022-00707-6 | oa_other |
==== Front
Artif Intell Rev
Artif Intell Rev
Artificial Intelligence Review
0269-2821
1573-7462
Springer Netherlands Dordrecht
10361
10.1007/s10462-022-10361-8
Article
An approach to prevent weight manipulation by minimum adjustment and maximum entropy method in social network group decision making
Sun Qi [email protected]
12
http://orcid.org/0000-0002-1482-9827
Wu Jian [email protected]
12
Chiclana Francisco [email protected]
34
Wang Sha [email protected]
12
Herrera-Viedma Enrique [email protected]
45
Yager Ronald R. [email protected]
6
1 grid.412518.b 0000 0001 0008 0619 School of Economics and Management, Shanghai Maritime University, Shanghai, 201306 China
2 grid.412518.b 0000 0001 0008 0619 Center for Artificial Intelligence and Decision Sciences, Shanghai Maritime University, Shanghai, 201306 China
3 grid.48815.30 0000 0001 2153 2936 Faculty of Computing, Engineering and Media, Institute of Artificial Intelligence, De Montfort University, Leicester, UK
4 grid.4489.1 0000000121678994 Department of Computer Science and AI, Andalusian Research Institute in Data Science and Computational Intelligence, University of Granada, 18071 Granada, Spain
5 grid.412125.1 0000 0001 0619 1117 Department of Electrical and Computer Engineering, Faculty of Engineering, King Abdulaziz University, Jeddah, 21589 Saudi Arabia
6 grid.419406.e 0000 0001 0087 8225 Machine Intelligence Institute, Iona College, New Rochelle, NY 10801 USA
13 12 2022
132
© The Author(s), under exclusive licence to Springer Nature B.V. 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
In social network group decision making (SN-GDM) problem, subgroup weights are mostly unknown, many approaches have been proposed to determine the subgroup weights. However, most of these methods ignore the weight manipulation behavior of subgroups. Some studies indicated that weight manipulation behavior hinders consensus efficiency. To deal with this issue, this paper proposes a theoretical framework to prevent weight manipulation in SN-GDM. Firstly, a community detection based method is used to cluster the large group. The power relations of subgroups are measured by the power index (PI), which depends on the subgroups size and cohesion. Then, a minimum adjustment feedback model with maximum entropy is proposed to prevent subgroups’ manipulation behavior. The minimum adjustment rule aims for ‘efficiency’ while the maximum entropy rule aims for ‘justice’. The experimental results show that the proposed model can guarantee the rationality of weight distribution to reach consensus efficiently, which is achieved by maintaining a balance between ‘efficiency’ and ‘justice’ in the mechanism of assigning weights. Finally, the detailed numerical and simulation analyses are carried out to verify the validity of the proposed method.
Keywords
Social network group decision making
Weight manipulation
Feedback mechanism
Minimum adjustment
Maximum entropy
http://dx.doi.org/10.13039/501100001809 NationalNaturalScienceFoundationofChina 71971135 Sun Qi
==== Body
pmcIntroduction
The rapid development of social network makes it convenient for the masses to participate in various group decision-making (GDM) problems, which can be seen as a stepping stone towards the realisation of electronic democracy. Recently, Wu and Chiclana et.al (2014,2015; 2017) proposed the Social network group decision making (SN-GDM), where trust relationship is regarded as a reliable resource to assign expert weighs and generate recommendation advice for the inconsistent expert to reach higher consensus level. It now becomes a hot issue in the field of science of policy making (Zhang et al. 2017; Wu and Xu 2018; Xu et al. 2019; Tang and Liao 2021; Wan et al. 2021; Li et al. 2022), a differentiation from traditional GDM problems when trust relationship is used to deal with the inconsistency among group (Xu et al. 2019; Zhao et al. 2019; Ding et al. 2020; Chao et al. 2021).
It is well known that differences between people’s background and professional opinion may lead to conflicts in GDM. The consensus reaching process (CRP) is an effective method to eliminate conflicts in GDM (Cabrerizo et al. 2014; Amirkhani et al. 2022). Different types of CRP models have been of interest to the research community: CRP in social networks (Zhang et al. 2020; Wu et al. 2022; Zhang et al. 2022), CRP in dynamic context (Pérez et al. 2018), CRP with preferred representative structure (Gong et al. 2018; Wu et al. 2020), adaptive CRP (Rodríguez 2018; Tian et al. 2021), CRP with minimum cost (Zhang et al. 2020; Chen et al. 2021; Gong et al. 2021), and CRP driven by behavior/attitude (Liu et al. 2019b; Sun et al. 2021; Cao et al. 2022). Conflicts in SN-GDM are likely to increase as the size of the group of people involved increases, which subsequently implies that SN-GDM require higher adjustment costs to reach consensus.
A feedback mechanism, as part of a CRP, is an effective method to eliminate conflicts among experts (Dong et al. 2017; Li et al. 2018; Cao et al. 2021; Wu et al. 2021). In the existing literature, the modifications of experts’ preferences are usually conducted by two types of feedback mechanisms (Liu et al. 2019; Zhang et al. 2020; Chen et al. 2021; Xing et al. 2022). The first type of feedback mechanisms are designed with the so-called identification rules and direction rules. The first type of rules identify both inconsistent experts and their inconsistent preferences, while the second type of rules provide the direction of preference modification required to increase consensus. The second type of feedback mechanisms are based on minimum adjustment/cost optimization modeling, which performs better than the former type of feedback mechanisms in changing as less as possible the original information of inconsistent experts. The minimum cost improves the consensus efficiency. However, it does not consider the fairness of weight distribution. Thus, in general, the second type of feedback processes have lower adjustment cost and lower computational complexity than the first type of feedback processes. However, they still suffer from the following limitations: Most of the existing feedback mechanisms for SN-GDM utilize subjective/objective methods to assign weights to subgroups (Zhang et al. 2017; Wu and Xu 2018; Ding et al. 2020; Wan et al. 2021). However, these methods ignore the weight manipulation behavior. Dong et al. (2018) and Liu et al. (2019a) argued that experts may strategically set attribute weights of alternatives to obtain benefits in the decision-making process. Therefore, weight manipulation is also an important issue in GDM since a decision maker may wish to attain a greater importance degree (weight) to increase his/her benefits. For example, a decision maker with lowest status is willing to adopt egalitarianism to assign weight, while a decision maker with highest status expects to use authoritarianism to distribute weight (see Fig. 2). Obviously, manipulation of weights may hinder consensus to a certain extent and affect the cost of reaching consensus, and then, it should be worth studying in SN-GDM.
There are few research studies on how to prevent weight manipulation. Examples of these studies include Yager’s study on the use of uninorm operators to prevent manipulation (Yager 2001; 2002), Wu et al.’s (2021a) studied on individual manipulation weights to gain benefit where an optimization model to prevent manipulation behavior with minimum cost was investigated. Although these studies are of interests and contributed towards manipulation prevention, they are established from an ‘efficiency’ policy, which achieves group consensus with minimal adjustment/costs from a group perspective. However, other policy point of view, such as ‘justice’, should also be considered to guarantee individual benefits when assigning weights to DMs. Therefore, a reasonable policy to prevent weight manipulation combining ‘justice’ and ‘efficiency’ can achieve a balance between individual benefit and group goal.
To deal with the above limitations, a novel approach to prevent weight manipulation in SN-GDM problems is designed in this paper. Thus, the main contribution of this paper are: A power index is proposed to measure the power relations of subgroups, which combines the subgroups size and cohesion. The power index is used as a reliable resource to determine the importance order relation of subgroups, and as such is embedded in the prevention weight manipulation mechanism.
A maximum entropy mechanism to assign weight to subgroups from the ‘justice’ policy point of view is developed. Combining the minimum adjustment feedback and the maximum entropy method, a comprehensive approach to prevent weight manipulation in SN-GDM problems is established, which has the merit of balancing the ‘efficiency’ and ‘justice’ policy. Specifically, the minimum cost solves the problem of ‘efficiency’, while the maximum entropy method solves the problem of ‘justice’ of weight distribution.
The rest of this article is organized as follows: Sect. 2 gives detailed literature review on conflicts problem in SN-GDM for large group, CRP with minimum cost and weight manipulation behavior. Section 3 introduces the concept of 2-tuple linguistic representation. In Sect. 4, a power index (PI) is proposed to determine the importance degree of subgroups. A minimum adjustment and maximum entropy based feedback mechanism to prevent weight manipulation in SN-GDM is proposed in Sect. 5. Section 6 provides a case study to verify the effectiveness of the proposed model. Finally, some conclusions are pointed out in Sect. 7. The main notations used in this paper are summarized in TABLE 1.Table 1 The main notations in this paper
Notations Meaning
M Set of alternatives
C Set of attributes
E Set of experts
SG Set of subgroups
LG Large-scale group
SM Social matrix
Ah=aijhm×n Decision matrix of expert eh
daijh,aijk Deviation degree between eh and ek
ϑaijh,aijk Similarity degree between eh and ek
CDhk Consensus degree between eh and ek
χhk Edge between eh and ek in network
ε Network consensus threshold
NSG The number of subgroups
PMr The number of experts in SGr
PCr Cohesion of subgroup SGr
PISGr Power index of subgroup SGr
σ Permutation
w=w1,w2,…,wr A weighting vector for subgroups
α Attitudinal parameter
ω=ω1,ω2,…,ωn A weighting vector for attributes
ACDSGr Consensus degree between SGr and the rest of subgroups
ACDSGr¯ Consensus degree between SGr and the rest of subgroups after feedback
ACAiSGr Consensus degree between SGr and the rest of subgroups on the alternative xi
ACEijSGr Consensus degree between SGr and the rest of subgroups on the alternative xi
with respect to the criterion cj
γ Consensus threshold for LG
δr Feedback parameter for inconsistent subgroup SGr
aijSGr~ Advice for SGr
MTC Minimum total cost of feedback process
Related work
This section presents a necessarily short overview of conflicts problem in SN-GDM for large group, CRP with minimum cost and weight manipulation behavior based on their relevance to our proposal.
Conflicts problem in SN-GDM for large group
As previously point out, there may be conflicts among decision makers due to their differences in knowledge and backgrounds (Del Moral 2018; Xiao et al. 2022). So it is necessary to reach a group consensus before their individual preferences are aggregated. Generally, there are two methods to resolve conflicts of large group in SN-GDM: The first way is to cluster the large group thereby reducing its dimension, which solves the conflict problem within subgroups. Many commonly used clustering method are utilized to solve SN-DGM problem. For instance, Wu and Xu (2018) identify the subgroups based on the k-means method under the fuzzy preference environment in LSGDM. Mandal et al. (2022) used the grey clustering algorithm to conduct clustering based on the similarity measure among the experts. Li et al. (2022) used the fuzzy cluster analysis method to divide large group into subgroups and integrate heterogeneous information. Wu et al. (2021) proposed a dynamic clustering method, which divides large groups by Louvain algorithm.
The second way is to reach a consensus among the subgroups after clustering, which resolves the conflict problem outside the subgroups. Wu et al. (2018) designed a local feedback policy with identification rules and direction rules to guide the CRP. Chao et al. (2021) presented a CRP model to address the heterogeneous with non-cooperative behaviors. Mandal et al. (2022) proposed a CRP model to manage non-cooperative behaviors by the cluster consensus index and group consensus index. Wang et al. (2022) proposed a two-stage consensus model with feedback mechanism considering different power structures in SN-GDM.
In general, the basic framework of SN-GDM process consists of the following parts, which are shown in Fig. 1. First, a set of alternatives about a decision problem is presented to a large group of experts. Experts provide their preferences about alternatives are then collected. Then the large group is clustered and the preferences of the clustered subgroups are aggregated. Next, if the consensus degree of subgroup reach a consensus threshold, the resolution process will be executed; otherwise, a feedback mechanism is activated to allow the inconsistent subgroups to modify their opinions, and re-aggregate the preferences until a consensus is reached.Fig. 1 Basic framework of SN-GDM process
CRP with minimum cost
In the feedback mechanism of CRP there is usually a moderator responsible for supervising the inconsistent experts and guiding them to modify their preferences to reach a consensus. Generally, the preferences-adjustment are the results of laborious negotiations, which escalate the cost of CRP. Ben-Arieh and Easton (2007) first proposed a consensus minimum cost optimization model for the multi-criteria decision-making consensus problem. In recent years, many minimum cost consensus models based on the method of Ben-Arieh and Easton (2007) have been proposed. For example, Gong et al. (2021) discussed the minimum cost consensus model under uncertain chance-constrained from the perspectives of moderators, individual decision makers, and non-cooperators. Chen et al. (2021) propose an approach to manage the consensus based on minimum adjustments with opinions evolution. Sun et al. (2021) proposed a attitudinal consensus threshold based dynamic minimum adjustment cost feedback model to resolve the GDM problems with different consistency requirements. Xiao et al. (2022) proposed a minimum adjustment element consensus model based on bounded confidences to help the failure mode and effect analysis team reach a consensus. Recently, Zhang et al. (2020) presented a state-of-the-art review of CRP models under minimum cost. By reviewing the research paradigm of minimum cost under classical and complex group decision problems, they pointed out the limitations and some new directions of the minimum cost consensus model.
Weight manipulation behavior
In the decision-making process, decision-makers may utilize some trick to set their own weights to gain benefits, which is usually called weight manipulation behavior. This behavior can help decision makers, to some extent make the decision result in their desired direction. For instance, Yager (2001,2002) studied the individuals weight manipulation behavior in the process of preference aggregation of group. Besides, considering that weight manipulation may lead to unreasonable decision results, the author proposed the management mechanism to prevent this behavior. Dong et al. (2018) studied the strategic weight manipulation in multiple attribute GDM problem. From an optimization point of view, Liu et al. (2019a) studied the strategic weight manipulation in a group GDM context with interval attribute weight information and proposed a minimum cost strategic weight manipulation model. Dong et al. (2021) investigates the clique-based strategies to manipulate trust relationships to gain the desired decision result. The aforementioned literatures on weigh manipulative behavior focus on the selection process stage in GDM. However, weight manipulation behavior may exist at any stage of GDM. Recently, Wu et al. (2021a) studied the effect of weight manipulation behavior on the efficiency of consensus reaching and proposed an optimization model to prevent weight manipulation to increase the efficiency of CRP. But it is necessary to consider the fairness of weight distribution. In general, the entropy weight method can reduce the uncertainty of weight distribution (Wang et al. 2022). The larger the entropy value, the fairer the weight distribution. Therefore, this paper uses a method based on maximum entropy (O’Hagan 1988) to determine the weight of subgroups.
Preliminary
In GDM problems, experts may prefer linguistic terms to numerical values when expressing their preferences (Herrera-Viedma et al. 2021; Li et al. 2021; Yu et al. 2021; Liu et al. 2021). Herrera and Martínez (2000) presented the below 2-tuple linguistic representation model to evaluate decision problems:
Definition 1
(2-tuple linguistic representation) Let S=s0,…,sf and β∈0,f be a linguistic term set and the result of a symbolic aggregation, respectively. Let i=roundβ∈0,…,f. The value α=β-i is called a symbolic translation, and si,α is called the 2-tuple linguistic representation of the symbolic aggregation α.
The 2-tuple linguistic representation of symbolic aggregation can be mathematically formalised as an strictly increasing continuous function:1 Δ:0,f→S×-0.5,0.5,
2 Δβ=si,α;i=roundβα=β-i,
with inverse function Δ-1:S×-0.5,0.5→0,f being Δ-1si,α=i+α=β, and following properties: Let si,α and sj,ψ be two 2-tuples linguistic. If i<j, then si,α is smaller than sj,ψ.
If i=j, then
i. If α=ψ, then si,α and sj,ψ are equivalent.
ii. If α<ψ, then si,α is smaller than sj,ψ.
2. The 2-tuple negation operator is Negsi,α=Δq-Δ-1si,α.
3. Given a set of 2-tuple linguistic labels s1,α,s2,α,…,sf,α and let λ1,λ2,…,λf be an associated weighting vector which satisfies λi∈0,1,∑i=1fλi=1. The 2-tuple weighted arithmetic average (WAA) is: 3 s~,α~=Δ∑i=1fλi·Δ-1si,α,
where s~∈S,α~∈-0.5,0.5.
For convenience, si,α is represented by a. Dong et al. (2010) defined the deviation measure between 2-tuple linguistic labels. Then, the following expert consensus degree functions are proposed:
Definition 2
(Consensus degree (CD)) The consensus degree between experts eh and ek with respect to their 2-tuples linguistic preferences on a set of m alternatives with respect a set of n criteria is:4 CDhk=1mn∑i=1m∑j=1nϑaijh,aijk,
where ϑaijh,aijk=1-daijh,aijk=1-Δ-1aijh-Δ-1aijkf+1.
The consensus degree between expert eh and the rest of experts in the group with respect to their preferences on a set of m alternatives with respect a set of n criteria is:5 ACDh=1p-1∑k=1,h≠kpCDhk.
Clustering analysis process based on community detection in SN-GDM
Clustering analysis process (CAP) is an effective method to simplify SN-GDM, since it classifies individuals with similar preferences or structures into the same subgroup (Wu and Xu 2018; Liu et al. 2019; Xu et al. 2019; Zhao et al. 2019; Rashidi et al. 2019; Ding et al. 2020). One key issues of CAP is the determination of subgroups’ weight.
Community detection based CAP
Community detection is an effective method of CAP to uncover the local characteristics of individual behaviors and the correlation between individuals in the network, i.e. to determine the network structure. Therefore, it can effectively detect the relationship between individuals in SN-GDM and then determine network subgroups (Wu et al. 2021). Raghavan et al. (2007) proposed the following label propagation algorithm (LPA) based community detection.
Definition 3
Assume a large network of experts LG=e1,e2,…,eq. The social network relation χhk linking eh and ek exists if their consensus degree CDhk is not lower than a network consensus threshold ε∈0,1. Thus, the network edge between eh and ek is denoted χhk=CDhk. Label propagation algorithm (LPA) main idea is as follows. Initial assignment to each expert a unique label, i.e. , ∀eh∈LGh=1,…,q:eh=lh.
Update the labels of all nodes one by one until reaching the convergence requirement. At each iterative round, the rule for updating labels is as follows: For each node, the label shared by most of its neighbors is assigned. If the most shared label is not unique, choose the label with largest sum of the edge weights of nodes connected to destination node: 6 ∀eh∈LGh=1,…,q:eh=argmaxb∑k∈Nhbχhk,
where Nhb is the set of all nodes labeled b among the neighbors of node h.
Algorithm 1 for the LPA based CAP in SN-GDM is given as follows.
Since only one node needs to be updated each time, the time complexity of each iteration of the LPA is linear Om. The higher the value of the network consensus threshold ε, the more sparse the relationship among nodes is and the greater the number of subgroups NSG detected.
Determining the importance order of subgroups
Algorithm 1 divides the large group into t subgroups SG=SG1,SG2,…,SGt of size PMr=NSGr r∈{1,…,t}. As per (5), the cohesion of subgroup SGr, denoted PCr, is based on the consensus degree of its experts:7 PCr=∑h=1PMrACDrhPMr,
where ACDrh is the consensus degree between expert eh∈SGr and the rest of experts in subgroup SGr.
The power index is obtained as a linear combination of the size and cohesion of the subgroups, and it determines their power relations (the permutation of weights) in the community.
Definition 4
(Power index (PI)) The power index of subgroup SGr r=1,…,t is:8 PISGr=λPMr∑r=1tPMr+1-λPCr∑r=1tPCr,
where λ∈0,1 is a power parameter used to balance the size and cohesion for different LSGDM problems.
The PI values can be used to determine the importance weight of subgroups in LG. Specifically, denoting by σ the permutation such that PISGσr is the r-th highest value of {PISG1,PISG2,…,PISGt}, it is9 PISGσr+1≤PISGσr⇒wσr+1≤wσr.
The consensus reaching process against weight manipulation in SN-GDM
Subgroups consensus measure
At the subgroup level, experts in the same subgroup are regarded as equal important. Meanwhile, each subgroup will be regarded as a new individual with 2-tuples linguistic defined as follows:
Definition 5
(2-tuples linguistic of subgroup and large group) The 2-tuples linguistic of subgroup SGrr=1,2 ,…,t are:10 aijSGr=Δ∑eh∈SGr1RM(r)·Δ-1aijh.
The linguistic 2-tuples of the large group are defined as follows:11 aijLG=Δ∑r=1twσr·Δ-1aijSGσr,
where σ is the permutation defined in (9). The subgroup weights, wσr∈0,1, used in (11), have associated orness value ∑r=1tt-rt-1wσr=α∈[0,1], which is herein called the attitudinal parameter.
Applying (5), the consensus degree between any subgroup and large group, ACDSGrr=1,…,t are computed. A subgroup SGr will be called consistent when its consensus degree is not below the consensus threshold, that is when ACDSGr≥γ; otherwise it is called inconsistent. When all subgroups are consistent, the selection process will be implemented. Otherwise the feedback mechanism will be activated to generate recommendations on the inconsistent elements of the members of the set of inconsistent subgroups: ACDSG={r|ACDSGr<γ}. Noteworthy, the threshold value can be set according to the consistency requirements of different decision-making problems.
Identification of inconsistent elements
Let r∈ACDSG, in other word, SGr is identified as an inconsistent subgroup. The value12 ACEijSGr=1t-1∑s=1,s≠rtϑaijSGr,aijSGs
is the consensus degree between the preferences of the subgroup SGr and the rest of subgroups of the large group at the alternative i and criterion j, which is known as the consensus at the element level. The value13 ACAiSGr=1n∑j=1nACEijSGr
is the consensus degree between the preferences of the subgroup SGr and the rest of subgroups of the large group at the alternative i, which is known as the consensus at the alternative level.
The sets of inconsistent alternatives of inconsistent subgroup SGr is: ACASGr={i|r∈ACDSGr∧ACAiSGr<γ}. The set of inconsistent elements of inconsistent subgroup SGr is: ACESGr={j|r∈ACDSGr∧i∈ACASGr∧ACEijSGr<γ}. Thus, the set of inconsistent elements to consider for feedback recommendation is:14 APS=(r,i,j)|r∈ACDSGr∧i∈ACASGr∧j∈ACESGr.
Advice generation for inconsistent subgroups
For all identified elements r,i,j∈APS, inconsistent subgroups receive advice based on the following rule:
Value aijSGr should be closer to:15 aijSGr~=1-δr·aijSGr+δr·aijLG,
where δr∈0,1 is the feedback parameter for inconsistent subgroup SGr that controls the degree of modification from the original evaluation aijSGr to the collective evaluation aijLG. The feedback parameter is given beforehand by the decision-making moderator.
Notice that ACDSGr is the consensus degree of subgroup SGr before the feedback process, while ACDSGr¯ is the consensus degree of subgroup SGr after the feedback advices to inconsistent groups (15) are implemented. The concept of cost, proposed by Ben-Arieh and Easton (2007), reflects the linear adjustment of the preference of inconsistent individuals required to reach consensus. Thus, the total cost for inconsistent subgroups is expressed as :16 TC=∑(r,i,j)∈APSaijSGr-aijSGr~=∑(r,i,j)∈APSδr·aijSGr-aijLG.
A minimum adjustment and maximum entropy based model to weight prevent manipulation
As aforementioned, experts may strategically set weights for their own benefit in (Dong et al. 2018; Wu et al. 2021a). This article proposes a novel method to prevent weight manipulation in SN-GDM. In detail, the following minimum adjustment feedback mechanism with maximum entropy method (O’Hagan 1988) is established to assign appropriate weights to subgroups.17 minTC=∑(r,i,j)∈APSδr·aijSGr-aijLGs.t.aijLG=∑r=1twr·aijSGraijSGr~=1-δr·aijSGr+δr·aijLGACDSGr<γ∧ACDSGr¯≥γ,r∈ACDSGmaxdisp(W)=-∑c=1twclnwcs.t.∑c=1tt-ct-1wc=α∈0.5,1wc∈0,1∑c=1twc=1
where c=σr.
Applying the Lagrange multiplier technique, the maximum entropy method can be transformed into the following geometric OWA operator proposed by Liu and Chen (2004):18 wc=aqc-1a>0,q≥0,c=1,2,…,t
The detailed derivation process is as follows. Let19 LW,ψ1,ψ2=-∑c=1twclnwc+ψ1wct-ct-1-α+ψ2∑c=1twc-1,
The necessary conditions of the solution are20 ∂L∂wc=-lnwc-1+ψ1t-ct-1+ψ2=0;
21 ∂L∂ψ1=∑i=1twct-ct-1-α=0;
22 ∂L∂ψ2=∑c=1twc-1=0.
From expression (20), we get that wc=eψ1t-ct-1+ψ2-1. Let eψ1=1/μ and t-ct-1-t-c+1t-1=1t-1 then, wcwc+1=μ-1t-1. Obviously, μ-1t-1 is a positive number. Therefore, the MEOWA weights is equal to GOWA weights. Since ∑c=1twc=1, it is23 wc=qc-1∑s=0t-1qs.
Since ∑c=1tt-ct-1wc=α, then q is the solution of the following equation:24 t-1αqt-1+∑c=2tt-1α-c+1qt-c=0.
Thus, optimization model (17) can be rewritten as:25 minTC=∑(r,i,j)∈APSδr·dijs.t.aLGij=∑r=1twr·aSGrijaSGrij~=1-δr·aSGrij+δr·aLGijACDSGr<γ∧ACDSGr¯≥γ,r∈ACDSGaijSGr-aijLG≤dijaijLG-aijSGr≤dijwc=qc-1∑s=0t-1qs,c=1,2,…,tt-1αqt-1+∑c=2tt-1α-c+1qt-c=0∑c=1tt-ct-1wc=α∈0.5,1∑c=1twc=1
Noteworthy, any feasible solution of model (25) that verifies dij>aijSGr-aijLG is not a solution of model (17). Thus, only the solutions of (25) that verifies dij=aijSGr-aijLG are solutions of model (17).
The detailed process of the proposed consensus model is shown in Algorithm 2.
The detailed execution flow charts of algorithm 1 and algorithm 2 are shown in Fig. 2a, b.Fig. 2 The flow chart of algorithms 1-2
The following properties guarantee that the subgroup consensus degrees are increasing and bounded above.
Proposition 1
For inconsistent subgroup SGrr∈APS, it isACDSGr¯≥ACDSGr.
Proof
Based on (5), it is26 ACDSGr=1t-1mn∑z=1,z≠rt∑i=1m∑j=1n1-aijr-aijz.
For simplify, aijSGr is denoted as aijr. From (15), ACDSGr¯ can be split into two valuesACDSGr¯=ACD1SGr¯+ACD2SGr¯,
computed using the set of inconsistent elements aijr1~(r,i,j)∈APS and consistent elements aijr2r,i,j∉ APS. Noteworthy, z∈int1,t consists of s and ro. When s∈int1,g in ACD1SGr¯, the consensus degree between the inconsistent subgroup SGr and other consistent subgroups in the large group after feedback process is given in (27) (at the top of the page 13). Notice that aijr=aijr1∪aijr2 and aijs=aijs1∪aijs2. While when ro∈intg+1,t in ACD2SGr¯, the consensus degree between the inconsistent subgroup SGr and other inconsistent subgroups in the large group after feedback process is given in (28) (at the top of the page 13) In this case, aijr=aijr1∪aijr2∪aijr3∪aijr4, and aijro=aijro1∪aijro2∪aijro3∪aijro4. Obviously, since δr and wr in the interval 0,1, so 1-δr1-wr∈0,1 in ACD1SGr¯. Additionally, since δr, δro and wr are in the interval 0,1, so 1-δr1-wr+δro·wr∈0,1, 1-δro1-wr∈0,1 and δr1-wr-1∈0,1 in ACD2SGr¯. Therefore, we get ACDSGr≤ACDSGr¯.27 ACD1SGr¯=1t-1mn∑s=1g∑i=1m∑j=1n1-aijr1~-aijs1+aijr2-aijs2=1t-1mn∑s=1g∑i=1m∑j=1n1-1-δr1-wr·aijr1-aijs1+aijr2-aijs2,
28 ACD2SGr¯=1t-1mn∑ro=g+1,ro≠rt×∑i=1m∑j=1n1-aijr1~-aijro1~+aijr2-aijro2~+aijr3~-aijro3+aijr4-aijro4=1t-1mn∑ro=g+1,ro≠rt∑i=1m∑j=1n1-1-δr1-wr+δro·wr·aijr1-aijro1+1-δro1-wr·aijr2-aijro2+δr1-wr-1·aijr3-aijro3+aijr4-aijro4.
29 ACDSGs=ACD1SGs+ACD2SGs=1t-1mn∑z′=1,z′≠st∑i=1m∑j=1n1-aijs-aijz′.
30 ACD1SGs¯=1t-1mn∑r=g+1t∑i=1m∑j=1n1-aijs1-aijr1~+aijs2-aijr2=1t-1mn∑r=g+1t∑i=1m∑j=1n1-δr1-wr-1·aijs1-aijr1+aijs2-aijr2,
31 ACD2SGs¯=1t-1mn∑so=1,so≠sg∑i=1m∑j=1n1-aijs-aijso.
□
Proposition 2
For consistent subgroup SGss∉APS, it isACDSGs¯≥ACDSGs.
Proof
Similar to Proposition 1, based on (5) it is (29) (at the top of the page 13), where z′∈int1,t consists of r and so. Then, when r∈intg+1,t in ACD1SGs¯ it is (30) (at the top of the page 13), while when so∈int1,g in ACD2SGs¯ it is (31) (at the top of the page 13). Since δr and wr in the interval 0,1, so δr1-wr-1∈0,1 in ACD1SGs¯. Therefore, we get ACDSGs≤ACDSGs¯. □
Numerical and simulation analysis
This section introduces a numerical and a simulation analysis to verify the effectiveness of the proposed model. With the frequent occurrence of emergencies in the world, large emergency group decision-making (LEGDM) has become a hot research issue. Generally, emergency decision-making problem is time-sensitive. In order to allocate resources efficiently and reduce losses, the location of emergency medical facility is a key issue in LEGDM. In response to emergencies brought about by COVID-19, many module hospitals have been established across China. Assume a modular hospital to be built in Lingang New Area, Shanghai, China. After pre evaluation, four emergency facilities M1,M2,M3,M4 have remained as alternatives for further evaluation. This paper collects the preference information of twenty experts ehh=1,…,20 from college emergency departments, hospital emergency departments and government departments through a questionnaire survey with respect to three decision criteria: C1: Geographical factor; C2: Traffic convenience; C3: Safety factor. The linguistic term set (LTS) for judging the location with regard to the three criteria is:S=s0=verybadVB,s1=badB,s2=littlebadLB,s3=mediumMs4=littlegoodLG,s5=goodG,s6=verygoodVG.
Numerical analysis
The twenty experts’ preferences are provided as follows. A1=C1C2C3M1s4,0s3,0s2,0M2s4,0s4,0s4,0M3s5,0s2,0s2,0M4s3,0s2,0s5,0A2=C1C2C3M1s1,0s1,0s5,0M2s4,0s0,0s6,0M3s2,0s4,0s3,0M4s0,0s5,0s1,0A3=C1C2C3M1s5,0s1,0s3,0M2s2,0s3,0s2,0M3s4,0s4,0s0,0M4s5,0s2,0s4,0A4=C1C2C3M1s2,0s4,0s1,0M2s2,0s4,0s0,0M3s4,0s2,0s2,0M4s2,0s2,0s6,0A5=C1C2C3M1s5,0s2,0s5,0M2s3,0s3,0s2,0M3s1,0s4,0s4,0M4s3,0s3,0s2,0A6=C1C2C3M1s3,0s4,0s5,0M2s2,0s2,0s4,0M3s4,0s3,0s0,0M4s4,0s0,0s6,0A7=C1C2C3M1s2,0s4,0s1,0M2s3,0s3,0s0,0M3s5,0s2,0s4,0M4s4,0s0,0s6,0A8=C1C2C3M1s2,0s4,0s1,0M2s5,0s1,0s3,0M3s4,0s3,0s1,0M4s6,0s3,0s4,0A9=C1C2C3M1s1,0s1,0s4,0M2s4,0s2,0s4,0M3s5,0s5,0s3,0M4s1,0s5,0s2,0A10=C1C2C3M1s1,0s4,0s5,0M2s4,0s0,0s4,0M3s3,0s4,0s1,0M4s3,0s2,0s2,0A11=C1C2C3M1s3,0s6,0s1,0M2s4,0s2,0s4,0M3s2,0s0,0s3,0M4s5,0s2,0s4,0A12=C1C2C3M1s2,0s4,0s3,0M2s2,0s4,0s2,0M3s5,0s5,0s4,0M4s3,0s2,0s1,0A13=C1C2C3M1s3,0s0,0s4,0M2s2,0s3,0s2,0M3s5,0s4,0s2,0M4s5,0s1,0s4,0A14=C1C2C3M1s4,0s5,0s3,0M2s2,0s4,0s0,0M3s4,0s2,0s1,0M4s6,0s1,0s4,0A15=C1C2C3M1s4,0s2,0s5,0M2s4,0s3,0s2,0M3s5,0s5,0s2,0M4s2,0s1,0s4,0
A16=C1C2C3M1s1,0s4,0s3,0M2s4,0s2,0s4,0M3s2,0s6,0s4,0M4s2,0s2,0s5,0A17=C1C2C3M1s1,0s1,0s4,0M2s3,0s4,0s5,0M3s1,0s4,0s5,0M4s2,0s4,0s4,0A18=C1C2C3M1s5,0s4,0s3,0M2s2,0s2,0s0,0M3s4,0s1,0s2,0M4s3,0s5,0s2,0A19=C1C2C3M1s6,0s4,0s1,0M2s5,0s4,0s4,0M3s3,0s3,0s4,0M4s3,0s2,0s4,0A20=C1C2C3M1s3,0s5,0s2,0M2s4,0s1,0s4,0M3s2,0s2,0s0,0M4s3,0s5,0s2,0 The social matrix SM=CDhkh,k=1,...q is constructed via (4):SM=00.643...0.7740.810...0.8690.7980.6430...0.7810.726...0.6310.750......0...............0.7740.781...00.750...0.7620.8330.8100.726...0.7500...0.8210.821...............0......0.8690.631...0.7620.821...00.7620.7980.750...0.8330.821...0.7620
Assuming a network consensus threshold of ε=0.8, the large network before CAP is established in Fig. 3(a). Applying Algorithm 1, the large network is divided into four subgroups, which are depicted in Fig. 3(b):SG1=e1,e4,e7,e14,e18,SG2=e2,e9,e10,e16,e17;SG3=e3,e5,e6,e12,e13,e15,SG4=e8,e11,e19,e20.
Assuming the power parameter λ=0.5 that considers size and the cohesion equally important criteria to measure the power relationship of subgroups, the RI of the four subgroup are computed based on (5) and (7)–(8).Subgroup SG1: ACD11=0.822; ACD14=0.848; ACD17=0.81; ACD114=0.821; ACD118=0.801; PM1=5 and PC1=0.821.
Subgroup SG2: ACD22=0.813; ACD29=0.827; ACD210=0.804; ACD216=0.804; ACD217=0.801; PM2=5 and PC2=0.81.
Subgroup SG3: ACD33=0.829; ACD35=0.79; ACD36=0.771; ACD312=0.79; ACD313=0.838; ACD315=0.829; PM3=6 and PC3=0.808.
Subgroup SG4: ACD48=0.813; ACD411=0.821; ACD419=0.798; ACD420=0.798; PM4=4 and PC4=0.808.
The power indices are:PISG1=0.252,PISG2=0.25,PISG3=0.274,PISG4=0.224.
Then, the power relations of subgroups are: w3≥w1≥w2≥w4.
where wσ1=w3,wσ2=w1,wσ3=w2,wσ4=w4 are obtained from expression (9). 2. Applying (10) the preference of four subgroups are: ASG1=C1C2C3M1s3,0.4s4,0s2,0M2s3,-0.4s3,0.4s1,-0.2M3s4,0.4s2,-0.2s2,0.2M4s4,-0.4s2,0s5,-0.4ASG2=C1C2C3M1s1,0s2,0.2s4,0.2M2s4,-0.2s2,-0.4s5,-0.4M3s3,-0.4s5,-0.4s3,0.2M4s2,-0.4s4,-0.4s3,-0.2ASG3=C1C2C3M1s4,-0.33s2,0.17s4,0.17M2s3,-0.5s3,0s2,0.33M3s4,0s4,0.17s2,0M4s4,-0.33s2,-0.5s4,-0.5ASG4=C1C2C3M1s4,-0.5s5,-0.25s1,0.25M2s5,-0.5s2,0s4,-0.25M3s3,-0.25s2,0s2,0M4s4,0.25s3,0s4,-0.5
The consensus degrees between each subgroup and the large group are:ACDSG1=0.81,ACDSG2=0.771,ACDSG3=0.833,ACDSG4=0.817.
3. With γ=0.8, the set of inconsistent elements is APS=2,1,1,2,1,2,2,1,3,2,2,3,2,4,1,2,4,2.
4. Implementing the feedback parameter δ2=0.3, the corresponding model (25) becomes: 32 min∑(r,i,j)∈APS0.3×dijs.t.aijLG=∑r=14wr·aijSGrACDSG2<0.8∧ACDSG2¯≥0.8aijSG2-aijLG≤dijaijLG-aijSG2≤dijwσr=qσr-11+q+q2+q3,σr=1,2,3,43αq3+3αq2-q2+3αq-2q+3α-3=00.5≤wσ1+23wσ2+13wσ3=α<1wσ1+wσ2+wσ3+wσ4=1
The solution of model (32) results in the following subgroup weights: w1=0.277,w2=0.177,w3=0.434,w4=0.113; with associated attitudinal parameter α=0.676, and MTC=0.391. And we have the feedback mechanism advices for SG2: Value a11SG2 should be closer to s2,-0.37.
Value a12SG2 should be closer to s2,0.43.
Value a13SG2 should be closer to s4,-0.09.
Value a23SG2 should be closer to s4,-0.04.
Value a41SG2 should be closer to s2,0.12.
Value a42SG2 should be closer to s3,0.17.
If SG2 accepts and implements the advice, the new decision matrix for such subgroup will be: Aδ2=0.3SG2=s2,-0.37s2,0.43s4,-0.09s4,-0.2s2,-0.4s4,-0.04s3,-0.4s5,-0.4s3,0.2s2,0.12s3,0.17s3,-0.2
The new ACDs after the feedback process would be: ACDSG1¯=0.821,ACDSG2¯=0.8,ACDSG3¯=0.84,ACDSG4¯=0.828.
Subgroup SG2 reaches the consensus threshold γ=0.8.
Fig. 3 The large network before and after CAP
Analysis of preventing manipulation behavior
Let PI=PISG1,…,PISGt be the set of subgroup PI values. Then the following attitude-OWA (AOWA) operator based on maximum entropy method (Yager 1988; O’Hagan 1988) is used to determine the dynamic subgroups weighting vector W=w1,…,wt:33 maxdisp(W)=-∑c=1twclnwcs.t.∑c=1tt-ct-1wc=α∈0.5,1wc∈0,1∑c=1twc=1
If α=0.5, from expression (33), we get weights w=1n,1n,⋯,1n, named as ‘egalitarianism’ weighting vector. If α=1, then w=1,0,…,0, named as ‘authoritarianism’ weighting vector. Since the power is too concentrated when α is too large, the value of α considered in this paper is not higher than 0.95. The manipulation behavior of subgroups from ‘egalitarianism’ to ‘authoritarianism’ is reflected by dynamic weights in Fig. 4.Fig. 4 The manipulation behavior of subgroups
To verify the prevention manipulation of the proposed model, we simulate the changes in cost with different attitudinal parameter α∈0.6,0.7,0.8,0.9 with δ2=0.3. To do that, we establish model (34), which reflects the weight manipulation of subgroups.34 maxdisp(W)=-∑c=1twclnwcs.t.aijLG=∑r=1twraijSGrACDSGr<γ∧ACDSGr¯≥γ,r∈ACDSG∑c=1tt-ct-1wc=α∈0.5,1wc∈0,1c=σr∑c=1twc=1
In model (34), the objective function has only maximum entropy and the attitudinal parameter is a variable. Then TC=∑(r,i,j)∈APSδr·aijSGr-aijLG can be calculated. In other words, model (34) maximises entropy with known attitudinal parameter. While model (17) obtains the attitudinal parameter by minimum TC. Noteworthy, when α∈0.5,0.676, the LG cannot reach consensus after one round of feedback. The model will have an optimal solution only after the inconsistent subgroup accept two rounds of advice, which means the ACDSGr¯ should be changed to ACDSGr¯¯ and TC will beTC2=TC1+TC2=∑r,i,j∈APSδr·aijSGr-aijLG+aijSGr~-aijLG~.
Figure 5 illustrates the results of TC in Table 2 obtained with weight manipulation (curve) by model (34) and prevention weight manipulation (blue dot) by model (32), respectively. As previously pointed out, subgroups may have weight manipulative behaviors. The reason for weight manipulation is that each subgroup wants to increase its own weight in LSGDM. When the subgroups with large discourse power (PI) are strong, they hope that the weights are concentrated on themselves, as shown in the curve on the right side of the blue dot in Figure 5. When the subgroups with large discourse power are not very strong, the distribution of weights will be more even, as shown in the curve on the left of the blue dots. But weight manipulative behavior will hinder consensus efficiency to a certain extent (Fig. 6). To this end, we designed an anti-manipulation model (17). Experimental results show that the TC value obtained with our proposed model (32) is the lowest. More specifically, the curve to the left of the blue dot in Fig.5 reflects the situation based on ‘justice’ policy, where the weight distribution is relatively balanced but requires more cost due to multiple rounds of feedback, so the efficiency is relatively low. While the curve to the right of the blue dot reflects the situation based on the efficiency policy, where the cost is relatively low, but the weight distribution is relatively unreasonable. Therefore, our policy is to ensure the justice of the weight distribution as much as possible under the condition of high efficiency.Table 2 TC with weight manipulation under different attitudinal parameter α
α Subgroups’ weights TC Maximum entropy ACDSG2¯
0.6 (0.272, 0.213, 0.347, 0.167) 0.677 1.35 0.817
0.7 (0.276, 0.165, 0.461, 0.098) 0.394 1.24 0.801
0.8 (0.252, 0.106, 0.596, 0.045) 0.405 1.03 0.802
0.9 (0.182, 0.043, 0.764, 0.01) 0.411 0.7 0.804
Model (32) (0.277, 0.177, 0.432, 0.114) 0.391 1.27 0.8
Fig. 5 The TC under weight manipulation and prevention weight manipulation
Sensitivity analysis
In this subsection, we simulate the change of feedback parameter in the interval 0,1 to analyze anti-manipulation behavior. Specifically, we randomly generate the feedback parameters of inconsistent experts from 0 to 1, and then substitute these parameters into model (25). By solving the minimum cost and subgroups’ weights, the large group can reach consensus by coordinating the ‘efficiency’ policy and ‘justice’ policy. To do that, the following three cases are considered: Case 1. When the value of feedback parameter is small, one round of feedback may not reach consensus, that is, the feedback parameter δr for inconsistent subgroup SGr in the interval 0,δrlow is required more iterations to reach consensus.
Case 2. When the value of feedback parameter δr for inconsistent subgroup SGr in the interval δrhigh,1, the LG can reach consensus after one round of feedback.
Case 3. While, when the feedback parameter is large enough for Case 2, there may be a set of fixed weights with the minimum cost, which is mainly due to the preferences are over-adjusted. Specifically, if the feedback parameter δr in the interval δrhigh,1,δrlow≤δrhigh, subgroups’ weights could be fixed values.
Noteworthy, due to Case 1 requires multiple rounds of iterations to reach consensus, it will incur a time cost, which is undesirable in the feedback mechanism based on minimum cost. Therefore, the model proposed in this article is mainly to analyze Case 2 and 3. The attitude parameter α is utilized to reflect the distribution of subgroups’ weights.
Randomly generate feedback parameters δ2∈0,1, from model (25), we get that when 0≤δ2<0.26, the model has no feasible solution, which is applicable to Case 1. Thus, this paper suggests that the value of the feedback parameter should be not lower than 0.26. When 0.26≤δ2≤1, the model has the optimal solution, which is applicable to Case 2. Specifically, when 0.26≤δ2≤0.335, the LG can reach consensus by different attitudinal parameter in the interval 0.5,1 with minimum cost. while when 0.335<δ2≤1, the LG can can reach consensus by the same attitudinal parameter α=0.5 with minimum cost, which is applicable to Case 3. This can be interpreted as adequate weight should be given when the inconsistent experts are fully willing to modify preferences. A visual simulation of the minimum cost and attitude parameters change with the feedback parameters is shown in Fig. 4.Fig. 6 Simulation of the minimum cost and attitude parameters change with the feedback parameters of inconsistent subgroup SG2 after feedback
The blue and red lines in Fig. 4 reflect the conditions after feedback of the feedback parameters in the interval 0.26≤δ2≤0.335 and 0.335<δ2≤1, respectively. The result of the blue line shows that when the weight distribution tends to be justice, the efficiency of CRP will decrease (due to the cost increase). The red line shows when the preferences are over-adjusted (ignore efficiency), the proposed model will give priority to the justice of weight distribution. Therefore, to better retain the original preferences of inconsistent subgroups, the feedback parameter δ2 should be selected in the interval 0.26,0,335 as the alternatives for inconsistent subgroup SG2, so that the LG can reach consensus by coordinating the ‘efficiency’ policy and ‘justice’ policy.
Ranking order of alternatives
Without loss of generality, it is assumed that the attribute weights of criteria are: ω=C1=1/3;C2=1/3; C3=1/3T. After the feedback process of the numerical analysis, we get the consensual collective decision matrix with α=0.676:ALG=C1C2C3M1s3,0.1s3,-0.03s3,0.24M2s3,-0.01s3,-0.25s2,0.47M3s4,-0.28s3,0.34s2,0.27M4s3,0.35s2,0.18s4,-0.32
Therefore, the overall consensual preference value of the four alternatives M1,M2,M3,M4 are:M1=s3,0.1,M2=s3,-0.26,M3=s3,0.11,M4=s3,0.07.
The final consensus ranking of alternatives is:M3≻M1≻M4≻M2.
In addition, we compared the rankings of the alternatives under different policy, as shown in Table 3. The result shows that the final alternative ranking result based on justice policy is different from our model, while the efficiency-based alternative ranking result is consistent with our model. As can be seen from Table 2, the weights obtained by the proposed method in this paper are fairer than the method based on the efficiency policy (the curve to the right of the blue dot in Fig. 5). Therefore, it can be verified that the model proposed in this article not only guarantees efficiency, but also guarantees justice as much as possible.Table 3 Rankings of the alternatives with different policy
α Type of policy Rankings of the alternatives
0.6 Justice M1≻M3≻M4≻M2.
0.9 Efficiency M3≻M1≻M4≻M2.
Model (32) both justice and efficiency M3≻M1≻M4≻M2.
Comparative analysis
To clearly differentiate between the proposed method and other methods on weight manipulation behavior, this section compares with other literatures from a theoretical perspective. The characteristic comparisons of our method with other methods are shown in Table 4. To simplify notion, weight manipulation behavior is denoted by WMB.Table 4 The comparative analysis among different models of weight manipulation behavior
References Research object Research perspective of WMB Prevent WMB
Yager (2001) Individual weight Selection process Yes
Yager (2002) Individual weight Selection process Yes
Dong et al. (2018) Attribute weight Selection process No
Liu et al. (2019a) Attribute weight Selection process No
Dong er al. (2021) Trust relationship Selection process No
Wu et al. (2021a) Individual weight Consensus process Yes
The proposed method Subgroup weight Consensus process Yes
In the aspect of research object: The manipulated objects mainly include attribute weight (Dong et al. 2018; Liu et al. 2019a), individual weight (Yager 2001; 2002; Wu et al. 2021a) and trust relationship (Its essence is to get more individual weight) (Dong et al. 2021). Attribute weight manipulation mainly affects the ranking of the alternatives. While the individual weight manipulation will not only affect the final ranking of the alternatives, but also affect the aggregating of individual preference.
In the aspect of research perspective: Weight manipulation behavior mainly exists in the consensus process and selection process in GDM. The purpose of attribute and individual weight manipulation in the selection process is mainly to obtain a desired ranking of alternatives by weight manipulation (Yager 2001; 2002; Dong et al. 2018, 2021; Liu et al. 2019a). But individual weight manipulation can also have an impact on the consensus process. The reason for this effect is that the weight of the individual will affect the aggregation preference of group and feedback mechanisms often suggest discordant individuals to make adjustments based on group preferences. Therefore, individual weight manipulation behavior may affect the efficiency of consensus (Wu et al. 2021a).
In the aspect of preventing manipulation: Although the behavior of weight manipulation can help decision makers to obtain the ideal ranking of alternative to a certain extent, it also has some adverse effects, such as unreasonable decision results or low consensus efficiency. To this end, Yager (2001; 2002) proposed a mechanism for modifying the construction of the group decision function to prevent weight manipulation behavior, but it doesn’t take into account consensus issues. Therefore, Wu et al. (2021a) proposed an efficiency policy-based mechanism to prevent individual weight manipulation behavior in CRP.
Here are some advantages of our methods compared with the above literatures: Compared with Yager (2001; 2002) and Wu et al. (2021a), we extend the research of weight manipulation from individual to subgroup. In general, subgroup behavior is more complex than individual behavior, so we define a power index to determine the importance ranking of subgroups, which provides a new research paradigm for weight manipulation behavior in LSGDM.
Most of articles focuses on the effect of weight manipulation behavior on ranking of alternatives in selection process (Yager 2001; 2002; Dong et al. 2018,2021; Liu et al. 2019a), while this paper mainly focus on the effect of manipulation behavior on consensus process, which is similar to Wu et al. (2021a). In contrast with method of Wu et al., in the design of the anti-manipulation model, we not only focus on the efficiency of CRP, but also consider the justice of weight distribution.
Discussion
In GDM, each decision maker wants to obtain greater importance degree (weight) for more benefits. For example, a decision maker with the lowest status is willing to adopt egalitarianism to assign weight. While a decision maker with the highest status expects to use authoritarianism to distribute weight. This behavior will increase the cost (efficiency) of the feedback process to a certain extent and hinder consensus. The existing studies on preventing weight manipulation behaviors have focused on group consensus efficiency: how to minimize interaction costs. However, these mechanisms often ignore the justice of weight distribution. In this article, we define a more reasonable policy to prevent weight manipulation by combining ‘justice’ and ‘efficiency’ simultaneously, which can achieve a balance between individual benefit and group goal.
Actually, under situation of weight manipulation, the total cost changes with the behavior (attitude parameters). While only using the maximum entropy method to assign weights cannot know how to assign weights appropriately because ‘egalitarianism’ or ‘authoritarianism’ policies are not the optimal choice, which is reflected by dynamic weights in Figs. (4 and 5). Therefore, it is more reasonable to combine the maximum entropy model and the minimum cost model as limiting conditions to assign weights, where (1): Entropy is the largest, which solves the problem of ‘justice’. (2): The feedback total cost is minimal, which solves the problem of ‘efficiency’. If all two points can be achieved, the weights of subgroups are not manipulated in our proposed model (17). Besides, we demonstrate that subgroup weight manipulation does also affect the final ranking of alternatives. We compare the effect of weight manipulative on the results in Table 3.
Conclusion
In SN-GDM, subgroups may exist weight manipulation behavior for certain benefit. This paper focuses on the impact of subgroup weight manipulation behavior on CRP. A method to prevent the weight manipulation in LSGDM problem by combining minimum adjustment and maximum entropy is investigated. The main contributions of the article are as follows. (i) It develops a new research paradigm to study the subgroup manipulation behavior in the process of preference aggregation by analyzing the network structure relationship of subgroups. To do that, a LPA based community detection method is introduced to cluster the large group into several subgroups. Then, a power index is defined to obtain the power relation of subgroup. An attitude-OWA based on maximum entropy method is introduced to simulate subgroups’ manipulation behavior from ‘egalitarianism’ to ‘authoritarianism’.
(ii) Considering that weight manipulation behavior may hinder the efficiency of consensus, it investigates a method to prevent subgroup weight manipulation and facility the convergence of consensus with minimum cost. So, a minimum adjustment feedback mechanism based on maximum entropy method is established to assign reasonable weights for subgroups. The minimum adjustment is used for ‘efficiency’ rule while the maximum entropy is used for ‘justice’ rule. Compared with the method of Wu et al. (2021a), the proposed method has the advantage of considering the ‘justice’ of the weight distribution.
The purposed method also has some problems that need to be further study. (i) This paper treats the subgroup as a whole during the feedback process, so the differential feedback within the subgroup is not considered. In addition, the selection of parameters such as consensus threshold, network consensus threshold and Power parameter in this paper is random. But in the actual decision-making environment, the choice of these parameters are often related to human behavior (Wu et al 2021, Sun et al 2021). Therefore, it is necessary to conduct further research on them.
(ii) This paper only considers weight manipulation, although there are also other types of manipulation behaviors that deserve more in-depth research. For example, the existing feedback mechanisms usually assume that inconsistent experts are willing to accept feedback suggestions with fixed feedback parameters, which means, more frequent than not, that more adjustment costs are incurred for them than necessary, thereby reducing the independence of inconsistent experts. This type of manipulative behavior is called as group manipulation behavior (Wu et al. 2021a).
It could be an interesting future research topic to discuss the trust relationship of social networks in SN-GDM when studying group manipulation behavior. The other research direction is to apply the proposed method to group recommender systems for social items, such as education recommendation, travel packages and TV shows, which tend to be consumed by groups rather than individuals. A key issue involved in group recommender systems is the consensus reaching process, which inevitably requires consideration of ‘justice’ and ‘efficiency’. Furthermore, a social network-based decision support system can be developed based on the proposed method.
Acknowledgements
This work was supported by National Natural Science Foundation of China (NSFC) (Nos. 71971135, 71910107002), Shanghai Foreign Experts Program (No. 22WZ2506100), Innovative Talent Training Project of Graduate Students in Shanghai Maritime University of China (No. 2019YBR017) and was supported in part by the Spanish State Research Agency under Project PID2019-103880RB-I00/AEI/10.13039/501100011033.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Amirkhani A Barshooi AH Consensus in multi-agent systems: a review Artif Intell Rev 2022 55 3897 3935 10.1007/s10462-021-10097-x
Ben-Arieh D Easton T Multi-criteria group consensus under linear cost opinion elasticity Decis Support Syst 2007 43 713 721 10.1016/j.dss.2006.11.009
Cabrerizo FJ Ureña R Pedrycz W Herrera-Viedma E Building consensus in group decision making with an allocation of information granularity Fuzzy Sets Syst 2014 255 115 127 10.1016/j.fss.2014.03.016
Cao MS Wu J Chiclana F A personalized feedback mechanism based on maximum harmony degree for consensus in group decision making IEEE Trans Syst Man Cybern Syst 2021 51 6134 6146 10.1109/TSMC.2019.2960052
Cao MS Liu YJ Gai TT A comprehensive star rating approach for cruise ships based on interactive group decision making with personalized individual semantics J Mar Sci Eng 2022 10 638 10.3390/jmse10050638
Chao XR Kou G Peng Y Herrera-Viedma E Large-scale group decision-making with non-cooperative behaviors and heterogeneous preferences: An application in financial inclusion Eur J Oper Res 2021 288 271 293 10.1016/j.ejor.2020.05.047
Chen X Ding ZG Dong YC Managing consensus with minimum adjustments in group decision making with opinions evolution IEEE Trans Syst Man Cybern Syst 2021 51 2299 2311 10.1109/TSMC.2019.2912231
Del Moral MJ Chiclana F Tapia JM Herrera-Viedma E A comparative study on consensus measures in group decision making Int J Intell Syst 2018 33 1624 1638 10.1002/int.21954
Ding RX Palomares I Wang XQ Large-scale decision-making: characterization, taxonomy, challenges and future directions from an artificial intelligence and applications perspective Inf Fusion 2020 59 84 102 10.1016/j.inffus.2020.01.006
Dong QX Zhü KY Copper O Gaining consensus in a moderated group: a model with a twofold feedback mechanism Expert Syst Appl 2017 71 87 97 10.1016/j.eswa.2016.11.020
Dong YC Xu YF Li HY Feng B The OWA-based consensus operator under linguistic representation models using position indexes Eur J Oper Res 2010 203 455 463 10.1016/j.ejor.2009.08.013
Dong YC Liu YT Liang HM Strategic weight manipulation in multiple attribute decision making Omega 2018 75 1339 1351 10.1016/j.omega.2017.02.008
Dong YC Zha QB Zhang HJ Herrera F Consensus reaching and strategic manipulation in group decision making with trust relationships IEEE Trans Syst Man Cybern Syst 2021 51 6304 6318 10.1109/TSMC.2019.2961752
Gong ZW Zhang N Chiclana F The optimization ordering model for intuitionistic fuzzy preference relations with utility functions Knowl Based Syst 2018 162 174 184 10.1016/j.knosys.2018.07.012
Gong ZW Xu XX Guo WW Minimum cost consensus modelling under various linear uncertain-constrained scenarios Inf Fusion 2021 66 1 17 10.1016/j.inffus.2020.08.015
Herrera F Martínez L A 2-tuple fuzzy linguistic representation model for computing with words IEEE Trans Fuzzy Syst 2000 8 746 752 10.1109/91.890332
Herrera-Viedma E Palomares I Li CC Revisiting fuzzy and linguistic decision making: scenarios and challenges for making wiser decisions in a better way IEEE Trans Syst Man Cybern Syst 2021 51 191 208 10.1109/TSMC.2020.3043016
Li GX Kou G Peng Y A group decision making model for integrating heterogeneous information IEEE Trans Syst Man Cybern Syst 2018 48 982 992 10.1109/TSMC.2016.2627050
Li CC Gao Y Dong YC Managing ignorance elements and personalized individual semantics under incomplete linguistic distribution context in group decision making Gr Decis Negot 2021 30 97 118 10.1007/s10726-020-09708-9
Li GX Kou G Peng Y Heterogeneous large-scale group decision making using fuzzy cluster analysis and its application to emergency response plan selection IEEE Trans Syst Man Cybern Syst 2022 52 3391 3403 10.1109/TSMC.2021.3068759
Liu XW Chen LH On the properties of parametric geometric OWA operator Int J Approx Reason 2004 35 163 178 10.1016/j.ijar.2003.09.001
Liu X Xu YJ Montes R Herrera F Social network group decision making: managing self-confidence-based consensus model with the dynamic importance degree of experts and trust-based feedback mechanism Inf Sci 2019 505 215 232 10.1016/j.ins.2019.07.050
Liu YT Dong YC Liang HM Multiple attribute strategic weight manipulation with minimum cost in a group decision making context with interval attribute weights information IEEE Trans Syst Man Cybern Syst 2019 49 1981 1992 10.1109/TSMC.2018.2874942
Liu YT Zhang HJ Wu YZ Dong YC Ranking range based approach to madm under incomplete context and its application in venture investment evaluation Technol Econ Dev Econ 2019 25 877 899 10.3846/tede.2019.10296
Liu JC Sheu JB Li DF Dai YW Collaborative profit allocation schemes for logistics enterprise coalitions with incomplete information Omega 2021 101 102237 10.1016/j.omega.2020.102237
Mandal P Samanta S Pal M Ranadive AS Three-way decision model under a large-scale group decision-making environment with detecting and managing non-cooperative behaviors in consensus reaching process Artif Intell Rev 2022 10.1007/s10462-021-10133-w
O’Hagan M Aggregating template or rule antecedents in real-time expert systems with fuzzy set logic Asilomar Conf Signals IEEE Comput Soc 1988 10.1109/ACSSC.1988.754637
Pérez IJ Cabrerizo FJ Alonso S On dynamic consensus processes in group decision making problems Inf Sci 2018 459 20 35 10.1016/j.ins.2018.05.017
Raghavan UN Albert R Kumara S Near linear time algorithm to detect community structures in large-scale networks Phys Rev E 2007 76 1 11 10.1103/PhysRevE.76.036106
Rashidi F Nejatian S Parvin H Rezaie V Diversity based cluster weighting in cluster ensemble: an information theory approach Artif Intell Rev 2019 52 1341 1368 10.1007/s10462-019-09701-y
Rodríguez RM Labella Á De Tré G Martnez L A large scale consensus reaching process managing group hesitation Knowl Based Syst 2018 159 86 97 10.1016/j.knosys.2018.06.009
Sun Q Wu J Chiclana F A dynamic feedback mechanism with attitudinal consensus threshold for minimum adjustment cost in group decision making IEEE Trans Fuzzy Syst 2021 30 1287 1301 10.1109/TFUZZ.2021.3057705
Tang M Liao HC From conventional group decision making to large-scale group decision making: what are the challenges and how to meet them in big data era? A state-of-the-art survey Omega 2021 100 102141 10.1016/j.omega.2019.102141
Tian ZP Nie RX Wang JQ Long RY Adaptive consensus-based model for heterogeneous large-scale group decision making: detecting and managing non-cooperative behaviors IEEE Trans Fuzzy Syst 2021 29 2209 2223 10.1109/TFUZZ.2020.2995229
Wan QF Xu XH Zhuang J Pan B A sentiment analysis-based expert weight determination method for large-scale group decision-making driven by social media data Expert Syst Appl 2021 185 115629 10.1016/j.eswa.2021.115629
Wang S Wu J Chiclana F Two stage feedback mechanism with different power structures for consensus in large-scale group decision-making IEEE Trans Fuzzy Syst 2022 10.1109/TFUZZ.2022.3144536
Wang Z Xiao FY Cao ZH Uncertainty measurements for Pythagorean fuzzy set and their applications in multiple-criteria decision making Soft Comput 2022 10.1007/s00500-022-07361-9
Wu J Chiclana F A social network analysis trust-consensus based approach to group decision-making problems with interval-valued fuzzy reciprocal preference relations Knowl Based Syst 2014 59 97 107 10.1016/j.knosys.2014.01.017
Wu ZB Xu JP A consensus model for large-scale group decision making with hesitant fuzzy information and changeable clusters Inf Fusion 2018 41 217 231 10.1016/j.inffus.2017.09.011
Wu J Chiclana F Herrera-Viedma E Trust based consensus model for social network in an incomplete linguistic information context Appl Soft Comput J 2015 35 827 839 10.1016/j.asoc.2015.02.023
Wu J Chiclana F Fujita H Herrera-Viedma E A visual interaction consensus model for social network group decision making with trust propagation Knowl Based Syst 2017 122 39 50 10.1016/j.knosys.2017.01.031
Wu ZB Jin BM Fujita H Xu JP Consensus analysis for AHP multiplicative preference relations based on consistency control: a heuristic approach Knowl Based Syst 2020 191 105317 10.1016/j.knosys.2019.105317
Wu J Cao MS Chiclana F An optimal feedback model to prevent manipulation behaviour in consensus under social network group decision making IEEE Trans Fuzzy Syst 2021 29 1750 1763 10.1109/TFUZZ.2020.2985331
Wu J Wang S Chiclana F Herrera-Viedma E Twofold personalized feedback mechanism for social network consensus by uninorm interval trust propagation IEEE Trans Cybern 2021 1 1 12 10.1109/TCYB.2021.3076420
Wu T Liu XW Qin JD Herrera F Balance dynamic clustering analysis and consensus reaching process with consensus evolution networks in large-scale group decision making IEEE Trans Fuzzy Syst 2021 29 357 371 10.1109/TFUZZ.2019.2953602
Wu J Chen J Liu W A calibrated individual semantic based failure mode and effect analysis and its application in industrial internet platform Mathematics 2022 10 2492 10.3390/math10142492
Xiao FY Wen JH Pedrycz W Generalized divergence-based decision making method with an application to pattern classification IEEE Trans Knowl Data Eng 2022 4347 1 10.1109/tkde.2022.3177896
Xiao J Wang XL Zhang HJ Exploring the ordinal classifications of failure modes in the reliability management: an optimization-based consensus model with bounded confidences Gr Decis Negot 2022 31 49 80 10.1007/s10726-021-09756-9
Xing YM Cao MS Liu YJ A Choquet integral based interval Type-2 trapezoidal fuzzy multiple attribute group decision making for sustainable supplier selection Comput Ind Eng 2022 165 107935 10.1016/j.cie.2022.107935
Xu Xh Du ZJ Chen XH Cai CG Confidence consensus-based model for large-scale group decision making: a novel approach to managing non-cooperative behaviors Inf Sci 2019 477 410 427 10.1002/int.22122
Xu YJ Gao PQ Martínez L An interindividual iterative consensus model for fuzzy preference relations Int J Intell Syst 2019 34 1864 1888 10.1016/j.ins.2019.07.050
Xu YJ Zhu SN Liu X Additive consistency exploration of linguistic preference relations with self-confidence Artif Intell Rev 2022 10.1007/s10462-022-10172-x
Yager RR On ordered weighted averaging aggregation operators in multicriteria decisionmaking IEEE Trans Syst Man Cybern 1988 18 183 190 10.1109/21.87068
Yager RR Penalizing strategic preference manipulation in multi-agent decision making IEEE Trans Fuzzy Syst 2001 9 393 403 10.1109/91.928736
Yager RR Defending against strategic manipulation in uninorm-based multi-agent decision making Eur J Oper Res 2002 141 217 232 10.1016/S0377-2217(01)00267-3
Yu GF Li DF Liang DC Li GX An intuitionistic fuzzy multi-objective goal programming approach to portfolio selection Int J Inf Technol Decis Making 2021 20 1477 1497 10.1142/S0219622021500395
Zhang Z Guo CH Martínez L Managing multigranular linguistic distribution assessments in large-scale multiattribute group decision making IEEE Trans Syst Man Cybern Syst 2017 47 3063 3076 10.1109/TSMC.2016.2560521
Zhang HJ Zhao SH Kou G An overview on feedback mechanisms with minimum adjustment or cost in consensus reaching in group decision making: Research paradigms and challenges Inf Fusion 2020 60 65 79 10.1016/j.inffus.2020.03.001
Zhang Z Gao Y Li ZL Consensus reaching for social network group decision making by considering leadership and bounded confidence Knowl Based Syst 2020 204 106240 10.1016/j.knosys.2020.106240
Zhang YJ Chen X Gao L Consensus reaching with trust evolution in social network group decision making Expert Syst Appl 2022 188 116022 10.1016/j.eswa.2021.116022
Zhao ZY Li C Zhang XJ An incremental method to detect communities in dynamic evolving social networks Knowl Based Syst 2019 163 404 415 10.1016/j.knosys.2018.09.002
| 0 | PMC9746597 | NO-CC CODE | 2022-12-15 23:21:57 | no | Artif Intell Rev. 2022 Dec 13;:1-32 | utf-8 | Artif Intell Rev | 2,022 | 10.1007/s10462-022-10361-8 | oa_other |
==== Front
Ann Oper Res
Ann Oper Res
Annals of Operations Research
0254-5330
1572-9338
Springer US New York
5101
10.1007/s10479-022-05101-8
Original Research
Improving data efficiency for analyzing global exchange rate fluctuations based on nonlinear causal network-based clustering
Choi Insu [email protected]
Yun Wonje [email protected]
http://orcid.org/0000-0001-8385-9598
Kim Woo Chang [email protected]
grid.37172.30 0000 0001 2292 0500 Department of Industrial and Systems Engineering, KAIST, Yuseong-gu Daehakro 291, Daejeon, 34141 Republic of Korea
13 12 2022
136
15 11 2022
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2022, Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
This article is made available via the PMC Open Access Subset for unrestricted research re-use and secondary analysis in any form or by any means with acknowledgement of the original source. These permissions are granted for the duration of the World Health Organization (WHO) declaration of COVID-19 as a global pandemic.
This study used information theory and network theory to predict the fluctuations of currency values of the machine learning model. For experiments, we calculate the causal relationships between currencies using loarithmic return (log-return) and entropic value-at-risk (EVaR) values of gold price per troy ounce in 48 currencies over 25 years. To quantify the causal relationships, we used the concept of transfer entropy. After quantifying their information flow, we modeled and analyzed those nonlinear causal relationships as a network. The network analysis results confirmed that information flow-based nonlinear causal relationships differed from the commonly-known key currency order. Then, we classified currencies using hierarchical clustering methods based on the configured networks. We predicted fluctuations in currency values using machine learning algorithms based on network topology-based information. As a result, we show that using the data columns in the same communities based on statistically significant nonlinear causal relationships can improve most machine-learning-based fluctuations of currency values for various countries from the perspective of data efficiency.
Keywords
Transfer entropy
Machine learning
Network theory
Currency value fluctuation
Exchange rate
==== Body
pmcIntroduction
During the past five years, the main issues of the global economy have been protectionism and economic stimulus. Issues regarding trade conflicts and own country first policy have triggered movements to enhance the domestic economy at the cost of foreign countries. Furthermore, with the global economic recession after the outbreak of COVID-19, economically significant countries worldwide have competitively implemented currency issuance to stimulate the economy. Such situations have increased the downside risk of the exchange rates within the global economic system. Due to the exchange rates being a factor influencing global investment and the economy, it is essential to find proper methods to hedge possible risks. One of the most straightforward methods to hedge downside risk is to predict the future price or risk fluctuations of financial assets.
In finance, prediction is a significant field. In other words, finding a model for predicting the objects of financial markets is still a highly researched topic despite major challenges, such as bankruptcy (Séverin & Veganzones, 2021), credit risk (Belhadi et al., 2021), and even financial products’ prices (Gupta et al., 2022; Jabeur et al., 2021). Especially predicting financial assets’ prices is very complicated due to their nonlinear, dynamic, confusing, and unpredictable characteristics.
However, among the various attempts to predict the financial time series, machine learning models have recently been most studied, given their ability to recognize complex patterns in various applications. Accordingly, many prior studies have used machine learning to predict economic indicators such as stock prices and exchange rates. For example, short-term exchange rate prediction has shown to be possible to some extent through machine learning techniques (Galeshchuk, 2016). In addition, forecasts through machine learning techniques have been argued to can be able to be more predictive compared to the Chaos model and the behavioral model (Lisi & Schiavo, 1999).
Based on this trend, we attempted to present and verify the entropy information flow and network theory to help predict the long-term exchange rates of the change via fluctuations of currency values. Here, as will mention in Chapter 2 in more detail, the current value refers to the amount of money required for the same commodity. To analyze the relationship between currencies as an information flow of fluctuations of currency values illustrated as network theory is a method suggested in studies by Liu et al. (2010) and Cao et al. (2017). Both studies attempted to intuitively model the global currency relationship through network analysis by recognizing logarithmic returns (log-returns) and their downside risks as the flow of information. In other words, this study aims to model the global trend of exchange rates by viewing the causal relationship between currency values’ logarithmic return and downside risk values between currencies as information flows between currencies and quantifying them through entropy measures. In addition, the modeled network is visualized to present the global currency order viewed from the information theory perspective based on PageRank. Finally, based on the hierarchical clustering results, we used machine learning techniques to forecast currency value using downside risk fluctuations and long-term exchange rates.
The remaining part of this study consists as follows. Chapter 2 confirms the justification for using the transfer entropy intended to be used in this paper through research data description and exploratory data analysis, and Chapter 3 explains the research methodology. Chapter 4 describes the results of a significant causal relationship according to the effective transfer entropy calculated based on the research methodology and includes the schematic results. In addition, network analysis was conducted using the configured nonlinear causal networks based on their log-returns and entropic value-at-risk (EVaR). Then we generated information flow-based communities with their network topology. In Chapter 5, based on the analysis in Chapter 4, whether the results of the nonlinear causal network analysis in this paper can derive improved results in predicting the fluctuations of currency values were analyzed. Finally, Chapter 6 describes the discussions and conclusions.
Data
Data description
The data in this paper are based on each country’s exchange rates against the US dollar. However, excluding the US dollar when performing a global currency network analysis would be futile. Thus, it was necessary to develop a unit encompassing the dollar, which has been used globally. Thus, this study converted all monetary units into the gold price per troy ounce. Gold is a globally recognized good and has been used as a global currency in the early days, making it the apt unit for our study. To this end, the currency value data were constructed by multiplying the dollar price of gold per troy ounce and the exchange rates of each currency in dollars, which means that the currency amount to buy a troy ounce amount of gold. We will call this currency amount as ‘currency value’ in this paper.
The source of the raw data in this paper is mainly the PACIFIC exchange rates Service, which is provided by Professor Werner Antweiler of the University of British Columbia in Canada (Antweiler, 2008). The above database provides exchange rates for 88 countries and three precious metals (gold, silver, and platinum) since 1971. This study extracted the exchange rate data from as many countries as possible from this database. Unfortunately, the available data decreased as the analysis time increased; thus, we compromised between periods and countries. As a result, the most suitable period and countries have been chosen to be from 1995 to 2020 in 48 countries.
The raw dataset also provides up-to-date exchange rates of countries integrated into the Eurozone. When included in the Eurozone, existing currencies are fixed their currencies to specific exchange rates and converted into euros. The current exchange rates were calculated by multiplying these fixed exchange rates by the current euro exchange rates. Among the raw dataset, a total of 0.3% were missing. We processed additional data from the central banks to complement the missing values. For Chile, Sri Lanka, and Kuwait, data were provided by the central bank, and for Brazil, the missing values were filled by the exchange rates data provided by the US Federal Reserve (Chile Banco Central, 2021; Central Bank of Kuwait, 2021; Board of Governors of the Federal Reserve System, 2021; Central Bank of Sri Lanka, 2021). The remaining missing values of Kuwait and the United Arab Emirates were supplemented through the ECOS database provided by the Bank of Korea (Korea Economic Statistics System, 2021). The remaining missing values were estimated using cubic spline interpolation.
As mentioned above, we used gold price data to create a global currency network. To create our datasets, data from the World Gold Council(WGC), which oversees the global gold industry, were used (World Gold Council, 2021). Through the data provided by the WGC, we obtained the dollar price of gold per troy ounce of gold from 1995 to 2020 for 48 countries. Within this dataset, Vietnam’s gold price per troy ounce could be obtained, which was added to the analysis target of this study. The names and ISO 4217 code of 48 selected currencies are as follows: United States dollar(USD), Australian dollar(AUD), Austrian schilling(ATS), Belgian franc(BEF), Brazilian real(BRL), Pound sterling(GBP), Canadian dollar(CAD), Chilean peso(CLP), Chinese Yuan Renminbi(CNY), Danish krone(DKK), Dutch Gulden(NLG), Egyptian pound(EGP), European euro(EUR), Finnish markka (FIM), French franc(FRF), German mark(DEM), Greek drachma(GRD), Hong Kong dollar(HKD), Hungarian forint(HUF), Icelandic krona(ISK), Indian rupee(INR), Indonesian rupiah(IDR), Irish pound(IEP), Israeli new shekel(ILS), Italian lira(ITL), Japanese yen(JPY), Kuwaiti dinar(KWD), Malaysian ringgit(MYR), Mexican peso(MXN), New Zealand dollar(NZD), Norwegian krone(NOK), Philippine peso(PHP), Polish zloty(PLN), Portugese Escudo(PTE), Russian ruble(RUB), Saudi Arabian riyal(SAR), Singapore dollar(SGD), South African rand(ZAR), South Korean won(KRW), Spanish peseta(ESP), Sri Lankan rupee(LKR), Swedish krona(SEK), Swiss franc(CFH), New Taiwan dollar(TWD), Thai baht(THB), Turkish lira(TRY), UAE dirham(AED), and Vietnamese dong(VND).
Exploratory data analysis
Based on the configured data, this study attempted to measure the causal relationships between currency values and exchange risk. After measuring them, we used them to predict the fluctunation of currency values. To this end, this study used the concept of the Entropic Value-at-risk (EVaR) (Ahmadi-Javid, 2012; Bohdalová, 2007). EVaR is a concept used to quantify downside risk in finance and probability optimization. At this time, EVaR results in an upper-bound value for Value-at-risk (VaR) and conditional Value-at-risk (CVaR). VaR and CVaR have commonly used risk measure indicators when measuring price-based risk by Chernoff inequality; therefore, this study used EVaR as a downside risk indicator for robustness. For a probability space (Ω,F,P), if the random variableX and LM+ is a Borel measurable function in X:Ω→R and has an moment generating function Mx(z), the EVar of X∈LM+ with the significance level 1-α is:1 EVaR1-α(X):=infz>0z-1lnMx(z))α
Some representative methods to calculate EVaR include the delta-normal method, Monte Carlo simulation method, and historical simulation method. In this paper, EVaR was estimated using a historical simulation method. We used the historical simulation method because the historical simulation method does not assume changes in risk factors of a particular distribution compared to the other two methods. Furthermore, as a nonparametric method, the historical simulations method does not include estimating statistical parameters such as variance or covariance and avoids inevitable estimation errors.
Consequently, we used the four primary datasets in this paper: 25 years’ worth (1996–2020) of gold price and the rates of change of 5 days, 10 days, and 20 days of EVaR. Each dataset has 48 columns(countries) and 6,761 rows (days). In this paper, basic statistics were identified for the four datasets. Then the normality of the data was verified through the Shapiro-Wilk test and the Jacque-Bera test. Also, the autocorrelation for the rates of change in the time series data of 5, 10, and 20 business days was examined through the Ljung-Box test. In addition, time-series stationarity was confirmed through three tests: the Augmented Decky-Fuller test (ADF), Phillips-Perron test (PP), and Kwiatkowski-Phillips-Schmidt-Shin test (KPSS). Homoscedasticity was confirmed through the White test. The calculation results are given in Tables 1, 2, 3, 4.
Tables 1, 2, 3, 4 represents the number of columns in every four datasets (at least 0, up to 48) satisfying normality, autocorrelation, stationarity, and homoscedasticity under significance levels of 0.1, 0.05, and 0.01, respectively. As seen in Tables 1, 2, 3, 4, all dataset used in this paper does not satisfy the normality at the significance level of α=0.01. In addition, it can be confirmed that the ADF, PP, and KPSS tests satisfy stationarity at all three significance levels. On the other hand, no data column satisfied homoscedasticity and autocorrelation due to the difference between data.Table 1 Statistical results of log-returns of currency value
Normality Stationarity Homoskedasticity 5d Autocorrelation 10d Autocorrelation 20d Autocorrelation
α=0.1 0 48 3 46 46 46
α=0.05 0 48 3 46 46 46
α=0.01 0 48 1 46 46 46
Table 2 Statistical results of 5-day EVaR
Normality Stationarity Homoskedasticity 5d Autocorrelation 10d Autocorrelation 20d Autocorrelation
α=0.1 0 48 46 1 1 0
α=0.05 0 48 46 1 0 0
α=0.01 0 48 46 1 0 0
Table 3 Statistical results of 10-day EVaR
Normality Stationarity Homoskedasticity 5d Autocorrelation 10d Autocorrelation 20d Autocorrelation
α=0.1 0 48 46 1 1 0
α=0.05 0 48 46 1 0 0
α=0.01 0 48 46 1 0 0
Table 4 Statistical results of 20-day EVaR
Normality Stationarity Homoskedasticity 5d Autocorrelation 10d Autocorrelation 20d Autocorrelation
α=0.1 0 48 46 1 1 0
α=0.05 0 48 46 1 0 0
α=0.01 0 48 46 1 0 0
Methodology
We used information theory-based causal relationship measures to analyze the causal relationship between currency values. Granger causality is the most used indicator, which measures linear causal relationships. In calculating the Granger causality, properties such as normality, stationarity, and linearity must be assumed. However, as seen in the results of Tables 1, 2, 3, 4, the datasets do not satisfy normality at all, which follows the results of prior studies that the rates of change in financial data generally do not satisfy normality (Dimpfl & Peter, 2013; Galeshchuk, 2016; Granger, 1969). Therefore, this study used a transfer entropy-based nonlinear causal measure, which measures causal relationships based on information theory to overcome the problems (Schreiber, 2000; Sheikh & Qiao, 2009; Tsai, 2011). Information theory has the advantage of using the model regardless of the characteristic of the data, making it possible to measure nonlinear causal relationships. Since transfer entropy is an entropy-based causal indicator, the advantage of the information theory is also valid. In this paper, the effective transfer entropy, a form of transfer entropy that has removed the finitesize effect, was used (Dimpfl & Peter, 2013; Boba et al., 2015).
Transfer entropy is a nonparametric measure for confirming the flow of information between two variables based on Shannon entropy. Transfer entropy quantifies the causal relationship in the system based on measuring the resolved uncertainty according to the flow of information and the magnitude of the causal relationship from the source to the target variable. Transfer entropy has been used in the financial sector to confirm the causal relationship between financial assets and financial markets in various studies (Marschinski & Kantz, 2002; Kwon & Yang, 2008; Dimpfl & Peter, 2013; Sensoy et al., 2014; Bekiros et al., 2017; Lim et al., 2017; Jang et al., 2019; Yue et al., 2020a, b; Choi & Kim, 2021). Based on the prior studies, this study chose transfer entropy to analyze exchange rates and exchange rates risk fluctuations. It is used as an indicator of causal relationships.
Entropy metrics
Shannon entropy
The Shannon entropy H(X) is a metric that represents the uncertainty of the discrete random variable X, and is calculated as:2 H(X)=-∑xp(x)log2p(x)
Conditional entropy
Conditional entropy can be defined as the expected value of a conditional entropy averaged on the conditional probability variable based on the Shannon entropy. In particular, the conditional entropy of the discrete probability variable Y, given the discrete probability variable X, can be expressed as follows:3 H(X∣Y)=-∑x∈Xy∈Yp(x,y)logp(x,y)p(y)
Transfer entropy
As aforementioned, transfer entropy is robust compared to the Granger causality and can be easier to analyze various types of causal relationships, such as nonlinear relationships.
The transfer entropy of two time series variables Xt(k)={Xt,Xt-1,⋯,Xt-k} and Yt(l)={Yt,Yt-1,⋯,Yt-l} is defined as follows (Schreiber, 2000):4 TEY→X(k,l)=H(Xt+1∣Xt(k))-H(Xt+1∣Xt(k),Yt(l))
The H(Xt+1∣Xt(k)) in (4) implies the estimated conditional entropy of Xt+1 given the time series data Xt(k). The H(Xt+1∣Xt(k),Yt(l)) in (4) implies the estimated conditional entropy of Xt+1 given the time series data Xt(k) and Yt(l). Transfer entropy has an asymmetric characteristic and acts as a causality measure by estimating the information flow that Yt(l) has on period Xt+1. Transfer entropy can be calculated after discretizing the probability distribution of the two time series data. Thus this study used the most commonly used time series data discretization method: histogram defined at equal intervals. The number of histogram intervals was referenced from prior studies (Granger, 1969; Hacine-Gharbi & Ravier, 2018; Hacine-Gharbi et al., 2012; Liu et al., 2010; Knuth, 2019; Zhukov & Popov, 2014).
Effective transfer entropy (ETE)
One of the problems with transfer entropy is that an insufficient number of samples can result in bias due to the finite-size effect, which is an effective transfer entropy. Therefore, this paper measured the causal relationship using the effective transfer entropy as shown in (5).5 ETEY→X(1,1)(t)=TEY→X(1,1)(t)-RTEY→X(1,1)(t)σRTERTEY→X(1,1)(t)=1M∑i=1MTEYshuffled→Xshuffled(1,1)(t)
Xshuffled and Yshuffled in the above equation denotes the randomly rearranged time series data, and M is the generated number of such Yshuffled. In this paper, we have run 300 random simulations (Dimpfl & Peter, 2013; Boba et al., 2015).
Network analysis
Network analysis can elucidate social phenomena by modeling them in networks. It can be convenient for describing and utilizing interactions between objects. Since the transfer entropy-based causal relationship of the exchange rates can be understood as an interaction between the national currencies, this study used network analysis to express and analyze the entropy causal relationship of log-returns and EVaR of gold price per ounce in currencies.
Network theory
In network theory, the main elements are nodes and links. A node refers to a subject that constitutes a network and becomes an actor of interaction. Moreover, a link or connection means an interaction or relationship between these subjects (Kwak, 2017). The type of network may be divided according to the characteristics of the connection. For example, networks can be divided into directional or directional networks depending on whether there is a direction. Also, if the network is weighted, it is called a weighted network, and if not, a binary network (Kwak, 2017).
In representing networks, graphs and matrices are used. The two methods each have advantages in mathematical processing and visual explanation (Kwak, 2017). Graph format is a way to visualize the network and intuitively show the shape of the network by assigning shape, color, size, label, and arrow to the nodes and links. Matrix format is a method of representing the properties of the network as a matrix, called an adjacency matrix. In the adjacent matrix X, the value xij indicates the connection from the ith node to the jth node and its corresponding properties (Lee, 2013).
Since this study quantified the causal relationship between log-returns and EVaR of gold price per ounce in currencies as transfer entropy, these information flows were expressed as a directed weighted network. The constructed directed weighted network has the direction and size of causality based on the value of the effective transfer entropy.
PageRank
Centrality measures are the most commonly used measures in network analysis and measure the degree of central position to measure the influence of each node. Measuring centrality can be primarily divided into degree centrality, betweenness centrality, harmonic centrality, closeness centrality, and eigenvector centrality. We focused on the eigenvector centrality of the constructed networks in this paper. Especially, PageRank is a centrality algorithm designed to measure the ranking of web pages in Google’s search engine and is the most famous eigenvector centrality measure (Page et al., 1999). PageRank is a directed network analysis method that can measure each node’s influence and current and future information flow.
The method to calculate the PageRank is as follows. For a network with N nodes, the network can be expressed in an adjacency matrix A∈RN×N. Through this adjacency matrix, we can mathematically calculate the column vector r∈RN that has the PageRank value of each node as its elements as in (6).6 r=(1-α)(I-αATD-1)-11
In (6), I∈RN×N is the identity matrix, 1∈RN is the all-ones vector, and D∈RN×N is a diagonal matrix with max(Kouti,1) as the ith element, where Kouti denotes the number of edges from the ith node. α is a damping coefficient ranging from 0 to 1, which in this paper, we have used the same value as in the original paper, 0.85 (Page et al., 1999). From (6), we can numerically calculate the PageRank value using the power method.
In this paper, PageRank was used as a criterion for determining whether the currency of each country has a characteristic of a key currency. This interpretation can be applied because we constructed networks based only on the currency values’ causal relationships. Therefore, a more considerable PageRank value of a currency would imply a stronger node’s centrality. Thus, we concluded that such currency would strongly influence the network, which can be translated as the key currency from the perspective of log-returns and downside risks.
Hierarchical clustering
Clustering is defined as grouping nodes with similar characteristics. Since network analysis enables the measurement of node similarity, we created a cluster of currencies. Among the clustering techniques in network analysis, we have used hierarchical clustering, generating communities by gradually combining the most similar communities.
For hierarchical clustering, the similarity between communities must be measured first. A primary method of measuring similarity is measuring the distance between communities. A path length is denoted as the number of connections within the path. Also, the distance between the nodes is defined as the shortest path among the various paths between those nodes.
With the node distances calculated, a proximity matrix is derived with the node distances as an element, and the clustering algorithm is executed on the proximity matrix. Methods of clustering algorithms include single connection, full connection, average connection, and word connection, where each algorithm differs in determining which nodes to cluster.
This study assumed that the clustering process results in a highly relevant community in the information flow-based network of log-returns and downside risks of currency values. Thus, in Chapter 5, we examined whether the uncertainty of predicting fluctuations of currency values can be improved using only intra-cluster information.
Machine learning models
In this paper, we have used multiple machine learning models to predict the fluctuations of currency values. We have used the following models for classification problems, which indicate the upward or downward trend of currency fluctuations: logistic regression (LR), decision tree (DT), K-means clustering (KM), XGBoost (XGB), LightGBM (LBM), support vector machine (SVM), random forest (RF), Gaussian naïve Bayes (GNB), and multilayer perceptron (MLP). Also, we have used a convolutional neural network (CNN), recurrent neural network (RNN), long short-term memory (LSTM), and gated recurrent unit (GRU) models for predicting log-returns of gold price per ounce in currencies.
Logistic regression (LR)
Logistic regression is a statistical approach used to estimate the probability of an event by combining independent variables linearly. The objective of logistic regression is to express the relationship between the dependent and independent variables as a specific function; it is primarily applied to binary classification models. Logistic regression is based on the sigmoid function, which transforms any real-valued number to a value between 0 and 1. For instance, the formula for the sigmoid function f(x) is as follows:7 f(x)=11+e-x
A classification problem can work better with logistic regression, and we can use 0.5 as a threshold. Many researchers have widely used logistic regression. However, it struggles with restrictive expressiveness (e.g., interactions must be added manually), and other models may have better predictive performance. Another disadvantage of the logistic regression model is that the interpretation is more difficult because the interpretation of the weights is multiplicative and not additive (Molnar, 2018).
Decision tree (DT)
A decision tree is a flowchart-like structure where an internal node represents a feature(or attribute), the branch represents a decision rule, and each leaf node represents the outcome. The topmost node in a decision tree is known as the root node. It learns to partition based on the attribute value. Then, it partitions the tree in a recursive manner called recursive partitioning. Its visualization is like a flowchart diagram that easily mimics human-level thinking (Fig. 1). Fig. 1 Illustration of decision tree
A decision tree is a white box type of machine learning algorithm. A decision tree can share the internal decision-making logic, which is unavailable in the black box type of algorithms. Also, the decision tree is a distribution-free or nonparametric method. The decision tree does not depend upon probability distribution assumptions. Decision trees can handle high-dimensional data with good accuracy for classification problems.
K-means clustering (KM)
K-Means clustering intends to partition n objects into k communities in which each object belongs to the cluster with the nearest mean. This method produces exactly k different communities of the greatest possible distinction. The best number of communities k leading to the most significant separation (distance) is not known as a priori and must be computed from the data. The objective of K-Means clustering is to minimize total intra-cluster variance or the squared error function J:8 J=∑j=1k∑i=1n‖xi(j)-cj‖2
The number k means the number of communities, n represents the number of cases, cj means the centroid for cluster j (j = 1, ..., k). K-Means clustering is a relatively efficient method. However, we need to specify the number of communities in advance, and the final results are sensitive to initialization and often terminate at a local optimum. Unfortunately, no global theoretical method exists to find the optimal number of communities. A practical approach compares the outcomes of multiple runs with different k and chooses the best one based on a predefined criterion. A large k probably decreases the error but increases the risk of overfitting.
eXtreme gradient boosting/XGBoost (XGB)
Recently, XGBoost has been utilized in various areas. Primarily, XGBoost is known as the powerful machine learning model that is predicting price in finance these days (Abbasi et al., 2019; Gumus & Kiran, 2017; Yun et al., 2021). XGBoost is developed by (Chen & Guestrin, 2016), and it is an algorithm that incorporates the boosting gradient model suggested by (Friedman, 2001). Moreover, normalization is utilized within the objective function to decrease complexity. Also, it can prevent the overfitting problem and create the learning process quicker. Especially, XGBoost is an ensemble model which comprises many decision trees. According to (Mo et al., 2018), the output function of XGBoost can be derived from the following equation:9 Y^iT=∑k=1Tfkxi=y^iT-1+fTxi
where Y^iT is the generated tree, and T is the total number of tree models. Also, fTxi is the newly created tree model.
Light gradient boosting machine/LightGBM (LBM)
Ke et al. (2017) proposed a gradient-boosting framework named LightGBM in 2017. Using a computing variance gain, the model employs gradient-based one-side sampling to mend the split point. The LightGBM algorithm built two novel approaches: exclusive feature bundling and gradient-based one-side sampling (Sun et al., 2020). Then, the estimated function of LightGBM results from the integration of T regression trees and can be defined as follows:10 Yt=∑i=1Mft(x)
where ft(x) denotes the regression trees. In LightGBM, Newton’s method was used to estimate the objective function. Several studies showed that the LightGBM provides more efficient and accurate performance than advanced machine learning algorithms. According to Sun et al. (2020), the advantages of LightGBM can be reflected in fast training speed, low memory consumption, and good model accuracy.
Support vector machine (SVM)
SVM is a particular linear classifier based on the margin maximization principle. The SVM accomplishes the classification task by constructing the hyperplane in a higher-dimensional space that optimally divides the data into two categories (Platt, 1999; Li, 2009; Jana et al., 2022; Hastie et al., 2009; Cristianini & Shawe-Taylor, 2000).
Random forest (RF)
Random forest is an ensemble model used for classification and regression analysis, which operates by outputting classification results from several decision trees configured during the training process. Random forests modify bagged decision trees that build an extensive collection of de-correlated trees to improve predictive performance further. As a result, they have become a very popular "out-of-the-box" or "off-the-shelf" learning algorithm that enjoys good predictive performance with relatively little hyperparameter tuning (Boehmke & Greenwell, 2019).
Gaussian Naïve Bayes classifier (GNB)
Let μk be the mean of the values in the continuous data x associated with class Ck, and let σk2 be the Bessel corrected variance of the values in x associated with class Ck. Then, one has collected some observation value v. Then the probability density of v given a class Ck, p(x=v∣Ck), can be computed by plugging v into the equation for a Gaussian distribution parameterized by μk.11 p(x=v∣Ck)=12πσk2e(ν-μk)22σk2
The Gaussian naïve Bayes classifier combines this model with a decision rule from the independent feature model. One common rule is to pick the most probable hypothesis, known as the maximum a priori (MAP) decision rule. The corresponding Gaussian naïve Bayes classifier is the function that assigns a class label y^=Ck for some k as follows:12 y^=*argmaxk∈{1,…,K}p(Ck)∏i=1np(xi∣Ck)
Multi-layer perceptron (MLP)
MLP is a supervised learning algorithm that learns a function f(·):Rm→Ro by training on a dataset, where m is the number of dimensions for input and o is the number of dimensions of output. Given a set of features X=x1,…,xm, and a target y, It is capable of learning a nonlinear function approximator for classification and regression. This model differs from logistic regression in that, between the input and output layers, one or more nonlinear layers, known as hidden layers, may exist. Figure 2 shows one hidden layer MLP with scalar output.Fig. 2 Illustration of multi-layer perceptron (MLP)
The input layers consist of a set of neurons {xi∣x1,…,xm} representing the input features. Each neuron in the hidden layer transforms the values from the previous layer with a weighted linear summation w1x1+w2x2+⋯+wmxm, followed by a nonlinear activation function g(·):R→R like ReLU function. Finally, the output layer receives the values from the last hidden layer and transforms them into output values (Pedregosa et al., 2011) (Fig. 3).Fig. 3 Illustration of convolutional neural network (CNN)
Convolutional neural network (CNN)
CNN is a type of feed-forward artificial neural network. The connectivity pattern between its neurons is inspired by the organization of the animal visual cortex, whose individual neurons are arranged to respond to overlapping regions tiling the visual field. CNN uses convolutional and downsampling layers as learnable feature extractors. Those layers allow feeding neural networks without sophisticated preprocessing to learn valuable features during the training. For example, CNN has many layers categorized into the input, convolutional, pooling, fully connected, and output layers (Sayavong et al., 2019) (Fig. 4).Fig. 4 Illustration of recurrent neural network (RNN)
Recurrent neural network (RNN)
RNN is an artificial neural network in which directed connections connect hidden nodes to form a recurrent structure. It is a model suitable for sequential data processing, such as voice and text. In particular, it is mainly used in prediction problems using time series data in the financial field. Since RNN is a network structure that can accept input and output regardless of sequence length, the most significant advantage is that it can make structures diverse and flexible depending on the experimenter’s needs.
Long short-term memory (LSTM)
LSTM is a machine learning model that complements the traditional RNN’s vanishing gradient problem. LSTM adds input gates, forgetting gates, and output gates to memory cells of RNN in the hidden layer to erase unnecessary previous results and determine the results to be maintained. In addition, LSTM is a model in which the equation for calculating hidden states has become a little more complex than vanilla RNNs, and a new value of cell states has been added. As a result, LSTM usually performs better in processing sequential order input forms than RNNs (Fig. 5).Fig. 5 Illustration of long short-term memory (LSTM)
Gated recurrent unit (GRU)
GRU is a model that reduces calculations that update hidden status while maintaining LSTM’s complementary points of RNN’s vanishing gradient problem. In particular, in the LSTM, there are three gates: output, input, and delete gates, whereas in the GRU, there are only two gates, an update gate, and a reset gate, so the parameters are less than those of the LSTM (Fig. 6).Fig. 6 Illustration of gated recurrent unit (GRU)
Performance metrics of machine learning models
Nine machine learning models for forecasting the rise or fall of the currency values are evaluated by two evaluation metrics: accuracy and macro F1 score. Also, the value forecasting performance of the four machine learning models is evaluated by three common evaluation metrics: root mean squared error (RMSE), mean absolute error (MAE), and median absolute error (MDAE) for robustness. These evaluation metrics were selected based on the following prior research of forecasting financial indices or prices: RMSE (He & Droppo, 2016, Ni et al., 2019), MAE (Chen etal., 2018, Zeng & Khushi, 2020), and MDAE (Rounaghi & Zadeh, 2016, Berkman et al., 2000). The illustration of the confusion matrix and the calculation of each metric are as follows (Fig. 7):Fig. 7 Illustration of a confusion matrix
13 Accuracy=TP+TNTP+FP+TN+FN
14 Macro F1 Score=2×Precision×RecallPrecision+Recall(Precision=TPTP+FP,Recall=TPTP+FN)
15 RMSE=∑i=1n(yi-y^i)2n
16 MAE=∑i=1n∣yi-y^i∣n
17 MDAE=Median(∣y1-y^1∣,…,∣yn-y^n∣)
Calculation and analysis results
This section aimed to derive two features that would enhance the machine learning models’ ability to predict fluctuations of currency values based on information theory and network analysis. First, we determined the centrality degree of currencies with PageRank values based on the ETE value that implies the causality between currencies. Second, we grouped similar currencies into communities with hierarchical clustering.
ETE calculation results
Table 5 Number of significant ETE connections of each network by significance level
Currency price change rate 5d EVaR 10d EVaR 20d EVaR
α=0.1 524(23.22%) 1033(45.79%) 1013(44.90%) 1018(45.12%)
α=0.05 496(21.99%) 729(32.31%) 718(31.83%) 714(31.65%)
α=0.01 431(19.0%) 276(12.23%) 295(13.08%) 289(12.81%)
We can measure the amount of reduction in information uncertainty, which means one currency has on another, and quantify the nonlinear causal relationship through ETE. With four datasets, among the total possible 48×47=2256 connections, this study used statistically significant transfer entropy values under significance levels α=0.01 for the stability of the results. The number of significant causal connections under the significance level α = 0.1, 0.05, and 0.01 for the four datasets are shown in Table 5. The value in parentheses refers to the percentage of the actual number of connection nodes compared to the total number of possible nodes and the network’s density.
Network analysis result
With the calculated ETE values, this study illustrated a directed weighted network among the currencies with the start node as a source, the end node as a target, and ETE values as weights, as we mentioned before.Fig. 8 Network representation of log-returns
Fig. 9 Network representation of 5-day EVaR
Network illustration
Figures 8, 9, 10, 11 displays the log-return, 5-day EVaR, 10-day EVaR, and 20-day EVaR networks of the total currencies. Each node is placed on the analyzed countries, thus representing the currencies. The width of the edges was multiplied by the transfer entropy values, thus representing the degree of correlation, i.e., the thicker edge means the higher correlation.Fig. 10 Network representation of 10-day EVaR
Fig. 11 Network representation of 20-day EVaR
Fig. 12 Network representation of log-returns based on PageRank
Comparing the edge thicknesses between the figures shows that the values of the ETE of EVaRs are relatively smaller than the value of the log-returns. In addition, the connection of the log-returns of gold price per ounce in the currency network plot is relatively sparse than that of the three rates of change of EVaR plots, indicating that fewer countries are in a causal relationship. What is noticeable is that the causal relationship of the US dollar, currently used as the significant key currency, is very low for every network. The number of connections in the US dollar is less than two, and the thickness of connections is also very lean, which is significantly different from other countries.
PageRank analysis
To analyze the most influential currency, we used PageRank to determine the centrality of each currency. The PageRank value was calculated by substituting the networks into the PageRank algorithm. The results indicate that the higher the PageRank value, the higher the centrality of the currency. Based on these calculation results, we visualized the network again to emphasize the centrality and classify currencies that act as key currencies.
Figures 12, 13, 14, 15 visualizes the networks with centrality taken into consideration. The size and color of each node were set to reflect the corresponding PageRank values. In the figures, the bluer to yellow the node becomes, and the larger the node becomes, the stronger the centrality. The noteworthy aspect of the analysis result is that the top three central countries were constant among EVaR networks. The first to third place was fixed as JPY, IDR, and KRW each. Thus, it could be seen that the currency values’ downside risks are forming a nonlinear causal network centered on JPY, IDR, and KRW. The log-return figure shows that KRW and IDR had the highest PageRank, while other countries had similar figures. This result implies that KRW and IDR act as a core in the global market fluctuation flow from the perspective of downside risks.
The key currencies and their corresponding PageRank values of each network are given in Table 6. Based on the PageRank values, we extracted key currencies from each network. The criterion for key currencies was set as the point at which the PageRank value began to converge. This point can be examined in Fig. 16, 17, 18, 19, which illustrates the PageRank values of each network in descending order.Fig. 13 Network representation of 5-day EVaR based on PageRank
Fig. 14 Network representation of 10-day EVaR based on PageRank
Fig. 15 Network representation of 20-day EVaR based on PageRank
Fig. 16 PageRank values of log-return network
Fig. 17 PageRank values of 5-day EVaR network
Fig. 18 PageRank values of 10-day EVaR network
Fig. 19 PageRank values of 20-day EVaR netw
Table 6 Key currencies of each network based on PageRank
Logarithmic returns 5-day EVaR 10-day EVaR 20-day EVaR
Currency Centrality Currency Centrality Currency Centrality Currency Centrality
KRW 0.2259 JPY 0.2263 JPY 0.1577 JPY 0.1669
IDR 0.1124 IDR 0.1240 IDR 0.1256 IDR 0.1202
ISK 0.0759 KRW 0.0934 KRW 0.0979 KRW 0.0915
EGP 0.0691 PHP 0.0673 EGP 0.0761
RUB 0.0442
Hierarchical clustering
Hierarchical clustering allows for determining the optimal cluster without predetermining the number of communities. We did not predetermine the particular degree of communities, making this method applicable flexibly. The feature in identifying the cluster is the distance of the currency network derived in Sect. 3.2.3. The distance between nodes was calculated into a distance matrix using the Dijkstra algorithm, commonly used to obtain the shortest distance of a weighted network. Then, the hierarchical clustering was performed using this distance matrix. This result grouped the log-returns, 5-day EVaR, 10-day EVaR, and 20-day EVaR into 10, 10, 14, and 15 communities, respectively. The communities within each network are presented in Tables 7, 8, 9, 10.Table 7 Communities of the log-return network
Currencies
Cluster 1 ATS RUB CHF AED THB MXN
Cluster 2 CNY DEM GRD ILS ZAR ESP SEK
Cluster 3 CLP DKK NLG EUR FIM FRF HKD
INR IEP ITL MYR PHP PTE LKR
VND
Cluster 4 ISK TRY KRW GBP HUF JPY SAR
Cluster 5 EGP BRL PLN
Cluster 6 USD AUD IDR NOK CAD
Cluster 7 BEF NZD
Cluster 8 KWD
Cluster 9 SGD
Cluster 10 TWD
Table 8 Communities of the 5-day EVaR network
Currencies
Cluster 1 HKD LKR USD
Cluster 2 AUD EGP THB BEF HUF FIM IEP
Cluster 3 SGD ZAR INR MXN SAR EUR ISK
Cluster 4 JPY TWD IDR KRW CLP CHF DKK
Cluster 5 NZD CAD NLG GRD
Cluster 6 GBP RUB VND ITL
Cluster 7 ATS BRL FRF
Cluster 8 AED ILS NOK SEK
Cluster 9 TRY PHP CNY MYR PTE
Cluster 10 KWD DEM ESP PLN
Table 9 Communities of the 10-day EVaR network
Currencies
Cluster 1 AUD EGP CNY HKD SLR USD
Cluster 2 SAR THB NOK
Cluster 3 TRY KRW CAD EUR
Cluster 4 INR CLP PTE
Cluster 5 IDR GRD
Cluster 6 TWD CHF
Cluster 7 ATS BEF DKK
Cluster 8 BRL FIM
Cluster 9 ZAR KWD HUF VND
Cluster 10 SGD PHP ILS IEP
Cluster 11 JPY RUB MXN AED NLG LSK
Cluster 12 FRF
Cluster 13 NZD GBP PLN DEM ITL ESP SEK
Cluster 14 MYR
Table 10 Communities of the 20-day EVaR network
Currencies
Cluster 1 AUD EGP USD HKD ZAR FIM ITL
Cluster 2 JPY SAR NLG
Cluster 3 MYR BEF CAD GRD
Cluster 4 SGD PHP DKK
Cluster 5 TYR NZD ISK
Cluster 6 THB KRW CLP CHF
Cluster 7 BRL
Cluster 8 INR RUB LKR IEP
Cluster 9 TWD FRF
Cluster 10 ATS GBP CNY VND
Cluster 11 KWD HUF
Cluster 12 MXN AED EUR
Cluster 13 PTE ESP NOK PLN SEK
Cluster 14 IDR DEM
Cluster 15 ILS
Prediction based on network analysis results
Prediction experiment layouts
Using the results in Sect. 4, we predicted the fluctuations and log-returns of gold price per ounce in currencies. As we mentioned above, we predicted the currency price of gold per troy ounce using four(log-return, 5-day EVaR, 10-day EVaR, and 20-day EVaR) constructed networks.
According to prior research, price prediction is mainly addressed in two types of problems: classification and regression. Thus, we considered both classification and regression problems to take into account the robustness of our experimental results.
In constructing these experiments, we referred to other prior research that planned experiments similar to this study and suggested improvements in the existing models and evaluation performances (Jana, 2021a; Liu et al., 2019; Jana, 2021b; Jana & Ghosh, 2018; Jana & Pal, 2021; Jana et al., 2021). These studies effectively improved decision-making and prediction models, which are similar to the proposed models in this study.
To check the prediction results based on our network, we performed four experiments by differing the use of results in Sect. 4. Experiment 1 was used as a baseline experiment by using the log-returns and EVaR of gold price per ounce in currencies data of the entire 48 countries in prediction. The other three experiments were designed to check the influence using the information flow network-based clustering results and PageRank-based key currency results. Experiment 2 used network-based clustering results, which means it used only the data columns in the same communities. Experiment 3 used the PageRank-based key currency results, and experiment 4 used both clustering and PageRank-based key currency results.
Clustering results refer to the result in Sect. 4.2.3, where similar currencies were grouped into communities based on the network distance. Experiment 2 used other currency data from the same community to predict currency prices. Centrality result refers to the outcomes from Sect. 4.2.2, where key currencies were chosen based on the PageRank of the network. Experiment 3 included the currency price data from the key currencies in predicting the currency price of one country, and Experiment 4 contained both clustering and centrality data in the prediction. Finally, we conducted prediction experiments of the fluctuations and log-return values, including 5-day EVaR, 10-day EVaR, and 20-day EVaR values, also using their centrality measures as we mentioned above.
Two sub-experiments were performed for each experiment: predicting the fluctuation and the value. Every experiment was performed on the four datasets, meaning 32 experiment iterations. For the stability of this study, each experiment iteration was conducted with 500 times simulations (5 optimization algorithms, 20 iterations, 5 cross-validations) with 1000 times learning iterations. We experimented using five optimizers, which are methods of finding optimal weights through machine learning. First, we selected the most widely used algorithms in recent studies: Stochastic Gradient Descent (SGD), AdaGrad, RMSProp, Adam, and AdaDelta. Then, we selected the best model from the experiments using different optimizers. The metrics for evaluating the models were used differently for two sub-experiments: in predicting the fluctuation, accuracy and macro F1 score were used, and for predicting the value, we used MAE, RMSE, and MDAE. The machine learning models for each sub-experiment are explained in Sect. 3.3. The summarized description of the experiment can be found in Table 11.Table 11 Experiment description
Category Experiment 1 Experiment 2 Experiment 3 Experiment 4
Subjects 48 countries
Dataset Logarithmic returns, 5-day EVaR, 10-day EVaR, 20-day EVaR
Types 1. Prediction of fluctuation
2. Prediction of log price rates of change
Simulations 500 times
Iterations 1000 times
Metrics 1. Accuracy, macro F1 score (for classification)
2. MAE, RMSE, MDAE (for regresssion)
Ex rate data Yes Yes Yes Yes
Clustering result No Yes No Yes
Centrality result No No Yes Yes
Table 12 Accuracy in predicting the change in log-returns of gold price per ounce in currencies (%)
ML Model Experiment 1 Experiment 2 Experiment 3 Experiment 4
LR 52.55±0.13 54.87±0.09 53.43±0.05 53.93±0.01
DT 51.09±0.05 52.55±0.05 51.34±0.08 51.70±0.01
KM 53.76±0.09 54.36±0.11 54.17±0.05 54.35±0.03
XGB 53.37±0.04 55.02±0.04 53.54±0.12 53.72±0.02
LBM 53.24±0.02 54.16±0.02 53.43±0.14 53.00±0.06
SVM 51.01±0.02 51.82±0.08 51.54±0.16 51.04±0.04
RF 53.54±0.06 54.84±0.03 54.34±0.09 54.27±0.03
GNB 52.11±0.09 52.59±0.04 51.86±0.18 52.14±0.05
MLP 53.70±0.04 55.23±0.10 54.29±0.11 53.47±0.09
Table 13 Macro F1 scores in predicting the change in log-returns of gold price per ounce in currencies (%)
ML Model Experiment 1 Experiment 2 Experiment 3 Experiment 4
LR 53.72±0.02 56.36±0.04 55.22±0.15 55.72±0.03
DT 52.92±0.02 54.40±0.07 52.92±0.14 53.58±0.04
KM 55.37±0.05 55.97±0.08 55.42±0.16 55.37±0.03
XGB 55.12±0.07 56.33±0.06 55.12±0.13 55.10±0.02
LBM 54.79±0.04 55.77±0.04 54.79±0.14 54.79±0.02
SVM 52.85±0.09 53.46±0.19 52.85±0.29 52.95±0.10
RF 55.48±0.08 55.92±0.12 55.48±0.14 55.42±0.09
GNB 53.52±0.05 54.00±0.04 53.45±0.16 53.52±0.08
MLP 55.19±0.09 56.49±0.15 55.54±0.08 55.37±0.06
Table 14 Average MAE scores in predicting log-returns of gold price per ounce in currencies (%)
ML Model Experiment 1 Experiment 2 Experiment 3 Experiment 4
CNN 3.55 4.20 3.49 2.44
(35,72.92%) (36,75.00%) (42,87.50%)
RNN 3.57 4.13 3.49 3.53
(35,72.92%) (36,75.00%) (42,87.50%)
LSTM 3.56 4.12 3.48 3.51
(35,72.92%) (36,75.00%) (42,87.50%)
GRU 3.57 4.12 3.48 3.51
(35,72.92%) (36,75.00%) (43,89.58%)
The first value in the parentheses refers to the number of improved currencies compared to the first experiment, and the second value refers to the improvement in percentage
Table 15 Average MSE scores in predicting log-returns of gold price per ounce in currencies (%)
ML Model Experiment 1 Experiment 2 Experiment 3 Experiment 4
CNN 5.05 5.73 5.08 5.09
(36,75.00%) (38,79.17%) (42,87.50%)
RNN 4.98 5.65 5.00 4.89
(36,75.00%) (38,79.17%) (42,87.50%)
LSTM 4.99 5.64 5.01 4.88
(36,75.00%) (38,79.17%) (42,87.50%)
GRU 4.98 5.64 5.01 4.89
(36,75.00%) (38,79.17%) (42,87.50%)
The first value in the parentheses refers to the number of improved currencies compared to the first experiment, and the second value refers to the improvement in percentage
Prediction result
The experiment results mentioned in Sect. 5.1 are as follows. First, the average value and the 95% confidence interval for predicting the currency price fluctuation are given in Tables 12, 13. Second, in prediction, a value of 1 was given if the currency price was rising or maintained and 0 if it was falling, making the experiment a case of classification. Third, comparing the metric values confirmed that the average accuracy and macro F1 scores in Experiments 2, 3, and 4 improved compared to Experiment 1 for every machine learning model. Moreover, statistically significant results were derived under the significance level α=0.01, under the paired t-sample test on the accuracy and macro F1 score results. The result implies that Experiment 2, Experiment 3, and Experiment 4 showed statistically better prediction results for predicting rise and fall of the currency values on the following day compared to Experiment 1.Table 16 Average MDAE scores in predicting log-returns of gold price per ounce in currencies (%)
ML Model Experiment 1 Experiment 2 Experiment 3 Experiment 4
CNN 3.04 3.53 2.93 3.09
(34,70.83%) (38,79.17%) (43,89.58%)
RNN 2.97 3.51 3.03 3.01
(34,70.83%) (38,79.17%) (43,89.58%)
LSTM 2.96 3.51 2.98 2.99
(34,70.83%) (37,77.08%) (43,89.58%)
GRU 2.96 3.50 2.99 2.99
(34,70.83%) (38,79.17%) (43,89.58%)
The first value in the parentheses refers to the number of improved currencies compared to the first experiment, and the second value refers to the improvement in percentage
Second, the average result value and the 95% confidence interval in the case of predicting the log-returns are given in Tables 14, 15, 16. Although all log-return forecasts do not indicate a strong statement as in the classification problem, deriving the number of currencies that showed improved metric scores showed meaningful results. In Experiment 2, it was confirmed that approximately 70% or more (34–36) of data displayed improved metric scores. Furthermore, by comparing the number of improved countries, Experiment 3 showed slightly better results than Experiment 2. Experiment 4, which combined the clustering and centrality results, showed the most improved prediction results: Over 80% improved currencies compared to Experiment 1. From Table 14, there were cases where the average MAE, RMSE, and MDAE values increased while the number of improved results increased. Such results could be understood as an outlier due to tremendous error values in some countries. Also, those currencies with very few constituents in their communities in EVaR got relatively worse results.
In sum, the results of this study can be translated as that predictive power can be improved like the results of prior research as mentioned at the top of Sect. 5.1. This result means that methods using network characteristics can improve predictive power through more efficient data utilization.
Conclusion
This study aimed to improve the exchange rates prediction performance using the causal relationships between log-returns and downside risks of currency values derived from gold and modeling them as networks. We have calculated the causal relationship of log-return, 5-day EVaR, 10-day EVaR, and 20-day EVaR values of currency values through effective transfer entropy. Based on the constructed networks, we generated communities based on their network topology and used them to improve the machine learning-based prediction performance of fluctuations in currency values.
As a result of visualizing the directed weighted networks with the transfer entropy as weights, we discovered the novel relationships based on log-return and downside risk values of currency values. For example, the primary currency based on global economic power and the major currency based on causality networks’ centrality measures differed. Using the clustering results derived from the topology of constructed networks for machine learning, we found that the performance statistically significantly improved despite decreasing the number of data columns. Thus, these results can help predict exchange rates in a roundabout way.
This study attempted to present a novel methodology for predicting exchange rates and analyzing exchange rates via log-returns and downside risks of gold price per ounce in currencies. However, factors affecting macroeconomic variables are vast and highly uncertain, making them difficult to quantify. Moreover, quantifying factors based only on currency-related variables may result in a bias, resulting in poor prediction performance as in Experiment 1 of Sect. 5.
However, this study’s novel contribution can be summarized as follows:We illustrated causal relationships based on log-returns and downside risks of gold price per ounce in currencies calculated by effective transfer entropy and analyzed through network theory.
Based on the illustrated log-return and downside risk networks, We used them to reduce information uncertainty in exchange rate prediction, resulting in better or almost similar results even though fewer data were used.
Since resolving information uncertainty is a significant factor in prediction, the results of this study can be expected to provide a different view on forecasting exchange rate prediction from the perspective of computation efficiency and feature selection.
It would be possible to apply more advanced machine learning algorithms for further studies. The models we used were basic machine learning algorithms utilized widely but not conducted based on the latest machine learning model. In addition, the relationship network we constructed only considered causal relationships between currencies and did not encompass other macroeconomic variables that would affect the exchange rates. The following research could derive more quantitatively developed results considering those improvements. Finally, it would be possible to combine the results with reinforcement learning-based foreign exchange investment methodology and expand in predicting the future price rates of change and risk fluctuations.
Acknowledgements
The authors sincerely appreciate the editors and anonymous reviewers for their feedback and suggestions for improving this study.
Availability of data and materials
The datasets generated or analyzed during the current study are available in the pacific exchange rates service repository, http://fx.sauder.ubc.ca/data.html. The datasets generated or analyzed during the current study are available in the World Gold Council repository, https://www.gold.org/goldhub/data/gold-prices. The datasets generated or analyzed during the current study are available in the Korea Economic Statistics System repository, https://ecos.bok.or.kr/. The datasets generated or analyzed during the current study are available in the Chile Banco Central repository, https://si3.bcentral.cl/Siete/ES/Siete/Cuadro/CAP_TIPO_CAMBIO/MN_TIPO_CAMBIO4/DOLAR_OBS_ADO/TCB_505. The datasets generated or analyzed during the current study are available in the Central Bank of Kuwait repository, https://www.cbk.gov.kw/en/monetary-policy/market-operations/exchange-rates. The datasets generated or analyzed during the current study are available in the Board of Governors of the Federal Reserve System repository, https://www.federalreserve.gov/releases/h10/hist/dat96_bz.htm. The datasets generated or analyzed during the current study are available in the Central Bank of Sri Lanka repository, https://www.cbsl.gov.lk/en/rates-and-indicators/exchange-rates.
Declarations
Conflict of interest
The authors have no relevant financial or non-financial interests to disclose.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
==== Refs
References
Abbasi, R. A., Javaid, N., Ghuman, M. N. J., Khan, Z. A., Ur Rehman, S., & Amanullah. (2019). Short term load forecasting using xgboost. L. Barolli, M. Takizawa, F. Xhafa, & T. Enokido (Eds.), Web, artificial intelligence and network applications (pp. 1120–1131). Cham: Springer International Publishing.
Ahmadi-Javid A Entropic value-at-risk: A new coherent risk measure Journal of Optimization Theory and Applications 2012 155 3 1105 1123 10.1007/s10957-011-9968-2
Antweiler, W. (2008). Pacific exchange rate service-database retrieval system. Available at http://fx.sauder.ubc.ca/data.html.
Bekiros S Nguyen DK Junior LS Uddin GS Information diffusion, cluster formation and entropy-based network dynamics in equity and commodity markets European Journal of Operational Research 2017 256 3 945 961 10.1016/j.ejor.2016.06.052
Belhadi, A., Kamble, S. S., Mani, V., Benkhati, I., & Touriki, F. E. (2021). An ensemble machine learning approach for forecasting credit risk of agricultural SMEs’ investments in agriculture 4.0 through supply chain finance. Annals of Operations Research, 1–29.
Berkman H Bradbury ME Ferguson J The accuracy of priceearnings and discounted cash flow methods of IPO equity valuation Journal of International Financial Management & Accounting 2000 11 2 71 83 10.1111/1467-646X.00056
Board of Governors of the Federal Reserve System (2021). H.10 brazil historical rates. Available at https://www.federalreserve.gov/releases/h10/hist/dat96 bz.htm (1999/12/31).
Boba P Bollmann D Schoepe D Wester N Wiesel J Hamacher K Efficient computation and statistical assessment of transfer entropy Frontiers in Physics 2015 3 10 10.3389/fphy.2015.00010
Boehmke, B., & Greenwell, B. (2019). Hands-on machine learning with R. Chapman and Hall/CRC.
Bohdalová, M. (2007). A comparison of value-at-risk methods for measurement of the financial risk (p. 10). Faculty of Management: Comenius University, Bratislava, Slovakia.
Cao G Zhang Q Li Q Causal relationship between the global foreign exchange market based on complex networks and entropy theory Chaos, Solitons & Fractals 2017 99 36 44 10.1016/j.chaos.2017.03.039
Central Bank of Kuwait (2021). Foreign currencies exchange rates. Available at https://www.cbk.gov.kw/en/monetary-policy/market-operations/exchange-rates.
Central Bank of Sri Lanka (2021). Exchange rates: Central bank of Sri Lanka. Available at https://www.cbsl.gov.lk/en/rates-and-indicators/exchange-rates.
Chen, T., & Guestrin, C. (2016). Xgboost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD international conference on knowledge discovery and data mining (pp. 785–794).
Chen W Yeo CK Lau CT Lee BS Leveraging social media news to predict stock index movement using RNN-boost Data & Knowledge Engineering 2018 118 14 24 10.1016/j.datak.2018.08.003
Chile Banco Central (2021). Exchange rates-observed dollar. Available at https://si3.bcentral.cl/Siete/ES/Siete/Cuadro/CAP_TIPO_CAMBIO/MN TIPO_CAMBIO4/DOLAR_OBS_ADO/TCB_505.
Choi I Kim WC Detecting and analyzing politically-themed stocks using text mining techniques and transfer entropy|focus on the republic of korea’s case Entropy 2021 23 6 734 10.3390/e23060734 34207887
Cristianini N Shawe-Taylor J An introduction to support vector machines and other kernel-based learning methods 2000 Cambridge Cambridge University Press
Dimpfl, T., & Peter, F. J. (2013). Using transfer entropy to measure information flows between financial markets. Studies in Nonlinear Dynamics and Econometrics, 17(1), 85–102.
Friedman, J. H. (2001). Greedy function approximation: A gradient boosting machine. Annals of Statistics, 1189–1232.
Galeshchuk S Neural networks performance in exchange rate prediction Neurocomputing 2016 172 446 452 10.1016/j.neucom.2015.03.100
Granger, C. W. (1969). Investigating causal relations by econometric models and cross-spectral methods. Econometrica: Journal of the Econometric Society, 424–438.
Gumus, M., & Kiran, M. S. (2017). Crude oil price forecasting using xgboost. In 2017 international conference on computer science and engineering (UBMK) (pp. 1100–1103).
Gupta S Modgil S Bhattacharyya S Bose I Artificial intelligence for decision support systems in the field of operations research: Review and future scope of research Annals of Operations Research 2022 308 215 274 10.1007/s10479-020-03856-6
Hacine-Gharbi A Ravier P A binning formula of bi-histogram for joint entropy estimation using mean square error minimization Pattern Recognition Letters 2018 101 21 28 10.1016/j.patrec.2017.11.007
Hacine-Gharbi A Ravier P Harba R Mohamadi T Low bias histogram-based estimation of mutual information for feature selection Pattern Recognition Letters 2012 33 10 1302 1308 10.1016/j.patrec.2012.02.022
Hastie, T., Tibshirani, R., Friedman, J. H., & Friedman, J. H. (2009). The elements of statistical learning: Data mining, inference, and prediction (Vol. 2). Springer.
He, T., & Droppo, J. (2016). Exploiting LSTM structure in deep neural networks for speech recognition. In 2016 IEEE international conference on acoustics, speech and signal processing (ICASSP) (pp. 5445–5449).
Jabeur, S. B., Mefteh-Wali, S., & Viviani, J.-L. (2021). Forecasting gold price with the xgboost algorithm and shap interaction values. Annals of Operations Research, 1–21.
Jana C Multiple attribute group decision-making method based on extended bipolar fuzzy mabac approach Computational and Applied Mathematics 2021 40 6 1 17 10.1007/s40314-021-01606-3
Jana C Multiple attribute group decision-making method based on extended bipolar fuzzy mabac approach Computational and Applied Mathematics 2021 40 6 1 17 10.1007/s40314-021-01606-3
Jana C Muhiuddin G Pal M Multi-criteria decision making approach based on svtrn dombi aggregation functions Artificial Intelligence Review 2021 54 5 3685 3723 10.1007/s10462-020-09936-0
Jana C Pal M A dynamical hybrid method to design decision making process based on GRA approach for multiple attributes problem Engineering Applications of Artificial Intelligence 2021 100 104203 10.1016/j.engappai.2021.104203
Jana DK Bhunia P Adhikary SD Bej B Optimization of effluents using artificial neural network and support vector regression in detergent industrial wastewater treatment Cleaner Chemical Engineering 2022 3 100039 10.1016/j.clce.2022.100039
Jana DK Ghosh R Novel interval type-2 fuzzy logic controller for improving risk assessment model of cyber security Journal of Information Security and Applications 2018 40 173 182 10.1016/j.jisa.2018.04.002
Jang SM Yi E Kim WC Ahn K Information flow between bitcoin and other investment assets Entropy 2019 21 11 1116 10.3390/e21111116
Ke G Meng Q Finley T Wang T Chen W Ma W Liu T-Y Lightgbm: A highly efficient gradient boosting decision tree Advances in Neural Information Processing Systems 2017 30 3146 3154
Knuth KH Optimal data-based binning for histograms and histogram-based probability density models Digital Signal Processing 2019 95 102581 10.1016/j.dsp.2019.102581
Korea Economic Statistics System (2021). The US dollar exchange rate of major currencies. Available at https://ecos.bok.or.kr/.
Kwak K-Y Social network analysis 2017 Seoul ChungRam
Kwon O Yang J-S Information flow between composite stock index and individual stocks Physica A: Statistical Mechanics and its Applications 2008 387 12 2851 2856 10.1016/j.physa.2008.01.007
Lee S Network analysis methodology 2013 Seoul NonHyung
Li, S. Z. (2009). Encyclopedia of biometrics: I-z. (Vol. 2). Springer Science & Business Media.
Lim K Kim S Kim SY Information transfer across intra/interstructure of CDS and stock markets Physica A: Statistical Mechanics and its Applications 2017 486 118 126 10.1016/j.physa.2017.05.084
Lisi F Schiavo RA A comparison between neural networks and chaotic models for exchange rate prediction Computational Statistics & Data Analysis 1999 30 1 87 102 10.1016/S0167-9473(98)00067-X
Liu J Wu C Li Y Improving financial distress prediction using financial network-based information and GA-based gradient boosting method Computational Economics 2019 53 2 851 872 10.1007/s10614-017-9768-3
Liu, L., Ye, Y.-t., Xie, Y., & Pu, L. (2010). Serial number extracting and recognizing applied in paper currency sorting system based on RBF network. In 2010 international conference on computational intelligence and software engineering (pp. 1–4).
Marschinski R Kantz H Analysing the information flow between financial time series The European Physical Journal B-Condensed Matter and Complex Systems 2002 30 2 275 281 10.1140/epjb/e2002-00379-2
Mo B Nie H Jiang Y Dynamic linkages among the gold market, us dollar and crude oil market Physica A: Statistical Mechanics and its Applications 2018 491 984 994 10.1016/j.physa.2017.09.091
Molnar, C. (2018). A guide for making black box models explainable. Available at https://christophm.github.io/interpretable-ml-book.
Ni L Li Y Wang X Zhang J Yu J Qi C Forecasting of forex time series data based on deep learning Procedia Computer Science 2019 147 647 652 10.1016/j.procs.2019.01.189
Page, L., Brin, S., Motwani, R., & Winograd, T. (1999). The pagerank citation ranking: Bringing order to the web. (Technical Report). Stanford InfoLab.
Pedregosa F Varoquaux G Gramfort A Michel V Thirion B Grisel O Duchesnay E Scikit-learn: Machine learning in Python Journal of Machine Learning Research 2011 12 2825 2830
Platt J Probabilistic outputs for support vector machines and comparisons to regularized likelihood methods Advances in Large Margin Classifiers 1999 10 3 61 74
Rounaghi MM Zadeh FN Investigation of market efficiency and financial stability between s &p 500 and London stock exchange: Monthly and yearly forecasting of time series stock returns using arma model Physica A: Statistical Mechanics and its Applications 2016 456 10 21 10.1016/j.physa.2016.03.006
Sayavong, L., Wu, Z., & Chalita, S. (2019). Research on stock price prediction method based on convolutional neural network. In 2019 international conference on virtual reality and intelligent systems (ICVRIS) (pp. 173–176).
Schreiber T Measuring information transfer Physical Review Letters 2000 85 2 461 10.1103/PhysRevLett.85.461 10991308
Sensoy A Sobaci C Sensoy S Alali F Effective transfer entropy approach to information flow between exchange rates and stock markets Chaos, Solitons & Fractals 2014 68 180 185 10.1016/j.chaos.2014.08.007
Séverin E Veganzones D Can earnings management information improve bankruptcy prediction models? Annals of Operations Research 2021 306 1 247 272 10.1007/s10479-021-04183-0
Sheikh AZ Qiao H Non-normality of market returns: A framework for asset allocation decision making The Journal of Alternative Investments 2009 12 3 8 35 10.3905/JAI.2010.12.3.008
Sun X Liu M Sima Z A novel cryptocurrency price trend forecasting model based on lightgbm Finance Research Letters 2020 32 101084 10.1016/j.frl.2018.12.032
Tsai, C.S.-Y. (2011). The real world is not normal. IL, USA: Morningstar Alternative Investments Observer:Chicago.
World Gold Council (2021). Gold price historical data: Gold price history. Available at https://www.gold.org/goldhub/data/gold-prices (2021/09).
Yue P Cai Q Yan W Zhou W-X Information flow networks of Chinese stock market sectors IEEE Access 2020 8 13066 13077 10.1109/ACCESS.2020.2966278
Yue P Fan Y Batten JA Zhou W-X Information transfer between stock market sectors: A comparison between the USA and China Entropy 2020 22 2 194 10.3390/e22020194 33285969
Yun KK Yoon SW Won D Prediction of stock price direction using a hybrid GA-xgboost algorithm with a three-stage feature engineering process Expert Systems with Applications 2021 186 115716 10.1016/j.eswa.2021.115716
Zeng, Z., & Khushi, M. (2020). Wavelet denoising and attention-based rnnarima model to predict forex price. In 2020 international joint conference on neural networks (IJCNN) (pp. 1–7).
Zhukov, M., & Popov, A. (2014). Bin number selection for equidistant mutual information estimaton. In 2014 IEEE 34th international scientific conference on electronics and nanotechnology (ELNANO) (pp. 259–263).
| 0 | PMC9746599 | NO-CC CODE | 2022-12-15 23:21:57 | no | Ann Oper Res. 2022 Dec 13;:1-36 | utf-8 | Ann Oper Res | 2,022 | 10.1007/s10479-022-05101-8 | oa_other |
==== Front
Int Bus Rev
Int Bus Rev
International Business Review (Oxford, England)
0969-5931
0962-9262
Elsevier Ltd.
S0969-5931(22)00111-1
10.1016/j.ibusrev.2022.102083
102083
Article
Convenient falsehoods and inconvenient truths: Not what leaders thought they would learn☆
Adler Nancy J. 1
McGill University, Desautels Faculty of Management, 1001 rue Sherbrooke ouest, Montreal, Quebec H3A 1G5, Canada
1 http://www.adlerartistry.com/galleries
13 12 2022
13 12 2022
10208322 9 2022
25 11 2022
© 2022 Elsevier Ltd. All rights reserved.
2022
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The COVID-19 pandemic and subsequent global economic meltdown severely challenged the world. What leadership lessons did we learn? What should we have learned? As global managers and international human-resource-management thought leaders, have we undervalued the role of humility? Have we overemphasized leaders’ impact while markedly underestimating the often-decisive influence of context? Have we embraced convenient illusions and rejected inconvenient truths? Whereas we are excellent at learning, are we equally good at unlearning—at dropping prior approaches and assumptions that either no longer work or have proven false? Have we succeeded in transcending the limiting vocabulary of economic efficiency and embraced a wider range of values and priorities to guide our most important strategies? How skilled are we at learning from each other, when ’the other’ differs markedly from us in what they look like, in the languages they speak, and in their most cherished beliefs? What roles are historic parochialism, ethnocentrism, and exceptionalism continuing to play in the 21st century? There is no single heroic expert who can give us the answers or guide us in reaching the future we yearn for. Rather, we need the best thinking, reflection, and creativity of all of us. This article opens that conversation with insights drawn from countries’ successes and failures during the pandemic. It then examines the process of learning—and unlearning—both during the pandemic and as it relates to the wider range of challenges currently confronting society. The article is an invitation to all of us to learn from each other by repeatedly unlearning convenient falsehoods and embracing novel, but inconvenient truths. It is an agenda that we avoid at our peril.
==== Body
pmc1 Introduction
Years ago, 5-year-old Kiara asked me what I did. I told her I was a teacher. She asked me who I taught. I answered, “Big people, like your parents and your friends’ parents.” This puzzled her. “Why did big people still need to learn?” She continued her questioning: “But Aunt Nancy, what do you teach the ‘big people’?” Try explaining global leadership and cross-cultural management to a five-year-old. I felt like I was failing altogether. Then Kiara, with all the seriousness she could muster, proclaimed, “Aunt Nancy, I know what you do. You try to get big people from around the world to play nicely with each other.” Spot on. That is exactly what we do, or at least it’s what we try to do; we try to get big people from around the world to play nicely with each other. And I would hazard to say that there has never been a more urgent need for us to succeed at our job. What we do, at our best, is exactly what the world needs.
2 Courage to see the world the way it is
To do our job, as scholars and thought leaders, we need courage (Palmer, 2007). We need the courage to see reality the way it is; not the way we would like it to be nor the way we have assumed it to be. We need to see the world with our own eyes and not through the lens of the supposed experts, including those in our own field.2 In International Business, for example, we have believed for decades that free trade and economic integration would make global wars unthinkable; that global economic interdependence would render conflict among nations too costly to consider, let alone instigate (Krugman, 2022). I too believed in the positive influence of globalization, or at least I wanted to believe in it. Yet today, in 2022, we need merely listen to the news to know that our assumption is wrong; and therefore, that one of our cherished beliefs is not true.3 If anything, the months following Russia’s invasion of Ukraine have forced us to recognize that global interdependence–such as Europe’s dependence on Russian oil and gas—has made war seemingly more likely, or at best it has made stopping wars more problematic. As the New York Times reported, Germany’s “…dependency [on Russian oil and gas] grew out of a German belief—embraced by a long succession of chancellors, industry leaders, journalists, [scholars,] and the public—that a Russia bound in trade would have too much to risk in conflict with Europe, … (Bennhold, 2022).” I too believed that global economic integration fostered world peace.4 When I reflect on my career, I used to feel good about the overarching direction and contributions of International Management, the field I had joined. Now I am less sure.TableOVERLY CONVENIENT FALSEHOODS: Think about a taken-for-granted assumption, perhaps implicit,that is commonly believed in our field that you are suspicious may not be true.
Could it be that many of our most cherished assumptions are wrong (Rich, 2022)?
2.1 Unlearning
Five-year-old Kiara asked, “Why do big people still need to learn?” We ‘big people’ not only need to continue learning; we also need to learn how to unlearn. Unlearning is key to being able to see the world the way it is. Only when we see the world accurately can we make meaningful contributions. In International Business, as in all disciplines, well-used methodology acts as a powerful lens for accurately seeing the world. According to Sri Zaheer, Dean of the University of Minnesota’s Carlton School of Management and Chair of the Federal Reserve Bank of Minneapolis, and Paul Vaaler, Chair in Law and Business at the University of Minnesota, methodologically informed perception may be one of the most valuable skills that scholars can contribute to public discourse, especially at a time when fact-free assertions have become commonplace.5
Adam Grant (2021), in his excellent book, Think Again: The Power of Knowing What You Don’t Know, explains why it is much harder to unlearn a belief we were certain was true than it is to learn something altogether new. He summarized the research on unlearning:“In our daily lives, too many of us favor the comfort of conviction over the discomfort of doubt. We listen to opinions that make us feel good, instead of ideas that make us think hard. … We surround ourselves with people who agree with our conclusions, when we should be gravitating toward those who challenge our thought process. We act too much like preachers defending our sacred beliefs …and too little like … [scholars] searching for truth. [Unfortunately] ”Intelligence is no cure…: Being good at thinking can make us worse at rethinking. The brighter we are, the blinder we can become to our own limitations.”6
It appears that scholars may be particularly bad at unlearning. Grant concludes by asking:“What would it mean to admit, on occasion, to being wrong …? If knowledge is power, knowing what we don’t know may be wisdom. (Think Again, Wikipedia, 2022)
Being blinded by our beliefs is not new. Sixty years ago, Thomas Kuhn, in The Structure of Scientific Revolutions, demonstrated that people will not let go of a cherished worldview simply because it has been shown to be wrong. They won’t let go until there is a new, more attractive belief to replace the original (Kuhn, 1962, Kuhn, 1962). I ask myself: Can I unlearn that increasing global economic integration and interdependence might undermine the possibility of world peace? If so, how? And if successful, what belief would I replace it with?
2.2 Convenient falsehoods - Inconvenient truths
To answer that question, we need to ask why it is so hard to see reality accurately. Why do we allow ourselves to be blinded by our beliefs? Why are our cherished, but incorrect, beliefs so very hard to let go of?
I recently asked a Russian colleague if people in Moscow were aware of the carnage being wrought in Ukraine. His answer: “Our friends know what is going on, but for most Russians it is more convenient not to know.”7 It is more convenient not to know! My colleague’s telling description of Russians’ current relationship to the truth echoes that of former U.S. Vice President Al Gore in his depiction of the inability of the world to see the ever-intensifying consequences of climate change. Already two decades ago, Gore referred to such knowledge in his Academy-Award winning documentary as An inconvenient Truth, a truth that we wish was not true.8 In May 2022, three-time Pulitzer-Prize-winning international author, reporter, and columnist Thomas Friedman (2022) suggested that the United States was no longer in an indirect war with Russia, but rather had edged toward a direct war. A highly inconvenient assessment. The question is not: Do we think Friedman is right? The question is: Are we willing and able to consider such a possibility?
What are the truths about the world and our field that are so inconvenient that we refuse to become aware of them? Which aspects of reality are simply too inconvenient for people in our culture to see? For people in our country to see? For people in the world to see? Let me give a classic example from academic publishing. My colleague, Academy-of-International-Business Fellow Anne-Wil Harzing, and I became suspicious that long publishing lag times had created an inconvenient truth not only for our field but for academia in general (Harzing & Adler, 2016). We raised the question first for ourselves, and then more broadly for our discipline: Is the lag time so long between initiating a study and having it published that most research is out-of-date even before the results are shared with the public and therefore of little use to the broader society? During the pandemic, for example, we saw the urgent need for knowledge that could immediately be converted into public-health policies and practices eclipse the very long traditional process for conducting, refereeing, and publishing scholarly work. We suspect that is true across domains and not just during global health crises.TableINCONVENIENT TRUTHS Which truths about our field are you suspicious are simply too inconvenient to see?
Let me give another example, this time from my own research. Years ago, I conducted a series of studies on global women leaders, women who were president or prime minister of their country or CEO of a global firm (Adler, 1996, Adler, 1997a, Adler, 1997b, Adler, 1998, Adler, 2002). I initiated the studies because, based on others’ research on women managers, I believed that women would bring more of the types of leadership the world most needed. To publish that research, I had to fight the prejudice and false assumptions that were widely held both within our field and among the broader public. Reflecting that prejudice, Canada’s Social Sciences and Humanities Research Council, our major funding source, rejected my research grant application with the succinct statement: What do women have to do with leadership? There was nothing hidden about the bias in their worldview, even if it was both illegal and prejudicial. Once I recovered from the shock, but not my anger at my grant application having been rejected, I found other ways to conduct the studies. Several years later, I submitted my first global-women-leaders article to Leadership Quarterly, the top leadership journal in our field. The article was rejected, with no invitation to revise or resubmit. The then editor, an eminent scholar, dismissed my research because, according to him, the sample was too small (even though there was no sampling in the study). I was again shocked and disappointment, especially because the reason he gave for rejecting my work was a blatant fabrication based on an overly convenient falsehood. To accept that senior leaders could be both men and women—and not just men—and that the historical research focus on male leaders in organizations had been consistently biased and self-limiting, and therefore inaccurate, remained invisible to the editor. As a reason for rejecting my article, it was simply too convenient a falsehood for the editor to consider that it might be wrong.
Kiara would be proud of me: I decided to try to “teach a big person.” The next day, I called the editor. As politely as possible, I suggested that perhaps the reviewers were ‘confused’, because the sample could not be too small given that I had included the entire population of the world’s most senior women political leaders during the past half century. I went on to say that I was certain that all of us want there to be more women leaders—so that in the future, among other benefits, a population would not be confused with a sample—but as scholars, we need to report what is factually accurate. We cannot simply report what we wish were true. In my case, my research needed to report the actual number, albeit paltry, of women leaders, not some inflated fantasy. Silence. Then, after what seemed like an interminable pause, the editor quietly asked me if he could re-read my paper. One week later, my first global-women-leaders article was officially accepted for publication in Leadership Quarterly without revision. Without revision! I give the editor credit for opening his mind. He had the courage to admit, albeit implicitly, that he had been wrong.
Although the situation forced the editor drop his overly convenient falsehood, it was not a particularly hard reality for me to face. It is one thing to see where others’ worldviews are in error, and why they need to drop their prior assumptions before the field can advance. It is quite another, and much harder, for us to face the ways in which our own worldview is limiting us, and therefore must change.
2.3 Humility: Letting go of convenient falsehoods
To let go of our own, often cherished, wrong assumptions, we first need to have the humility to admit that we could possibly be wrong (Grant, 2021; Rich, 2022). Only then can we begin to unlearn what we had previously thought to be true. Humility is neither particularly valued nor an oft-talked-about virtue in academia or in leadership.9
A moment for such humility and unlearning came to me later during that same series of studies when I had to face that the women who had led their countries were by no means all good leaders. Just like male leaders, some women leaders were terrific, and some were awful. Some were competent and some were not. I had so wanted to believe, and be able to report, that every woman leader was making a positive difference. Most do, but some unequivocally do not. The moment when I finally accepted that inconvenient truth was when I read an excellent paper by my graduate students entitled “Corrupt Global Women Leaders”. So much for my assumption that all senior women leaders—in both business and government—would bring a more caring, inclusive, and overall better approach to senior leadership than would men.
Once I was forced to accept the inconvenient truth that some women leaders were neither excellent nor admirable, I was caught in yet another dilemma. For me, the existence of corrupt, incompetent, egotistical women leaders was exceedingly inconvenient, even if these embarrassing women by no means represented the majority. There was no question that I did not see my role as the person who would let the world know the extent to which some very senior women leaders are incompetent or corrupt. Yet as a scholar, I could not possibly publish work that, to the best of my ability, was not accurate.
My Russian colleagues were right. Truth is all too often very inconvenient.
Seeing the world the way it is takes keen observation skills, humility, the ability to unlearn, and the ability to recognize both overly convenient falsehoods and inconvenient truths. Seeing the world accurately often requires the courage to see what other have yet to see and what most do not yet want to see. It all too frequently requires, at least initially, the courage to stand alone.
3 Using a cultural lens: Escaping the trap of overly convenient falsehoods
In times of great uncertainty, such as the during the pandemic, assumptions that were implicit–that had been taken-for-granted–are exposed. In Spring 2020, Warren Buffet, the legendary investment guru, famously said: “Only when the tide goes out do you discover who’s been swimming naked (Tchir, 2020).” As perceived threats multiply, we desperately try to make sense of the situations we find ourselves in, while simultaneously holding onto optimistic scenarios that we hope remain true. All too frequently, we hold onto prior beliefs that have been exposed as false. Our sense-making remains trapped in what we want to believe—in overly convenient falsehoods. We continue to believe in that which benefits us and reject that which we see as potentially detrimental.
At such times, how do we escape the trap of overly convenient falsehoods? In a literal sense, how do we unlearn and let go of that which benefits us, but is untrue? The pandemic provides us with numerous telling examples. In 2021, COVID raged around the globe and the United States was performing worse than all other economically advantaged countries (see Rabin, 2022, among others). A world-renowned cross-cultural physician in Texas asked me why the U.S. faced a more contagious and deadly variant of the virus than did other countries. I was stunned by the question. This person is an extremely intelligent, experienced public-health professional who should have known that viruses do not carry passports, and that all countries faced similar strains of the virus. The reality, that the U.S.’s extraordinarily high hospitalization and death rates might be caused by Americans’ own behavior, was simply too ’inconvenient’ for my colleague to accept. Instead, he invented the fiction that Americans were fighting a unique, more infectious, and more deadly U.S.-only variant. He created a convenient falsehood and stubbornly continued to believe in it. Sadly, that false belief kept him from advocating for the very behaviors that were mitigating serious disease and COVID-related deaths in most other countries. Given its life-and-death consequences, the fact that Americans’ own behavior might be causing the United States to under-perform the rest of the world so markedly was simply too inconvenient to accept.
From the beginning of the pandemic, I worked with a group of twenty cross-cultural management scholars from around the world to try to understand why some countries were more effective, and others dramatically less so, in managing the public health and economic consequences of COVID-19 (Adler & Sackmann, 2022; Adler, 2022). There is much we can learn, and need to unlearn, from what happened. Consider, for example, some of the assumptions that we, along with many of the best leaders worldwide, made that turned out to be wrong, often leading to unnecessarily high infection, hospitalization, and death rates. Those learnings have implications that go far beyond those triggered by health crises Interestingly, most of the more powerful assumptions are rooted in concepts that cross-cultural management scholars know well, including xenophobia, individualism, collectivism, high and low context, and exceptionalism.
3.1 Xenophobia
Xenophobia is the dislike of, or prejudice against, foreigners. During the pandemic xenophobia protected the public from confronting the inconvenient truth that if people in a foreign country were dying from the coronavirus, people in their own country could also die. Xenophobia erases our appreciation of our common humanity by artificially separating us into distinct subcultures. During the pandemic, xenophobia allowed people to see foreigners as the problem rather than recognizing the virus to be at fault. Rising nationalism exacerbated such xenophobia in countries around the world.
How xenophobia undermined the United States, one of the most sophisticated countries in the world, is illustrative. From very early in the pandemic, many Americans, including national leaders, labelled COVID-19 as ‘the China Virus’. The implication was that the Chinese were the source and cause of infection. This assumption led to seriously flawed policies. On February 2nd, 2020, the U.S. closed its border to Chinese, but not to Americans, returning home from China. The U.S. government implicitly assumed that Chinese were dangerous disease carriers, but that Americans returning from China were not. The medical reality, of course, is that one’s compatriots can infect home-country residents just as readily as can foreigners. Viruses don’t read passports. An additional inconvenient truth was that no matter what the source of the original virus, most transmission rapidly became domestic, not cross-border transmission from China or any other foreign location.
Perhaps this convenient, albeit false, xenophobic belief, that only foreigners could infect Americans, was somewhat understandable early in the pandemic when so little was known about COVID-19, but certainly not a year later when the U.S. repeated the same xenophobic-border-closing procedure in 2021 against South Africa after their scientists announced that they had identified the Delta variant.
Neither xenophobia nor such xenophobic border-closing behavior is unique to the United States. What exposes it as a widely believed, convenient falsehood is that “…45 nations imposed travel restrictions on China before the United States did. …The U.S. travel restrictions came a [full] month after China first announced its outbreak and at a point when the United States and more than 20 additional countries had already reported [their own domestic] coronavirus cases. Several of those countries, including Germany and the United States, were already reporting local transmission of COVID-19.10 Between the first official report of an outbreak in China and the announcement of U.S. travel restrictions, more than 40,000 travelers from China were estimated to have entered the United States. [Still more problematic] scientists [now] believe the virus likely emerged and began circulating a month or more before it was first recognized in China, which may have allowed it to spread beyond the countries where cases were initially recognized (Eder et al., 2020, updated 2021).” If world leaders and the general public would have been able to ‘unlearn’ their xenophobia and more accurately see what was taking place, the world may have saved hundreds of thousands of lives.
3.2 Context matters
How did other cultural assumptions influence popular beliefs, leadership behavior, and critical outcomes during the pandemic? Below are examples from Africa, Asia, Europe, and North America of the influence of culture-based assumptions that often undermined the crafting and implementation of effective policy.
3.2.1 The United States: A highly individualistic, low-context culture
As a culture, Americans are extremely individualistic and very low context. During the pandemic, those two cultural characteristics led most Americans to overemphasize leadership and underemphasize context. Initially, and perhaps not surprisingly, many people blamed the poor performance in the U.S. on an individual, then President Trump, and his policies (Bollykys & Nuzzo, 2020). When Biden assumed the presidency, many thought the situation would immediately, markedly improve. In both cases, the cultural assumption was the same; that is, that the difference between good and bad outcomes rested with the leadership of an individual. The extreme emphasis on individualism blinded most Americans on both sides of the political spectrum from considering the considerable influence of context. Today, most analysts believe that the main factors leading to the poor performance of the United States were primarily contextual, not presidential behavior. The U.S.’s lack of a national healthcare system and the large number of Americans who are overweight are seen as key causal factors. Neither of these contextual factors can be changed quickly by any individual leader, no matter how powerful, competent, or well-intentioned. Sadly, if countries do not learn to recognize culture-based assumptions and their consequences, they will not alter the underlying contextual factors that render them vulnerable to equally bad outcomes in the future.TableCONTEXT MATTERS Which contextual factors influenced pandemic outcomes in your country? When were the reasons for success or failure attributed more to the behavior of individual leaders versus to enduring cultural and contextual factors?
3.2.2 The Japanese context: High collectivism
The cultural context of all countries, not just that of the United States, influenced their responses during the pandemic. Japan, for example, whose highly collectivist cultural contrasts markedly with Americans’ extreme individualism, found that their citizens readily followed public-health recommendations, even when they were not codified into rules nor enforced with government sanctions. Collective pressure to fit in and to take care of one’s neighbor, along with cultural norms to avoid being shamed for not adhering to collective behavior, led Japan to achieve the lowest COVID death rate among the world’s wealthiest nations, one-twelfth of that in the United States (Rich & Dooley, 2022). Given their strongly collectivist cultural norms, Japan’s infection rates consistently ranked among the lowest in the world and their vaccination rates among the highest (Rich & Dooley, 2022).TableCULTURE MATTERS: INDIVIDUALISM VERSUS COLLECTIVISM Think about cultures you know well that are more individualistic versus those that are more collective? How did those two cultural dimensions—individualism versus collectivism—impact each country’s success or failure in dealing with the pandemic?
The highly inconvenient truth is that, in the short term, context is almost impossible to change. To the extent that context determines outcomes, the influence of leaders is limited. An uncomfortable determinism accompanies contextual assumptions and realities that is particularly unsettling when confronting crisis situations that demand immediate action, such as imminent mass infection, possible death, and potential economic collapse. Understandably, people want to believe that they can protect themselves, their neighbors, and their country. But what if that is not possible, especially in the near term? Several additional examples highlight the extent to which context determined pandemic outcomes.
3.2.3 The Italian context: Demography matters
Italy was the second country after China to be severely stricken by COVID. Already by February 2020, Italian hospitals were overwhelmed and people were dying. Initially the world press drew on stereotypes of ‘creative, disorganized Italians’ to suggest that Italy’s extremely bad outcomes were due to less-than-competent, 'disorganized’ leadership and ‘creative’ citizens who refused to follow necessary rules (see Levenstein, 2020; Viola, 2022, among others). Reality could not have been farther from the stereotype (see Braw, 2020). None of the stereotypes accurately described Italians during the initial months of the pandemic when Italy instituted some of the strictest lockdowns (initially in the north and around Milano) and most Italians rallied to follow and enforce public-health rules.
Both stereotypic attributions, however, were notably convenient for people living in other countries. They allowed people living outside of Italy to believe that few citizens in their own country would become sick, that their hospitals would not be overwhelmed, and that neither they nor their neighbors would die. The single most salient factor causing Italy’s high hospitalization and death rates, however, was not a failure of leadership (nor Italians’ ‘overly zealous creativity’), but rather a contextual factor. Italy has one of the oldest populations in the world and COVID disproportionately infects and kills older people. Age demographics are a contextual factor. No change in leaders or leadership style could have changed the average age of Italians. Age demographics, not poor leadership, was the primary culprit.
3.2.4 Context in Africa: demographics matters
Similarly, but in reverse, Africa experienced a very low incidence of COVID deaths in the initial phase of the pandemic (as did other of the poorest countries in South Asia). The same public pundits who misunderstood the situation in Italy, misunderstood it in Africa. Media commentators and other public figures again relied on cultural stereotypes to supposedly explain why people remained healthy in countries whose populations were so poor and that had inadequate healthcare infrastructures. Once more starting with an attribution to leadership, commentators reported that African leaders were so corrupt that they were not reporting accurate health statistics. They then offered other invented explanations, including that African countries had such poor healthcare systems that their leaders simply did not know how many people were sick and dying, and thus were under reporting the hospitalizations and deaths. Without any factual basis, the pundits went on to assert that, because Africans had previously been exposed to other coronaviruses (such as SARS), they had acquired a higher level of immunity to COVID-19 than had people in other parts of the world. Each explanation supported the belief—a ‘convenient falsehood’—of many people outside of Africa that poor countries would fare the worst (and implicitly that richer countries would do much better), thus implicitly assuring people living in rich countries that they were not at great risk and thus had little to fear.
However, during that first year of the pandemic, Africa’s good outcomes, that had, in fact, been accurately reported, were not caused primarily by any of the overly convenient explanations offered by the pundits, but rather by a hard-to-change contextual factor—age demographics. African countries have some of the youngest populations in the world. Africa has a median age under 20, with Niger having a median age of just 15.4 years. By comparison, Italy’s median age of 45½ years is three times older than that of Niger.11 Younger people, at least during the first years of the pandemic, tended neither to be hospitalized nor to die from COVID. Both horrible and fantastic African leaders benefited from the youth of their respective country’s population.
In summary, context matters, not just leadership. Over emphasizing leadership protects us from the inconvenient truth that leaders, whether good or bad, have less power that we generally imagine. I, along with other management scholars, have tended to believe that leadership is extremely important and have published numerous articles describing the nuances of various types of leadership. As our group of cross-cultural COVID scholars observed, approaches to managing the pandemic varied widely, yet the factors making the most difference repeatedly remained contextual, not primarily the behavior of individual leaders. Yet another inconvenient truth.
4 Toward a successful tomorrow: Moving beyond convenient falsehoods and inconvenient truths
There are many more examples of how convenient falsehoods and inconvenient truths keep leaders and society stuck in the past and incapable of moving forward. The most important question for society is how we move beyond accepting convenient falsehoods and blinding ourselves to inconvenient truths.
4.1 Stepping off the pedestal of certainty: Reclaiming confident humility
Within the group of cross-cultural scholars who observed and analyzed responses to the pandemic, one unavoidable pattern emerged. As a group, no matter how smart and experienced we were, we were repeatedly humbled, not only by what we did not know, but more consequentially, by what we thought we knew that turned out to be wrong. Moreover, the more we recognized our errors in making sense of the pandemic, the more we realized that the problem of hubris—false certainty—was not confined to the pandemic. The following question invites you to step off the pedestal of certainty and observe more closely those instances when you are consciously or blindly wrong.TableSTEPPING OFF THE PEDESTAL OF CERTAINTY: RECLAIMING CONFIDENT HUMILITY Recall an assumption you made that turned out to be false. When did you first realize that your assumption was wrong? What caused you to open your eyes? How have you thought or behaved differently since changing your belief?
4.1.1 Strategic humility versus over confidence: The Danish context
In almost every country, confidence—not humility—is associated with leadership. Confident leaders are viewed as capable. They are trusted, followed, and liked. The problem, however, in the volatile, uncertain, complex, ambiguous (VUCA) (Bennett & Lemoine, 2014) world we live in, is that overconfidence has become an all-too-common trap. Overconfidence is the opposite of what we currently need. One of the best examples of a leader who inoculated herself against the penalty of being wrong is the Prime Minister of Denmark Mette Frederiksen. In her address to the nation in March 2020 at the beginning of the pandemic, surrounded by medical and public health experts:
the Prime Minister showed humility when faced with the overwhelming task of managing the COVID crisis: ‘We are entering an unknown territory. The situation we find ourselves in is not similar to anything we have previously experienced. Will we make any mistakes? Yes. Will I make any mistakes? Yes.” (Søderberg, 2022).
The Danish Prime Minister continued by explaining that she would tell the public as soon as the experts identified new or better directions to take. She warned Danes to expect mistakes, changes, and learning on the job. As did almost all national leaders, Denmark’s Prime Minister altered policies based on new information. However, unlike most leaders, she was not penalized during that initial year of the pandemic for having been wrong. Strategic humility is core to understanding her leadership.
4.2 Unlearning: Letting go of overly convenient falsehoods
To unlearn a convenient falsehood, we must first recognize how we benefit from continuing to believe something that is either unproven or untrue. In my case, as I flew home from Italy to Montreal on March 13th, 2020, I wanted to believe that the global public health officials knew what they were doing and could protect me against the threats of a deadly virus. I wanted to believe that the airport authorities knew how to keep passengers safe—myself included. I wanted to believe that my home country, Canada, one of the most advantaged countries on earth, would succeed in keeping its residents healthy. They didn’t and couldn’t. My desire for safety overwhelmed my discernment. Fear blinded me. I failed to recognize that what I believed (and wanted to believe) was convenient, but false. I was seemingly incapable of recognizing that neither I nor the general public was safe. I failed to appreciate that there were no experts anywhere in the world who, at that time, knew how to keep us safe. The morning after I arrived home, I woke up with COVID. I was sick, but lucky. I was not hospitalized; I did not die. The health professionals told me I had the flu, not COVID. A week later when the first tests became available, my COVID-positive diagnosis was confirmed.TableUNLEARNING: LETTING GO OF OVERLY CONVENIENT FALSEHOODS Consider a belief you hold that you are suspicious is no longer true if it ever was. How do you benefit from continuing to hold this untrue belief? Why is this false belief so convenient for you?
4.3 Recognizing inconvenient truths in new ideas
In the same way that overly convenient falsehoods harbor benefits that we do not want to let go of, new ideas often contain implications that we fear will be detrimental to us. Emerging ideas frequently have inconvenient truths embedded within them. To identify such inconvenient truths, ask yourself:TableRECOGNIZING INCONVENIENT TRUTHS What new ideas are trying to get your attention? How might accepting them disadvantage you? What is most inconvenient for you about these new ideas? How are you blinding yourself from seeing their inherent possibilities?
4.4 Courageous questions: Not what leaders thought they would learn
The fundamental question for leaders and leadership scholars is: Are we willing to ask big questions that have the potential to make a significant difference in the world? In asking those questions, are we willing not just to learn, but also to unlearn what we and the field have always assumed to be true but, in reality, is false? Such learning is never easy.
What might great leadership look like? Thought leadership. Political leadership? Organizational leadership? Against the backdrop of today’s volatile, dangerous, distrustful, self-interested, uncertain, complex, ambiguous world, Ukraine’s President Zelensky has given the world an inspiring image of leadership. Zelensky did not study for an MBA, nor was he fast-tracked into senior leadership in a major global company’s high-potential program. Working initially as an artist, it’s doubtful if he read many leadership books or scholarly management articles. What he has given the world though is a startingly positive image of what leadership could and does look like. Courageous, passionate, honest, articulate, caring, strategic, constantly evolving, and indefatigable. On February 24th, 2022, Zelensky now famously responded to the Americans’ offer of asylum, by asserting, “The fight is here; I need ammunition, not a ride (Zelensky, 2022b). Zelensky challenges us to think about the world and leadership differently. With his constant cadence of learning and unlearning, he demonstrates how powerful humility can be. He invites us to passionately care. He invites us to be smart, strategic, and honest. He pleads with us, as he so eloquently did in his speech at the Grammy’s on April 3rd, 2022, not to be silent (Zelensky, 2022a).“The war. What is more opposite to music? The silence of ruined cities and killed people,” Zelensky said. “Our children draw swooping rockets, not shooting stars. Over 400 children have been injured and 153 children died. And we’ll never see them drawing. Our parents are happy to wake up in the morning in bomb shelters – but alive.
“Our loved ones don’t know if we will be together again. The war doesn’t let us choose who survives and who stays in eternal silence. Our musicians wear body armors instead of tuxedos. They sing to the wounded in hospitals. Even to those who can’t hear them. But the music will break through anyway. We defend our freedom. To live. To love. To sound.
“On our land, we are fighting Russia which brings horrible silence with its bombs. The dead silence. Fill the silence with your music. Fill it today to tell our story. Tell the truth about this war on your social networks, on TV. Support us, in any way you can. Any – but not silence. And then peace will come.
“To all our cities the war is destroying. Chernihiv, Kharkiv, Volnovakha, Mariupol and others. They are legends already. But I have a dream of them living, and free – free like you on the Grammy stage.”
Zelensky’s invitation is not convenient. In the mirror of his urging, we know we can do more; that we must do more. We know that we must ask bigger questions that have the potential to make a more meaningful difference. Even when it scares us, we are forced to recognize that we are capable of acting on that bigger stage that he is inviting us onto. We know that the world needs who we are as well as what we know. The pull to return to ‘normal’ – whatever that means – reveals itself to be a bankrupt concept. As scholars, we cannot merely publish another article in an academic journal that has little impact on any ‘real world’ issue; an article that often seems to simply serve to lengthen our C.V.
Perhaps Martin Luther King, Jr. (King, 1964) best captured our purpose as international scholars and global leaders in his Nobel Peace Prize lecture:
Some years ago a famous novelist died. Among his papers was found a list of suggested plots for future stories, the most prominently underscored being this one: A widely separated family inherits a house in which they have to live together. This is the great new problem of humankind.12 We have inherited a large house, a great world house in which we have to live together, black and white, Easterner and Westerner, Gentile and Jew, Catholic and Protestant, Moslem and Hindu—a family unduly separated in ideas, culture and interest[s], who, because we can never again live apart, must learn somehow to live with each other in peace.
Let’s make 5-year-old Kiara proud of what we ‘big people’ have learned, are learning, and will learn. Let’s let Kiara respect us ‘big people’ for what we have had the courage and humility to unlearn. Let’s leave Kiara excited and in awe of the future that we ‘big people’ have bequeathed, not just to her generation but to her children and her children’s children.
5 Afterword
The special issue of International Business Review, in which this article appears, on Looking Back to Look Forward: Disruption, Innovation and Future Trends in International Human Resource Management (Farndale et al., 2023 ) is critically important. It asks all of us to rethink our role as human beings and the role that human systems could play in creating a better world. The fundamental question for scholars in the field of international human resource management (IHRM), as for scholars in all disciplines, is: Are we willing to ask big questions that have the potential to make a significant difference in the world? As a part of our methodology and overall research process, are we willing not just to learn, but also to unlearn what we and the field have previously assumed to be true but, in reality, is false? Chris Brewster, a leader in the field of IHRM, exemplifies such questioning, and the need for unlearning, in his most recent research. He challenges the hierarchical orthodoxy of the field and its embedded assumptions, in his current questioning of the “elitism in international human resource studies” (Brewster, 2022). Brewster asserts that
International Human Resources management (like business and management studies…) has operated from an elitest base that not only led to a distortion in our research but has led to some unwanted, negative, and even nefarious outcomes. This has been caused by our focus on the interests of owners of business (when the vast majority of people are not owners); on large international organizations (when 95% of people work in small local firms); on managers and executives (when 80% of people are neither); on talent management (when 95% of people are not classified as talent); [and] on western organizations (when most of the world does not fit into that category) …. What this means is that we have largely ignored the problems and issues of most people in the world. I suggest that focusing on the non-elite will avoid contributing further to the negative impact our work has created so far and positively address some of the Grand Challenges that the world is facing (Brewster, 2022).
The challenge for us as scholars, in all domains, is to be willing to question the orthodoxies of our field and our own worldview. It is to have the courage and confident humility to let go of convenient falsehoods and to embrace inconvenient truths.
Data availability
No data was used for the research described in the article.
☆ PRE-PUBLICATION: To be published in International Business Review in the Special Issue on Looking Back to Look Forward: Disruption, Innovation and Future Trends in International Human Resource Management. It was originally presented as a keynote address entitled Not What We thought We Would Learn: Global Leadership Lessons in a World of Grand Challenges at the Fifth Global Conference on International Human Resource Management in New York City on May 20th 2022.
2 For a discussion of arts-based approaches to seeing reality as it is, see Adler (2006).
3 According to Dani Rodrik, professor of international political economy at Harvard’s Kennedy School, as cited in Wong and Swanson (2022) “Your interdependence can be weaponized against you” … “That’s a lesson that I imagine many countries are beginning to internalize.”
4 For a discussion of Russia’s invasion of Ukraine and how it has brought into question core concepts within the field of International Business, see the Academy of International Business Central and Eastern European Chapter Webinar on “War in Ukraine: Post-invasion Geopolitics and Implications for IB” held on March 31, 2022 with Michael Witt (INSEAD based in Singapore), Timothy Devinney (Alliance Manchester Business School (United Kingdom), Taras Danko (National Technical University, Ukraine), Andreja Jaklič (University of Ljubljana, Slovenia), Piotr Trapczynski (Poznan University of Economics and Business, Poland), and Dana Minbaeva, Kings College London (United Kingdom).
5 The value that scholars bring to public discourse, beyond their content knowledge, was presented by Paul Vaaler and Sri Zaheer at the Academy of International Business Meetings in Miami on July 7, 2022, in the session on “Doing Impactful Research.”
6 Think Again by Adam Grant Publisher’s Description as found on July 24, 2022 at Book Queen at https://www.booklistqueen.com/think-again-adam-grant/#:∼:text=In%20our%20daily%20lives%2C%20too,than%20an%20opportunity%20to%20learn. Accessed July 7th, 2022.
7 For the reality of that inconvenience, see MacFarquharil & Lobzina (2022)
8 The film, An Inconvenient Truth, can be viewed at: https://www.youtube.com/watch?v=8ZUoYGAI5i0
9 For exceptions, see Hoekstra et al. (2008) and Morris et al. (2005) among others.
10 See Press Release: CDC Confirms Person-to-Person Spread of New Coronavirus in the United States, Centers for Disease Control and Prevention: CDC Newsroom, January 30, 2020. Retrieved from https://www.cdc.gov/media/releases/2020/p0130-coronavirus-spread.html. Accessed on July 26th, 2022.
11 Worldwide, Monaco has the oldest population with a median age of 53; and Japan’s median age is more than 47 years old.
12 ‘Mankind’ in the original has been changed to humankind.
==== Refs
References
Adler N.J. Global women political leaders: An invisible history, an increasingly important future Leadership Quarterly 7 1 1996 133 161
Adler N.J. Global leaders: A dialogue with future history International Management 1 2 1997 21 33
Adler N.J. Global leadership: Women leaders Management International Review 37 1 1997 171 196
Adler N.J. Did you hear? Global leadership in charity’s world Journal of Management Inquiry 7 2 1998 135 143
Adler N.J. Global managers: No longer men alone International Journal of Human Resource Management 13 5 2002 743 760
Adler N.J. The arts and leadership: Now that we can do anything, what will we do? Academy of Management Learning and Education 5 4 2006 466 499
Adler N.J. The United States: Elusive success. In N. J. Adler & S. Sackmann, The grand challenge none of us chose: Succeeding (and failing) against the global pandemic Osland J.S. Reiche B.S. Szkudlarek B. Mendenhall M.E. Advances in Global Leadership Vol. 14 2022 Emerald Bingley, UK 69 72
Adler N.J. Sackmann S.A. The grand challenge none of us chose: Succeeding (and failing) against the global pandemic. With S. Arieli, M. Akter, C. Barmeyer, C. Barzantny, D. V. Caprar, Y. Lee, L.A. Liu, G. Magnani, J. Marcus, C. Miska, F. Moore, S.H. Park, B. S. Reiche, A-M Søderberg, J. Solomons, and Z.-X. Zhang Osland J.S. Reiche B.S. Szkudlarek B. Mendenhall M.E. Advances in Global Leadership Vol. 14 2022 Emerald Bingley, UK 3 85
Bennett, Nate & Lemoine, G.James (2014). What VUCA Really Means for You. Harvard Business Review, January-February. Retrieved from 〈https://hbr.org/2014/01/what-vuca-really-means-for-you〉 on August 1 2022.
Bennhold, Katrin (2022). The former chancellor who became Putin’s man in Germany. New York Times, April 23.
Bollyky, Thomas J. & Nuzzo, Jennifer B.. (2020). Trump’s ‘early’ travel ‘bans’ weren’t early, weren’t bans and didn’t work. The Washington Post, October 1st. Retrieved from 〈https://www.washingtonpost.com/outlook/2020/10/01/debate-early-travel-bans-china/〉 Accessed May 8th 2022.
Braw, Elisabeth (2020). How Italy snatched health from the jaws of death, Foreign Policy, September 23. Retrieved from 〈https://foreignpolicy.com/2020/09/23/italy-mismanagment-covid-19-health-pandemic/〉 Accessed on July 26 2022.
Brewster, Chris (2022). Questioning Elitism in International Human Resource Management Studies. IHRM Webinar Series organized on Zoom by the Centre for Global Workforce Strategy at Simon Fraser University (Canada), the Center for International Human Resource Studies at the Pennsylvania State University (USA), Pennsylvania Western University (USA), and ESCP Business School and RIT Croatia. September 15th.
Eder, Steve; Fountain, Henry; Keller, Michael H.; Xiao, Muyi; and Stevenson, Alexandra (2020). 430,000 people have traveled from China to U.S. since coronavirus surfaced. New York Times, April 4, 2020, update September 21, 2021.
Farndale, Elaine; Horak, Sven; Piyanontalee, Rakoon; Maja Vidović, Maja. (2023). Looking Back to Look Forward: Disruption, Innovation and Future Trends in International Human Resource Management. International Business Review.
Friedman, Thomas (2022). The war is getting more dangerous for America, and Biden knows it. New York Times, May 6.
Grant A. Think again: The power of knowing what you don’t know 2021 Penguin Random House LLC, New York
Harzing Anne-Wil Adler N.J. Disseminating Knowledge: From Potential to Reality–New Open-Access Journals Collide with Convention Academy of Management Learning and Education 15 1 2016 140 156
Hoekstra E. Bell A. Peterson S.R. Humility in leadership: abandoning the pursuit of unattainable perfection Quatro S.A. Sims R.R. Executive ethics: Ethical dilemmas and challenges for the C-suite 2008 Information Age Publishing Greenwich, CT
King, Martin Luther, Jr. (1964). The World House. Chapter in his last book, Where Do We Go from Here: Chaos or Community? taken from his Noble Prize Lecture at the University of Oslo on December 11, 1964. As reported in “Martin Luther King Jr. on the presidential election of 1964 and the path to healing”, Fierce Urgency: The Journal and Blog of the Institute for Nonviolence and Social Justice. Retrieved from 〈https://usfblogs.usfca.edu/fierce-urgency/2020/11/09/martin-luther-king-jr-on-the-presidential-election-of-1964-and-the-path-to-healing/#:∼:text=Some%20years%20ago%20a%20famous,great%20new%20problem%20of%20mankind〉. Posted November 9, 2020 by Jonathan Greenberg. Accessed May 15, 2022.
Kuhn T. The Structure of Scientific Revolutions 2012 1962 The University of Chicago Press Chicago
Kuhn T. The Structure of Scientific Revolutions 50th Anniversary edition 1962 The University of Chicago Press Chicago
Levenstein, Susan (2020). Coronavirus: ‘Italy is notorious for disorganization, but this time it’s bested the US’. The Local-IT, March 6. Retrieved from 〈https://www.thelocal.it/20200306/coronavirus-italy-is-notorious-for-disorganization-but-this-time-its-bested-the-us/〉. Accessed July 26 2022.
MacFarquhar, Neil & Lobzina, Alina (2022). Thousands swept up as Kremlin clamps down on war criticism. New York Times, June 3.
Morris J.Andrew Brotheridge C.éleste M. Urbanski John C. Bringing humility to leadership: Antecedents and consequences of leader humility Human Relations 58 10 2005 1323 1350 10.1177/0018726705059929
Palmer P.J. The Courage To Teach: Exploring The Inner Landscape Of A Teacher's Life 10th anniversary edition 2007 Jossey Bass San Francisco
Rabin, Roni Caryn (2022). U.S. Life Expectancy Falls Again in ‘Historic’ Setback. New York Times. August 30.
Rich, Motoko & Dooley, Ben (2022). Japan’s secret to taming the coronavirus: Peer pressure. New York Times. July 2.
Rich, Nathaniel (2022). Everything you thought you knew, and why you’re wrong. New York Times, May 11.
Søderberg A.-M. Denmark: Political communication and crisis management based on a worst-case scenario. In N. J. Adler & S. Sackmann, The grand challenge none of us chose: Succeeding (and failing) against the global pandemic Osland J.S. Reiche B.S. Szkudlarek B. Mendenhall M.E. Advances in Global Leadership Vol. 14 2022 Emerald Bingley, UK 53 57
Tchir, Peter (2020). What if Buffett is the one swimming naked? Forbes, May 4.
Viola, Lorella (2022). Italy, for example, is just incredibly stupid now. European crisis narrations in relation to Italy's response to COVID-19. Frontiers in Communication, February 2. Retried from 〈https://www.frontiersin.org/articles/10.3389/fcomm.2022.757847/full〉. Accessed on July 26, 2022.
Wong, Edward & Swanson, Ana (2022). Ukraine war and pandemic force nations to retreat from globalization. New York Times, March 22.
Zelensky, Volodymyr (2022a). April 3rd 2022 Grammy Speech. As reported by Josh Sanchez (2022) Volodymyr Zelensky Grammy Speech: Full Transcript. Coed, April 4th. Retrieved from 〈https://coed.com/2022/04/04/volodymyr-zelensky-grammy-speech-full-transcript/〉. Accessed on May 16th.
Zelensky, Volodymyr (2022b). as cited by the National Post staff, in “‘I need ammunition, not a ride’: Zelensky turns down U.S. evacuation offer. The National Post, Ottawa, Canada, February 26th 2022. Retrieved from 〈https://nationalpost.com/news/world/i-need-ammunition-not-a-ride-zelensky-turns-down-u-s-evacuation-offer〉. Accessed on May 12 2022.
| 0 | PMC9746791 | NO-CC CODE | 2022-12-15 23:21:57 | no | Int Bus Rev. 2022 Dec 13;:102083 | utf-8 | Int Bus Rev | 2,022 | 10.1016/j.ibusrev.2022.102083 | oa_other |
==== Front
J Allergy Clin Immunol
J Allergy Clin Immunol
The Journal of Allergy and Clinical Immunology
0091-6749
1097-6825
American Academy of Allergy, Asthma & Immunology
S0091-6749(22)01568-8
10.1016/j.jaci.2022.11.010
Review Article
Impact of SARS-CoV-2 infection and COVID-19 on patients with inborn errors of immunity
Tangye Stuart G. PhD ∗
for the COVID Human Genetic Effort consortium∗
Garvan Institute of Medical Research, Darlinghurst, Darlinghurst, Australia
St Vincent’s Clinical School, University of New South Wales Sydney, Randwick, Randwick, Australia
Clinical Immunogenomics Research Consortium of Australasia (CIRCA)
∗ Corresponding author: Stuart G. Tangye, PhD, Garvan Institute of Medical Research, 384 Victoria St, Darlinghurst, NSW, 2010 Australia.
∗ Members of the COVID Human Genetic Effort consortium are listed in the Acknowledgments at the end of the article.
13 12 2022
13 12 2022
29 9 2022
2 11 2022
4 11 2022
© 2022 American Academy of Allergy, Asthma & Immunology.
2022
American Academy of Allergy, Asthma & Immunology
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Since the arrival of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in December 2019, its characterization as a novel human pathogen, and the resulting coronavirus disease 2019 (COVID-19) pandemic, over 6.5 million people have died worldwide—a stark and sobering reminder of the fundamental and nonredundant roles of the innate and adaptive immune systems in host defense against emerging pathogens. Inborn errors of immunity (IEI) are caused by germline variants, typically in single genes. IEI are characterized by defects in development and/or function of cells involved in immunity and host defense, rendering individuals highly susceptible to severe, recurrent, and sometimes fatal infections, as well as immune dysregulatory conditions such as autoinflammation, autoimmunity, and allergy. The study of IEI has revealed key insights into the molecular and cellular requirements for immune-mediated protection against infectious diseases. Indeed, this has been exemplified by assessing the impact of SARS-CoV-2 infection in individuals with previously diagnosed IEI, as well as analyzing rare cases of severe COVID-19 in otherwise healthy individuals. This approach has defined fundamental aspects of mechanisms of disease pathogenesis, immunopathology in the context of infection with a novel pathogen, and therapeutic options to mitigate severe disease. This review summarizes these findings and illustrates how the study of these rare experiments of nature can inform key features of human immunology, which can then be leveraged to improve therapies for treating emerging and established infectious diseases.
Key words
SARS-CoV-2
COVID-19
inborn errors of immunity
primary immune deficiencies
immune dysregulation
type I IFN signaling
cytokine storm
Abbreviations used
APECED, Autoimmune polyendocrinopathy candidiasis ectodermal dystrophy
AR, Autosomal recessive
BTK, Bruton tyrosine kinase
CFR, Case fatality rate
COVID-19, Coronavirus disease 2019
CVID, Common variable immunodeficiency
GOF, Gain of function
ICU, Intensive care unit
IEI, Inborn errors of immunity
JAK, Janus kinase
mAb, Monoclonal antibody
SARS-CoV-2, Severe acute respiratory syndrome coronavirus 2
XL, X linked
XLA, XL agammaglobulinemia
==== Body
pmcInborn errors of immunity (IEI) are diseases caused by germline pathogenic variants, typically in single genes.1, 2, 3, 4 IEI have an incidence of ∼1 per 5,000 to 10,000 individuals.1, 2, 3, 4, 5 Currently, pathogenic variants in more than 480 genes have been identified that cause IEI. These variants can lead to loss of expression, complete (null) or partial (hypomorphic) loss of function, gain of function (GOF; hypermorphic), haploinsufficiency, or dominant negative function of the encoded protein. IEI can present as autosomal dominant (AD; heterozygous variants), autosomal recessive (AR; homozygous/compound heterozygous variants), or X-linked (XL) recessive (hemizygous in male subjects; homozygous or heterozygous with skewed X inactivation in female subjects) conditions.4 , 6 However, some IEI have incomplete penetrance, with a significant proportion of individuals carrying some pathogenic variants compromising protein function remaining unaffected.7 The mechanism or mechanisms underlying incomplete penetrance remain unclear but may involve epistatic effects of modifier genes, epigenetics, and/or variants in additional genes.7 It is also worth noting that a monogenic cause for the most common IEI—common variable immunodeficiency (CVID)—has only been determined for ∼20-30% of affected individuals,8 thus suggesting that most cases of CVID are likely to be oligo- or polygenic.
IEI are characterized by defects in immune cell development, and/or impaired innate and adaptive immune function of hematopoietic and nonhematopoietic cells. Consequently, affected individuals are highly susceptible to severe, recurrent, and sometimes fatal infections.4 , 6 As a result of this immunodeficient state, vaccine efficacy can also be compromised in IEI, resulting in affected individuals having modest, if any, vaccine-induced immunity against infectious diseases. Thus, IEI patients continue to be susceptible to infection as well as being vulnerable to disease as a result of live-attenuated vaccines.9
Although historically considered to be immune deficiencies manifesting as infections, the clinical spectrum of IEI is extremely broad, with autoimmunity, autoinflammatory diseases, allergy, bone marrow failure, and/or malignancy also being common maladies of patients.1 , 3 , 4 , 6 , 10 , 11 Although most are individually rare, IEI are collectively common5 and have enabled the delineation of fundamental roles of individual genes, proteins, signaling pathways, and cell types in immune cell development; immune homeostasis and regulation; antitumor immunity; and host defense against infectious diseases.1, 2, 3 Thus, IEI provide insights into the molecular pathogenesis of more common diseases and have led to the development of targeted therapies for various immune dyscrasias.1, 2, 3 , 12
SARS-CoV-2 and the COVID-19 pandemic
Coronaviruses have caused pandemics in the human population for decades.13 Certainly we would have a short memory if we failed to recall the deadly toll of the original SARS coronavirus outbreak in 2002-3.13 In December 2019, the novel coronavirus severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) emerged from Wuhan, China, and then spread rapidly to cause a catastrophic global pandemic.14 At the time of writing, more than 650 million people have been infected and at least 6.6 million people have died from SARS-CoV-2 infection (www.covid19.who.int/, www.worldometers.info/coronavirus/). The clinical spectrum of coronavirus disease 2019 (COVID-19) due to SARS-CoV-2 infection ranges from asymptomatic to life-threatening disease. The global case fatality rate (CFR) due to SARS-CoV-2 infection is currently ∼1.1%, but this varies widely across different countries, ranging from 0.1% to 5%, and even up to 10% to 15% for some regions (www.ourworldindata.org/grapher/deaths-covid-19-vs-case-fatality-rate). Importantly, early during the pandemic, when viral screening was restricted to symptomatic individuals and vaccines were still 12 to 18 months away, the average global CFR was 5% to 7%, and as high as 10% to 20% in the United Kingdom and some European countries15 , 16 (www.ourworldindata.org/grapher/deaths-covid-19-vs-case-fatality-rate).
Several risk factors have been identified for developing severe disease, as defined by the World Health Organization. These include primarily age, with the frequency of severe cases/death escalating with each decade of increasing age. For example, the mortality rate for people aged <50 years was <1.0%; for individuals aged 60-80 or more years, the mortality rate was ∼4% to 25%. Male sex as well as comorbidities such as cardiovascular/pulmonary disease, obesity, diabetes, and liver/kidney dysfunction also have an impact, albeit less than age.16, 17, 18, 19 Correlates of severe disease and mortality include lymphopenia, increased levels of inflammatory mediators, cytokines, chemokines,18 , 20, 21, 22, 23, 24, 25, 26 and complement components,27, 28, 29 which indicate the intense immune activation and inflammation that can lead to severe and potentially fatal SARS-CoV-2–induced cytokine storm and consequent tissue pathology.
In healthy individuals, SARS-CoV-2 infection induces functional CD4+ and CD8+ T cells and memory B cells specific for viral epitopes, as well as neutralizing antibodies.30, 31, 32, 33, 34, 35, 36, 37, 38, 39 These correlates of protective immunity are detectable 1 or 2 weeks after infection and persist at peak levels for 3 to 4 months. However, in most cases, levels of neutralizing IgG and of SARS-CoV-2–specific memory B cells and T cells dramatically wane 8 to 12 months after infection,32, 33, 34 , 36, 37, 38, 39, 40 potentially compromising host defense against subsequent infections. Furthermore, several SARS-CoV-2 variants that have acquired mutations in the immunodominant spike domain, thus rendering these variants less susceptible to antibody-mediated neutralization, have emerged.41 Waning of acquired immunity after natural infection, combined with immune-escape variants, are a significant challenge in controlling SARS-CoV-2 infection, resulting in COVID-19 continuing to represent a significant global health risk.
SARS-CoV-2 infection, COVID-19, and IEI
Since the beginning of the pandemic, it was recognized that people diagnosed with an IEI were potentially at risk of developing severe COVID-19. Over the past 2 years, outcomes of SARS-CoV-2 infection have been reported for ∼1330 individuals with IEI. These studies range from reports of single cases or small numbers of patients42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, 72, 73, 74, 75, 76, 77, 78, 79, 80, 81, 82, 83, 84, 85, 86, 87, 88 to cohort studies from Iran,89, 90, 91 Turkey,92, 93, 94 Brazil,95 , 96 Israel,97 Italy,98, 99, 100, 101 Spain,102 the United Kingdom,15 , 103 , 104 Mexico,105 Denmark,106 , 107 Poland,108 the Czech Republic,109 France,110 and the United States,111, 112, 113, 114 as well as an international survey of 94 patients followed in 12 countries.115 These studies have revealed key outcomes of SARS-CoV-2 infection in IEI and defined fundamental requirements for host defense against infection.
Patients with IEI infected with SARS-CoV-2
Affected patients have been found to represent most, if not all, categories of IEI as defined by the International Union of Immunological Societies Committee (Table I ).6 Of the ∼1330 patients reported so far, approximately 60% have antibody deficiencies, consistent with antibody deficiency being the most common IEI.6 , 8 This includes CVID, hypogammaglobulinemia, and specific antibody and immunoglobulin subclass deficiencies due to unknown genetic causes8 (∼600 cases), as well as XL (BTK pathogenic variants) and AR (eg, TCF3 pathogenic variants) agammaglobulinemia (∼110 cases) and a series of patients with pathogenic variants in single genes known to disrupt B-cell function and humoral immunity, such as NFKB1, NFKB2, PIK3CD, or PIK3R1 (Table I). Outcomes of SARS-CoV-2 infections have also been reported for patients with the following:• Severe combined (JAK3, RAG, IL7RA, DCLRE1C) or combined (CD40LG, RASGRP1, RELB, STK4, WAS, ICOS, ATM, IKBKG, STAT3 DN, PGM3) immunodeficiencies.
• Immune dysregulatory disorders (STAT3 GOF, AIRE, CTLA4, CD70, LRBA, RAB27A, SH2D1A, XIAP, RLTPR/CARML2, CD137, STXBP2, ALPS).
• Phagocytic defects (chronic granulomatous disease, GATA2).
• Innate immune defects (IFNGR1, IFNGR2, IFNAR1, IFNAR2, IL12RB1, IRAK4, MYD88, STAT1 GOF, CXCR4, TBK1, TLR3, TLR7, IRF3, IRF7, IRF9).
• Autoinflammatory disorders (MEFV, TNFAIP3, IL36R, ADA2).
• Complement deficiencies.
• Phenocopies of IEI.
Table I SARS-CoV-2 infection in defined IEI
Type of IEI Gene defect/IEI Approximate no. of patients Study or studies
Severe combined immunodeficiency (n = 25) JAK3 1 70
RAG 3 92, 97, 115
IL7RA 1 91, 94
DCLRE1C 1 49
IL2RG 4 77, 95, 115
CD3D 1 105
Not specified 15 95, 99, 108
Combined immunodeficiency (n = 91) STAT3 DN 7 103, 109, 115, 176
PGM3 1 102, 115
ARPC1B 1 47, 105, 115
WAS 8 47, 48, 95, 99, 100, 103, 105, 108, 109, 115
ZAP70 1 115
CD40L 9 94, 95, 97, 103, 109, 111, 116, 143
RASGRP1 1 92
CARD11 1 92, 103
RELB 3 97, 116
STK4 1 89
DNMT3B/NBS1 4 89, 91, 94
ICOS 1 15, 103
IKBKG (NEMO) 3 72, 78, 94
ATM 11 91, 92, 94, 99, 100, 102, 103, 108
Di George syndrome 16 99, 100, 108
Not specified 23 89, 92, 94, 95, 99, 103, 108
Predominantly antibody deficient (n = 714) CVID∗ 589 51, 52, 58, 71, 75, 83, 92, 94, 95, 97, 98, 99, 100, 102, 103, 104, 105, 106, 107, 108, 109, 111, 112, 113, 114, 115, 143
BTK 98 15, 46, 51, 53, 55, 60, 61, 66, 73, 85, 86, 91, 92, 94, 95, 97, 98, 99, 100, 102, 103, 104, 105, 108, 109, 111, 115, 116, 139, 140, 143
AR agammaglobulinemia 9 99, 100, 115
PIK3R1/PIK3CD GOF 7 64, 82, 91, 95, 99, 100, 115
NFKB1 4 15, 91, 103, 111, 115
NFKB2 3 43, 103, 115, 143
IKZF1 1 91
Immune dysregulation (n = 64) AIRE (APS1/APECED) 29 57, 84, 94, 118, 122, 149
CTLA4 7 15, 97, 103, 115, 177
LRBA 3 92, 97, 115
SOCS1 1 76
STAT3 GOF 1 111
RAB27A 1 89
CD70 1 89
ALPS 5 95, 99, 102, 108
XLP (XIAP, SH2D1A) 4 63, 95, 108, 109, 115
PRKCD 1 115
RLTPR/CARMIL2 2 94
CD137 1 94
STXBP2 2 88, 94
Not specified/other 6 92, 99, 105, 108
Phagocytic defects, bone marrow failure (n = 36) Chronic granulomatous disease (CYBB; NCF2) 28 15, 59, 89, 95, 97, 102, 103, 105, 108, 115
GATA2 2 15, 103, 115
DNAJC21 1 115
Not specified/other 5 92, 99
Innate immune defects (n = 75) TLR3/UNC93B/TRIF/IRF3/IRF7/IRF9/TBK1 23 65, 68, 69, 120, 123
IFNAR1/2 7 42, 56, 87, 126
STAT1/TYK2 2 126
TLR7 22 90, 124, 125, 126
MYD88/IRAK4 8 45, 81, 95, 99, 102
IFNGR1/IFNGR2/IL12RB1 5 54, 79, 95, 111, 115
STAT1 GOF 6 50, 92, 95, 102, 109, 115
CXCR4 GOF 2 94, 95
Autoinflammatory disorders (n = 96) MEFV 68 93, 95, 110, 115
IL1RN 1 89
Aicardi-Goutières syndrome (RNASEH2B, SAMHD1) 5 15, 99, 100, 115
TNFAIP3 1 15
NLRP1, NLRP3, NLRP12 3 91, 95
IL36RN 1 74
ADA2 1 94
Not specified/other 16 95, 108
Complement deficiencies (n = 55) Hereditary angioedema (pathogenic SERPING variants), C3 deficiency, other 55 15, 91, 95, 96, 109
Phenocopies of IEI Good syndrome 13 83, 100, 103, 105, 109
Autoantibodies to type I IFNs Many! 128, 129, 130, 131, 132, 133, 134, 135, 136
∗ Including hypogamma, immunoglobulin subclass deficiency, and specific antibody deficiency.
A complete reference listing is provided in Table I.
Clinical features in IEI after SARS-CoV-2 infection
The clinical presentation of SARS-CoV-2 infection in patients with IEI resembled that of the general population16 , 19 inasmuch that symptoms frequently include fever, cough, headache, upper respiratory symptoms, fatigue, and dyspnea.15 , 89 , 91 , 92 , 94 , 95 , 97 , 99 , 102 , 103 , 105 , 106 , 108 , 109 , 115 Similarly, risk factors for hospital/intensive care unit (ICU) admission and developing severe and/or fatal disease were also consistent with those determined from studies of the general population. Thus, the most severe disease was observed in older patients with IEI as well as those with pre-existing comorbidities, such as previous infection; lung, kidney, heart, or gut disease, diabetes, and obesity; or after solid organ or hematopoietic stem cell transplantation.43 , 58 , 89 , 91 , 92 , 94 , 95 , 98 , 103 , 105 , 106 , 108 , 111 , 112 , 115 Other predictors of severe disease in IEI patients included leukopenia (reduced numbers of B, CD4+ T, and natural killer cells) and hypogammaglobulinemia/low IgG trough levels before infection, and increased levels of markers of systemic inflammation after infection.43 , 46 , 58 , 66 , 103 , 109 , 111 , 114 Interestingly, and similar to the general population, ∼10% to 20% of infected IEI patients were asymptomatic, and up to another ∼30% to 50% developed only mild disease.15 , 89 , 90 , 92, 93, 94, 95 , 97, 98, 99, 100, 101, 102, 103, 104, 105, 106 , 108, 109, 110, 111, 112, 113, 114, 115
Despite such similarities in disease presentation and risk factors for the general population and IEI patients, there were notable differences. First, the age of affected IEI patients was markedly younger than the general population (∼28 years vs ∼50-plus years).16 , 44 , 79 , 81 , 84 , 89 , 91 , 93, 94, 95 , 100 , 102, 103, 104 , 108, 109, 110 , 115 There were also differences in age at infection for different IEI. Thus, SARS-CoV-2–infected patients with CVID, periodic fevers, or complement defects were generally older, and patients with defects in innate immune cell signaling due to pathogenic variants in IRAK4, MYD88, or IFNAR1/IFNAR2 were generally younger, than the entire cohort of published IEI patients (Fig 1 , A). Second, the proportion of IEI patients admitted to ICU—including younger individuals—was substantially higher than the general population (10-30% vs 2-5%).15 , 16 , 44 , 81 , 91 , 92 , 94 , 102 , 105 , 109 , 111 , 115 Third, duration of disease—likely a result of prolonged viremia and virus shedding—was longer (1-6 months vs 1-2 weeks), and the likelihood of reinfection was greater, than observed for the general population.46 , 55 , 59 , 77 , 83 , 84 , 92 , 99 , 100 , 104 , 106 , 107 , 116 Thus, COVID-19 generally manifests clinically at a younger age, runs a more protracted course, and has a more severe outcome requiring hospitalization and/or ICU admission in many individuals with IEI compared to the epidemiology of SARS-CoV-2 infection in the general population (Fig 2 ).16 , 44 This is reminiscent of findings for SARS-CoV-2 infection in patients with cystic fibrosis. Here, it was found that many cystic fibrosis patients had mild disease and common risk factors such as diabetes and previous solid organ transplantation, but subgroups of patients exhibited increased hospitalization rates and younger age at presentation relative to the general population.117 Fig 1 Features of cohorts of patients with IEI and SARS-CoV-2 infection. (A) Age of patients with the indicated IEI. Data are shown as median ages and quartiles for each patient group. Values above each data set represent the mean ages of patients with the indicated IEI. (B) CFR for all IEI patients, as well as range for the CFR in the general population (www.covid19.who.int/, www.worldometers.info/coronavirus/). Values in each patient group represent the number of deaths/total number of patients with the indicated IEI. Agamma, Agammaglobulinemia; AIRE, patients with APECED; ATM, ataxia telangiectasia; C’ def, complement deficiency; CGD, chronic granulomatous disease; FMF, familial Mediterranean fever; IFNAR1/2, patients with pathogenic variants in type I IFN receptors; IRF7, MYD88/IRAK4, patients with pathogenic variants in IRF7 or MYD88/IRAK4 that disrupt type I IFN signaling.
Fig 2 Consequences and outcomes of SARS-CoV-2 infection in patients with IEI.
Mortality due to SARS-CoV-2 infection in IEI
Depending on the country or region where different studies have been performed, as well as the size of the cohort being investigated, the CFR after SARS-CoV-2 infection in patients with IEI is highly variable, being 0,97 , 102 , 106 , 113 , 114 2% to 5%,99 , 101 , 107, 108, 109 5% to 10%,92 , 95 , 103 , 104 15% to 20%,105 , 112 20% to 30%,94 , 111 , 118 and >30%.15 , 89 , 91 From all available published studies, 113 of 1328 patients with IEI died after SARS-CoV-2 infection, resulting in an overall CFR of 8.5% (Fig 1, B). Remarkably, this is highly similar to the CFR reported by Meyts et al115 for an international survey of 94 patients with a broad range of IEI recruited from 12 countries (9.4%). The significant variability in CFR reported for many studies likely reflects the type of cohort being analyzed (eg, children vs adults; predominantly CVID due to unknown genetic defects vs severe combined immunodeficiency/combined immunodeficiency),108 the predominant SARS-CoV-2 variant at the time of study,41 the burden of SARS-CoV-2 infection in different countries and the relative impact this had on the respective health care systems, and the differences in screening for SARS-CoV-2 infection across the population. It is also important to note that the ∼500 IEI described exhibit enormous diversity6—so much so that it is challenging to draw conclusions when assessing these patient cohorts with limited granularity. It is also likely that some IEI will result in greater predisposition to severe COVID-19, while others may even be protective,119 thereby obscuring the overall severity of some IEI.
While it is difficult to make a direct comparison between CFR for IEI and the general population, this has been addressed for some countries. In Brazil,95 Italy,95 , 99, 100, 101 and the United Kingdom,103 the CFR in IEI was ∼2- to 4-fold greater than the general population. More strikingly, though, were findings from Iran, Italy, the United Kingdom, and an international study that the CFR for IEI patients aged 20-60 years or 60-75 years was 20-50 times or 2.5-5 times greater, respectively, than the general population.91 , 99 , 103 , 115 Furthermore, while the absolute number of patients analyzed is relatively small, the CFR for IEI patients aged 0-19 years is also much greater—possibly up to 100 times—than this age group in the general population.91 , 99 , 103 , 115 Consequently, the overall average age at death due to SARS-CoV-2 infection in IEI patients is much younger than the general population (Fig 2; ∼50 years vs ∼80 years).16 , 44 , 79 , 81 , 84 , 89 , 91 , 93, 94, 95 , 100 , 102, 103, 104 , 109 , 110 , 115 Thus, in addition to IEI patients’ generally presenting with COVID-19 at a younger age and a greater proportion requiring admission to ICU than the general population, the mortality rate of SARS-CoV-2 infection is greater in IEI, especially at ages where SARS-CoV-2 has a very low—even negligible—CFR in the general population (Fig 2).16 , 91 , 99 , 103 , 115
Innate immune defects predispose to severe and fatal SARS-CoV-2 infection
When comparing different IEI, there was often no correlation between the type of IEI and severity of disease/death after SARS-CoV-2 infection. For instance, the CFR for CVID, agammaglobulinemia, or chronic granulomatous disease were 7.2%, 6.2%, and 7.7%, respectively, compared to 8.5% for all IEI patients reported to date (Fig 1, B). However, there were several striking exceptions. First, although only few individuals have been identified, AR-pathogenic variants in IFNAR1 or IFNAR2, encoding individual receptor subunits for type I IFNs, resulted in lethal COVID-19 in 4 (57%) of 7 patients (Fig 1, B) and an average age at death of 11.8 years.42 , 56 , 87 , 120 Second, SARS-CoV-2 infection was severe in most patients with autoimmune polyendocrinopathy candidiasis ectodermal dystrophy (APECED) as a result of biallelic pathogenic AIRE variants. These individuals develop neutralizing autoantibodies against a range of cytokines, including type I IFN.121 In the setting of SARS-CoV-2 infection of APECED patients, rates of hospitalization (72%, 21/29), ICU admission (59%, 17/29), and death (13.8%, 4/29)57 , 84 , 115 , 118 , 122 were higher than all IEI patients as well as the general population (Fig 1, B).16 , 19 Third, patients with biallelic pathogenic variants in MYD88, IRAK, or IRF7—which function downstream of virus-sensing Toll-like receptors to induce production of type I IFNs by dendritic cells—experience severe COVID-19, with 5 of 8 MYD88/IRAK-deficient and all 5 IRF7-deficient SARS-CoV-2–infected individuals developing COVID-19 pneumonia, requiring hospitalization and/or admission to ICU; 1 of 5 IRF7-deficient patient died (Fig 1, B).45 , 81 , 95 , 99 , 102 , 120 , 123 Thus, genetic lesions or autoantibodies that compromise innate immunity by disrupting production or function of type I IFNs underpin severe, life-threatening, and often fatal SARS-CoV-2 infection (Fig 3 ).Fig 3 Critical roles of innate and adaptive immune cells in host defense against SARS-CoV-2 infection and disease pathogenesis.
These findings have been validated by a forward genetics approach. Whole-exome and -genome sequencing of adults and children who developed severe and/or life-threatening SARS-CoV-2 infection/COVID-19 identified pathogenic variants in genes involved in type I IFN signaling. These include genes required for the production of (TLR3, TLR7, UNC93B1, TICAM1, TBK1, IRF3, IRF7) or responses to (IFNAR1, IFNAR2, TYK2, STAT2, IRF7) type I IFN produced by plasmacytoid dendritic cells or respiratory epithelial cells after viral infection.65 , 90 , 120 , 123, 124, 125, 126 Overall, genetic variants in the type I IFN signaling pathway were the cause of severe COVID-19 in ∼3% of adults and ∼10% of children (Fig 3).120 , 126 , 127
Parallel to these genetic studies was the discovery that neutralizing autoantibodies specific for type I IFNs cause severe COVID-19 in 10% to 20% of otherwise healthy individuals infected with SARS-CoV-2.128, 129, 130, 131, 132, 133, 134, 135, 136, 137 Interestingly, these autoantibodies were: (1) predominantly directed against IFN-α and IFN-ω but not IFN-β; (2) found in increasing proportions of affected patients with each decade of life; (3) associated with disease severity, prolonged virus clearance, and admission to ICU; (4) inversely related to serum levels of type I IFNs and interferon-stimulated gene signatures in myeloid cells;78 , 128, 129, 130, 131, 132, 133, 134, 135, 136, 137 and (5) enriched in affected male subjects compared to female subjects across different age intervals. This, together with XL TLR7 deficiency, may contribute to the increased incidence of hospitalization and severe COVID-19 in male versus female subjects. These genetic and serologic studies unequivocally identified a fundamental nonredundant role for type I IFN–dependent immunity against SARS-CoV-2 infection, with 20% to 25% of cases of severe and life-threatening COVID-19 resulting from defective type I IFN production or function (Fig 3).
Additional anecdotal data have also linked impaired type I IFN–dependent immunity with susceptibility to SARS-CoV-2 infection. First, the CFR for autoinflammatory conditions such as Aicardi-Goutières syndrome or familial Mediterranean fever was lower than that for all reported cases of IEI (4.4% vs 8.5%; Fig 1, B).15 , 93 , 95 , 99 , 110 , 115 Thus, increased basal type I IFN signaling in these conditions may enable prompt host defense against SARS-CoV-2. Second, a recent study of patients with systemic lupus erythematosus, which is characterized by overproduction of type I IFNs, found that a subset of these patients also produced autoantibodies against type I IFNs. Remarkably, while these autoantibody-positive patients were less likely to develop active lupus disease, members of this same group were at increased risk of severe viral infections and sequelae including COVID-19 pneumonia.138
B cells and protective IgG in host defense against SARS-CoV-2
The study of COVID-19 in IEI provides an elegant opportunity to define redundant and nonredundant requirements for host defense against SARS-CoV-2. Initial studies found that patients with congenital B-cell deficiency and agammaglobulinemia had relatively mild disease and prompt recovery after SARS-CoV-2 infection.51 , 73 , 92 , 97 , 98 This led to a suggestion that B cells and neutralizing IgG may not be necessary for controlling SARS-CoV-2 infection and preventing severe COVID-19.98 Consistent with this, the CFR for XL/AR agammaglobulinemia patients is lower than all IEI patients (6.2%, 6/97, vs 8.5%, Fig 1, B). However, COVID-19 and SARS-CoV-2 viremia/virus shedding are prolonged in many B-cell–deficient/agammaglobulinemia patients, resulting in pneumonia requiring extended or multiple hospital stays, as well as numerous treatments to control viral infection.46 , 55 , 60 , 61 , 85 , 86 , 99 , 100 , 104 , 116 , 139 There have also been reports of chronic and/or repeated infections with worse outcomes than primary infection before vaccination, as well as breakthrough infections after vaccination in some XL agammaglobulinemia (XLA) patients.100 , 104 , 109 , 116 , 140 Similar observations in terms of relapsing COVID-19, as well as reinfection and/or sustained infection with SARS-CoV-2, have been made for patients with primary antibody deficiencies,82 , 104 , 116 further underscoring an important role for secreted immunoglobulin in controlling and clearing viral infection and attenuating disease. These findings from analysis of SARS-CoV-2 infection in individuals with congenital B-cell deficiency are also supported by studies of patients with rheumatic/musculoskeletal autoimmune diseases (rheumatoid arthritis, vasculitis, Sjögren syndrome, systemic lupus erythematosus) who are treated with B-cell–depleting therapies such as rituximab. In these cases, therapeutic B-cell depletion can result in high rates of hospital admissions, severe COVID-19 including protracted pneumonia and acute respiratory distress syndrome, and death after SARS-CoV-2 infection.141 , 142 Thus, the inability to generate specific IgG responses to novel antigens as a result of a lack of naive B cells can have dire consequences in the setting of SARS-CoV-2 infection (Fig 3).
This apparent paradox of prolonged illness and viremia but often-milder disease and lower CFR in XLA patients who completely lack B cells may be explained by the nature of the genetic defect. On the one hand, agammaglobulinemia in these patients highlights a key role for specific immunoglobulins in controlling and clearing viral infection, even when responses of innate immune cells and CD4+ and CD8+ T cells are intact.46 , 83 Indeed, administration of convalescent plasma isolated from previously infected healthy donors or anti–SARS-CoV-2–specific monoclonal antibodies (mAbs) led to rapid reductions in virus load and recovery in XLA—more so than observed with antiviral treatments alone (Fig 3).46 , 53 , 55 , 61 , 85 , 86 , 104 , 139 , 143 Although convalescent plasma or anti–SARS-CoV-2 mAbs are a logical treatment for XLA patients, similar results have also been reported for other IEI patients who have near-normal B cells and serum immunoglobulin levels but defects in generating functional and protective IgG-dependent humoral immunity. For instance, passive IgG therapy led to dramatic improvements in the clinical course of SARS-CoV-2 infection in patients with pathogenic variants in NFKB2,43 IL2RG,77 IKBKG (NEMO),72 and PIK3CD GOF,82 as well as many cases of CVID.83 , 103 , 104 , 109 , 143 In fact, anti–SARS-CoV-2 mAb or convalescent plasma greatly improved virus clearance and disease outcomes when combined with antivirals (eg, remdesivir).43 , 104 , 143 Thus, while type I IFN–mediated innate immunity is indispensable for containing acute SARS-CoV-2 infection, antibodies are necessary to mitigate prolonged viral infection, minimize disease, and prevent reinfections (Fig 3).
On the other hand, Bruton tyrosine kinase (BTK) deficiency—the genetic cause of XLA—compromises production of inflammatory cytokines by myeloid cells.144 Thus, relatively mild pulmonary disease in XLA may result from a lessened cytokine storm after SARS-CoV-2–induced activation of BTK-deficient myeloid cells. This is consistent with findings that some SARS-CoV-2–infected XLA patients have lower serum IL-6 levels than infected individuals in the general population,111 observations of mild COVID-19 in patients with B-cell malignancies who were treated with BTK inhibitors,145 and rapid clinical improvement in COVID-19 patients treated with a BTK inhibitor as a therapeutic intervention.146 These findings reveal dual roles for BTK in host defense and tissue pathology after SARS-CoV-2 infection. First, B cells and virus-specific antibodies are important for controlling prolonged infection. Second, BTK in myeloid cells may drive the SARS-CoV-2–induced cytokine storm characteristic of severe COVID-19. These findings provide a rationale for the use of passive immunoglobulin serotherapy (intravenous immunoglobulin, mAbs) to expedite virus clearance in IEI characterized by impaired humoral immunity, as well as of BTK inhibitors, Janus kinase (JAK) inhibitors, and tocilizumab (anti–IL-6R)146, 147, 148 to quell SARS-CoV-2–induced production of inflammatory cytokines by myeloid cells. However, it needs to be emphasized that timing of the delivery of these treatments can also influence outcome and efficacy. For instance, if administered too early, JAK inhibits may attenuate the protective effect of type I IFNs, while delayed treatment with tocilizumab may be ineffectual. Similarly, these interventions may be better suited for some specific types of IEI, particularly as results from clinical trials of these inhibitors in the general population have been variable.
Gene-directed therapies for COVID-19 in some IEI
Delineation of the genetic and serologic causes of severe COVID-19 has led to the implementation of specific therapies in some IEI. For instance, the discovery that inborn errors in type I IFN signaling are a risk factor for severe COVID-19 inspired the use of IFN-α2a or IFN-β, anti–SARS-CoV-2 mAbs, or convalescent plasma to treat SARS-CoV-2 infection in individuals with pathogenic variants in TLR3, IRF3, IRF7, or IRF9,68 , 69 , 123 which genetically disrupt type I IFN function, or patients with pathogenic AIRE variants or incontinentia pigmenti due to pathogenic IKBKG variants that result in production of neutralizing anti–type I IFN autoantibodies.78 , 84 , 149 However, convalescent plasma has also been found to contain neutralizing anti–type I IFN autoantibodies,137 which obviously could impact the efficacy of this treatment.
Similarly, plasma exchange was effective at reducing serum levels of neutralizing anti–type I IFN autoantibodies in an APECED patient.57 While it is difficult to draw specific conclusions regarding possible therapies for SARS-CoV-2 infection in IEI from these anecdotal investigations, most treated patients exhibited mild disease, experienced rapid resolution of symptoms, and made a full recovery.57 , 68 , 69 , 78 , 84 , 149 This contrasts with those IEI patients who did not receive specific treatments and experienced severe and even fatal COVID-19.120 , 126 , 127 Thus, early provision of type I IFN or antibody against SARS-CoV-2 may represent an immunotherapeutic approach to prevent critical pneumonia in patients who are most vulnerable to severe SARS-CoV-2 infection due to disrupted type I IFN–mediated immunity. Furthermore, because anti–type I IFN autoantibodies are mostly directed against IFN-α and IFN-ω, IFN-β can still be used therapeutically for severe COVID-19 in individuals who develop these neutralizing autoantibodies.
Vaccines against SARS-CoV-2
The global rollout of several different SARS-CoV-2 vaccines (mRNA, adenoviral based, inactivated virus, viral proteins) has dramatically attenuated COVID-19–associated mortality.150 These vaccines induce SARS-CoV-2–specific CD4+ and CD8+ T cells, memory B cells, and neutralizing serum IgG in >95% of healthy donors. Readouts of vaccine-induced immunity generally peaked 2 or 3 weeks after receipt of the second vaccine dose and then either significantly declined (specific IgG titers, CD8+ T cells), plateaued (CD4+ T cells), or even increased (memory B cells).150, 151, 152 Regardless of these trajectories, SARS-CoV-2–specific adaptive cellular and humoral immunity remained detectable ∼6 months after vaccination.150, 151, 152 The magnitude of these vaccine-induced correlates of immunity in healthy individuals was generally comparable to or greater than those observed in convalescent individuals recovering from natural SARS-CoV-2 infection.150, 151, 152
While these findings are encouraging, several challenges remain in controlling SARS-CoV-2. First, vaccine efficacy declines from 85-95% at 2 to 4 weeks after full vaccination to 20-50% 6 months later, thus revealing an inability to completely resist future infection and highlighting the need for vaccine boosters.150 , 153 , 154 Second, while successfully reducing disease severity, hospital admissions, and mortality, current vaccines do not effectively prevent SARS-CoV-2 transmission.150 , 155 Third, the emergence of variants of concern—which can arise in immunocompromised individuals156—compromise vaccine efficacy, with vaccine-induced immunity being significantly reduced against several SARS-CoV-2 variants.41 , 154 , 157 Thus, COVID-19 continues to represent a significant health risk despite the availability of several SARS-CoV-2 vaccines. Furthermore, findings from studies of IEI have established the importance of SARS-CoV-2–specific neutralizing IgG in preventing severe and prolonged disease as well as reinfection, so it is critical to continue encouraging vaccine and booster uptake in the general population.
Efficacy of SARS-CoV-2 vaccines in IEI patients
Many studies have initially assessed the immunogenicity and effectiveness of SARS-CoV-2 vaccines in IEI. The general findings from these studies were that (1) fewer patients mounted SARS-CoV-2–specific IgG (30-75%) and T-cell responses (∼50-70%) compared to healthy donors (∼95-100%), (2) titers of SARS-CoV-2–specific IgG, efficacy of virus neutralization, and magnitude of T-cell responses were reduced in patients compared to healthy donors, and (3) poor vaccine-induced immunity in patients correlated with reduced numbers of CD4+ T cells or memory B cells, low serum IgG and IgA, and older age.80 , 158, 159, 160, 161, 162, 163, 164, 165, 166, 167, 168, 169, 170, 171, 172, 173 Importantly, IEI that disrupt type I IFN–mediated immunity or autoantibodies against type I IFN do not impair humoral immune responses to RNA vaccines.174 Furthermore, despite normal levels of neutralizing IgG, some patients with anti–type I IFN autoantibodies develop breakthrough COVID-19 pneumonia.175
Overall, these studies established that SARS-CoV-2 vaccines are safe and well tolerated in people with IEI, and that they can induce specific adaptive immune responses, albeit at reduced levels compared to the general population. However, several significant unknowns remain. First, most studies assessed immune responses 2 to 8 weeks after the second vaccine dose. Thus, sustained durability of vaccine-induced immunity in IEI patients against SARS-CoV-2 has not been determined. Second, while almost all vaccine studies measured SARS-CoV-2–specific IgG, only a few determined virus neutralization. Thus, it is unknown whether vaccine-induced immunoglobulin in IEI patients can neutralize the original SARS-CoV-2 strain and emerging variants. Third, specific CD4+ and CD8+ T-cell responses in vaccinated IEI patients were not assessed in most studies. The paucity of data relating to responses of T-cell subsets impacts our ability to predict vulnerability of individuals with intrinsic T-cell defects to SARS-CoV-2 infection. Fourth, how waning immunity and SARS-CoV-2 variants impact host defense, as well as the capacity of vaccine booster doses to amplify immunity, in IEI patients is unexplored. Fifth, ∼80% of all IEI patients assessed in these studies did not have a molecular diagnosis; most had CVID. Thus, it is difficult to (1) delineate cellular and molecular mechanisms underlying impaired immunity in IEI patients, (2) extrapolate these findings from predominantly CVID and antibody-deficient patients to IEI in general, (3) identify which pathways are necessary to elicit robust and long-lived immune responses, and (4) leverage these findings to develop methods to target specific key molecules/pathways to improve host defense against infectious diseases induced by next-generation vaccines. These are issues that need to be addressed in ongoing and future studies.
Conclusion
Analysis of individuals with single-gene defects that result in immune dysregulation have defined the fundamental requirements for immune homeostasis and host defense against a broad range of infectious agents. It was upon this foundation that the fields of genetics/genomics, basic and clinical immunology, and infectious diseases combined to make profound advances in unraveling the complexity of SARS-CoV-2 infection and severe COVID-19. Indeed, some of the key discoveries over the past 2 or 3 years have arisen from studying severe COVID-19 in otherwise healthy individuals, as well as in individuals with IEI. These studies established the framework to further define host factors necessary for early innate and sustained adaptive immune-mediated protection against SARS-CoV-2 infection and the establishment of immunologic memory, as well as mechanisms of severe disease and identifying opportunities for therapeutic intervention to manage COVID-19. Despite these breakthrough findings, there remains significant uncertainty regarding SARS-CoV-2 and IEI patients. These include the impact of standard treatments for IEI on immunity against SARS-CoV-2 infection and vaccination (eg, JAK inhibitors, TNF inhibitors, abatacept, rapamycin), long-term effects of SARS-CoV-2 infection/reinfection on IEI patients with autoimmunity and/or malignancy, whether long COVID and neurologic impacts are more prevalent in IEI compared to the general population, and the protective effect of neutralizing antibodies that are accumulating in donor blood products used for immunoglobulin replacement therapy. However, with the rapid pace of the advances already made since we first became aware of SARS-CoV-2, there is no doubt that answers to these questions—and more—will be delivered as we move into the third year (and, I hope, the last frontier) of this pandemic.
I would like to acknowledge all the clinicians, nurses, caregivers, parents, and patient advocacy groups who have ensured the safety and well-being of patients with IEI during the COVID-19 pandemic. I also want to thank Jean-Laurent Casanova and Helen Su for their leadership of the COVID-19 Human Genetic Effort consortium (www.covidhge.com), as well as their insightful discussions and inspiration since the onset of the pandemic; and Isabelle Meyts and many members of the COVID-19 Human Genetic Effort for their constructive feedback.
Members of the COVID Human Genetic Effort consortium are as follows: Laurent Abel (INSERM U1163, University of Paris, Imagine Institute, Paris, France); Salah Al-Muhsen (Immunology Research Lab, Department of Pediatrics, College of Medicine, King Saud University, Riyadh, Saudi Arabia); Alessandro Aiuti (San Raffaele Telethon Institute for Gene Therapy, IRCCS Ospedale San Raffaele, and Vita Salute San Raffaele University, Milan, Italy); Saleh Al-Muhsen (Immunology Research Laboratory, Department of Pediatrics, College of Medicine and King Saud University Medical City, King Saud University, Riyadh, Saudi Arabia); Fahd Al-Mulla (Dasman Diabetes Institute, Department of Genetics and Bioinformatics, Dasman, Kuwait); Mark S. Anderson (Diabetes Center, University of California, San Francisco, Calif); Evangelos Andreakos (Biomedical Research Foundation of the Academy of Athens, Athens, Greece); Antonio Novelli (Laboratory of Medical Genetics, IRCCS Bambino Gesù Children’s Hospital, Rome, Italy); Andrés A. Arias (Group of Primary Immunodeficiencies, University of Antioquia UdeA, Medellin, Colombia); Hagit Baris Feldman (The Genetics Institute, Tel Aviv Sourasky Medical Center and Sackler Faculty of Medicine, Tel Aviv University, Tel Aviv, Israel); Alexandre Belot (Pediatric Nephrology, Rheumatology, Dermatology, HFME, Hospices Civils de Lyon, National Referee Centre RAISE, and INSERM U1111, Université de Lyon, Lyon, France); Catherine M. Biggs (Department of Pediatrics, British Columbia Children’s Hospital, University of British Columbia, Vancouver, British Columbia, Canada); Ahmed A. Bousfiha (Clinical Immunology Unit, Department of Pediatric Infectious Disease, CHU Ibn Rushd, and LICIA, Laboratoire d’Immunologie Clinique, Inflammation et Allergie, Faculty of Medicine and Pharmacy, Hassan II University, Casablanca, Morocco); Petter Brodin (SciLifeLab, Department of Women’s and Children’s Health, Karolinska Institutet, Stockholm, Sweden); John Christodoulou (Murdoch Children’s Research Institute and Department of Paediatrics, University of Melbourne, Melbourne, Australia); Antonio Condino-Neto (Department of Immunology, Institute of Biomedical Sciences, University of São Paulo, São Paulo, Brazil); Clifton L. Dalgard (Department of Anatomy, Physiology, and Genetics, Uniformed Services University of the Health Sciences, Bethesda, Md); Sara Espinosa-Padilla (National Institute of Pediatrics, Mexico City, Mexico); Jacques Fellay (School of Life Sciences, Ecole Polytechnique Fédérale de Lausanne; Precision Medicine Unit, Lausanne University Hospital; and University of Lausanne, Lausanne, Switzerland); Carlos Flores (Genomics Division, Instituto Tecnológico y de Energías Renovables [ITER], Santa Cruz de Tenerife; Research Unit, Hospital Universitario NS de Candelaria, Santa Cruz de Tenerife; and CIBER de Enfermedades Respiratorias, Instituto de Salud Carlos III, Madrid, Spain); José Luis Franco (Group of Primary Immunodeficiencies, University of Antioquia UdeA, Medellin, Colombia); Antoine Froidure (Pulmonology Department, Cliniques Universitaires Saint-Luc, and Institut de Recherche Expérimentale et Clinique [IREC], Université Catholique de Louvain, Brussels, Belgium); Filomeen Haerynck (Department of Paediatric Immunology and Pulmonology, Centre for Primary Immunodeficiency Ghent [CPIG], PID Research Laboratory, Jeffrey Modell Diagnosis and Research Centre, Ghent University Hospital, Ghent, Belgium); Rabih Halwani (Sharjah Institute of Medical Research, College of Medicine, University of Sharjah, Sharjah, United Arab Emirates [UAE]); Lennart Hammarström (Department of Biosciences and Nutrition, Karolinska Institutet, Stockholm, Sweden); Sarah E. Henrickson (Department of Pediatrics, Division of Allergy Immunology, Children’s Hospital of Philadelphia; and Department of Microbiology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, Pa); Elena W. Y. Hsieh (Departments of Pediatrics, Immunology, and Microbiology, University of Colorado, School of Medicine, Aurora, Colorado); Yuval Itan (Institute for Personalized Medicine, Icahn School of Medicine at Mount Sinai, and Department of Genetics and Genomic Sciences, Icahn School of Medicine at Mount Sinai, New York, NY); Timokratis Karamitros (Bioinformatics and Applied Genomics Unit, Hellenic Pasteur Institute, Athens, Greece); Yu-Lung Lau (Department of Paediatrics and Adolescent Medicine, University of Hong Kong, Hong Kong); Davood Mansouri (Department of Clinical Immunology and Infectious Diseases, National Research Institute of Tuberculosis and Lung Diseases, Clinical Tuberculosis and Epidemiology Research Center, National Research Institute of Tuberculosis and Lung Diseases [NRITLD], Masih Daneshvari Hospital, Shahid Beheshti, University of Medical Sciences, Tehran, Iran); Isabelle Meyts (Department of Pediatrics, University Hospitals Leuven, Department of Microbiology, Immunology, and Transplantation, and Laboratory for Inborn Errors of Immunity, KU Leuven, Leuven, Belgium); Trine H. Mogensen (Department of Biomedicine, Aarhus University, Aarhus, Denmark); Tomohiro Morio (Tokyo Medical and Dental University Hospital, Tokyo, Japan); Lisa F. P. Ng (A∗STAR Infectious Disease Labs, Agency for Science, Technology and Research; and Lee Kong Chian School of Medicine, Nanyang Technology University, Singapore); Luigi D. Notarangelo (National Institute of Allergy and Infectious Diseases [NIAID], National Institutes of Health [NIH], Bethesda, Md); Giuseppe Novelli (Department of Biomedicine and Prevention, Tor Vergata University of Rome, Rome, Italy); Satoshi Okada (Department of Pediatrics, Graduate School of Biomedical and Health Sciences, Hiroshima University, Hiroshima, Japan); Tayfun Ozcelik (Department of Molecular Biology and Genetics, Bilkent University, Bilkent, Ankara, Turkey); Qiang Pan-Hammarström (Department of Biosciences and Nutrition, Karolinska Institutet, Stockholm, Sweden); Rebeca Perez de Diego (Laboratory of Immunogenetics of Human Diseases, Innate Immunity Group, IdiPAZ Institute for Health Research, La Paz Hospital, Madrid, Spain); Carolina Prando (Faculdades Pequeno Príncipe, Instituto de Pesquisa Pelé Pequeno Príncipe, Curitiba, Brazil); Aurora Pujol (Neurometabolic Diseases Laboratory, Bellvitge Biomedical Research Institute [IDIBELL]), L’Hospitalet de Llobregat; Catalan Institution of Research and Advanced Studies [ICREA]; and Center for Biomedical Research on Rare Diseases [CIBERER], ISCIII, Barcelona, Spain); Laurent Renia (A∗STAR Infectious Disease Labs, Agency for Science, Technology and Research; and Lee Kong Chian School of Medicine, Nanyang Technology University, Singapore); Igor Resnick (Department of Medical Genetics, Medical University; and Department of Hematology and BMT, University Hospital St Marina, Varna, Bulgaria); Carlos Rodríguez-Gallego (Department of Immunology, University Hospital of Gran Canaria Dr Negrín, Canarian Health System; and Department of Clinical Sciences, University Fernando Pessoa Canarias, Las Palmas de Gran Canaria, Spain); Vanessa Sancho-Shimizu (Department of Paediatric Infectious Diseases and Virology, Imperial College London; and Centre for Paediatrics and Child Health, Faculty of Medicine, Imperial College London, London, United Kingdom); Mikko R. J. Seppänen (Adult Immunodeficiency Unit, Infectious Diseases, Inflammation Center, University of Helsinki and Helsinki University Hospital; Rare Diseases Center and Pediatric Research Center, Children’s Hospital, University of Helsinki; and Helsinki University Hospital, Helsinki, Finland); Anna Shcherbina (Department of Immunology, Dmitry Rogachev National Medical Research Center of Pediatric Hematology, Oncology, and Immunology, Moscow, Russia); Andrew L. Snow (Department of Pharmacology and Molecular Therapeutics, Uniformed Services University of the Health Sciences, Bethesda, Md); Pere Soler-Palacín (Pediatric Infectious Diseases and Immunodeficiencies Unit, Vall d’Hebron Barcelona Hospital Campus, Barcelona, Spain); András N. Spaan (St Giles Laboratory of Human Genetics of Infectious Diseases, Rockefeller Branch, Rockefeller University, New York, NY; and Department of Medical Microbiology, University Medical Center Utrecht, Utrecht, The Netherlands); Ivan Tancevski (Department of Internal Medicine II, Medical University of Innsbruck, Innsbruck, Austria); Stuart G. Tangye (Garvan Institute of Medical Research, Darlinghurst; and St Vincent’s Clinical School, Faculty of Medicine, University of New South Wales Sydney, Sydney, Australia); Ahmad Abou Tayoun (Al Jalila Children’s Hospital, Dubai, UAE); Sehime G. Temel (Bursa Uludag University, Medical Faculty, Department of Medical Genetics, Bursa, Turkey); Stuart E. Turvey (BC Children’s Hospital, The University of British Columbia, Vancouver, Canada); Mohammed J. Uddin (College of Medicine, Mohammed Bin Rashid University of Medicine and Health Sciences, Dubai, UAE; and Cellular Intelligence [CI] Lab, GenomeArc Inc, Toronto, Ontario, Canada); Donald C. Vinh (Department of Medicine, Division of Infectious Diseases, McGill University Health Centre; and Infectious Disease Susceptibility Program, Research Institute, McGill University Health Centre, Montreal, Quebec, Canada); Mayana Zatz (Biosciences Institute, University of São Paulo, São Paulo, Brazil); Keisuke Okamoto (Tokyo Medical and Dental University, Tokyo, Japan); David S. Pelin (Center for Discovery and Innovation, Hackensack Meridian Health, Nutley, NJ); Graziano Pesole (Department of Biosciences, Biotechnology, and Biopharmaceutics, University of Bari A. Moro, Bari, Italy); Diederik van de Beek (Department of Neurology, Amsterdam Neuroscience, Amsterdam University Medical Center, University of Amsterdam, Amsterdam, The Netherlands); Roger Colobran (Hospital Universitari Vall d’Hebron, Barcelona, Spain); Joost Wauters (Department of General Internal Medicine, Medical Intensive Care Unit, University Hospitals Leuven, Leuven, Belgium); Helen C. Su (NIAID, NIH, Bethesda, Md); Jean-Laurent Casanova (Rockefeller University and Howard Hughes Medical Institute, New York, NY; and Necker Hospital for Sick Children and INSERM, Paris, France).
S.G.T. is supported by an Investigator Grant awarded by the 10.13039/501100000925 National Health and Medical Research Council of Australia, the Allergy & Immunology Foundation of Australia, the Jeffrey Modell Foundation, and a University of New South Wales Sydney COVID Rapid Response Initiative grant. H.C.S. and L.D.N. (listed under the COVID Human Genetic Effort consortium) are supported by the Division of Intramural Research of the National Institute of Allergy and Infectious Diseases, NIH. J.L.C. (listed under the COVID Human Genetic Effort consortium) is supported by the NIH (R01AI088364, R01AI163029, UL1TR001866), Fisher Center for Alzheimer’s Research Foundation, Meyer Foundation, JPB Foundation, French National Research Agency (ANR-10-IAHU-01, ANR-10-LABX-62-IBEID, ANR-20-CE93-003, ANR-20-CO11-0001, French Foundation for Medical Research (EQU201903007798), the ANRS-COV05, European Union’s Horizon 2020 Research and Innovation Program (824110; EASI-genomics), HORIZON-HLTH-2021-DISEASE-04 Program (01057100; UNDINE), the ANR-RHU COVIFERON Program (ANR-21-RHUS-08), the Square Foundation, Grandir - Fonds de solidarité pour l’enfance, the Fondation du Souffle, the SCOR Corporate Foundation for Science, and French Ministry of Higher Education, Research, and Innovation (MESRI-COVID-19).
Disclosure of potential conflict of interest: The author declares no relevant conflicts of interest.
==== Refs
References
1 Casanova J.L. Abel L. Human genetics of infectious diseases: unique insights into immunological redundancy Semin Immunol 36 2018 1 12 29254755
2 Casanova J.L. Abel L. From rare disorders of immunity to common determinants of infection: following the mechanistic thread Cell 185 2022 3086 3103 35985287
3 Fischer A. Rausell A. What do primary immunodeficiencies tell us about the essentiality/redundancy of immune responses? Semin Immunol 36 2018 13 16 29229549
4 Notarangelo L.D. Bacchetta R. Casanova J.L. Su H.C. Human inborn errors of immunity: an expanding universe Sci Immunol 2020
5 Zhang Q. Frange P. Blanche S. Casanova J.L. Pathogenesis of infections in HIV-infected individuals: insights from primary immunodeficiencies Curr Opin Immunol 48 2017 122 133 28992464
6 Tangye S.G. Al-Herz W. Bousfiha A. Cunningham-Rundles C. Franco J.L. Holland S.M. Human inborn errors of immunity: 2022 update on the classification from the International Union of Immunological Societies Expert Committee J Clin Immunol 2022
7 Rieux-Laucat F. Casanova J.L. Immunology. Autoimmunity by haploinsufficiency Science 345 2014 1560 1561 25258064
8 Ramirez N.J. Posadas-Cantera S. Caballero-Oteyza A. Camacho-Ordonez N. Grimbacher B. There is no gene for CVID—novel monogenetic causes for primary antibody deficiency Curr Opin Immunol 72 2021 176 185 34153571
9 Poyhonen L. Bustamante J. Casanova J.L. Jouanguy E. Zhang Q. Life-threatening infections due to live-attenuated vaccines: early manifestations of inborn errors of immunity J Clin Immunol 39 2019 376 390 31123910
10 Thalhammer J. Kindle G. Nieters A. Rusch S. Seppanen M.R.J. Fischer A. Initial presenting manifestations in 16,486 patients with inborn errors of immunity include infections and noninfectious manifestations J Allergy Clin Immunol 148 2021 1332 1341.e5 33895260
11 Fischer A. Provot J. Jais J.P. Alcais A. Mahlaoui N. Autoimmune and inflammatory manifestations occur frequently in patients with primary immunodeficiencies J Allergy Clin Immunol 140 2017 1388 1393.e8 28192146
12 Leiding J.W. Ballow M. Redefining precision medicine in disorders of immune dysregulation J Allergy Clin Immunol Pract 7 2019 2801 2803 31706494
13 Weiss S.R. Forty years with coronaviruses J Exp Med 2020
14 Zhu N. Zhang D. Wang W. Li X. Yang B. Song J. A novel coronavirus from patients with pneumonia in China, 2019 N Engl J Med 382 2020 727 733 31978945
15 Shields A.M. Burns S.O. Savic S. Richter A.G. UK PIN COVID-19 Consortium COVID-19 in patients with primary and secondary immunodeficiency: the United Kingdom experience J Allergy Clin Immunol 147 2021 870 875.e1 33338534
16 Stokes E.K. Zambrano L.D. Anderson K.N. Marder E.P. Raz K.M. El Burai Felix S. Coronavirus disease 2019 case surveillance—United States, January 22-May 30, 2020 MMWR Morb Mortal Wkly Rep 69 2020 759 765 32555134
17 Grasselli G. Zangrillo A. Zanella A. Antonelli M. Cabrini L. Castelli A. Baseline characteristics and outcomes of 1591 patients infected with SARS-CoV-2 admitted to ICUs of the Lombardy region, Italy JAMA 323 2020 1574 1581 32250385
18 Li X. Xu S. Yu M. Wang K. Tao Y. Zhou Y. Risk factors for severity and mortality in adult COVID-19 inpatients in Wuhan J Allergy Clin Immunol 146 2020 110 118 32294485
19 Russell T.W. Hellewell J. Jarvis C.I. van Zandvoort K. Abbott S. Ratnayake R. Estimating the infection and case fatality ratio for coronavirus disease (COVID-19) using age-adjusted data from the outbreak on the Diamond Princess cruise ship February 2020 Euro Surveill 25 2020 2000256
20 Abers M.S. Delmonte O.M. Ricotta E.E. Fintzi J. Fink D.L. de Jesus A.A.A. An immune-based biomarker signature is associated with mortality in COVID-19 patients JCI Insight 6 2021 e144455
21 Merad M. Martin J.C. Pathological inflammation in patients with COVID-19: a key role for monocytes and macrophages Nat Rev Immunol 20 2020 355 362 32376901
22 Herold T. Jurinovic V. Arnreich C. Lipworth B.J. Hellmuth J.C. von Bergwelt-Baildon M. Elevated levels of IL-6 and CRP predict the need for mechanical ventilation in COVID-19 J Allergy Clin Immunol 146 2020 128 136.e4 32425269
23 Del Valle D.M. Kim-Schulze S. Huang H.H. Beckmann N.D. Nirenberg S. Wang B. An inflammatory cytokine signature predicts COVID-19 severity and survival Nat Med 26 2020 1636 1643 32839624
24 Hadjadj J. Yatim N. Barnabei L. Corneau A. Boussier J. Smith N. Impaired type I interferon activity and inflammatory responses in severe COVID-19 patients Science 369 2020 718 724 32661059
25 Quartuccio L. Sonaglia A. Pecori D. Peghin M. Fabris M. Tascini C. Higher levels of IL-6 early after tocilizumab distinguish survivors from nonsurvivors in COVID-19 pneumonia: a possible indication for deeper targeting of IL-6 J Med Virol 92 2020 2852 2856 32515499
26 Ruan Q. Yang K. Wang W. Jiang L. Song J. Clinical predictors of mortality due to COVID-19 based on an analysis of data of 150 patients from Wuhan, China Intensive Care Med 46 2020 846 848 32125452
27 Holter J.C. Pischke S.E. de Boer E. Lind A. Jenum S. Holten A.R. Systemic complement activation is associated with respiratory failure in COVID-19 hospitalized patients Proc Natl Acad Sci U S A 117 2020 25018 25025 32943538
28 Ramlall V. Thangaraj P.M. Meydan C. Foox J. Butler D. Kim J. Immune complement and coagulation dysfunction in adverse outcomes of SARS-CoV-2 infection Nat Med 26 2020 1609 1615 32747830
29 Chen G. Wu D. Guo W. Cao Y. Huang D. Wang H. Clinical and immunological features of severe and moderate coronavirus disease 2019 J Clin Invest 130 2020 2620 2629 32217835
30 Hou H. Zhang Y. Tang G. Luo Y. Liu W. Cheng C. Immunologic memory to SARS-CoV-2 in convalescent COVID-19 patients at 1 year postinfection J Allergy Clin Immunol 148 2021 1481 1492.e2 34536418
31 Breton G. Mendoza P. Hagglof T. Oliveira T.Y. Schaefer-Babajew D. Gaebler C. Persistent cellular immunity to SARS-CoV-2 infection J Exp Med 2021
32 Dan J.M. Mateus J. Kato Y. Hastie K.M. Yu E.D. Faliti C.E. Immunological memory to SARS-CoV-2 assessed for up to 8 months after infection Science 2021
33 Bilich T. Nelde A. Heitmann J.S. Maringer Y. Roerden M. Bauer J. T cell and antibody kinetics delineate SARS-CoV-2 peptides mediating long-term immune responses in COVID-19 convalescent individuals Sci Transl Med 2021
34 Moriyama S. Adachi Y. Sato T. Tonouchi K. Sun L. Fukushi S. Temporal maturation of neutralizing antibodies in COVID-19 convalescent individuals improves potency and breadth to circulating SARS-CoV-2 variants Immunity 54 2021 1841 1852.e4 34246326
35 Sherina N. Piralla A. Du L. Wan H. Kumagai-Braesch M. Andrell J. Persistence of SARS-CoV-2–specific B and T cell responses in convalescent COVID-19 patients 6-8 months after the infection Med (N Y) 2 2021 281 295.e4 33589885
36 Hartley G.E. Edwards E.S.J. Aui P.M. Varese N. Stojanovic S. McMahon J. Rapid generation of durable B cell memory to SARS-CoV-2 spike and nucleocapsid proteins in COVID-19 and convalescence Sci Immunol 5 2020 eabf8891
37 Wajnberg A. Amanat F. Firpo A. Altman D.R. Bailey M.J. Mansour M. Robust neutralizing antibodies to SARS-CoV-2 infection persist for months Science 370 2020 1227 1230 33115920
38 Balachandran H. Phetsouphanh C. Agapiou D. Adhikari A. Rodrigo C. Hammoud M. Maintenance of broad neutralizing antibodies and memory B cells 1 year post-infection is predicted by SARS-CoV-2–specific CD4+ T cell responses Cell Rep 38 2022 110345 35090598
39 Abayasingam A. Balachandran H. Agapiou D. Hammoud M. Rodrigo C. Keoshkerian E. Long-term persistence of RBD+ memory B cells encoding neutralizing antibodies in SARS-CoV-2 infection Cell Rep Med 2 2021 100228 33748788
40 Chia W.N. Zhu F. Ong S.W.X. Young B.E. Fong S.W. Le Bert N. Dynamics of SARS-CoV-2 neutralising antibody responses and duration of immunity: a longitudinal study Lancet Microbe 2 2021 e240 e249 33778792
41 Hirabara S.M. Serdan T.D.A. Gorjao R. Masi L.N. Pithon-Curi T.C. Covas D.T. SARS-COV-2 variants: differences and potential of immune evasion Front Cell Infect Microbiol 11 2021 781429 35118007
42 Abolhassani H. Landegren N. Bastard P. Materna M. Modaresi M. Du L. Inherited IFNAR1 deficiency in a child with both critical COVID-19 pneumonia and multisystem inflammatory syndrome J Clin Immunol 42 2022 471 483 35091979
43 Abraham R.S. Marshall J.M. Kuehn H.S. Rueda C.M. Gibbs A. Guider W. Severe SARS-CoV-2 disease in the context of a NF-kappaB2 loss-of-function pathogenic variant J Allergy Clin Immunol 147 2021 532 544.e1 33007327
44 Beccuti G. Ghizzoni L. Cambria V. Codullo V. Sacchi P. Lovati E. A COVID-19 pneumonia case report of autoimmune polyendocrine syndrome type 1 in Lombardy, Italy: letter to the editor J Endocrinol Invest 43 2020 1175 1177 32519200
45 Bucciol G. Moens L. Corveleyn A. Dreesman A. Meyts I. A novel kindred with MyD88 deficiency J Clin Immunol 42 2022 885 888 35286514
46 Buckland M.S. Galloway J.B. Fhogartaigh C.N. Meredith L. Provine N.M. Bloor S. Treatment of COVID-19 with remdesivir in the absence of humoral immunity: a case report Nat Commun 11 2020 6385 33318491
47 Castano-Jaramillo L.M. Yamazaki-Nakashimada M.A. Scheffler Mendoza S.C. Bustamante-Ogando J.C. Espinosa-Padilla S.E. Lugo Reyes S.O. A male infant with COVID-19 in the context of ARPC1B deficiency Pediatr Allergy Immunol 32 2021 199 201 32683750
48 Cenciarelli S. Calbi V. Barzaghi F. Bernardo M.E. Oltolini C. Migliavacca M. Mild SARS-CoV-2 infection after gene therapy in a child with Wiskott-Aldrich syndrome: a case report Front Immunol 11 2020 603428 33329599
49 Gabryszewski S.J. England R.N. Sun D. Gentile T.L. Hochgertel W. Jyonouchi S. Self-limited COVID-19 in a patient with artemis hypomorphic SCID J Clin Immunol 2021
50 Guisado Hernandez P. Blanco Lobo P. Villaoslada I. de Felipe B. Lucena J.M. Martin Gutierrez G. SARS-CoV-2 infection in a pediatrics STAT1 GOF patient under ruxolitinib therapy—a matter of balance? J Clin Immunol 2021
51 Gupta S. Agrawal S. Sandoval A. Su H. Tran M. Demirdag Y. SARS-CoV-2–specific and functional cytotoxic CD8 cells in primary antibody deficiency: natural infection and response to vaccine J Clin Immunol 2022
52 Gupta S. Su H. Narsai T. Agrawal S. SARS-CoV-2–associated T-cell responses in the presence of humoral immunodeficiency Int Arch Allergy Immunol 182 2021 195 209 33486489
53 Iaboni A. Wong N. Betschel S.D. A patient with X-linked agammaglobulinemia and COVID-19 infection treated with remdesivir and convalescent plasma J Clin Immunol 41 2021 923 925 33547548
54 Jin H. Moss R. Reed J.C. Hertzberg E. Cruz M.R. Akkoyun E. IFN-gamma receptor 2 deficiency initial mimicry of multisystem inflammatory syndrome in children (MIS-C) J Allergy Clin Immunol Pract 9 2021 989 992.e1 32916324
55 Jin H. Reed J.C. Liu S.T.H. Ho H.E. Lopes J.P. Ramsey N.B. Three patients with X-linked agammaglobulinemia hospitalized for COVID-19 improved with convalescent plasma J Allergy Clin Immunol Pract 8 2020 3594 3596.e3 32947026
56 Khanmohammadi S. Rezaei N. Khazaei M. Shirkani A. A case of autosomal recessive interferon alpha/beta receptor alpha chain (IFNAR1) deficiency with severe COVID-19 J Clin Immunol 42 2022 19 24 34713375
57 Lemarquis A. Campbell T. Aranda-Guillen M. Hennings V. Brodin P. Kampe O. Severe COVID-19 in an APS1 patient with interferon autoantibodies treated with plasmapheresis J Allergy Clin Immunol 148 2021 96 98 33892926
58 London J. Boutboul D. Lacombe K. Pirenne F. Heym B. Zeller V. Severe COVID-19 in patients with B cell alymphocytosis and response to convalescent plasma therapy J Clin Immunol 41 2021 356 361 33219491
59 Mantravadi V. Nguyen S.T. Morley S.C. Bednarski J.J. Kitcharoensakkul M. Cooper M.A. Recovery from COVID-19 in a child with chronic granulomatous disease and T cell lymphopenia J Clin Immunol 41 2021 23 25 33108551
60 Milosevic I. Jovanovic J. Stevanovic O. Atypical course of COVID-19 in patient with Bruton agammaglobulinemia J Infect Dev Ctries 14 2020 1248 1251 33296336
61 Mira E. Yarce O.A. Ortega C. Fernandez S. Pascual N.M. Gomez C. Rapid recovery of a SARS-CoV-2–infected X-linked agammaglobulinemia patient after infusion of COVID-19 convalescent plasma J Allergy Clin Immunol Pract 8 2020 2793 2795 32652231
62 Mullur J. Wang A. Feldweg A. A fatal case of coronavirus disease 2019 in a patient with common variable immunodeficiency Ann Allergy Asthma Immunol 126 2021 90 92 32818593
63 Narahari P.G. Gebbia J. Alperstein W. Kleiner G. Gans M. Post–SARS-CoV-2 atypical inflammatory syndrome in a toddler with X-linked inhibitor of apoptosis deficiency after stem cell transplant J Clin Immunol 2022
64 Sanchez Clemente N. Penner J. Breuer J. Ip W. Booth C. Case report: a severe paediatric presentation of COVID-19 in APDS2 immunodeficiency Front Immunol 13 2022 881259 35707532
65 Schmidt A. Peters S. Knaus A. Sabir H. Hamsen F. Maj C. TBK1 and TNFRSF13B mutations and an autoinflammatory disease in a child with lethal COVID-19 NPJ Genom Med 6 2021 55 34210994
66 Soresina A. Moratto D. Chiarini M. Paolillo C. Baresi G. Foca E. Two X-linked agammaglobulinemia patients develop pneumonia as COVID-19 manifestation but recover Pediatr Allergy Immunol 31 2020 565 569 32319118
67 Vignesh P. Mondal S. Sudhakar M. Sharma Y.K. Bansal A. Singh M. SARS-CoV-2 infection in a child with severe congenital neutropenia J Clin Immunol 41 2021 1165 1168 33942228
68 Levy R. Bastard P. Lanternier F. Lecuit M. Zhang S.Y. Casanova J.L. IFN-alpha2a therapy in two patients with inborn errors of TLR3 and IRF3 infected with SARS-CoV-2 J Clin Immunol 41 2021 26 27 33392853
69 Levy R. Zhang P. Bastard P. Dorgham K. Melki I. Hadchouel A. Monoclonal antibody–mediated neutralization of SARS-CoV-2 in an IRF9-deficient child Proc Natl Acad Sci U S A 2021
70 Al-Saud B. Hazzazi K.M. Mohammed R. Al Najjar A. Al Hazmi T. Monies D. SARS-CoV-2–related acute respiratory distress syndrome uncovers a patient with severe combined immunodeficiency disease J Clin Immunol 2021
71 Aljaberi R. Wishah K. Positive outcome in a patient with coronavirus disease 2019 and common variable immunodeficiency after intravenous immunoglobulin Ann Allergy Asthma Immunol 125 2020 349 350 32505720
72 Alkan G. Artac H. Oz S.K.T. Emiroglu M. Management of COVID-19 pneumonia in a child with NEMO deficiency Immunol Res 69 2021 391 393 34061330
73 Almontasheri A. Al-Husayni F. Alsuraihi A.K. Binyahib H. Albanna A.S. The clinical course of COVID-19 pneumonia in a 19-year-old man on intravenous immunoglobulin replacement therapy for X-linked agammaglobulinemia Am J Case Rep 22 2021 e929447
74 Bozonnat A. Assan F. LeGoff J. Bourrat E. Bachelez H. SARS-CoV-2 infection inducing severe flare up of deficiency of interleukin thirty-six (IL-36) receptor antagonist (DITRA) resulting from a mutation invalidating the activating cleavage site of the IL-36 receptor antagonist J Clin Immunol 2021
75 Fill L. Hadney L. Graven K. Persaud R. Hostoffer R. The clinical observation of a patient with common variable immunodeficiency diagnosed as having coronavirus disease 2019 Ann Allergy Asthma Immunol 125 2020 112 114 32387167
76 Lee P.Y. Platt C.D. Weeks S. Grace R.F. Maher G. Gauthier K. Immune dysregulation and multisystem inflammatory syndrome in children (MIS-C) in individuals with haploinsufficiency of SOCS1 J Allergy Clin Immunol 146 2020 1194 1200.e1 32853638
77 van Oers N.S.C. Hanners N.W. Sue P.K. Aquino V. Li Q.Z. Schoggins J.W. SARS-CoV-2 infection associated with hepatitis in an infant with X-linked severe combined immunodeficiency Clin Immunol 224 2021 108662 33412294
78 Bastard P. Levy R. Henriquez S. Bodemer C. Szwebel T.A. Casanova J.L. Interferon-beta therapy in a patient with incontinentia pigmenti and autoantibodies against type I IFNs infected with SARS-CoV-2 J Clin Immunol 41 2021 931 933 33763778
79 Penafiel Vicuna A.K. Yamazaki Nakashimada M. Leon Lara X. Mendieta Flores E. Nunez Nunez M.E. Lona-Reyes J.C. Mendelian susceptibility to mycobacterial disease: retrospective clinical and genetic study in Mexico J Clin Immunol 2022
80 Bloomfield M. Parackova Z. Hanzlikova J. Lastovicka J. Sediva A. Immunogenicity and safety of COVID-19 mRNA vaccine in STAT1 GOF patients J Clin Immunol 42 2022 266 269 34718945
81 Mahmood H.Z. Madhavarapu S. Almuqamam M. Varying Illness severity in patients with MyD88 deficiency infected with coronavirus SARS-CoV-2 Pediatrics 147 2021 453 454
82 Rivalta B. Amodio D. Giancotta C. Santilli V. Pacillo L. Zangari P. Case report: successful treatment with monoclonal antibodies in one apds patient with prolonged SARS-CoV-2 infection not responsive to previous lines of treatment Front Immunol 13 2022 891274 35799775
83 Steiner S. Schwarz T. Corman V.M. Gebert L. Kleinschmidt M.C. Wald A. SARS-CoV-2 T cell response in severe and fatal COVID-19 in primary antibody deficiency patients without specific humoral immunity Front Immunol 13 2022 840126 35359967
84 Schidlowski L. Iwamura A.P.D. Covid S.U.D. Condino-Neto A. Prando C. Diagnosis of APS-1 in two siblings following life-threatening COVID-19 pneumonia J Clin Immunol 42 2022 749 752 35305203
85 Palomba E. Carrabba M. Zuglian G. Alagna L. Saltini P. Fortina V. Treatment of SARS-CoV-2 relapse with remdesivir and neutralizing antibodies cocktail in a patient with X-linked agammaglobulinaemia Int J Infect Dis 110 2021 338 340 34332085
86 Hovey J.G. Tolbert D. Howell D. Bruton’s agammaglobulinemia and COVID-19 Cureus 12 2020 e11701
87 Duncan C.J.A. Skouboe M.K. Howarth S. Hollensen A.K. Chen R. Borresen M.L. Life-threatening viral disease in a novel form of autosomal recessive IFNAR2 deficiency in the Arctic J Exp Med 2022 219
88 Reiff D.D. Zhang M. Smitherman E.A. Mannion M.L. Stoll M.L. Weiser P. A rare STXBP2 mutation in severe COVID-19 and secondary cytokine storm syndrome Life (Basel) 12 2022
89 Delavari S. Abolhassani H. Abolnezhadian F. Babaha F. Iranparast S. Ahanchian H. Impact of SARS-CoV-2 pandemic on patients with primary immunodeficiency J Clin Immunol 41 2021 345 355 33263173
90 Abolhassani H. Vosughimotlagh A. Asano T. Landegren N. Boisson B. Delavari S. X-linked TLR7 deficiency underlies critical COVID-19 pneumonia in a male patient with ataxia-telangiectasia J Clin Immunol 42 2022 1 9
91 Abolhassani H. Delavari S. Landegren N. Shokri S. Bastard P. Du L. Genetic and immunological evaluation of children with inborn errors of immunity and severe or critical COVID-19 J Allergy Clin Immunol 2022
92 Esenboga S. Ocak M. Akarsu A. Bildik H.N. Cagdas D. Iskit A.T. COVID-19 in patients with primary immunodeficiency J Clin Immunol 2021
93 Guven S.C. Erden A. Karakas O. Armagan B. Usul E. Omma A. COVID-19 outcomes in patients with familial Mediterranean fever: a retrospective cohort study Rheumatol Int 41 2021 715 719 33611656
94 Karakoc Aydiner E. Bilgic Eltan S. Babayeva R. Aydiner O. Kepenekli E. Kolukisa B. Adverse COVID-19 outcomes in immune deficiencies: inequality exists between subclasses Allergy 77 2022 282 295 34314546
95 Goudouris E.S. Pinto-Mariz F. Mendonca L.O. Aranda C.S. Guimaraes R.R. Kokron C. Outcome of SARS-CoV-2 infection in 121 patients with inborn errors of immunity: a cross-sectional study J Clin Immunol 2021
96 Grumach A.S. Goudouris E. Dortas Junior S. Marcelino F.C. Alonso M.L.O. Martins R.O. COVID-19 affecting hereditary angioedema patients with and without C1 inhibitor deficiency J Allergy Clin Immunol Pract 9 2021 508 510 33271349
97 Marcus N. Frizinsky S. Hagin D. Ovadia A. Hanna S. Farkash M. Minor clinical impact of COVID-19 pandemic on patients with primary immunodeficiency in Israel Front Immunol 11 2020 614086 33519822
98 Quinti I. Lougaris V. Milito C. Cinetto F. Pecoraro A. Mezzaroma I. A possible role for B cells in COVID-19? Lesson from patients with agammaglobulinemia J Allergy Clin Immunol 146 2020 211 223.e4 32333914
99 Milito C. Lougaris V. Giardino G. Punziano A. Vultaggio A. Carrabba M. Clinical outcome, incidence, and SARS-CoV-2 infection-fatality rates in Italian patients with inborn errors of immunity J Allergy Clin Immunol Pract 9 2021 2904 2906.e2 33894392
100 Giardino G. Milito C. Lougaris V. Punziano A. Carrabba M. Cinetto F. The impact of SARS-CoV-2 infection in patients with inborn errors of immunity: the experience of the Italian Primary Immunodeficiencies Network (IPINet) J Clin Immunol 2022
101 Milito C. Cinetto F. Palladino A. Garzi G. Punziano A. Lagnese G. Mortality in severe antibody deficiencies patients during the first two years of the COVID-19 pandemic: vaccination and monoclonal antibodies efficacy 2022 Biomedicines
102 Deya-Martinez A. Garcia-Garcia A. Gonzalez-Navarro E.A. Yiyi L. Vlagea A. Jordan I. COVID-19 in children and young adults with moderate/severe inborn errors of immunity in a high burden area in pre-vaccine era Clin Immunol 230 2021 108821 34391937
103 Shields A.M. Anantharachagan A. Arumugakani G. Baker K. Bahal S. Baxendale H. Outcomes following SARS-CoV-2 infection in patients with primary and secondary immunodeficiency in the UK Clin Exp Immunol 2022
104 Brown L.K. Moran E. Goodman A. Baxendale H. Bermingham W. Buckland M. Treatment of chronic or relapsing COVID-19 in immunodeficiency J Allergy Clin Immunol 149 2022 557 561.e1 34780850
105 Castano-Jaramillo L.M. Yamazaki-Nakashimada M.A. O’Farrill-Romanillos P.M. Muzquiz Zermeno D. Scheffler Mendoza S.C. Venegas Montoya E. COVID-19 in the context of inborn errors of immunity: a case series of 31 patients from Mexico J Clin Immunol 2021
106 Drabe C.H. Hansen A.E. Rasmussen L.D. Larsen O.D. Moller A. Mogensen T.H. Low morbidity in Danish patients with common variable immunodeficiency disorder infected with severe acute respiratory syndrome coronavirus 2 Infect Dis (Lond) 53 2021 953 958 34375571
107 Katzenstein T.L. Rasmussen L.D. Drabe C.H. Larsen C.S. Hansen A.E. Staerkind M. Outcome of SARS-CoV-2 infection among patients with common variable immunodeficiency and a matched control group: a Danish nationwide cohort study Front Immunol 13 2022 994253 36211430
108 Koltan S. Zietkiewicz M. Grzesk E. Becht R. Berdej-Szczot E. Cienkusz M. COVID-19 in unvaccinated patients with inborn errors of immunity-polish experience Front Immunol 13 2022 953700 36211407
109 Milota T. Sobotkova M. Smetanova J. Bloomfield M. Vydlakova J. Chovancova Z. Risk factors for severe COVID-19 and hospital admission in patients with inborn errors of immunity—results from a multicenter nationwide study Front Immunol 13 2022 835770 35296097
110 Bourguiba R. Delplanque M. Vinit C. Ackermann F. Savey L. Grateau G. Clinical course of COVID-19 in a cohort of 342 familial Mediterranean fever patients with a long-term treatment by colchicine in a French endemic area Ann Rheum Dis 2020
111 Ho H.E. Mathew S. Peluso M.J. Cunningham-Rundles C. Clinical outcomes and features of COVID-19 in patients with primary immunodeficiencies in New York City J Allergy Clin Immunol Pract 9 2021 490 493.e2 33039649
112 Cohen B. Rubinstein R. Gans M.D. Deng L. Rubinstein A. Eisenberg R. COVID-19 infection in 10 common variable immunodeficiency patients in New York City J Allergy Clin Immunol Pract 9 2021 504 507.e1 33217615
113 Greenmyer J.R. Joshi A.Y. COVID-19 in CVID: a case series of 17 patients J Clin Immunol 42 2022 29 31 34669142
114 Kuster J.K. Unlu S. Makin T.A. Par-Young J. Simonov M. Shafi S. Low IgG trough and lymphocyte subset counts are associated with hospitalization for COVID-19 in patients with primary antibody deficiency J Allergy Clin Immunol Pract 10 2022 633 636.e3 34929372
115 Meyts I. Bucciol G. Quinti I. Neven B. Fischer A. Seoane E. Coronavirus disease 2019 in patients with inborn errors of immunity: an international study J Allergy Clin Immunol 147 2021 520 531 32980424
116 Marcus N. Ashkenazi-Hoffnung L. Ovadia A. Dalal I. Yoffe S. Kropach N. SARS-CoV-2 symptomatic reinfection among patients with primary antibody deficiency J Allergy Clin Immunol Pract 10 2022 1907 1909 35598866
117 Terlizzi V. Motisi M.A. Pellegrino R. Padoan R. Chiappini E. Risk factors for severe COVID-19 in people with cystic fibrosis: a systematic review Front Pediatr 10 2022 958658 36003489
118 Bastard P. Orlova E. Sozaeva L. Levy R. James A. Schmitt M.M. Preexisting autoantibodies to type I IFNs underlie critical COVID-19 pneumonia in patients with APS-1 J Exp Med 2021 218
119 Tangye S.G. Bucciol G. Meyts I. Mechanisms underlying host defense and disease pathology in response to severe acute respiratory syndrome (SARS)-CoV2 infection: insights from inborn errors of immunity Curr Opin Allergy Clin Immunol 21 2021 515 524 34494617
120 Zhang Q. Bastard P. Liu Z. Le Pen J. Moncada-Velez M. Chen J. Inborn errors of type I IFN immunity in patients with life-threatening COVID-19 Science 370 2020 eabd4570 32972995
121 Puel A. Bastard P. Bustamante J. Casanova J.L. Human autoantibodies underlying infectious diseases J Exp Med 2022
122 Meisel C. Akbil B. Meyer T. Lankes E. Corman V.M. Staudacher O. Mild COVID-19 despite autoantibodies against type I IFNs in autoimmune polyendocrine syndrome type 1 J Clin Invest 2021
123 Campbell T.M. Liu Z. Zhang Q. Moncada-Velez M. Covill L.E. Zhang P. Respiratory viral infections in otherwise healthy humans with inherited IRF7 deficiency J Exp Med 2022
124 van der Made C.I. Simons A. Schuurs-Hoeijmakers J. van den Heuvel G. Mantere T. Kersten S. Presence of genetic variants among young men with severe COVID-19 JAMA 324 2020 663 673 32706371
125 Asano T. Boisson B. Onodi F. Matuozzo D. Moncada-Velez M. Maglorius Renkilaraj M.R.L. X-linked recessive TLR7 deficiency in ∼1% of men under 60 years old with life-threatening COVID-19 Sci Immunol 2021
126 Zhang Q. Matuozzo D. Le Pen J. Lee D. Moens L. Asano T. Recessive inborn errors of type I IFN immunity in children with COVID-19 pneumonia J Exp Med 2022
127 Zhang Q. Bastard P. Effort C.H.G. Cobat A. Casanova J.L. Human genetic and immunological determinants of critical COVID-19 pneumonia Nature 2022
128 Bastard P. Rosen L.B. Zhang Q. Michailidis E. Hoffmann H.H. Zhang Y. Autoantibodies against type I IFNs in patients with life-threatening COVID-19 Science 2020
129 Troya J. Bastard P. Planas-Serra L. Ryan P. Ruiz M. de Carranza M. Neutralizing autoantibodies to type I IFNs in >10% of patients with severe COVID-19 pneumonia hospitalized in Madrid, Spain J Clin Immunol 41 2021 914 922 33851338
130 al e Bastard P. Gervais A. Le Voyer T. Rosain J. Philippot Q. Autoantibodies neutralizing type I IFNs are present in ∼4% of uninfected individuals over 70 years and account for ∼20% of COVID-19 deaths Sci Immunol 2021
131 de Prost N. Bastard P. Arrestier R. Fourati S. Mahevas M. Burrel S. Plasma exchange to rescue patients with autoantibodies against type I interferons and life-threatening COVID-19 pneumonia J Clin Immunol 41 2021 536 544 33616813
132 Abers M.S. Rosen L.B. Delmonte O.M. Shaw E. Bastard P. Imberti L. Neutralizing type-I interferon autoantibodies are associated with delayed viral clearance and intensive care unit admission in patients with COVID-19 Immunol Cell Biol 2021
133 Manry J. Bastard P. Gervais A. Le Voyer T. Rosain J. Philippot Q. The risk of COVID-19 death is much greater and age dependent with type I IFN autoantibodies Proc Natl Acad Sci U S A 119 2022 e2200413119
134 Eto S. Nukui Y. Tsumura M. Nakagama Y. Kashimada K. Mizoguchi Y. Neutralizing type I Interferon autoantibodies in Japanese patients with severe COVID-19 J Clin Immunol 2022
135 van der Wijst M.G.P. Vazquez S.E. Hartoularos G.C. Bastard P. Grant T. Bueno R. Type I interferon autoantibodies are associated with systemic immune alterations in patients with COVID-19 Sci Transl Med 13 2021 eabh2624 34429372
136 Goncalves D. Mezidi M. Bastard P. Perret M. Saker K. Fabien N. Antibodies against type I interferon: detection and association with severe clinical outcome in COVID-19 patients Clin Transl Immunology 10 2021 e1327 34429968
137 Raadsen M.P. Gharbharan A. Jordans C.C.E. Mykytyn A.Z. Lamers M.M. van den Doel P.B. Interferon-alpha2 auto-antibodies in convalescent plasma therapy for COVID-19 J Clin Immunol 42 2022 232 239 34767118
138 Mathian A. Breillat P. Dorgham K. Bastard P. Charre C. Lhote R. Lower disease activity but higher risk of severe COVID-19 and herpes zoster in patients with systemic lupus erythematosus with pre-existing autoantibodies neutralising IFN-alpha Ann Rheum Dis 2022
139 Nguyen H. Salkeld J. Agarwal S. Goodman A. Compassionate use of REGN-COV2 in the treatment of COVID-19 in a patient with impaired humoral immunity Clin Infect Pract 12 2021 100089 34426799
140 Loh S.Y. Bassett J. Hoodless E.J. Walshaw M. Possible COVID-19 reinfection in a patient with X-linked agammaglobulinaemia BMJ Case Rep 2021
141 Daoussis D. Leonidou L. Kalogeropoulou C. Paliogianni F. Tzouvelekis A. Protracted severe COVID-19 pneumonia following rituximab treatment: caution needed Rheumatol Int 41 2021 1839 1843 34409510
142 Loarce-Martos J. Garcia-Fernandez A. Lopez-Gutierrez F. Garcia-Garcia V. Calvo-Sanz L. Del Bosque-Granero I. High rates of severe disease and death due to SARS-CoV-2 infection in rheumatic disease patients treated with rituximab: a descriptive study Rheumatol Int 40 2020 2015 2021 32945944
143 Lang-Meli J. Fuchs J. Mathe P. Ho H.E. Kern L. Jaki L. Case series: convalescent plasma therapy for patients with COVID-19 and primary antibody deficiency J Clin Immunol 42 2022 253 265 34893946
144 Lougaris V. Baronio M. Vitali M. Tampella G. Cattalini M. Tassone L. Bruton tyrosine kinase mediates TLR9-dependent human dendritic cell activation J Allergy Clin Immunol 133 2014 1644 1650.e4 24612681
145 Treon S.P. Castillo J.J. Skarbnik A.P. Soumerai J.D. Ghobrial I.M. Guerrera M.L. The BTK inhibitor ibrutinib may protect against pulmonary injury in COVID-19–infected patients Blood 135 2020 1912 1915 32302379
146 Roschewski M. Lionakis M.S. Sharman J.P. Roswarski J. Goy A. Monticelli M.A. Inhibition of Bruton tyrosine kinase in patients with severe COVID-19 Sci Immunol 2020 5
147 Roumier M. Paule R. Vallee A. Rohmer J. Ballester M. Brun A.L. Tocilizumab for severe worsening COVID-19 pneumonia: a propensity score analysis J Clin Immunol 41 2021 303 314 33188624
148 Jamilloux Y. Henry T. Belot A. Viel S. Fauter M. El Jammal T. Should we stimulate or suppress immune responses in COVID-19? Cytokine and anti-cytokine interventions Autoimmun Rev 19 2020 102567
149 Ferré E.N.N. Schmitt M.M. Ochoa S. Rosen L.B. Shaw E.R. Burbelo P.D. SARS-CoV-2 spike protein–directed monoclonal antibodies may ameliorate COVID-19 complications in APECED patients Front Immunol 12 2021 720205
150 Barouch D.H. COVID-19 vaccines—immunity, variants, boosters N Engl J Med 387 2022 1011 1020 36044620
151 Sette A. Crotty S. Immunological memory to SARS-CoV-2 infection and COVID-19 vaccines Immunol Rev 310 2022 27 46 35733376
152 Zhang Z. Mateus J. Coelho C.H. Dan J.M. Moderbacher C.R. Galvez R.I. Humoral and cellular immune memory to four COVID-19 vaccines Cell 185 2022 2434 2451.e17 35764089
153 Chemaitelly H. Tang P. Hasan M.R. AlMukdad S. Yassine H.M. Benslimane F.M. Waning of BNT162b2 vaccine protection against SARS-CoV-2 infection in Qatar N Engl J Med 385 2021 e83 34614327
154 Andrews N. Stowe J. Kirsebom F. Toffa S. Rickeard T. Gallagher E. Covid-19 vaccine effectiveness against the Omicron (B.1.1.529) variant N Engl J Med 386 2022 1532 1546 35249272
155 Siddle K.J. Krasilnikova L.A. Moreno G.K. Schaffner S.F. Vostok J. Fitzgerald N.A. Transmission from vaccinated individuals in a large SARS-CoV-2 Delta variant outbreak Cell 185 2022 485 492.e10 35051367
156 Wilkinson S.A.J. Richter A. Casey A. Osman H. Mirza J.D. Stockton J. Recurrent SARS-CoV-2 mutations in immunodeficient patients Virus Evol 8 2022 veac050 35996593
157 Harvey W.T. Carabelli A.M. Jackson B. Gupta R.K. Thomson E.C. Harrison E.M. SARS-CoV-2 variants, spike mutations and immune escape Nat Rev Microbiol 19 2021 409 424 34075212
158 Amodio D. Ruggiero A. Sgrulletti M. Pighi C. Cotugno N. Medri C. Humoral and cellular response following vaccination with the BNT162b2 mRNA COVID-19 vaccine in patients affected by primary immunodeficiencies Front Immunol 12 2021 727850
159 Antoli A. Rocamora-Blanch G. Framil M. Mas-Bosch V. Navarro S. Bermudez C. Evaluation of humoral and cellular immune responses to the SARS-CoV-2 vaccine in patients with common variable immunodeficiency phenotype and patient receiving B-cell depletion therapy Front Immunol 13 2022 895209
160 Arroyo-Sanchez D. Cabrera-Marante O. Laguna-Goya R. Almendro-Vazquez P. Carretero O. Gil-Etayo F.J. Immunogenicity of anti–SARS-CoV-2 vaccines in common variable immunodeficiency J Clin Immunol 42 2022 240 252 34787773
161 Barmettler S. DiGiacomo D.V. Yang N.J. Lam T. Naranbhai V. Dighe A.S. Response to severe acute respiratory syndrome coronavirus 2 initial series and additional dose vaccine in patients with predominant antibody deficiency J Allergy Clin Immunol Pract 10 2022 1622 1634.e4 35381395
162 Bergman P. Blennow O. Hansson L. Mielke S. Nowak P. Chen P. Safety and efficacy of the mRNA BNT162b2 vaccine against SARS-CoV-2 in five groups of immunocompromised patients and healthy controls in a prospective open-label clinical trial EBioMedicine 74 2021 103705
163 Bergman P. Wullimann D. Gao Y. Wahren Borgstrom E. Norlin A.C. Lind Enoksson S. Elevated CD21low B cell frequency is a marker of poor immunity to Pfizer-BioNTech BNT162b2 mRNA vaccine against SARS-CoV-2 in patients with common variable immunodeficiency J Clin Immunol 42 2022 716 727 35290571
164 Carrabba M. Baselli L.A. Consonni D. Ceriotti F. Fabio G. Responses to SARS-CoV-2 vaccines of patients with common variable immune deficiencies and X-linked agammaglobulinemia J Clin Immunol 2022
165 Delmonte O.M. Bergerson J.R.E. Burbelo P.D. Durkee-Shock J.R. Dobbs K. Bosticardo M. Antibody responses to the SARS-CoV-2 vaccine in individuals with various inborn errors of immunity J Allergy Clin Immunol 148 2021 1192 1197 34492260
166 Hagin D. Freund T. Navon M. Halperin T. Adir D. Marom R. Immunogenicity of Pfizer-BioNTech COVID-19 vaccine in patients with inborn errors of immunity J Allergy Clin Immunol 2021
167 Pham M.N. Murugesan K. Banaei N. Pinsky B.A. Tang M. Hoyte E. Immunogenicity and tolerability of COVID-19 messenger RNA vaccines in primary immunodeficiency patients with functional B-cell defects J Allergy Clin Immunol 149 2022 907 911.e3 34952033
168 Ponsford M.J. Evans K. Carne E.M. Jolles S. COVID-19 vaccine uptake and efficacy in a national immunodeficiency cohort J Clin Immunol 42 2022 728 731 35149962
169 Pulvirenti F. Fernandez Salinas A. Milito C. Terreri S. Piano Mortari E. Quintarelli C. B cell response induced by SARS-CoV-2 infection is boosted by the BNT162b2 vaccine in primary antibody deficiencies 2021 Cells
170 Romano C. Esposito S. Donnarumma G. Marrone A. Detection of neutralizing anti-severe acute respiratory syndrome coronavirus 2 antibodies in patients with common variable immunodeficiency after immunization with messenger RNA vaccines Ann Allergy Asthma Immunol 127 2021 499 501 34352358
171 Salinas A.F. Mortari E.P. Terreri S. Quintarelli C. Pulvirenti F. Di Cecca S. SARS-CoV-2 vaccine induced atypical immune responses in antibody defects: everybody does their best J Clin Immunol 41 2021 1709 1722 34669144
172 Shields A.M. Faustini S.E. Hill H.J. Al-Taei S. Tanner C. Ashford F. SARS-CoV-2 vaccine responses in individuals with antibody deficiency: findings from the COV-AD study J Clin Immunol 2022
173 van Leeuwen L.P.M. GeurtsvanKessel C.H. Ellerbroek P.M. de Bree G.J. Potjewijd J. Rutgers A. Immunogenicity of the mRNA-1273 COVID-19 vaccine in adult patients with inborn errors of immunity J Allergy Clin Immunol 149 2022 1949 1957 35421449
174 Sokal A. Bastard P. Chappert P. Barba-Spaeth G. Fourati S. Vanderberghe A. Human type I IFN deficiency does not impair B cell response to SARS-CoV-2 mRNA vaccination J Exp Med 2022
175 Bastard P. Vazquez S. Liu J. Laurie M.T. Wang C.Y. Gervais A. Vaccine breakthrough hypoxemic COVID-19 pneumonia in patients with auto-Abs neutralizing type I IFNs Sci Immunol 2022
176 Khalid M. Urban A. Darnel D. Freeman A. Clinical outcomes of SARS-CoV2 infection in STAT3 deficiency J Clin Immunol 41 2021 abstract 83
177 Ochoa S. Rosen L. Lionakis M. Uzel G. Suez D. COVID-19 in 3 patients with CTLA4 haploinsufficiency and absence of autoantibodies to type 1 interferons J Clin Immunol 2021
| 36522221 | PMC9746792 | NO-CC CODE | 2022-12-15 23:21:57 | no | J Allergy Clin Immunol. 2022 Dec 13; doi: 10.1016/j.jaci.2022.11.010 | utf-8 | J Allergy Clin Immunol | 2,022 | 10.1016/j.jaci.2022.11.010 | oa_other |
==== Front
Environ Health Perspect
Environ Health Perspect
EHP
Environmental Health Perspectives
0091-6765
1552-9924
Environmental Health Perspectives
EHP11374
10.1289/EHP11374
Research Letter
Detection of Neonicotinoid Insecticides and Their Metabolites in Human Cerebrospinal Fluid
Li Adela Jing 1 *
Si Mengya 2 *
Yin Renli 1
Qiu Rongliang 1
Li Huashou 1
Yao Fen 3
Yu Yunjiang 4
Liu Wenhua 5 6
Wang Zhen 5 6
https://orcid.org/0000-0002-7568-0595
Jiao Xiaoyang 7
1 College of Natural Resources and Environment, South China Agricultural University, Guangzhou, China
2 First Affiliated Hospital of Shantou University Medical College, Shantou, China
3 Department of Pharmacology, Shantou University Medical College, Shantou, China
4 South China Institute of Environmental Sciences, Ministry of Ecology and Environment, Guangzhou, China
5 Guangdong Provincial Key Laboratory of Marine Disaster Prediction and Prevention, Shantou University, Shantou, China
6 Guangdong Provincial Key Laboratory of Marine Biotechnology, Shantou University, Shantou, China
7 Department of Cell Biology and Genetics, Shantou University Medical College, Shantou, China
Address correspondence to Zhen Wang. Email: [email protected]. And, Xiaoyang Jiao. Email: [email protected]
13 12 2022
12 2022
130 12 12770207 4 2022
16 10 2022
29 11 2022
https://ehp.niehs.nih.gov/about-ehp/license EHP is an open-access journal published with support from the National Institute of Environmental Health Sciences, National Institutes of Health. All content is public domain unless otherwise noted.
* These authors contributed equally to this work.
The authors declare they have no actual or potential competing financial interest.
Note to readers with disabilities: EHP strives to ensure that all journal content is accessible to all readers. However, some figures and Supplemental Material published in EHP articles may not conform to 508 standards due to the complexity of the information being presented. If you need assistance accessing journal content, please contact [email protected]. Our staff will work with you to assess and meet your accessibility needs within 3 working days.
==== Body
pmcIntroduction
Neonicotinoid insecticides (NEOs) are now the most widely used neurotoxic insecticides on 140 crops in 120 countries.1 Global use of NEOs has caused growing concerns for decline of pollinators,2 risk to nontarget species,3 and potential adverse effects on human health.4 A systematic review in 2017 reported associations between human chronic NEO exposure and adverse neurological outcomes based on interview data.4 A previous study in mice reported that NEOs and their metabolites could be distributed to the cerebral cortex, hippocampus, and striatum after oral intake of the NEOs.5 Validated biomarker investigations of NEOs were generally measured in human urine4 or mouse blood,5 which may not be reflective of central nervous system (CNS) exposure.
Cerebrospinal fluid (CSF) is an integral CNS component emerging in parallel with the developing CNS. Several biochemical markers in CSF have been used for diagnosis and evaluation of neurological diseases.6 The present study aimed to explore whether NEOs and their metabolites could be evident in CSF.
Methods
This study recruited 314 donors from 4,410 patients available for CSF analysis in the First Affiliated Hospital of Shantou University, Shantou, China, from April 2019 to January 2021. CSF specimens were collected by a senior physician using the method of clinical lumbar puncture. A total of 314 CSF samples were collected from patients experiencing symptoms with different diagnoses (mostly viral encephalitis, encephalitis other than viral encephalitis, leukemia, cerebral hemorrhage, cerebral laceration, urinary tract infection, respiratory failure, pulmonary tuberculosis, and posterior circulation ischemia). Informed consent forms were completed by all participants. The research protocol was approved by the clinical research ethical committee from the First Affiliated Hospital of Shantou University Medical College.
CSF samples of 250μL were prepared using the same protocol as previously reported for serum extractions.7 The method for the analysis of NEOs and their metabolites entailed acidification by 2% formic acid in water, solid phase extraction (Bond Elut Plexa, 3mL, 60mg; Agilent), and high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) detection [ABSCIEX 6500 electrospray ionization (ESI)-MS/MS; Applied Biosystems] connected with a Acquity ultra-performance liquid chromatography ethylene-bridged-hybrid (UPLC BEH) C18 column (100×2.1mm, 1.7μm; Waters). Nine NEOs were measured: nitenpyram (NIT), thiamethoxam, imidacloprid, acetamiprid (ACE), thiacloprid, clothianidin, flonicamid, imidaclothiz, and sulfoxaflor; along with six metabolites: N-desmethyl-thiamethoxam, olefin-imidacloprid, 5-hydroxy-imidacloprid, N-desmethyl-acetamiprid (N-dm-ACE), thiacloprid-amide, and 6-chloronicotinic acid. The solvent-based calibration curves resulted in linear regression coefficients (r) of >0.99. The relative recoveries of the 15 NEOs and metabolites spiked into CSF at 1, 10, and 20 ng/mL ranged from 85% to 114%. The relative standard deviations of repeated analyses of samples were ≤18.3%.
The instrument-derived concentrations of NEOs and metabolites below the limits of detection (LODs) were assigned values of the LODs divided by the square root of 2. Data are presented as medians with interquartile ranges (IQRs) given that all data were skewed. Differences in median CSF levels of study analytes among the five age groups were tested by a nonparametric Kruskal-Wallis H-test, whereas differences in median CSF levels between genders were examined by a Mann-Whitney U-test. Data were analyzed using SPSS (version 19.0; SPSS Inc.) with values of p<0.05 denoting statistical significance.
Results and Discussion
Overall, 99% of all CSF samples contained quantifiable amounts of at least one NEO. Nine percent (28/314) of all samples had a single compound, 84% (265/314) had between 2 and 6, and 6% (19/314) had between 7 and 10 (Table 1).
Table 1 Descriptive statistics of concentrations (in ng/mL) of neonicotinoids and their metabolites in cerebrospinal fluid samples (n=314).
Analyte LOD (ng/mL) n >LOD DF (%) Median (ng/mL) IQR (ng/mL) Min. (ng/mL) Max. (ng/mL)
NIT 0.015 143 45.5 <LOD <LOD–0.152 <LOD 102
THX 0.009 73 23.3 <LOD <LOD–<LOD <LOD 2.19
IMI 0.011 126 40.1 <LOD <LOD–0.230 <LOD 0.357
ACE 0.002 124 39.5 <LOD <LOD–0.006 0.002 2.07
THI 0.004 4 1.27 <LOD <LOD–<LOD <LOD 0.009
CLO 0.094 23 7.32 <LOD <LOD–<LOD <LOD 1.50
FLO 0.004 52 16.6 <LOD <LOD–<LOD <LOD 0.045
N-DMT 0.032 50 15.9 <LOD <LOD–<LOD <LOD 0.367
TA 0.007 50 15.9 <LOD <LOD–LOD <LOD 0.196
IMZ 0.022 105 33.4 <LOD <LOD–0.026 <LOD 0.185
N-dm-ACE 0.006 268 85.4 0.049 0.015–0.139 <LOD 2.38
6-CN 0.052 1 0.32 <LOD <LOD–<LOD <LOD 0.060
SUF 0.004 8 2.55 <LOD <LOD–<LOD <LOD 0.085
Of-IMI 0.003 123 39.2 <LOD <LOD–0.004 <LOD 0.094
5-OH-IMI 0.172 25 7.96 <LOD <LOD–<LOD <LOD 4.20
Note: Measurements below the LODs were assigned values of the LODs divided by the square root of 2. 5-OH-IMI, 5-hydroxy-imidacloprid; 6-CN, 6-chloronicotinic acid; ACE, acetamiprid; CLO, clothianidin; DF, detection frequency; FLO, flonicamid; IMI, imidacloprid; IMZ, imidaclothiz; IQR, interquartile range; LOD, limit of detection; max., maximum value; min., minimum value; N-dm-ACE, N-desmethyl-acetamiprid; N-DMT, N-desmethyl-thiamethoxam; NIT, nitenpyram; Of-IMI, olefin-imidacloprid; SUF, sulfoxaflor; TA, thiacloprid-amide; THI, thiacloprid; THX, thiamethoxam.
As a specific metabolite from ACE, N-dm-ACE was found at the highest concentration (median=0.049 ng/mL) and at the highest frequency of occurrence (85.4%, n=268) among the 15 analytes tested. Detection rates for the other 14 compounds were in the range of 0.32%–45.5%. Studies done using in vitro models have reported that some metabolites are more potent than the corresponding parent NEO because of their higher affinity to nicotinic α4β2 nicotinic acetylcholine receptors.8 A recent study reported that N-dm-ACE concentrations in 93% of CSF samples (13/14) collected from children treated for leukemia and lymphomas were within an order of magnitude (median=0.012 ng/mL)9 of the concentrations detected in our study. The potential neurotoxic effects of N-dm-ACE to the human CNS should be considered comprehensively.
In our study, the two NIT detections at 102 and 39.2 ng/mL in CSF were considered outliers, with reported concentrations well over three times the IQR. The first participant was a 72-y-old male at the sampling time in 2020. The latter participant was a 60-y-old female at the sampling time in 2021. Both were diagnosed with cerebral hemorrhage at the sampling point.
The enrolled 314 patients were from 1 month to 89 years of age, with 180 (57.3%) males (Table 2). Median CSF concentrations of N-dm-ACE, solely, differed significantly among age groups, having an increasing trend with age (Table 2). A recent study reported that the aging process diminished copper clearance from the CSF of rats by disrupting copper transporting proteins in the choroid plexus.10 In the present study, no significant difference was found for median CSF concentrations of target analytes between genders (Table 2).
Table 2 Median concentrations (in ng/mL) of neonicotinoids and their metabolites with overall detection frequency (DF) of >30% in all cerebrospinal fluid samples, by age and sex with individual DF in parentheses (%).
Characteristics ACE N-dm-ACE IMI Of-IMI NIT IMZ
Total (n=314) <LOD (39.5)** 0.049 (85.4) <LOD (40.1) <LOD (39.2) <LOD (45.5) <LOD (33.4)
Age (y)
0–14 (n=29) <LOD (41.4) 0.011 (55.2) <LOD (31.0) <LOD (44.8) <LOD (37.9) <LOD (37.9)
15–24 (n=41) <LOD (31.7) 0.045 (90.2) <LOD (48.8) <LOD (29.3) <LOD (43.9) <LOD (19.5)
25–44 (n=92) <LOD (47.8) 0.062 (89.1) <LOD (37.0) <LOD (34.8) <LOD (37.0) <LOD (26.1)
45–64 (n=97) <LOD (36.1) 0.077 (86.6) <LOD (41.2) <LOD (40.2) 0.038 (54.6) <LOD (39.2)
≥65 (n=55) <LOD (36.4) 0.058 (89.1) <LOD (41.8) <LOD (49.1) <LOD (49.1) <LOD (43.6)
Sex
Male (n=180) <LOD (41.1) 0.047 (84.4) <LOD (39.4) <LOD (37.8) <LOD (42.2) <LOD (35.6)
Female (n=134) <LOD (37.3) 0.051 (86.6) <LOD (41.0) <LOD (41.0) <LOD (50.0) <LOD (30.6)
Note: Measurements below the LODs were assigned values of the LODs divided by square root of 2 prior to calculating median concentrations. Differences in median CSF levels among the five age groups tested by a non-parametric Kruskal-Wallis H-test. Differences in median CSF levels between genders tested by a Mann-Whitney U-test. ACE, acetamiprid; CSF, cerebrospinal fluid; IMI, imidacloprid; IMZ, imidaclothiz; LOD, limit of detection; N-dm-ACE, N-desmethyl-acetamiprid; NIT, nitenpyram; Of-IMI, olefin-imidacloprid. **, p=0.002 denoting statistically significant differences in CSF levels of N-dm-ACE among the five age groups.
This was an exploratory study to identify detectable concentrations of nine NEOs and six metabolites in the CSF of 314 patients. For continued global use of NEOs, mechanisms of toxicity, especially to the CNS in humans, need to be more rigorously investigated.
Acknowledgments
The authors appreciate the technical support of Z. Lin and the comments of W.W. Au. Research funding was provided by Start-Up Funds for Recruitment of Talents (4200/221068 to A.J.L.) and the Youth Cultivation (4200/221321 to A.J.L.) at South China Agricultural University, the National Natural Science Foundation of China (42277427 to A.J.L., 42177264 to Z.W., and 21906067 to Z.W.) and Shantou University Scientific Research Foundation for Talents (NTF19044 to Z.W.), and 2020 Li Ka Shing Foundation Cross-Disciplinary Research Grant (2020LKSFG03E to Z.W.). The views expressed are those of the authors and not necessarily those of funding organizations.
==== Refs
References
1. Bass C, Denholm I, Williamson MS, Nauen R. 2015. The global status of insect resistance to neonicotinoid insecticides. Pestic Biochem Physiol 121 :78–87, PMID: , 10.1016/j.pestbp.2015.04.004.26047114
2. Woodcock BA, Bullock JM, Shore RF, Heard MS, Pereira MG, Redhead J, et al. 2017. Country-specific effects of neonicotinoid pesticides on honey bees and wild bees. Science 356 (6345 ):1393–1395, PMID: , 10.1126/science.aaa1190.28663502
3. Eng ML, Stutchbury BJM, Morrissey CA. 2019. A neonicotinoid insecticide reduces fueling and delays migration in songbirds. Science 365 (6458 ):1177–1180, PMID: , 10.1126/science.aaw9419.31515394
4. Cimino AM, Boyles AL, Thayer KA, Perry MJ. 2017. Effects of neonicotinoid pesticide exposure on human health: a systematic review. Environ Health Perspect 125 (2 ):155–162, PMID: , 10.1289/EHP515.27385285
5. Hirai A, Sugio S, Nimako C, Nakayama SMM, Kato K, Takahashi K, et al. 2022. Ca2+ imaging with two-photon microscopy to detect the disruption of brain function in mice administered neonicotinoid insecticides. Sci Rep 12 (1 ):5114, PMID: , 10.1038/s41598-022-09038-7.35332220
6. Fame RM, Lehtinen MK. 2020. Emergence and developmental roles of the cerebrospinal fluid system. Dev Cell 52 (3 ):261–275, PMID: , 10.1016/j.devcel.2020.01.027.32049038
7. Li AJ, Banjabi AA, Takazawa M, Kumosani TA, Yousef JM, Kannan K. 2020. Serum concentrations of pesticides including organophosphates, pyrethroids and neonicotinoids in a population with osteoarthritis in Saudi Arabia. Sci Total Environ 737 :139706, PMID: , 10.1016/j.scitotenv.2020.139706.32512300
8. Loser D, Grillberger K, Hinojosa MG, Blum J, Haufe Y, Danker T, et al. 2021. Acute effects of the imidacloprid metabolite desnitro-imidacloprid on human nACh receptors relevant for neuronal signaling. Arch Toxicol 95 (12 ):3695–3716, PMID: , 10.1007/s00204-021-03168-z.34628512
9. Laubscher B, Diezi M, Renella R, Mitchell EAD, Aebi A, Mulot M, et al. 2022. Multiple neonicotinoids in children’s cerebro-spinal fluid, plasma, and urine. Environ Health 21 (1 ):10, PMID: , 10.1186/s12940-021-00821-z.35016674
10. Liu LL, Du D, Zheng W, Zhang Y. 2022. Age-dependent decline of copper clearance at the blood–cerebrospinal fluid barrier. Neurotoxicology 88 :44–56, PMID: , 10.1016/j.neuro.2021.10.011.34718061
| 36512430 | PMC9746793 | NO-CC CODE | 2022-12-16 23:23:55 | no | Environ Health Perspect. 2022 Dec 13; 130(12):127702 | utf-8 | Environ Health Perspect | 2,022 | 10.1289/EHP11374 | oa_other |
==== Front
Environ Health Perspect
Environ Health Perspect
EHP
Environmental Health Perspectives
0091-6765
1552-9924
Environmental Health Perspectives
EHP12418
10.1289/EHP12418
Erratum
Erratum: “Evaluation of Neurotoxicity in BALB/c Mice following Chronic Exposure to Polystyrene Microplastics”
Jin Haibo
Yang Chen
Jiang Chengyue
Li Luxi
Pan Mengge
Li Dongmei
Han Xiaodong
Ding Jie
13 12 2022
12 2022
13 12 2022
130 12 12900110 11 2022
15 11 2022
https://ehp.niehs.nih.gov/about-ehp/license EHP is an open-access journal published with support from the National Institute of Environmental Health Sciences, National Institutes of Health. All content is public domain unless otherwise noted.
Environ Health Perspect. 130(10):107002 (2022), https://doi.org/10.1289/EHP10255
==== Body
pmcThe author affiliations were incorrectly listed. The correct affiliations are:
Haibo Jin,1,2 Chen Yang,3 Chengyue Jiang,1,2 Luxi Li,1,2 Mengge Pan,1,2 Dongmei Li,1,2 Xiaodong Han,1,2 and Jie Ding1,2
1Immunology and Reproduction Biology Laboratory & State Key Laboratory of Analytical Chemistry for Life Science, Medical School, Nanjing University, Nanjing 210093, China
2Jiangsu Key Laboratory of Molecular Medicine, Nanjing University, Nanjing 210093, China
3State Key Laboratory of Analytical Chemistry for Life Science, School of Chemistry and Chemical Engineering, Nanjing University, Nanjing, Jiangsu, China
In addition, the Abstract incorrectly stated “Meanwhile, exposed mice also exhibited disruption of the blood–brain barrier, higher level of dendritic spine density, and an inflammatory response in the hippocampus.” This sentence should read “Meanwhile, exposed mice also exhibited disruption of the blood–brain barrier, lower level of dendritic spine density, and an inflammatory response in the hippocampus.”
The authors regret the error.
| 36512431 | PMC9746794 | NO-CC CODE | 2022-12-16 23:23:55 | no | Environ Health Perspect. 2022 Dec 13; 130(12):129001 | utf-8 | Environ Health Perspect | 2,022 | 10.1289/EHP12418 | oa_other |
==== Front
Appl Acoust
Appl Acoust
Applied Acoustics. Acoustique Applique. Angewandte Akustik
0003-682X
1872-910X
The Authors. Published by Elsevier Ltd.
S0003-682X(21)00144-4
10.1016/j.apacoust.2021.108051
108051
Article
How much COVID-19 face protections influence speech intelligibility in classrooms?
Caniato Marco ⁎
Marzi Arianna
Gasparella Andrea
Free University of Bozen Bolzano, Italy
⁎ Corresponding author at: Piazza Università, 39100 Bozen, Italy.
23 3 2021
7 2021
23 3 2021
178 108051108051
14 12 2020
9 3 2021
11 3 2021
© 2021 The Authors
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The ongoing pandemic caused by the COVID-19 virus is challenging many aspects of daily life. Several personal protective devices have become essential in our lives. Face protections are mostly used in order to stop the air aerosol coming out of our mouths. Nevertheless, this fact may also have a negative effect on speech transmission both in outdoor and indoor spaces. After a severe lockdown, classes have now started again. The adoption of face protection by teachers is either recommended or mandatory even though this is affecting speech intelligibility and thus students’ comprehension. This study aims to understand how protections may affect the speech transmission in classrooms and how this could be influenced by the several typologies of face protections. An experimental campaign was conducted in a classroom in two different reverberant conditions, measuring and comparing the variation in speech transmission and sound pressure level at different receiver positions. Furthermore, a microphone array was used to investigate the distribution of the indoor sound field, depending on the sound source. Results clearly show how different types of personal protection equipment do affect speech transmission and sound pressure level especially at mid-high frequency and that the source emission lobes vary when wearing certain types of personal devices.
Keywords
COVID-19
Pandemic
Face mask
Surgical
Acoustic
Hearing
==== Body
pmc1 Introduction
At present, the world is dealing with the SARS-CoV-2 or COVID-19 pandemic. Several precautions were taken to limit the spread of the virus. Social distances and lockdown were imposed in all countries, leading to a decrease in contagious behaviour and situations. When these actions were mitigated, governments made individuals use protection at least in indoor environments. Thus, face masks, shields and respirators were adopted everywhere, becoming a necessary element of everybody life. However, personal facial protections are already widely used in medical care and in hospitals to prevent healthcare workers from virus infections by aerosol particles [1]. They have also demonstrated to be efficient personal devices that slow down the spread of the Corona virus [2]. However, there are several different types of protective equipment available on the market: (i) surgical masks, (ii) cloth or community masks [3] or respirators (also known as Filtering Face Piece) and (iii) face shields [4].This variety of protections not only has an impact on the filtering of aerosol particles [5], but also on communication capabilities. Moreover, face masks are a visible obstacle in body language, impacting: (i) verbal communication [6], (ii) people’s emotional expressivity, the signs of understanding [7], (iii) lip reading [8], which is a very useful aid in ordinary communication among people and extremely useful for people with hearing loss and finally (iv) voice [9].
For the above-mentioned reasons, it is therefore essential to study how face protections may influence noise source in the built learning environment. The adoption of these devices is certainly needed for the pandemic [10], but they could negatively influence communication. Indeed, several studies have already demonstrated the impact of these devices on speech. Aranaz Andrés et al. [11], in their study on the secondary effects of face masks, highlighted the difficulties related to speech understanding, when speakers are wearing face protection. Hulzen and Fabry [8] investigated their effects on COVID-19 infection protection and the discomfort caused to individuals with hearing loss due to the use of face masks. Accordingly, face protections pose two problems for these people: they cannot receive any clues from lip reading [12] and furthermore, other people’s voices may be attenuated or in some way distorted.
Many available studies focus on communication difficulties in healthcare environments. Wittum et al. [13] investigated the side-effects of wearing surgical masks in hospitals. Goldin et al. [14] compared, in a controlled environment, a few face masks applied to a white noise simulator through an artificial mouth. They compared the difference in sound pressure levels between all the masks offered by their healthcare providers, highlighting that each type of device essentially reduced the speech transmission in this environment. Other studies have demonstrated the impact of face masks on communication and on speech. Corey et al. [15] measured the sound reduction caused by several masks, focusing on sound pressure levels in a laboratory with sound-absorbing walls. Conversely, Palmiero et al. [16] studied the Speech Transmission Index of three masks to understand how these may compromise sound transmission.
During the pandemic, in several countries the use of masks inside educational environments has either become mandatory for teachers and/or for students [17], [18], or recommended [19], in order to protect teachers, students and their families [20].
From this perspective, the investigation of how these protections may influence learning and communication performance assumes a key role. There are various types of educational environments. In some countries there are standards for acoustic conditions in the classes that have thresholds for reverberation times, noise levels and sound pressure levels at receivers’ positions [21]. However, generally, the classroom can have a high reverberation time (reverberant conditions) [22] or may feature sound absorbing material to reduce the reverberation [23], [24].
Moreover, the educational environment is strongly influenced by the indoor acoustic conditions [25]. Indeed, in classrooms more than one factor is generally taken into account: (i) background noise [26], (ii) acoustic field [27], (iii) effects of noise on the speaker/teacher [28] and (iv) effects on the listener/students [29] . Skarlatos and Manatakis [30] found that reverberation time is one of the main causes of high noise levels in classrooms. As also mentioned by Hodgson et al. [31], this parameter is of paramount importance and it should be taken into account when characterizing indoor spaces. It refers to the reflection of sound waves [32] and it has been shown to have a strong impact on classroom soundscape [33] and on students’ achievements [34]. Indeed, control of the indoor sound field improves pupils’ speech comprehension [35] and it has been shown that a good acoustic environment helps to increase voice and words understanding [36] . Generally, in order to assess the acoustic quality of a room, specific objective parameters are used, coupled with reverberation time. When focusing on rooms dedicated to speech, clarity is one of the most utilized parameters [37]. Clarity considers reverberation time and acoustic strength as a single index and thus it is very useful in determining how the indoor sound field varies when moving from one student’s desk to another [38]. However, these parameters are not affected by possible source signal amplitude variations. Thus, they can effectivly assess indoor sound field, but they cannot be used to investigate source-receiver sound wave transmission.
Accordingly, Zannin and Zwirtis [39] measured the sound pressure levels in three Brazilian classes and found that this parameter is not only related to the quality of verbal communication, but also to the listening effort, which can impact students’ activities. Astolfi et al. [40] in their study highlighted how (i) sound pressure levels measured inside the classroom, (i) reverberation time and (iii) numerical models should be integrated to have a systematic procedure to approach this complex indoor environment. Within the learning environment, Prodi et al. [41] found that the Speech Transmission Index (STI) is also a relevant parameter strongly related to Clarity and the measured sound pressure levels at the receivers’ positions. Seetha et al. [42] reported in their research that the STI measurement coupled with the analysis of the sound pressure levels is enough to have a complete overview of the sound field in a classroom. Peng et al. studied STI in 28 Chinese schools, finding that the speech intelligibility score is a very important parameter for all ages. Radosz [43] illustrated a method to assess the acoustic quality of classrooms including all the factors mentioned above with a unique global index.
The importance of students’/listeners’ distance from the source/teacher has also been assessed. Bilzi et al. [44] highlighted how in classrooms STI could significantly vary from one position to another, while Stewart and Cabrera [45] found that the directivity of the source strongly influences the STI results in noise-free conditions. Thus, reverberation time clarity and STI are the parameters most often used to characterize indoor sound field, while STI and sound pressure levels at receivers’ positions are needed to investigate the quality of the transmitted speech.
Among all these medical and acoustic studies, there is a lack of focus on the impact of the face protections within educational environments. Students in classrooms are careful listeners [46], [47] . Wearing masks may impact the communication between teachers and students, affecting linguistic and non-verbal information [48] . Transparent masks or shields may represent one solution as they may help by making lip reading possible and helping to overcome speech degradation [49]. However, to the authors’ knowledge, no studies have been conducted on these issues.
For these reasons, the aim of this research is to understand to what extent face protections influence (i) speech intelligibility, (ii) indoor sound pressure level distribution and (iii) source directivity within a classroom. Another aim of this research is to understand if the impact of these devices varies, according to diverse indoor sound field conditions.
2 Materials and methods
This study was conducted in the Free University of Bozen living lab, which provides reconfigurable educational environments that are used for everyday activities. The facility includes a classroom, wherein it is possible to change indoor sound absorption and thus to study diverse acoustic conditions.
Two different scenarios were built:• Scenario A (Higher reverberant room), where a reverberant classroom was recreated, emulating a traditional school environment;
• Scenario B (Lower reverberant room), where 16-square-meter sound absorbing panels were installed on the room ceiling. Their sound absorption coefficient is reported in Fig. 1 . Panels were composed of 5 cm polyester fibre panels with a porosity of 0.99. The air gap between them and the room ceiling is 10 cm.Fig. 1 Panels sound absorption coefficient.
Indoor sound field acoustic characterizations were performed in accordance with the ISO 3382 standard [50], [51], [52] using an omnidirectional (dodecahedral) sound source and a logarithmic sine sweep as a signal, as depicted in Fig. 2 . One excitation per one receiver position was played, recorded and then post produced by means of a convolution, in order to obtain the impulse response. The source position was chosen very close to the teacher’s desk. The positions of the receivers were uniformly distributed along the rows of desks of the existing students. Reverberation time T30 and clarity C50 were derived from measurements and used to characterize different scenarios.Fig. 2 hanging microphones and dodecahedral source positioning.
In order to understand how face protections may influence sound source (teachers), speech intelligibility tests were carried out using a repeatable and robust procedure so as to ensure reliable results. In particular, a constant and directive sound source is needed to ensure comparable values. This procedure determines the Speech Transmission Index (STI). It assesses the quality of the received speech in different student positions and it is normally used to determine whether a classroom is suitable for teaching and learning or not. STI evaluation is based on IEC 60268-16 [53] which provides a gender specific octave band weighting and redundancy factors. Gender-related factors are expressed differently due to signal spectra and different weighting factors.
The STI measurement is strongly influenced by several factors related to the acoustic conditions of the classroom [54]: the reverberation time, background noise, clarity and sound pressure levels. Evaluations connected to speech were assessed using speech intelligibility tests [19], developed using a directive MLS-equipped noise source, positioned on the teacher’s desk and receivers located at the students’ positions [55]. Unlike clarity and reverberation time, which depend exclusively on (i) the classroom volume, (ii) the internal reverberation time and (iii) the source-receiver position, the STI also relies on the sound source frequency level. If this is altered, the final result varies.
In terms of background noise, it was regularly checked so as not to disturb or influence the acoustic tests. In this way, especially for STI measurements, it was possible to consider the uncertainty to be lower than a JND [56]
Since the face protections are not normally attached to real mouths, special accessories were plugged on a directive speaker surface in order to maintain the masks’ membrane at a distance of about 4 mm to 7 mm from the directive speaker in order to simulate a mouth-nose appearance and profile (Fig. 3 ).Fig. 3 used mouth and nose adapters.
Protections were then mounted on the directive speaker, using this configuration. In order to analyze as many options as possible within the two described scenarios A and B, ten kinds of individual face protections were tested. They were mounted on the directive sound source, covering the noise emission so as to simulate real mouth and nose overlapping. The devices used are listed in Table 1 and examples of applications are reported in Fig. 4 . Complete identification of the face masks under consideration and transparent shields are reported in the Appendix tables (see Fig. 5 ).Table 1 List and description of all masks and combination.
ID number Name of the mask Description Available in pharmacy Filter type Certified PPE
1 Surgical Surgical masks Yes – No
2 Surgical Masks featuring stiff transparent insertion Surgical mask with a transparent insertion dedicated to lip reading. Transparent insertion is made of Polyethylene Terephthalate Glycol No – No
3 Surgical Masks featuring flexible transparent insertion Surgical mask with a transparent insertion dedicated to lip reading. Transparent insertion is made of Transparent acrylic sheet No – No
4a Community Mask - no filter [57] Two layers of cotton fabric, featuring a possibility of filter insertion. No filter No – No
4b Community Mask - with filter [57] Two layers of cotton fabric, featuring a possibility of filter insertion. Paper filter No Non-woven polypropylene (PP) and paper filter No
5 Hand-made face Mask, kid size Two layers of cotton fabric, featuring a possibility of filter insertion No – No
6 FFP2 safety mask Filtering face mask made of polypropylene melt blown cloth Yes Removes at least 94% of all particles starting from 0.3 µm in diameter Yes
7 NCN MR2 Certified organic cotton and treated with exclusive Carbon Nanoclusters. No NCN materials with high anti-particle, antimicrobial and bacteriostatic properties Yes
8 Face shield Polyethylene Terephthalate Glycol Yes – No
9 Face shield coupled with surgical mask Face shield made of Polyethylene Terephthalate Glycol, or PETG combined with a surgical mask. Yes (face shield – No
Yes (face mask)
10 Face shield coupled with community Mask – with paper filter Face shield made of Polyethylene Terephthalate Glycol combined with mask made with two layers of cotton fabric. Yes (face shield) – No
No (face mask) Paper filter
Fig. 4 examples of complete tests. Left: cotton mask. Right: transparent shield.
Fig. 5 Face protections investigated useful for COVID 19 spread prevention.
Tests were performed using a B&K 4720 directive noise source and a B&K 2270 sound level meter, connected to a sound card controlled by B&K Dirac 6.0 software. The receiving points were placed at 1.4 m from the ground in six representative positions, according to the class layout (Fig. 6 ). The room is box-shaped, 7.3 m × 7.6 m × 3.6 m, for a volume of 196 m3. The source was placed at 1.5 m from the ground and near the teacher’s desk. As consistent with COVID guidelines, the source (teacher) and the positions of the receivers (students) are not expected to change [58]. For every scenario, firstly a traditional STI measurement was performed, i.e. without any face protection (case 0), as a starting point for the comparison. For the sake of simplicity, only frequency average results will be presented. These are calculated by taking into consideration the octave band measured valued of 500 Hz, 1000 Hz and 2000 Hz. An arithmetic average is then performed, in order to obtain an average STI result.Fig. 6 Classroom layout: positions of the noise source and receivers for STI measurements. Locations of students are depicted by chair occupancy.
Reverberation time, clarity and speech transmission index variations are evaluated using the Just Noticeable Difference (JND) approach [50]. Seraphim [59] stated that a reverberation time variation is perceived when its variation is at about 5%. Karjalaine and Jarvelainen [60] reported that a 10% variation is more accurate than 5%. For this reason, in this paper a variation of 10% is used to consider a reverberation time variation to be significant. For clarity and STI, a variation of 1 dB for clarity and 0.03 for STI are to be classified as minimum JNDs, according to Bradley et al. [61].
In order to better understand indoor sound field variations, synchronous sound pressure level measurements were developed by means of an omnidirectional microphones model ECM ½” 999 hung on the roof and controlled by a Zoom F8 sound card. The positions are depicted in Fig. 6. They were located at a height of 2 m (Fig. 2).
Furthermore, since source directivity may be affected by mouth obstruction, acoustic camera high frequency analyses were carried out using a planar 40 cm × 40 cm microphone array, equipped with 40 MEMS noise sensors coupled with an A/D converter. Every device features a linear frequency response from 60 Hz to 15000 Hz. Software analysis was performed using a Robust Asymptotic Functional Beamforming algorithm, providing high quality 5 megapixels real time images and videos. The acoustic camera was situated in receiver 4 position (Fig. 6), in order to better monitor the phenomenon.
3 Results and discussion
In terms of the acoustic characterization of the room scenarios, Tables 2 and 3 report reverberation time and clarity for mid-high frequencies (500 Hz – 2000 Hz) and STI results for case 0 (no device).Table 2 Reverberation time and clarity frequency results.
Position ID Scenario A Scenario B
T30 C50 T30 C50
500 Hz 1000 Hz 2000 Hz 500 Hz 1000 Hz 2000 Hz 500 Hz 1000 Hz 2000 Hz 500 Hz 1000 Hz 2000 Hz
1 1.22 1.05 1.05 0.42 0.25 0.48 0.74 0.70 0.72 4.85 3.96 3.89
2 1.16 1.08 1.06 −1.28 0.41 −1.12 0.69 0.69 0.70 0.25 3.12 1.86
3 1.17 1.06 1.07 −2.37 0.16 0.81 0.72 0.68 0.73 2.37 3.77 3.64
4 1.28 1.10 1.07 −1.11 −0.31 −0.55 0.66 0.67 0.73 0.48 2.18 2.89
5 1.12 1.08 1.04 −0.67 −1.25 0.02 0.69 0.69 0.72 3.29 2.14 2.65
6 1.21 1.11 1.04 1.60 0.47 0.84 0.73 0.7 0.69 4.02 2.21 2.46
Table 3 – Speech transmission index results for case 0 (no device).
Scenario A Scenario B
STI Male STI Female STI Male STI Female
1 0.51 0.50 0.56 0.56
2 0.48 0.48 0.54 0.54
3 0.51 0.50 0.60 0.59
4 0.48 0.48 0.52 0.52
5 0.48 0.48 0.50 0.51
6 0.48 0.48 0.50 0.51
It can be seen that in Scenario A (higher reverberant room), most of the parameters present values which are usually conducive to a good indoor learning environment [62] at most considered frequencies. Accordingly, reverberation time is over 1 s in all frequency bands in all positions, while clarity results are close to zero or below in almost all cases. As for the STI results, it is clear that Scenario A provides poor speech intelligibility [63] in all positions, regardless of the teacher gender, except in position 1 (very close to the source) and in position 3 (in front of the source).
In terms of Scenario B (lower reverberant room), it can be seen that indoor sound field quality greatly improved. Reverberation time is always lower than 0.75 s, highlighting a variation of up to 4 JNDs. Clarity is always higher than 2 dB, except in position 2 and 4 at 500 Hz, where, however, the values are always higher than zero. This shows a positive variation up to 4 JNDs. The value of the STI also improved, gaining up to 3 JND per position per device type.
As an overall result, it is possible to state that the indoor sound field significantly varied its conditions and thus STI measurements using face masks in scenario A and scenario B start from clearly different indoor conditions.
3.1 STI results
The results of STI measurements performed using individual protection were grouped based on the receiver position and catalogued based on the wearable protective equipment. For the sake of brevity, all tables are reported in appendix A. As a reference, the case without any face protections (case 0 – no device) was used, which as expected presents the best STI values. As an overall result (Table 4 ), it can be highlighted that many individual protections highly limit sound wave propagation. Considering the maximum JND results, up to 2 JNDs were found in all positions. Maximum values show how STI male values are 2 to 3 JNDs for scenario A and 2 to 5 JNDs for scenario B. For STI female values are lower: 1 to 2 JNDs, for the first case, while 1 to 3 for the second case.Table 4 – maximum and minimum STI JNDs related to different face protections.
Scenario A Scenario B
Position STI male max JNDs STI Female max JNDs STI male max JNDs STI Female max JNDs
1 3 ID 10 1 ID 10 4 ID 10 3 ID 10
2 3 ID 6 2 ID 6 4 ID 10 2 ID 10
3 3 ID10 2 ID 17 5 ID 10 3 ID 10
4 2 ID 10 1 ID 10 4 ID 8 2 ID 8
5 2 ID 4b 1 ID 4b 2 ID 4b 1 ID 4b
6 3 ID 7 2 ID 7 4 ID 4b 3 ID 4b
STI male min JNDs STI female min JNDs STI male min JNDs STI Female min JNDs
1 1 ID5 0 ID5 0 ID1 0 ID1
2 0 ID1 0 ID1 1 ID1 0 ID1
3 1 ID1 0 ID1 2 ID1 1 ID1
4 0 ID5 0 ID5 1 ID5 0 ID5
5 1 ID1 0 ID1 1 ID8 0 ID8
6 1 ID8 1 ID8 0 ID8 0 ID8
In Fig. 7 , a representative background noise level is reported. This was measured during the test. As it can be seen, levels are very low in every frequency. Since the noise source level has to be 60 dB (A) at 1 m from the noise source, it can be concluded that background noise cannot influence the final results.Fig. 7 Representative background noise level during the measurements.
Another general outcome is that differences are higher in scenario B (lower reverberant room), rather than scenario A (higher reverberant room). It is also evident how the STI male in both cases suffers more compared to the female STI. Positions 3 and 6 are where maximum variation are verified for both scenarios, followed by Position 2 and Position 1, which have a similar variation of STI in both scenarios. In addition, as expected, the face shield coupled with the community mask featuring a paper sheet is the worst combination, followed by the community mask featuring a paper filter (ID 4b). The maximum JNDs assessed are mostly referable to the face shield coupled with the community mask (ID 10), followed by the community mask with a paper filter (ID 4b), NCN MR2 mask (ID 7), the face shield (ID 8) and the FFP2 safety mask (ID 6). This is attributable to the fact that they are all medical or high efficiency protection devices.
Referring to minimum differences in JNDs, it can be seen that surgical masks (ID 1) are the best masks, followed by kids’ masks (ID 5). Values show that very few JNDs are caused by surgical masks in most receivers’ positions in both scenarios and for both male and female sources.
3.2 Sound pressure level results
In order to understand how the indoor acoustic field is affected by source protection, sound pressure level measurements were taken. For the sake of brevity, results for the receivers 2, 3 and 6 are shown in Figs. 8 and 9 , reporting on the x-axis the difference between the reference case (ID 0 – no face protection) and the investigated device (e.g. ID 0 minus ID 1). In order to present an overview of both the least and the most impacting face protection, the results are listed in Table 5 . In this table, the sum of the sound pressure levels at voice frequency (630–5000 Hz) Lp and the difference between the reference case (ID 0 – no face protection) and each device Δ are reported.Fig. 8 SPL relative differences frequency trends for face protections which least affect indoor sound field. Positions 2, 3 and 6.
Fig. 9 SPL relative differences frequency trends for face protections which most affect indoor sound field. Positions 2, 3 and 6.
Table 5 Sound pressure level of voice frequency (630–5000 Hz) Lp and relative case difference Δ. Grey: minimum differences; plain: medium differences; bold maximum differences.
Position n.2 Position n.3 Position n.6
Scenario A Scenario B Scenario A Scenario B Scenario A Scenario B
ID Lp (dB) Δ (dB) Lp (dB) Δ (dB) Lp (dB) Δ (dB) Lp (dB) Δ (dB) Lp (dB) Δ (dB) Lp (dB) Δ (dB)
0 56.8 – 57.6 – 58.4 – 56.7 – 55.0 – 52.2 –
1 55.9 1.0 56.9 0.7 57.9 0.5 55.9 0.8 54.4 0.6 52.2 0.1
2 56.0 0.9 56.6 1.0 57.9 0.5 55.9 0.8 54.6 0.4 51.8 0.4
3 55.7 1.1 56.7 0.9 58.1 0.3 55.9 0.7 55.0 0.0 51.9 0.3
4a 54.7 2.1 56.4 1.1 56.9 1.5 54.5 2.2 53.5 1.5 50.9 1.3
4b 54.9 1.9 56.0 1.6 57.4 1.0 53.9 2.7 53.8 1.2 50.6 1.6
5 55.1 1.7 56.5 1.1 56.7 1.7 54.6 2.1 53.5 1.5 50.8 1.5
6 55.1 1.7 56.8 0.8 57.3 1.2 54.7 2.0 54.1 0.9 51.2 1.0
7 55.3 1.5 56.6 1.0 57.6 0.8 54.8 1.9 53.9 1.1 51.3 0.9
8 60.7 −3.9 60.6 −3.0 62.8 −4.4 61.0 −4.3 59.8 −4.8 57.3 −5.1
9 60.1 −3.3 59.6 −2.0 62.5 −4.1 60.2 −3.5 58.9 −3.9 56.8 −4.5
10 60.2 −3.4 59.7 −2.1 62.2 −3.8 59.6 −2.9 59.3 −4.3 56.5 −4.3
It is evident that surgical masks, even those featuring a transparent window, do not significantly modify sound pressure levels at the receivers. Other types of masks provide significant deviations of up to 2.7 dB. Face shields clearly alter the sound wave transmission. Negative differences mean that visors reflect a significant part of the sound waves to the upper part of the room, thus considerably affecting the source-receiver propagation path.
On the frequency domain, some results offer interesting insights. In Figs. 8 and 9, the results of respectively the least and the most influencing devices are reported. Accordingly, it can be seen that all individual protections provide differences at mid-high frequencies. Accordingly, volcano-shaped trends at 2000 Hz can easily be identified in both scenarios. In the cases reported in Fig. 8, levels can vary from 5 to 10 dB in both scenarios, while the ones included in Fig. 9 fluctuate from 8 dB to 11 dB in scenario A and from 9 dB to 14 dB in scenario B. Thus, these protections may modify the original emitted signal in a range where human voices are very active. A reduction of up to 10 dB in both reverberant and absorbing rooms was found, thus implying that sound pressure reductions do not depend on indoor sound field conditions.
On higher frequencies, some differences are still present. They linearly decrease to 4000 Hz, where they stop and in some cases start to slightly increase again. Accordingly, both scenarios present the same overall trends. The face shields (ID 8 and ID 10) offer a positive effect at low and high frequency, but when it is coupled with a mask, they negatively influence soundwave propagation only at high frequency.
In this light, it is interesting to present tridimensional relative difference trends of sound pressure levels related to the single most influencing face protection in order to understand if there are some significant differences among the diverse devices. Fig. 10 shows scenario A, while Fig. 11 shows scenario B. For the first scenario, it can be deduced that all face protections provide similar 3D distribution. Differences are present in all receiver positions and always show significant values, starting from the closest to the most distant positions. Specifically, it can be highlighted that position 6, which is also the furthest from the noise source, always shows the highest value. Conversely, locations 1, 2 and 3 very often provide the same results. As an overall result, we can deduce that the majority of the face protections, which clearly influence sound propagation in a reverberant classroom, will provide almost the same sound field variation.Fig. 10 3D SPL relative difference distribution for the 5 most impacting face protections at 2000 Hz. Scenario A.
Fig. 11 3D SPL relative difference distribution for the 5 most impacting face protections at 2000 Hz. Scenario B.
In the case of a sound absorbing environment, differences are always distributed using the same pattern. The worst location (except mask ID 10) is position 3, where the highest difference can be found. In the first row of desks, values are often the lowest ones and in position 3 results are the lowest for mask ID 4b and 6. Furthermore, in scenario B, we can conclude that all face protections, which regularly affect noise propagation, present common sound field differences.
In terms of sound pressure level, as a global result it can be determined that differences are significantly high in both scenarios and they provide the same difference ranges. The asymmetry could be caused by the sound absorbing panels distribution on the ceiling. This implies that there is no significant difference provided by the indoor environment conditions, but the sound limitations are intrinsically related to the face mask. Another paramount result is that face protections mostly act in a difference range of 8–14 dB at 2000 Hz. This means that 8 dB of source reduction is granted even if a receiver (student) stands close to the sound source (teacher), regardless of the position and of the indoor acoustic conditions. Furthermore, the last receivers’ row, which is the furthest away from the source (teacher), present the worst difference values, up to 14 dB. This means that wearing one of these face protections may reduce the sound wave emitted by the speaker by up to 14 dB.
3.3 Acoustic camera results
Acoustic camera results are reported in Fig. 12 for scenario A and in Fig. 13 for scenario B, both for 2000 Hz, for the sake of brevity. When comparing the two different indoor acoustic fields, as an overall result it is evident that in a more reverberant room the source emission presents a wider emission lobe. The application of face masks modifies the received sound pressure levels but does not significantly variate the lobes’ dimensions. On the other hand, face shields do change emission patterns, splitting them into two distinct parts.Fig. 12 Acoustic camera analyses in scenario A for all types of individual protections at 2000 Hz. Color legend represent 10 dB range. Center purple represents the maximum level while external yellow-orange represents the minimum. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 13 Acoustic camera analyses in scenario B for all types of individual protections at 2000 Hz. Center purple represents the maximum level while external yellow-orange represents the minimum. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
This effect is caused by the sudden reflection of the direct sound wave against the polymeric shield. However, these frequencies are related to the human voice. If teachers wore a face shield, the direct sound would represent a minor contribution, compared to the reflected sound. This may not be perceived as a comfortable situation by the receivers, since they would be able to see the teacher in a definite point in the indoor space but hear his/her voice coming from other directions.
3.4 Overall considerations
Masks negatively affect voice propagation in classrooms, regardless of the indoor acoustic conditions, but with differences linked to characteristic male and female emissions. Furthermore, the results clearly identify how face shields split the emitted noise into two different directions, thus compromising the direct propagation.
The results above (Table 4) demonstrate that wearing masks influences the male voice more than the female voice. Accordingly, in the first case (male voice), more JNDs are found to be present rather than in the second case (female voice) in both scenarios. This depends on how the face masks behave when a sound wave propagates through them. A membrane-like model most appropriately explains the above results, with the mask stretched from nose to mouth. In this case, a sound absorbing effect can be predicted when the resonance frequency of the system is verified, in accordance with Eq. (1):(1) f0=60d∙σ
where f0 is the resonance frequency [Hz], d is the distance between the membrane and the solid behind it [m] and σ is the superficial mass of the membrane [kg/m2]. In this case, the membrane is the mask, the solid is the mouth and the typical weight range is about 2–8 g. Masks are worn differently and this causes differences in the membrane free area, however balanced by mass variations. In this calculation, it is also difficult to define a perfect distance from the mouth as well as the free vibrating areas related to single masks due to the morphological diversity in human faces. However, ranges may be determined based on real mask weights and average facial differences, retrieved from the literature [64], [65], [66]. Thus, a range starting from about 1600 Hz to about 2200 Hz can be assessed. This is clearly consistent with what is shown in Fig. 9, where the major difference peak starts to rise at 1600 Hz, presents its maximum at 2000 Hz and then decreases. The noise limitation effect observed at higher frequency ranges could be caused by sound absorption of the textiles (masks) or transmission loss produced by transparent polymer (face shields).
In this light, the morphology, shape and components of the face protections act together to limit the sound wave generated by the source. When analyzing the male–female voice difference, it may be helpful to consider the A-weighted reference speech levels used to calculate STI male and STI female. In Fig. 14 , the STI male and female -weighting trends (expressed in percentage of reduction) are reported and combined with the range where the face protections mostly act.Fig. 14 STI male and female percentage A-weighted trends, with reference to the most impacting face protection frequency range (grey region).
It is evident how face protections act in a range where the female voice presents a higher percentage reduction than the male voice, with a difference of around 5% (3 dB in the standard). Thus, the STI female is influenced less by wearing face protections than the male STI, because the weighting penalizes the female voices more, where the masks provide the most efficient sound reduction.
4 Conclusions
In this study, the influence of face protections on sound propagation in classrooms was experimentally investigated. Two different scenarios were considered, featuring very different indoor sound fields (more reverberant, less reverberant). In both scenarios, 10 different masks and face shield were tested. Speech intelligibility index, sound pressure levels and emission lobes were assessed.
The main findings are summarised as follows:1) face protections worn by teachers in classrooms do affect sound propagation. Considering differences related to the kind of protection device, speech transmission index could vary up to 5 Just Noticeable Differences. In the sound absorbing scenario (Scenario B), differences are higher than in the reverberant one (Scenario A), both for the male voice and all the receiver positions.
2) Face protections defined as high medical protections are the ones that mostly reduce source emission, except for surgical masks.
3) Surgical masks, also featuring transparent plastic windows on the mouth area for lip reading, are not shown to significantly influence speech transmission in both classroom scenarios.
4) Speech Transmission Index for male and female voices present different reduction patterns. In particular, female speech is less influenced by face protection compared to male speech. This is due to the membrane resonance natural frequency and textile sound absorption.
5) The most affected frequencies lay in a range, which starts form 1600 Hz and ends at 6300 Hz. The most affected octave band is 2000 Hz for all face protections. In this frequency, a minimum of 8 dB and a maximum of 14 dB reduction was recorded, depending on the type of device worn and chosen position in the classroom.
6) All of the most influencing face protections present similar 3D reduction patterns along the indoor acoustic field in both scenarios, but with different values. It thus can be concluded that the indoor sound field will be affected in the same way in most of the cases (different face protection), but with some differences in noise reduction.
7) Source emission and directivity is affected by face protection, especially by face shields. In this case, mostly at high frequencies, the emission lobe is always split into two parts in both scenarios. This negatively affects sound perceptions, as the source is not clearly identified in the space.
Finally, it can be concluded that in classrooms surgical masks should be adopted as they protect from mouth spray and do not significantly reduce speech intelligibility and indoor sound field, both without and with a transparent window for lip reading. Other types of tested facial masks significantly influence speech transmissions and thus teacher-student communication in learning environments.
CRediT authorship contribution statement
Marco Caniato: Conceptualization, Methodology, Measurements, Paper Writing, Supervision. Arianna Marzi: Measurements, Paper Writing. Andrea Gasparella: Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A In Table A1, Table A2, Table A3, Table A4, Table A5, Table A6 , in green the minimum STI value and in grey the maximum STI value for each position are depicted, in accordance with the methodology depicted in Section 3.1.Table A1 Speech transmission index in reverberating and absorbing condition in position one.
ID Position ID Face Mask Scenario A Scenario B
STI Male STI Female STI Male STI Female
1 0 0.51 0.50 0.56 0.56
1 1 0.48 0.48 0.54 0.55
1 2 0.44 0.46 0.50 0.53
1 3 0.46 0.48 0.50 0.54
1 4a 0.46 0.47 0.48 0.52
1 4b 0.44 0.47 0.46 0.50
1 5 0.49 0.49 0.53 0.54
1 6 0.44 0.46 0.49 0.52
1 7 0.44 0.46 0.46 0.50
1 8 0.46 0.47 0.53 0.54
1 9 0.44 0.47 0.46 0.51
1 10 0.41 0.45 0.42 0.47
Table A2 Speech transmission index in reverberating and absorbing condition in position two.
ID Position ID Face Mask Scenario A Scenario B
STI Male STI Female STI Male STI Female
2 0 0.48 0.48 0.54 0.54
2 1 0.46 0.48 0.50 0.53
2 2 0.42 0.45 0.44 0.49
2 3 0.43 0.46 0.46 0.50
2 4a 0.41 0.44 0.46 0.51
2 4b 0.39 0.43 0.43 0.48
2 5 0.45 0.47 0.48 0.51
2 6 0.37 0.41 0.43 0.47
2 7 0.43 0.46 0.48 0.51
2 8 0.46 0.46 0.50 0.52
2 9 0.41 0.44 0.46 0.50
2 10 0.42 0.45 0.42 0.47
Table A3 Speech transmission index in reverberating and absorbing condition in position three.
ID Position ID Face Mask Scenario A Scenario B
STI Male STI Female STI Male STI Female
3 0 0.51 0.50 0.60 0.59
3 1 0.48 0.48 0.54 0.55
3 2 0.44 0.46 0.49 0.52
3 3 0.44 0.46 0.48 0.52
3 4a 0.47 0.46 0.51 0.53
3 4b 0.42 0.45 0.45 0.49
3 5 0.48 0.47 0.53 0.54
3 6 0.46 0.47 0.50 0.52
3 7 0.43 0.44 0.48 0.51
3 8 0.47 0.46 0.51 0.53
3 9 0.43 0.46 0.46 0.50
3 10 0.42 0.45 0.43 0.48
Table A4 Speech transmission index in reverberating and absorbing condition in position four.
ID Position ID Face Mask Scenario A Scenario B
STI Male STI Female STI Male STI Female
4 0 0.48 0.48 0.52 0.52
4 1 0.46 0.47 0.48 0.51
4 2 0.43 0.45 0.44 0.48
4 3 0.42 0.44 0.43 0.47
4 4a 0.43 0.45 0.47 0.50
4 4b 0.42 0.45 0.41 0.46
4 5 0.47 0.47 0.49 0.52
4 6 0.44 0.46 0.45 0.49
4 7 0.42 0.44 0.43 0.47
4 8 0.45 0.45 0.40 0.44
4 9 0.40 0.44 0.43 0.48
4 10 0.40 0.43 0.43 0.47
Table A5 Speech transmission index in reverberating and absorbing condition in position five.
ID Position ID Face Mask Scenario A Scenario B
STI Male STI Female STI Male STI Female
5 0 0.48 0.48 0.50 0.51
5 1 0.45 0.46 0.47 0.50
5 2 0.43 0.45 0.44 0.49
5 3 0.42 0.44 0.44 0.48
5 4a 0.44 0.46 0.44 0.48
5 4b 0.40 0.44 0.43 0.48
5 5 0.44 0.45 0.47 0.51
5 6 0.42 0.45 0.44 0.49
5 7 0.41 0.44 0.43 0.48
5 8 0.43 0.44 0.49 0.52
5 9 0.41 0.44 0.43 0.48
5 10 0.41 0.45 0.45 0.49
Table A6 Speech transmission index in reverberating and absorbing condition in position six.
ID Position ID Face Mask Scenario A Scenario B
STI Male STI Female STI Male STI Female
6 0 0.48 0.48 0.50 0.51
6 1 0.43 0.45 0.47 0.50
6 2 0.42 0.45 0.41 0.46
6 3 0.41 0.44 0.40 0.45
6 4a 0.40 0.43 0.43 0.48
6 4b 0.41 0.44 0.37 0.42
6 5 0.43 0.45 0.43 0.47
6 6 0.40 0.43 0.40 0.45
6 7 0.38 0.42 0.38 0.43
6 8 0.46 0.47 0.48 0.51
6 9 0.41 0.44 0.43 0.47
6 10 0.41 0.44 0.43 0.48
Acknowledgements
Authors would like to thank Christian Platzgummer for his precious help (mask supports realization and panel movements).
Authors would also like to thank Chiara Da Rous for her kind and wonderful work (mask realization and customization). This work was supported by the Open Access Publishing Fund of the Free University of Bozen-Bolzano.
==== Refs
References
1 Coia JE, Ritchie L, Adisesh A, Makison Booth C, Bradley C, Bunyan D, Carson G, Fry C, Hoffman P, Jenkins D, Healthcare Infection Society Working Group on Respiratory and Facial Protection. Guidance on the use of respiratory and facial protection equipment. J Hospital Infect 85 (n.d.); 2013: 170–182.
2 Organization W.H. Advice on the use of masks in the context of COVID-19: interim guidance 2020 World Health Organization
3 Qaseem A. Etxeandia-Ikobaltzeta I. Yost J. Miller M.C. Abraham G.M. Obley A.J. Use of N95, surgical, and cloth masks to prevent COVID-19 in health care and community settings: living practice points from the American college of physicians (Version 1) Ann Intern Med 2020
4 Fathizadeh H. Maroufi P. Momen-Heravi M. Dao S. Köse Ş. Ganbarov K. Protection and disinfection policies against SARS-CoV-2 (COVID-19) Infez Med 28 2020 185 191 32275260
5 Brosseau L.M. McCullough N.V. Vesley D. Mycobacterial aerosol collection efficiency of respirator and surgical mask filters under varying conditions of flow and humidity Appl Occup Environ Hyg 12 6 1997 435 445
6 Marler H. Ditton A. “I’m smiling back at you”: exploring the impact of mask wearing on communication in healthcare Int J Lang Commun Disorders 56 2021
7 Carbon C.-C. Wearing face masks strongly confuses counterparts in reading emotions Front Psychol 11 2020 2526
8 Ten Hulzen R.D. Fabry D.A. Impact of hearing loss and universal face masking in the COVID-19 era Mayo clinic proceedings 2020 Elsevier 2069 2072
9 Ribeiro VV, Dassie-Leite AP, Pereira EC, Santos ADN, Martins P, de Alencar Irineu R. Effect of wearing a face mask on vocal self-perception during a pandemic. J Voice (2020).
10 Eikenberry S.E. Mancuso M. Iboi E. Phan T. Eikenberry K. Kuang Y. To mask or not to mask: Modeling the potential for face mask use by the general public to curtail the COVID-19 pandemic Infect Disease Model 2020
11 Andrés J.M.A. de Castro M.T.G.V. Vicente-Guijarro J. Peribáñez J.B. Haro M.G. Valencia-Martín J.L. Mascarillas como equipo de protección individual durante la pandemia de COVID-19: cómo, cuándo y cuáles deben utilizarse J Healthcare Qual Res 35 2020 245 252
12 Roy David Masks method and impact in the classroom Creative Educ 11 05 2020 710 734
13 Wittum KJ, Feth L, Hoglund E. The effects of surgical masks on speech perception in noise. In: Proceedings of meetings on acoustics ICA2013, Acoustical Society of America; 2013. p. 060125.
14 Goldin A. Weinstein B. Shiman N. How do medical masks degrade speech perception Hear Rev 27 2020 8 9
15 Corey RM, Jones U, Singer AC. Acoustic effects of medical, cloth, and transparent face masks on speech signals. ArXiv Preprint ArXiv:2008.04521. (2020)
16 Palmiero A.J. Symons D. Morgan J.W. III Shaffer R.E. Speech intelligibility assessment of protective facemasks and air-purifying respirators J Occup Environ Hygiene 13 12 2016 960 968
17 Government of United Kingdom, Guidance Face coverings in education Updated 5 November 2020, Education and Childcare during Coronavirus (n.d.). https://www.gov.uk/government/publications/face-coverings-in-education/face-coverings-in-education (accessed November 12, 2020).
18 The Victorian Government, Face masks in schools All school-based staff and students must wear face masks at school and when travelling to and from school., Face Masks in Schools (n.d.). https://www.coronavirus.vic.gov.au/face-masks-in-schools#use-of-personal-protective-equipment-in-education (accessed November 12, 2020).
19 U.S. Department of Health & Human Services, Cloth face masks in schools, Public Healt Emergency (2020). https://www.phe.gov/facecovering/Pages/cloth-face-masks-in-school.aspx (accessed November 12, 2020).
20 Spitzer M. Masked education? The benefits and burdens of wearing face masks in schools during the current Corona pandemic. Trends Neurosci Educ 20: 2020; 100138.
21 GBS Buildings, D. Unit, Acoustic design of schools, Stationery Office; 2003.
22 Klatte M. Lachmann T. Meis M. Effects of noise and reverberation on speech perception and listening comprehension of children and adults in a classroom-like setting Noise Health 12 2010 270 20871182
23 Shield B. Conetta R. Dockrell J. Connolly D. Cox T. Mydlarz C. A survey of acoustic conditions and noise levels in secondary school classrooms in England J Acoust Soc Am 137 1 2015 177 188 25618049
24 D. for E. and S. United Kingdom, Building Bulletin 93 : Acoustic Design of Schools., DfES London; 2003
25 Lubman David Sutherland Louis C. Good classroom acoustics are a good investment for America J Acoust Soc Am 106 4 1999 2283–2283
26 Pekkarinen Eeva Viljancn Vesa Acoustic conditions for speech communication in classrooms Scand Audiol 20 4 1991 257 263 1842300
27 Ljung R. Sörqvist P. Kjellberg A. Green A.-M. Poor listening conditions impair memory for intelligible lectures: implications for acoustic classroom standards Build Acoust 16 2009 257 265
28 Garnier M. Henrich N. Speaking in noise: How does the Lombard effect improve acoustic contrasts between speech and ambient noise? Comput Speech Lang 28 2014 580 597
29 Sarampalis A. Kalluri S. Edwards B. Hafter E. Objective measures of listening effort: Effects of background noise and noise reduction J Speech Lang Hear Res 2009
30 Skarlatos D. Manatakis M. Effects of classroom noise on students and teachers in Greece Percept Mot Skills 96 2 2003 539 544 12776837
31 Hodgson M. Rempel R. Kennedy S. Measurement and prediction of typical speech and background-noise levels in university classrooms during lectures J Acoust Soc Am 105 1999 226 233
32 Crandell C.C. Smaldino J.J. Flexer C.A. Sound field amplification: applications to speech perception and classroom acoustics 2005 Singular
33 Bradley J. Classroom acoustics to support student learning. In: Jamieson D, Rvachew S, Carhart R, Tillman TW, (1970). Interaction of competing speech signals with hearing losses. Arch Otolaryn 91: 2007; 273–279.
34 Ronsse LM, Wang LM. Acoustics of elementary school classrooms: Correlations of varying background noise level and reverberation time to student achievement. In: 38th International congress and exposition on noise control engineering 2009, INTER-NOISE 2009;2009. p. 1875–1879
35 Shield B, Conetta R, Cox T, Mydlarz C, Dockrell J, Connolly D. Acoustics and noise in English secondary schools. In: INTER-NOISE and NOISE-CON congress and conference proceedings, Institute of Noise Control Engineering; 2013. p. 5672–5678.
36 Mikulski W. Radosz J. Acoustics of classrooms in primary schools–results of the reverberation time and the speech transmission index assessments in selected buildings Arch Acoust 36 2011 777 793
37 Rościszewska T. Miśkiewicz A. Rogala T. Rudzki T. Fidecki T. Concert hall sound clarity: a comparison of auditory judgments and objective measures Arch Acoust 2012 41 46
38 Cabrera D. Acoustic clarity and auditory room size perception. In: Cairns. Australia 9-12 July, 2007; 2007.
39 Zannin P.H.T. Zwirtes D.P.Z. Evaluation of the acoustic performance of classrooms in public schools Appl Acoust 70 4 2009 626 635
40 Astolfi A. Corrado V. Griginis A. Comparison between measured and calculated parameters for the acoustical characterization of small classrooms Appl Acoust 69 2008 966 976
41 Prodi N. Visentin C. Feletti A. On the perception of speech in primary school classrooms: Ranking of noise interference and of age influence J Acoust Soc Am 133 2013 255 268 23297900
42 Seetha P, Karmegam K, Ismail MY, Sapuan SM, Ismail N, Moli L. Effects to teaching environment of noise level in school classrooms; 2008.
43 Radosz J. Global index of the acoustic quality of classrooms Arch Acoust 38 2013 159 168
44 Bilzi P, Bozzoli F, Farina A. Influence of artificial mouth’s directivity in determining Speech Transmission Index, in: Audio Engineering Society Convention 119, Audio Engineering Society; 2005
45 Stewart K. Cabrera D. Effect of acoustic environment on the sensitivity of speech transmission index to source directivity Arch Sci Rev 52 2009 71 78
46 Tobey E.A. Rekart D. Buckley K. Geers A.E. Mode of communication and classroom placement impact on speech intelligibility Arch Otolaryngol-Head Neck Surg 130 2004 639 15148190
47 Van Engen K.J. Peelle J.E. Listening effort and accented speech Front Hum Neurosci 8 2014 577 25140140
48 de Nobrega M. Opice R. Lauletta M.M. de Nobrega C.A. How face masks can affect school performance Int J Pediatr Otorhinolaryngol 2020
49 Atcherson SR, Finley ET, McDowell BR, Watson C. More Speech Degradations and Considerations in the Search for Transparent Face Coverings During the COVID-19 Pandemic, Audiology Today. July/August 2020 (n.d.).
50 I.S.O. Standard, 3382-1. Acoustics-measurement of room acoustic parameters-part 1: performance spaces. International Organization for Standardization; 2009
51 Tronchin L. Merli F. Dolci M. Virtual acoustic reconstruction of the Miners’ Theatre in Idrija (Slovenia) Appl Acoust 172 2021 107595 10.1016/j.apacoust.2020.107595
52 Tronchin L. Merli F. Manfren M. On the acoustics of the Teatro 1763 in Bologna Appl Acoust 172 2021 107598 10.1016/j.apacoust.2020.107598
53 IEC 60268: Sound system equipment–Part 16: Objective rating of speech intelligibility by speech transmission index, (2003).
54 P. CODE, Sound system equipment–Part 16: Objective rating of speech intelligibility by speech transmission index; 2003.
55 Steeneken H.J.M. Houtgast T. A physical method for measuring speech-transmission quality J Acoust Soc Am 67 1980 318 326 7354199
56 Mapp P. Uncertainties in speech transmission index measurements J Acoust Soc Am 135 2014 2236–2236
57 Word Health organization, Mask use in the context of COVID-19: Interim guidance, Word Health Organization. (2020). https://www.who.int/publications/i/item/advice-on-the-use-of-masks-in-the-community-during-home-care-and-in-healthcare-settings-in-the-context-of-the-novel-coronavirus-(2019-ncov)-outbreak (accessed December 1, 2020).
58 W. World Health Organization, Coronavirus disease (COVID-19): Schools, World Health Organization. (2020). https://www.who.int/emergencies/diseases/novel-coronavirus-2019/question-and-answers-hub/q-a-detail/q-a-schools-and-covid-19?gclid=CjwKCAjw8-78BRA0EiwAFUw8LB_KDmwWvvY7HKs42SwpEJnSRVige8tiwLix2-MNwPbrZ3UJ_sGsKRoC7GgQAvD_BwE.
59 Seraphim H.-P. Untersuchungen über die Unterschiedsschwelle exponentiellen Abklingens von Rauschbandimpulsen Acta Acust United Acust 8 1958 280 284
60 Karjalainen M, Jarvelainen H. More about this reverberation science: perceptually good late reverberation. In: Audio engineering society convention 111, Audio Engineering Society; 2001
61 Bradley J.S. Reich R. Norcross S.G. A just noticeable difference in C50 for speech Appl Acoust 58 2 1999 99 108
62 Mealings K. Classroom acoustic conditions: Understanding what is suitable through a review of national and international standards, recommendations, and live classroom measurements; 2016.
63 Barnett P.W. Acoustics A.M. Overview of speech intelligibility Proc Inst Acoust 21 1999 1 16
64 Heidari Z. Sagheb H.-R.M. Morphological evaluation of head and face in 18–25 years old J Med Sci 6 2006 462 466
65 Abbas H. 3D face morphology classification for medical applications, Cardiff University; 2018.
66 Tronchin L. Variability of room acoustic parameters with thermo-hygrometric conditions. Appl Acoust 177: 2021; 107933
| 0 | PMC9746872 | NO-CC CODE | 2022-12-15 23:21:57 | no | Appl Acoust. 2021 Jul 23; 178:108051 | utf-8 | Appl Acoust | 2,021 | 10.1016/j.apacoust.2021.108051 | oa_other |
==== Front
Appl Acoust
Appl Acoust
Applied Acoustics. Acoustique Applique. Angewandte Akustik
0003-682X
1872-910X
The Authors. Published by Elsevier Ltd.
S0003-682X(21)00473-4
10.1016/j.apacoust.2021.108379
108379
Article
Indoor soundscapes at home during the COVID-19 lockdown in London – Part II: A structural equation model for comfort, content, and well-being
Torresin Simone ab⁎
Albatici Rossano a
Aletta Francesco c
Babich Francesco b
Oberman Tin c
Stawinoga Agnieszka Elzbieta d
Kang Jian c
a Department of Civil Environmental and Mechanical Engineering, University of Trento, Italy
b Institute for Renewable Energy, Eurac Research, Bozen/Bolzano, Italy
c UCL Institute for Environmental Design and Engineering, The Bartlett, University College London, London, UK
d Management and Committees, Eurac Research, Bozen/Bolzano, Italy
⁎ Corresponding author at: Department of Civil Environmental and Mechanical Engineering, University of Trento, Italy.
8 9 2021
1 1 2022
8 9 2021
185 108379108379
8 6 2021
16 7 2021
25 8 2021
© 2021 The Authors
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The present work constitutes the sequel to the analysis of data from an online survey administered to 464 home workers in London in January 2021 during the COVID-19 lockdown. Perceived affective quality of indoor soundscapes has been assessed in the survey through a previously developed model, as the combination of two perceptual dimensions, one related to comfort (a comfortable – annoying continuum) and the other to content (a full of content – empty continuum). Part I of the study reported on differences in comfort, content, and soundscape appropriateness based on the activity performed at home during the lockdown, i.e. working from home (WFH) and relaxation. Moreover, associations between soundscape dimensions and psychological well-being have been highlighted. Part II of the study deals with the exploration of the influences of several acoustical, building, urban and person-related factors on soundscape dimensions and well-being. A mixed-method approach has been adopted by combining multivariate regression of questionnaire scores with the qualitative analysis of spontaneous descriptions given by respondents. Results showed that several sound sources, urban features, housing characteristics, working modes and demographic factors can influence (positively and negatively) soundscape dimensions differently depending on the task at hand. Notably, the perceived dominance of neighbours’ noises during relaxation, moderated by noise sensitivity, and the number of people at home were common factors negatively affecting both comfort and well-being, that partially explained the association between comfortable indoor soundscapes and better mental health. The discussion points out the importance of considering the different impacts that acoustical factors (e.g. sound typology), building (e.g., house size), urban (e.g., availability of a quiet side), situational (e.g., number of people at home), and person-related factors (e.g., noise sensitivity) can provide on building occupants depending on the specific activity people are engaged with at home and the opportunities to foster people’s well-being through building, urban and acoustic design.
Keywords
Indoor soundscape
Indoor environmental quality
Acoustic design
Well-being
COVID-19
WFH
==== Body
pmc1 Introduction
The work reports on the second part of the analysis of data gathered from an online survey conducted within the project ‘Home as a place of rest and work: the ideal indoor soundscape during the Covid-19 pandemic and beyond’. The general objective of the project was to assess the acoustic environment in relation to two main activities performed at home during the pandemic, i.e. relaxing and working from home (WFH), and to link soundscape evaluation with the psychological well-being of participants and a number of acoustical, building, urban, and person-related factors that are known to potentially affect acoustic perception in residential buildings [1]. The study constituted a first application of the indoor soundscape model developed in a previous laboratory investigation [2] for the assessment of the acoustic environment in residential buildings. The model allows to represent the affective responses to the indoor acoustic environment in a two-dimensional model where the main dimension is related to how comfortable or annoying the environment was judged, and therefore noted as comfort. The second dimension is related to the saturation of the environment with sounds or events and is represented by how empty or full of content the acoustic environment is perceived to be, therefore noted as content.
Part I of the study [3] revealed a difference in soundscape evaluation according to the activity carried out at home. Environments were rated as more comfortable and more content-rich when considered for relaxation than for WFH. Moreover, acoustic environments perceived as more appropriate for WFH and for relaxation were characterized by higher comfort scores and lower content scores than those that were assessed as inappropriate. Soundscapes that are appropriate for relaxation were evaluated as comfortable and either full of content (i.e., engaging) or empty (i.e., private and under control), while spaces that are more appropriate to home working were comfortable but also poor in content, i.e., perceived as private and under control. Interestingly, the analysis showed an association between soundscapes and psychological well-being of respondents, evaluated through the WHO-5 well-being index [4]. Psychological well-being was positively associated with comfortable soundscapes both in relation to WFH and relaxation. As regards content, a weak negative correlation was found between content scores and psychological well-being in relation to WFH, but not for relaxation.
But what are the factors underlying psychological well-being and the two perceptual dimensions of indoor soundscapes? Part II of the study addresses this research question by assessing the influence of several acoustical, building, urban and person-related factors on the well-being of building occupants and on soundscape dimensions (i.e., comfort and content) evaluated according to two main activities performed at home during the COVID-19 lockdown. The information derived will allow to gain insights on factors to control through building and urban design for the creation of healthy and supportive acoustic environments at home [5].
2 Methods
An online survey was carried out in January 2021 via Prolific participant pool [6], [7], targeting home workers living in UK (London) and Italy during the COVID-19 lockdown. The following analysis will focus on London as a first case study. The survey involved 464 Londoner respondents and was composed of five main sections addressing: (1) the WFH activity; (2) leisure activities performed at home; (3) housing features; (4) the urban context; and (5) person-related characteristics. The questionnaire included both closed and open-ended questions, that were analyzed through a mixed-method approach in order to increase result validity via methodological triangulation [8]. A detailed description of the study design and of the questions included in the survey (Q1 – Q29) is reported in Part I of the study [3].
In the following, details on quantitative analysis of data from closed-ended questions and qualitative analysis on verbal descriptions from open-ended questions are provided. Statistical analyses were run in IBM SPSS Statistics 26 [9] and in R [10], while qualitative analyses have been conducted in NVivo 12 software.
2.1 Quantitative analysis: Multivariate regression
Multivariate regression was employed as a special case of structural equation modelling (SEM) to investigate patterns of effects within a system of observed variables [11] and to visually display relationships in a path diagram. Please consider that the word “effect” is widely accepted in SEM but should not be taken to indicate claims of causality. We fitted two models, one for each of the two investigated uses (WFH and relaxation). The computation was performed with the lavaan package [12]. Models included the direct effect of acoustic, housing, urban context and person-related variables on comfort and content and on psychological well-being (cf. Fig. 1 ). Based on previous findings on the association between positive soundscapes and enhanced well-being [3], [13], [14], [15], [16], we hypothesize that the affective response to the acoustic environment, expressed by the comfort and content dimensions, and the psychological well-being have common predictors among the investigated variables. Moreover, we tested main and interaction effects between noise sensitivity and perceived dominance of sound sources on comfort, content and well-being. The hypothesis is that the relationship between how loud a sound source is heard and the investigated outcomes (comfort, content and well-being) is modulated by people’s sensitivity to noise.Fig. 1 Path diagrams depicting hypothesized pathways between variables in relation to working (a) and relaxing (b) at home. Rectangles represent measured variables. Single-headed arrows represent a direct effect of one variable on another. Double-headed arrows depict the covariance of residuals of the endogenous variables. Covariances between exogenous variables are not displayed to enhance readability.
Variables included in the path models are described in Appendix A and in Fig. 1. All the regression paths between the exogenous variables (on the left in Fig. 1) and the three endogenous variables (on the right) have been tested. Variables expressed in Likert scales were considered as continuous. In order to reduce the model complexity, nominal variables were in general recoded to reduce the number of categories. Variables on house ownership, house size, and gender were reduced to dichotomous (cf. Appendix A). Two binary variables were derived by combining information on the rooms chosen respectively for WFH and relaxation (Q3, Q9, cf. Appendix A in Part I [3]), and the self-reported quietness of the urban areas outside those rooms (Q18). The variables describe whether the rooms overlooked a quiet or noisy urban area. Another binary variable was extracted from Q19 and is related to the presence of children at home. Due to the lockdown situation, we assume that children experiencing home schooling might have resulted into disrupting and comforting reactions respectively while working and relaxing at home. Variables related to the typology of building services at home were reduced into dichotomous variables specifying the presence of air systems for heating, cooling and ventilation (e.g. air conditioners, mechanical ventilation). While the extended data collected by those questions will be analyzed elsewhere, we included here only the information about air systems as they might provide new source of noise inside buildings. The variable describing the type of urban area (Q25) was dichotomized (0 = suburban, rural; 1 = urban), due to the few occurrences on the “rural” category (N = 1). Housing type variable, having more than two categories, was dummy coded (cf. Appendix A). Responses to “other” option were firstly inspected. When responses could not be included into existing or new categories, “other”, “not applicable” and incongruent responses (e.g. participants giving conflicting information across different questions) were generally treated as missing values and deleted listwise. Exceptions are described in Appendix A. In questions related to the relevance of activities performed while WFH (Q1.1 – Q1.8), “not applicable” and “not at all” responses were collapsed. Similarly, “not applicable” answers to questions about sound dominance from other people at home and from neighbours were treated as “not at all” responses, as the information about people living alone and not having neighbours was already included elsewhere (Q20 and Q17).
Covariances between exogenous variables were modelled to account for correlations (e.g. between the type of urban area and the perceived dominance of certain sound sources). The covariance of residuals for the endogenous variables was included in the model (depicted with the double headed arrow in Fig. 1) and represent a correlation of unexplained variance from the two variables.
As endogenous variables (comfort, content, and well-being) failed to exhibit multivariate normality, a maximum likelihood estimation with robust standard errors and a Satorra-Bentler scaled test statistic (MLM) [17] was utilized for both parameter estimates and goodness-of-fit statistics. Model fit was evaluated through the Robust Comparative Fit Index CFI (≥0.95 for good fit [18]), the robust Root Mean Square Error of Approximation RMSEA (≤0.05 for good fit [18]), the Standardized Root Mean Square Residual SRMR (≤0.80 for good fit [18]), and the relative chi-square (a ratio χ2/df between 2 and 3 is indicative of a good or acceptable fit [19]). The relative chi-square has been preferred, as chi-square is known to be sensitive to large sample sizes, generally above 200 [20]. The statistical significance threshold was set at 0.05.
2.2 Qualitative analysis
Participants’ responses to open-ended questions were analyzed using the method of constant comparison of data [21]. Occurrences within each theme were summed across the participants, and then analyzed through descriptive statistics. Only codes with more than five occurrences have been retained.
The analysis was intended to assess the frequency of negative, neutral, and positive evaluations of specific sound sources that emerged from the analysis of free-format responses to questions Q6 [“In your view, how is the sound environment currently (positively and negatively) affecting your working activity from home? – e.g. heard noises and sounds, building characteristics, urban environment”] and Q12 [“In your view, how is the sound environment currently (positively and negatively) affecting your leisure activities at home? – e.g. heard noises and sounds, building characteristics, urban environment” (While watching TV, reading, listening to music)” – cf. Appendix A, Part I [3]]. While questions Q4 and Q10 showed the perceived dominance of specific sound sources specified by the researcher without an affective evaluation, the qualitative analysis of verbal data allowed to infer judgments for the sound sources that were spontaneously expressed by respondents, as previously done in the soundscape literature [22], [23]. Furthermore, the frequency with which a sound source was mentioned can provide a first clue about potential factors that negatively and positively influence indoor soundscapes.
3 Results
3.1 Evaluation of sound sources from qualitative analysis
Fig. 2 shows the frequency of negative, neutral, and positive evaluations of specific sound sources in relation to the WFH (Fig. 2a) and relaxation (Fig. 2b) activities.Fig. 2 Percentage of negative, neutral and positive evaluations of perceived sound sources that emerged from the analysis of free-format responses to Q6 and Q12, in relation to (a) WFH (N = 505) and (b) relaxation (N = 319).
As regards WFH, results showed that the most frequent negative judgments were associated to noise from people at home, followed by traffic noise, neighbours, construction works, and noise generated by people outside. Positive judgments referred primarily to natural sounds, followed by sounds from music and TV controlled by the respondents themselves.
When assessing the impacts on relaxation, neighbours’ noise featured as the most frequently mentioned source with a negative connotation, followed by traffic noise, and people at home. On the positive side, music and sounds from TV were most frequently reported, followed by natural sounds.
3.2 Path models: Influences of acoustical, building, urban and person-related factors on well-being and on soundscape dimensions
This section addresses the investigation of the influence of person-related variables, building features and variables related to the acoustic and urban contexts on soundscape assessment and well-being. The tested path model was described in Fig. 1.
We hypothesized that noise sensitivity would moderate the relationship between sound source dominance and the investigated outcomes (comfort, content, and well-being). However, as the saliency of different sound sources was not derived by objective measures but by the appraisal made by respondents, we firstly tested whether perceived sound dominance was correlated to the noise sensitivity of respondents. The hypothesis was that people more sensitive to noise would report sound sources as more dominant. The only statistically significant relations were between noise sensitivity and the perceived dominance of sounds from neighbours both while working, rs = 0.186, p < .0005, and relaxing, rs = 0.199, p < .0005, and between noise sensitivity and the perceived dominance of sounds from people outside during relaxation, rs = 0.099, p = .033. Those sound sources, and in particular neighbours’ noise, were thus rated as more dominant by people more sensitive to noise. For other sound sources, the associations with noise sensitivity were not statistically significant.
3.2.1 Working from home model
The model exhibited a good fit, with robust CFI = 0.957, robust RMSEA = 0.047, SMRM = 0.077, and χ2/df = 1.95. Significant regression paths are described in Table 1 and depicted in Fig. 3 . The model explained 29% of the variance in comfort scores, 23% in content, and 24% in well-being.Table 1 Parameter estimates for significant regression paths in the WFH model. Single-headed arrows represent a direct effect of one variable on another. Double-headed arrows depict the covariance of residuals of the endogenous variables, as also indicated in Fig. 3.
Regression and covariance paths Estimate SE p-value St. estimate
Other noise from outside → Comfort −0.199 0.064 0.002 −0.568
(Other noise from outside) × NS → Comfort 0.247 0.099 0.013 0.598
Natural sounds from outside → Comfort 0.128 0.063 0.041 0.361
Music or TV played by you → Comfort 0.073 0.034 0.033 0.280
Room where WFH facing a quiet area → Comfort 0.081 0.026 0.002 0.130
Age → Comfort 0.003 0.002 0.034 0.096
(Sounds from neighbors) × NS → Comfort −0.230 0.097 0.018 −0.603
Frequency of headphone use → Comfort −0.021 0.010 0.045 −0.081
Number of people present at home → Comfort −0.042 0.013 0.002 −0.154
Frequency of headphone use → Content 0.028 0.009 0.003 0.132
Number of people present at home → Content 0.046 0.011 0.000 0.202
Online meetings → Content −0.026 0.010 0.007 −0.121
Thinking/creative thinking → Content 0.025 0.011 0.028 0.108
Urban area → Content 0.059 0.027 0.029 0.105
Gender → Content 0.068 0.023 0.003 0.127
Gender → Well-being −0.035 0.016 0.026 −0.091
Individual focused work away from your desk → Well-being 0.027 0.009 0.003 0.150
Sounds from other human beings present in your house → Well-being −0.068 0.031 0.027 −0.364
Window view: vegetation → Well-being 0.021 0.008 0.007 0.124
Comfort ↔ Content −0.017 0.003 0.000 −0.287
Comfort ↔ Well-being 0.012 0.002 0.000 0.269
Fig. 3 Path diagram depicting significant pathways between variables in relation to working at home. Single-headed arrows represent a direct effect of one variable on another. Standardized regression estimates with their significance level are given for each path. Positive associations are depicted by continuous green, while negative associations are given by red dotted arrows. R2 represents the proportion of variance explained in endogenous variables. Double-headed arrows depict the covariance of residuals of the endogenous variables. Covariances between exogenous variables are not displayed to enhance readability. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Comfort was found to increase with natural sounds, and with music played by respondents themselves. The effect of noise from sirens, industry and construction works on comfort was less straightforward, as it was given by the simple main (negative) effect of the perceived noise and the (positive) interaction term with noise sensitivity. Notably, the impact of those outdoor sounds on comfort was moderated by noise sensitivity, with a higher comfort for people exhibiting high noise sensitivity. Noise sensitivity also moderated the effect of the perceived noise from neighbours on comfort. The higher the noise sensitivity, the more negative the effect of neighbours’ noise on comfort. Higher comfort was associated with a less frequent use of headphones, working in a room facing a quiet area, the presence of fewer people at home, and older respondents.
An increase in content scores was correlated to a more frequent use of headphones, a higher number of people living at home, to living in an urban area (compared to a suburban or rural area), to a lower relevance of online meetings and to a higher importance of creative thinking in daily work. As regards gender, women were more likely to report higher content and lower psychological well-being than men.
Being more engaged with individual focused work away from the desk and seeing vegetation from windows was associated with higher well-being, while hearing more sounds from other people at home provided a negative effect. A residual negative correlation resulted between comfort and content, while a residual positive correlation resulted between comfort and well-being, due to unspecified factors.
3.2.2 Relaxing at home model
The model for relaxation had a good fit, with robust CFI = 0.972, robust RMSEA = 0.057, SMRM = 0.071, and χ2/df = 2.42. Significant regression paths are reported in Table 2 and shown in Fig. 4 . The model explained 32% of the variance in comfort, 22% in content and 21% in well-being. Higher comfort was associated with rooms overlooking quiet areas, with bigger houses (>80 m2), and fewer people present at home. The effect of noise from neighbours on comfort was given by the simple main (positive) effect of the perceived noise and by the negative interaction term with noise sensitivity. Noise sensitivity thus moderated the relationship between perceived neighbourś noise and comfort: the higher the noise sensitivity, the more negative the effect of neighbours’ noise on comfort.Table 2 Parameter estimates for significant regression paths in the relaxation model. Single-headed arrows represent a direct effect of one variable on another. Double-headed arrows depict the covariance of residuals of the endogenous variables, as also indicated in Fig. 4.
Regression and covariance paths Estimate SE p-value St. estimate
Room where relaxing facing a quiet area → Comfort 0.059 0.028 0.035 0.097
House size > 80 m2 → Comfort 0.084 0.027 0.002 0.131
Sounds from neighbors → Comfort 0.151 0.068 0.026 0.404
(Sounds from neighbors) × NS → Comfort −0.290 0.099 0.003 −0.727
Number of people present at home → Comfort −0.035 0.014 0.014 −0.127
Room where relaxing facing a quiet area → Comfort −0.053 0.021 0.012 −0.117
House size > 80 m2 → Content −0.043 0.022 0.050 −0.091
(Sounds from neighbors) × NS → Content 0.170 0.077 0.028 0.578
Number of people present at home → Content 0.045 0.011 0.000 0.223
Mechanical ventilation → Content −0.064 0.032 0.049 −0.095
Housing type: Apartment block → Content 0.074 0.033 0.025 0.155
Gender → Content 0.080 0.020 0.000 0.171
(Sounds from neighbours) × NS → Well-being −0.137 0.059 0.021 −0.561
Number of people present at home → Well-being −0.018 0.009 0.036 −0.108
Gender → Well-being −0.044 0.016 0.007 −0.112
Sounds from human beings from outside →Well-being −0.120 0.039 0.002 −0.512
(Sounds from human beings from outside) × NS 0.153 0.061 0.012 0.607
Music or TV played by you −0.074 0.027 0.007 −0.405
Comfort ↔ Content −0.006 0.003 0.020 −0.116
Comfort ↔ Well-being 0.011 0.002 0.000 0.253
Fig. 4 Path diagram depicting significant pathways between variables in relation to relaxation activities. Single-headed arrows represent a direct effect of one variable on another. Standardized regression estimates with their significance level are given for each path. Positive associations are depicted by continuous green, while negative associations are given by red dotted arrows. R2 represents the proportion of variance explained in endogenous variables. Double-headed arrows depict the covariance of residuals of the endogenous variables. Covariances between exogenous variables are not displayed to enhance readability. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Higher content scores were correlated to rooms facing noisy areas, to smaller houses (<80 m2), apartments (compared to detached houses), absence of mechanical ventilation, presence of more people at home, and female respondents. Noise sensitivity moderated the effect of sounds from neighbours on content: the higher the noise sensitivity, the higher the content scores at increasing dominance of neighbours’ noises.
Results showed that lower psychological well-being was associated with the presence of more people at home, female respondents, TV sounds and music played by the respondents themselves. The effect of sounds from people outside on well-being was given by the simple main (negative) effect of the perceived noise and the (positive) interaction term with noise sensitivity: the higher the noise sensitivity, the more positive the effect of outdoor human sounds on well-being. Noise sensitivity also moderated the relationship between the perceived dominance of neighbours’ noise and well-being, with a more negative effect on well-being for people highly sensitive to noise. A residual negative correlation resulted between comfort and content, while a residual positive correlation resulted between comfort and well-being, likely due to factors not included in the model.
4 Discussion
The study presented the results of an online survey conducted in London with the purpose of exploring the relationships between a number of contextual, building, urban, and person-related variables on psychological well-being and indoor soundscapes, when evaluated according to working and relaxation activities at home. In the following, the effects on soundscape dimensions and well-being are discussed by triangulating the results from rating scales with those from the qualitative analysis of free format responses, also with reference to previous literature and findings from Part I of the study [3].
4.1 Comfort
Higher comfort was associated with working and relaxing in a room overlooking a quiet urban area. Findings are consistent with the existing literature on the beneficial effects of the availability of a quiet side of the dwelling in terms of reduced annoyance, increased health and quality of life [24], [25], [26], [27]. Notably, the present study suggests that people who have access to a quiet side perceive the acoustic environment as more comfortable in relation to both relaxing and working at home, thus extending the previous findings.
As regards the contribution of specific sound sources, listening to natural sounds resulted in improved comfort conditions while WFH (cf. Fig. 3), while the direct effect on comfort during relaxation was not statistically significant (cf. Fig. 4). This was also confirmed by the analysis of open-ended questions, as natural sounds were most often mentioned among the positively perceived sounds in relation to remote working (cf. Fig. 2). If the literature has repeatedly reported on the positive perception of natural sounds and their enhancing effect towards comfort and pleasantness [2], [28], the present study suggests a prevailing effect in relation to WFH compared to relaxation, that might be worth investigating in future research.
Higher comfort while working was correlated to listening to sounds from TV and to music played by the people themselves (cf. Fig. 3). Music and TV sounds have been mentioned most often as pleasant sounds in relation to relaxation and among the most beneficial for working, behind natural sounds (cf. Fig. 4). According to the findings reported in Part I of the study [3], listening to music, watching TV, and wearing noise cancelling headphones allowed occupants to take control over the acoustic environment and to shape their wanted soundscapes against the available ones, thus resulting in improved comfort conditions.
While positively perceived sounds (e.g., natural sounds, music) had a direct effect on comfort, the effect of more disrupting sounds (e.g., construction works, neighbours’ noise, cf. Fig. 2) was generally moderated by noise sensitivity. In the literature, higher annoyance to indoor and outdoor noise sources is generally reported by individuals that are more sensitive to noise [29], [30], [31]. In the present study, an association between noise sensitivity and the perceived dominance of sounds from neighbours and from people outside was observed. Building occupants exhibiting higher noise sensitivity reported neighbours’ noise, both while working and relaxing, and sounds from people outside during relaxation as more dominant. No significant direct effect of noise sensitivity on comfort was observed.
Traffic noise has been frequently mentioned as a disturbing noise source both in relation to home working, behind the effect of other people present in the dwelling, and in relation to relaxation, behind neighbours (cf. Fig. 2). Nevertheless, comfort was not directly affected by the perceived dominance of traffic noise neither in relation to working or relaxation. This is likely due to a drop in traffic flow and related noise levels in London [32], [33], that resulted in a reduction of noise annoyance from outdoor sources during the COVID-19 lockdown compared to the pre-pandemic period [33]. Differently, construction works were still allowed during the lockdown and the noise generated could stand out louder due to the reduced traffic flows [33]. As a result, an increased number of tweets and noise complaints related to construction and building works was revealed in London during the period of confinement [33], [34]. In the present study construction works were mentioned among the sound sources negatively impacting working activities (cf. Fig. 2). From the multivariate model, construction works, industry and sirens had an impact on comfort while WFH through the direct (negative) effect of perceived noise dominance and the (positive) interaction effect with noise sensitivity (cf. Fig. 3). Counterintuitively, all else equal, the negative impact of those noises on comfort was reduced in people exhibiting high noise sensitivity. The effect was not significant in relation to leisure activities, as those are often performed at times when construction and industrial works have ceased, and traffic and sirens are reduced [3].
As regards neighbours’ noise, noise sensitivity was found to moderate the effect of the perceived noise from neighbours on comfort both in relation to WFH and relaxation: the higher the noise sensitivity, the more negative the effect of neighbours’ noise on comfort. Depending on the studies and investigated contexts, neighbours’ noises are reported to be sources of annoyance in residential settings that can be more [35] or less disruptive [36] than outdoor sounds. In a pre-COVID study in London, neighbours’ and outdoor noises perceived inside dwellings were found to be equally annoying [37]. During the lockdown in London, the perceived neighbours’ noise level, the related annoyance and noise complaints significantly increased [33], [34]. The present study provides complementary information on the perceived comfort in relation to the performed activity at home. Noises from neighbours featured as the third most frequently mentioned disruptive sound sources in relation to WFH and the most frequently mentioned when considering relaxation (cf. Fig. 2).
The number of people at home was found to be negatively correlated to comfort both in relation to working and relaxing at home (Fig. 3 and Fig. 4). Notably, people at home were the most frequently mentioned source of disturbance in relation to WFH (cf. Fig. 2). Similarly, a Canadian study reported that noise generated by occupants in the same suite (e.g., roommates and family) was the biggest issue for those working from home [38].
Increased use of headphones while working from home has been shown to correlate to reduced comfort. More frequent use of headphones can be considered as a proxy for poor acoustic conditions. While this variable was included in the survey only in the WFH section, the analysis of open-ended questions revealed that headphones (including noise-cancelling headphones) were often employed to cope with unfavourable acoustic conditions both while working and relaxing at home [3].
House size resulted a relevant variable in defining comfort during relaxation. Living in larger dwellings (>80 m2) was associated with greater comfort during relaxation, whereas the effect was not statistically significant when considered for working at home.
As regards age, in the present study higher comfort scores were associated with older respondents but only in relation to WFH. Other studies reported higher annoyance by older people (e.g. [36]) and it must be noticed that no clear trend can be found in the literature regarding the effect of socio-demographic parameters, such as age, on comfort and noise annoyance [30].
4.2 Content
Studies on urban soundscapes have been recently investigating the impact of several acoustical and non-acoustical factors on the two main perceptual dimensions that define the affective response to the outdoor acoustic environment, i.e. pleasantness and eventfulness [28], [39]. Differently, previous research in indoor built environments has traditionally dealt with the impact of several factors on valence-related constructs only (i.e., acoustic comfort, satisfaction, and noise annoyance). The present study constitutes the first attempt to explore factors associated with content, that is the eventfulness-equivalent dimension for indoor residential settings [2].
Relaxing in rooms overlooking a quiet urban area was related to lower content (cf. Fig. 4), while the effect was not significant in relation to home working (cf. Fig. 3). As regards the sound source typology, the effect of neighbours’ noise on content during relaxation was moderated by noise sensitivity, with higher dominance of neighbours’ noises resulting in higher content for respondents that were more sensitive to noise.
More frequent use of headphones while working was correlated to increased content scores. This is likely due to a coping mechanism in presence of soundscapes rich in content that were generally judged inappropriate to WFH [3]. Interestingly, the type of task at hand influenced the perceived soundscape content. Being more engaged in online meetings was associated with a lower perceived content, whereas creative thinking was correlated to increased content. The lower sensitivity to content during online meetings and the higher sensitivity to content during creative thinking might be linked to different cognitive functions and listening modes involved. Given the exploratory nature of the study, future research might help provide further groundwork.
The higher the number of people in the dwelling, the higher the content due to the increased amount of human activities at home. Moreover, content was explained by several building and urban-related features. Living in larger houses (>80 m2) resulted in reduced content scores during relaxation (cf. Fig. 4), likely because activities are “diluted” over a larger space. Compared to detached houses, flats were associated with higher content during relaxation (cf. Fig. 4), and dwellings in urban areas were associated with increased content when WFH compared to suburban and rural areas (cf. Fig. 3). This might be due to the fact that in apartments and urban contexts dwellings are more saturated with outdoor and neighbours’ sounds, thus resulting in increased content values. Moreover, houses equipped with mechanical ventilation were related to lower content during relaxation. In the absence of mechanical ventilation, opening windows to ventilate the dwellings (i.e., natural ventilation) can result in increased content values due to the access of outdoor sounds.
As regards demographic parameters, only gender was associated with content, with higher values reported by female respondents both in relation to home working and relaxation.
4.3 Connecting comfort and content results: privacy, control, and engagement
By intersecting the factors that influence the two main perceptual dimensions, comfort and content, it was possible to derive the variables that contribute to describe the alternative dimensions related to privacy, control and engagement [2]. The result is conceptually depicted in Fig. 5 , where acoustical, building, and urban-related features are reported in relation to soundscape evaluation during working and relaxation. Variables involving both main and interaction effects have not been included as their interpretation would not be straightforward and suitable to this simplified representation.Fig. 5 Conceptual representation of acoustical, building and urban-related variables contributing to privacy, control and engagement dimensions in relation to (a) WFH and (b) relaxation.
By way of example, soundscapes perceived as private and under control, corresponding to high comfort and low content scores, were perceived as more appropriate to home working [3]. As depicted in Fig. 5 a, soundscapes characterized by increased privacy and perceived control can be thought as a combination of the following factors generating either greater comfort or lower content: the presence of more natural sounds, music and sounds from TV played by the respondents themselves, lower neighbours’ noise and/or by people less sensitive to noise, less frequent use of headphones, fewer people present in the dwelling, houses mainly located in a suburban or rural area and facing a quiet area outdoor.
When evaluated in relation to leisure activities, a private and controlled soundscape was more likely found in spaces facing a quiet outdoor area, having lower neighbours’ noise and/or people less sensitive to noise, and could be more frequently found in larger, detached houses, equipped with mechanical ventilation and with a lower number of people at home (Fig. 5 b).
Differently, acoustic environments characterized by a combination of natural sounds, music and TV sounds, with lower neighbours’ noise and/or by people less sensitive to noise, within rooms located in urban areas and having a quiet side were conductive to engaging soundscapes for WFH. Flats (compared to detached houses) and rooms in which windows are opened for ventilation are highly associated with higher content. If coupled with factors providing higher comfort this can result in soundscapes that are engaging during relaxation, thus confirming the potential effect of natural ventilation to provide positive soundscapes as previously suggested [40].
4.4 Well-being
The results showed an association between the perceived dominance of sounds from specific sources and building occupants’ psychological well-being. Noise sensitivity was found to moderate the relationship between well-being and the perceived dominance of outdoor human sounds and neighbours’ noises during relaxation (cf. Fig. 4). The effect of outdoor human sounds on well-being was given by the sum of the simple main negative effect of the perceived sounds and the positive interaction effect with noise sensitivity, meaning a more positive effect of outdoor human sounds on well-being for people that were more sensitive to noise. Outdoor sounds might contribute in some cases to an enhanced well-being by creating a connection with the outdoor environment and alleviating the sense of loneliness, especially during the confinement period, as reported by participants [3]. As regards neighbours’ noises during relaxation, a more negative effect on well-being was observed in people that were more sensitive to noise. The exposure to neighbours’ noise can result in annoyance and in a lack of privacy, with a detrimental effects on health and well-being of building occupants [41], [42], [43], [44]. The perceived dominance of TV sounds and music played by the respondents themselves during relaxation were found to be negatively correlated with well-being. This might be a proxy for an excessively noisy acoustic environment, as people in some cases reported about the need to turn up the volume of TV or playback devices in order to overpower the existing background noise. Previous literature reported some evidence on the relationship between the exposure to indoor noise pollution in living environments and depressive moods [45]. Lower well-being was correlated to higher dominance of sounds from other human beings present at home while WFH. The intrusion of noises from other people at home while working could induce negative mental states, such as reduced comfort, that might increase the risks of mental health issues. In relation to relaxation, the number of people present in the dwelling was negatively associated with participants’ well-being. The effect of crowding on depression, poor mental health and social well-being has been frequently reported in the literature [46], [47], and a study on children suggested that the noise generated at home might be one of the mechanisms underpinning the impact of crowding on well-being [48]. The number of people at home and the dominance of neighbours’ noises while relaxing, moderated by noise sensitivity, were common factors negatively affecting both comfort and well-being and this can partially explain the association between uncomfortable conditions and poor well-being.
Among the investigated working modes, performing individual focused work away from the desk was associated with improved well-being. This might be explained by reduced cognitive loads or by a higher flexibility provided by those activities (e.g. allowing to choose a more appropriate space where to work), but dedicated investigations would be needed before conclusions could be drawn.
The view of vegetation from a window where WFH was positively correlated to the mental well-being of respondents (cf. Fig. 3). This is in line with the findings of previous literature, showing an impact of poor-quality views on depressive symptoms during the COVID-19 lockdown [49]. On the contrary, windows with access to nature can elicit positive emotions, recover from stressful experiences, restore attentional capacity after cognitive fatigue, and improve well-being [50], [51], [52].
As regards gender effect, lower psychological well-being was associated with female participants.
5 Conclusions
In this second part of the study, results of an online survey conducted on 464 home workers in London in January 2021 during the COVID-19 lockdown have been analyzed in order to explore the influences of several acoustical, building, urban and person-related factors on occupants’ well-being and on soundscape dimensions (i.e., comfort and content), according to two activities performed at home during the pandemic, i.e., working and relaxing.
The analysis of data collected from open-ended questions revealed that, as regards WFH, the most frequent negative judgments were associated to noise from people at home, whereas positive judgments referred primarily to natural sounds. When considering relaxation, neighbours’ noise featured as the most frequently mentioned source with a negative connotation, while, on the positive side, music and sounds from TV were most frequently reported. Higher comfort scores while WFH were related to a higher dominance of natural sounds, music and TV sounds, lower neighbours’ noises and/or low sensitivity to noise, working in a room overlooking a quiet area, a less frequent use of headphones, a lower number of people at home, and being older. With reference to relaxation, increased comfort was correlated to the availability of a quiet side, a lower number of people at home, and a larger dwelling (>80 m2). Higher content while WFH was associated with a more frequent use of headphones, a lower relevance of online meetings and a higher relevance of creative thinking, a higher number of people at home, being female, and living in an urban area. As regards relaxation, higher perceived dominance of neighbours’ noises and/or high sensitivity to noise, sharing the house with more people, the absence of mechanical ventilation, living in a flat and being female was associated with higher content scores. Finally, increased well-being resulted in those performing individual focused work away from the desk, male respondents, people less exposed to sounds from other people at home while working, having access to vegetation through the window while WFH, lower dominance of neighbours’ noises while relaxing and/or lower noise sensitivity, a lower dominance of music and TV sounds while relaxing, and a lower number of people at home. The dominance of neighbours’ noises during relaxation, moderated by noise sensitivity, and the number of people at home were common factors negatively affecting both comfort and well-being, and this might partially explain the association between acoustic comfort and positive mental states.
Results pointed out the importance of the availability of a quiet side of the dwelling and the detrimental effect of crowding and neighbours’ noise on comfort. Content was related to the saturation of the environment with indoor (e.g., people at home, neighbours) and outdoor sounds, and on space availability depending on housing features. Different sound typologies were found to influence comfort and content, and such relation was often moderated by building occupants’ noise sensitivity.
Overall, the study confirmed the importance of considering the different impacts that acoustical, building, urban and person-related factors can provide on building occupants depending on the specific activity people are engaged with at home. The lack of data in the pre-pandemic period does not allow to determine at this stage the impacts, if any, in the observed associations due to the psychological status of participants in this emergency period. Therefore, the associations highlighted in the present study might be further assessed in future longitudinal studies. The study highlighted an association between positively perceived soundscapes and psychological well-being that could only be partially explained by the variables included in the present study but that reinforce the role of building, urban and acoustic design to promote healthy conditions for their occupants. Future investigations might help to explain the direction and the mechanisms linking comfort to psychological well-being (or vice versa). Lastly, the future analysis of the Italian dataset will help generalize the results presented in the present paper also considering possible cultural differences [53], thus informing about actions to be applied at a broad scale in the post-pandemic design of healthy buildings.
CRediT authorship contribution statement
Simone Torresin: Conceptualization, Methodology, Formal analysis, Investigation, Writing – original draft, Visualization, Writing - review & editing. Rossano Albatici: Conceptualization, Writing - review & editing. Francesco Aletta: Conceptualization, Methodology, Writing - review & editing. Francesco Babich: Conceptualization, Writing - review & editing. Tin Oberman: Conceptualization, Writing - review & editing. Agnieszka Elzbieta Stawinoga: Formal analysis, Writing - review & editing. Jian Kang: Conceptualization, Supervision, Funding acquisition, Writing - review & editing.
Declaration of Competing Interest
The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Jian Kang reports financial support was provided by Chartered Institution of Building Services Engineers.
Appendix A – Variables
Summary of variables included in the path models. The table presents reference questions for the variables (cf. Part I of the study), a short description, variable treatment in lavaan, missing values treatment, and the model in which the variables are included (WFH stands for working from home, REL stands for relaxation).Reference questions Variable description Type of variable Missing values, “other”, “not applicable” responses Model
Q1.1 – Q1.8 9 variables on the relevance of different activities to WFH Continuous “Not applicable” responses treated as “not at all” WFH
Q2 Frequency of headphone use Continuous None WFH
Q4.1 – Q4.9 9 variables on the perceived dominance of different sound sources Continuous “Not applicable” responses removed listwise.
Only for Q4.5 – Q4.6:
“Not applicable” responses treated as “not at all” WFH
Q5.1 – Q5.3 3 variables on the perceived dominance of components in window view Continuous “Not applicable” responses removed listwise WFH
Q8.1 – Q8.8 Comfort and content scores Continuous None WFH
Q10.1 – Q10.9 9 variables on the perceived dominance of different sound sources Continuous “Not applicable” responses removed listwise.
Only for Q10.5 – Q10.6:
“Not applicable” responses treated as “not at all” REL
Q11.1 – Q11.3 3 variables on the perceived dominance of components in window view Continuous “Not applicable” responses removed listwise REL
Q14.1 – Q14.8 Comfort and content scores Continuous None REL
Q15 House ownership Dichotomous (0 = Rent – not owned, 1 = Owned) “Other” responses removed listwise WFH / REL
Q16 House size Dichotomous (0 = ≤ 80 m2; 1 = > 80 m2) None WFH / REL
Q17 Housing type Dummy coded: Semi-detached or terraced house; Apartment block [ref. Detached single family] None WFH / REL
Q3, Q18 Quietness of area outside the room where WFH Dichotomous (0 = noisy; 1 = quiet) Missing values removed listwise WFH
Q9, Q18 Quietness of area outside the room where WFH Dichotomous (0 = noisy; 1 = quiet) Missing values removed listwise REL
Q19 Children presence at home Dichotomous (0 = no children present; 1 = children present) None WFH / REL
Q20 Number of people present at home Continuous None WFH / REL
Q21 Mechanical ventilation Dichotomous (0 = no mechanical ventilation; 1 = mechanical ventilation) None WFH / REL
Q22 Air system for heating Dichotomous (0 = no air system; 1 = air system for heating) None WFH / REL
Q23 Air system for cooling Dichotomous (0 = no air system; 1 = air system for cooling) None WFH / REL
Q25 Type of urban area Dichotomous (0 = suburban, rural; 1 = urban) None WFH / REL
Q26 Noise sensitivity Continuous None WFH / REL
Q27 Well-being Continuous None WFH / REL
Q28 Age Continuous None WFH / REL
Q29 Gender Dichotomous (0 = male; 1 = female) “Other” responses removed listwise WFH / REL
Acknowledgments
The authors would like to thank prof. Stefano Siboni (DICAM, University of Trento) for his advice on the statistical analysis. The authors thank the Department of Innovation, Research and University of the Autonomous Province of Bozen/Bolzano for covering the Open Access publication costs.
Funding
This work was funded by the Chartered Institution of Building Services Engineers (CIBSE) within the project ‘Home as a place of rest and work: the ideal indoor soundscape during the Covid-19 pandemic and beyond’. This work was supported by the Programma di cooperazione Interreg V-A Italia-Svizzera 2014–2020, project QAES [ID no. 613474]; and the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme [grant agreement No. 740696].
==== Refs
References
1 Torresin S. Albatici R. Aletta F. Babich F. Kang J. Assessment methods and factors determining positive indoor soundscapes in residential buildings: a systematic review Sustain 11 2019 5290 10.3390/su11195290
2 Torresin S. Albatici R. Aletta F. Babich F. Oberman T. Siboni S. Indoor soundscape assessment: a principal components model of acoustic perception in residential buildings Build Environ 182 2020 107152 10.1016/j.buildenv.2020.107152
3 Torresin S. Albatici R. Aletta F. Babich F. Oberman T. Stawinoga A.E. Indoor soundscapes at home during the COVID-19 lockdown in London – Part I: associations between the perception of the acoustic environment, occupantś activity and well-being Appl Acoust 183 2021 10.1016/j.apacoust.2021.108305
4 Topp C.W. Østergaard S.D. Søndergaard S. Bech P. The WHO-5 well-being index: a systematic review of the literature Psychother Psychosom 84 3 2015 167 176 10.1159/000376585 25831962
5 Torresin S. Aletta F. Babich F. Bourdeau E. Acoustics for supportive and healthy buildings: emerging themes on indoor sustainability Sustain 6054 2020 10.3390/su12156054
6 Peer E. Brandimarte L. Samat S. Acquisti A. Beyond the Turk: Alternative platforms for crowdsourcing behavioral research J Exp Soc Psychol 70 2017 153 163 10.1016/j.jesp.2017.01.006
7 Palan S. Schitter C. Prolific.ac—A subject pool for online experiments J Behav Exp Financ 17 2018 22 27 10.1016/j.jbef.2017.12.004
8 Fiebig A. Soundscape standardization dares the impossible - Case studies valuing current soundscape standards. ICA, 23rd Int Congr Acoust 2019:8p.
9 IBM Corp. IBM SPSS Statistics for Windows, Version 26.0 2019.
10 Team RC, others. R: A language and environment for statistical computing 2013.
11 Duncan O.D. Path analysis: Sociological examples Causal Model Soc Sci 72 2017 55 80 10.4324/9781315081663
12 Rosseel Y. lavaan: An R Package for Structural Equation Modeling J Stat Softw 48 2012 1 36
13 Aletta F. Oberman T. Mitchell A. Erfanian M. Lionello M. Kachlicka M. Associations between soundscape experience and self-reported wellbeing in open public urban spaces: a field study Lancet 394 2019 S17 10.1016/S0140-6736(19)32814-4
14 Aletta F, Molinero L, Astolfi A, Di Blasio S, Shtrepi L, Oberman T, et al. Exploring associations between soundscape assessment, perceived safety and well-being: A pilot field study in Granary Square, London. Proc Int Congr Acoust 2019;2019-Septe:7946–53. doi: 10.18154/RWTH-CONV-238876.
15 Aletta F. Oberman T. Kang J. Associations between positive health-related effects and soundscapes perceptual constructs: a systematic review Int J Environ Res Public Health 15 2018 2392 10.3390/ijerph15112392 30380601
16 Erfanian M. Mitchell A.J. Kang J. Aletta F. The psychophysiological implications of soundscape: a systematic review of empirical literature and a research agenda Int J Environ Res Public Health 16 19 2019 3533 10.3390/ijerph16193533 31546577
17 Satorra A, Bentler PM. Corrections to test statistics and standard errors in covariance structure analysis. 1994.
18 Kline R.B. Principles and Practices of Structural Equation Modelling 2015 Guiford Press Fourth. New York
19 Schermelleh-Engel K. Moosbrugger H. Müller H. Evaluating the fit of structural equation models: Tests of significance and descriptive goodness-of-fit measures MPR-Online 8 2003 23 74
20 Schumacker R.E. Lomax R.G. A Beginner’s Guide to Structural Equation Modeling 2004 psychology press
21 Hallberg L.-M. The “core category” of grounded theory: Making constant comparisons Int J Qual Stud Health Well-Being 1 3 2006 141 148 10.1080/17482620600858399
22 Guastavino C. The ideal urban soundscape: Investigating the sound quality of French cities Acta Acust United with Acust 92 2006 945 951
23 Tarlao C. Fernandez P. Frissen I. Guastavino C. Influence of sound level on diners’ perceptions and behavior in a montreal restaurant Appl Acoust 174 2021 107772 10.1016/j.apacoust.2020.107772
24 Van Kamp I, Klaeboe R, Brown AL, Lercher P. Soundscapes, human restoration and quality of life. In: Kang J, Schulte-Fortkamp B, editors. Soundscape Built Environ., Boca Raton, FL, USA: CRC Press; 2016, p. 43–68. https://doi.org/10.1201/b19145-4.
25 Kluizenaar Y.d. Salomons E.M. Janssen S.A. van Lenthe F.J. Vos H. Zhou H. Urban road traffic noise and annoyance: the effect of a quiet façade J Acoust Soc Am 130 4 2011 1936 1942 21973348
26 de Kluizenaar Y. Janssen S. Vos H. Salomons E. Zhou H. van den Berg F. Road traffic noise and annoyance: a quantification of the effect of quiet side exposure at dwellings Int J Environ Res Public Health 10 6 2013 2258 2270 10.3390/ijerph10062258 23736655
27 Van Renterghem T. Botteldooren D. Focused study on the quiet side effect in dwellings highly exposed to road traffic noise Int J Environ Res Public Health 9 2012 4292 4310 10.3390/ijerph9124292 23330222
28 Axelsson Ö. Nilsson M.E. Berglund B. A principal components model of soundscape perception J Acoust Soc Am 128 5 2010 2836 2846 10.1121/1.3493436 21110579
29 Job R.F.S. Noise sensitivity as a factor influencing human reaction to noise Noise Heal 1 1999 57
30 Marquis-Favre C. Premat E. Aubrédue D. Noise and its effects - A review on qualitative aspects of sound. Part II: Noise and annoyance Acta Acust United with Acust 91 2005 626 642
31 Ryu J.K. Jeon J.Y. Influence of noise sensitivity on annoyance of indoor and outdoor noises in residential buildings Appl Acoust 72 6 2011 336 340 10.1016/j.apacoust.2010.12.005
32 Aletta F. Oberman T. Mitchell A. Tong H. Kang J. Assessing the changing urban sound environment during the COVID-19 lockdown period using short-term acoustic measurements Noise Mapp 7 2020 1 12 10.1515/noise-2020-0011
33 Lee P.J. Jeong J.H. Attitudes towards outdoor and neighbour noise during the COVID-19 lockdown: a case study in London Sustain Cities Soc 67 2021 102768 10.1016/j.scs.2021.102768 33585168
34 Tong H. Aletta F. Mitchell A. Oberman T. Increases in noise complaints during the COVID-19 lockdown in Spring 2020: a case study in Greater London, UK Sci Total Environ 2021
35 Jeon J.Y. Ryu J.K. Lee P.J. A quantification model of overall dissatisfaction with indoor noise environment in residential buildings Appl Acoust 71 10 2010 914 921 10.1016/j.apacoust.2010.06.001
36 Andargie M.S. Touchie M. O’Brien W. A survey of factors the impact noise exposure and acoustic comfort in multi-unit residential buildings Can Acoust 48 2020 25 41
37 Lee P.J. Hopkins C. Penedo R. Attitudes to noise inside dwellings in three megacities: Seoul, London, and São Paulo Int J Environ Res Public Health 17 2020 1 27 10.3390/ijerph17166005
38 Andargie MS, Touchie M, O’Brien W. Case study: A survey of perceived noise in Canadian multi-unit residential buildings to study long-term implications for widespread teleworking. Build Acoust 2021:1351010X2199374. doi: 10.1177/1351010x21993742.
39 ISO TS 12913-3:2019 – Acoustics - Soundscape part 3: Data analysis 2019.
40 Torresin S. Albatici R. Aletta F. Babich F. Oberman T. Kang J. Acoustic design criteria in naturally ventilated residential buildings: New research perspectives by applying the indoor soundscape approach Appl Sci 9 2019 1 25 10.3390/app9245401
41 Jensen H.A.R. Rasmussen B. Ekholm O. Neighbour and traffic noise annoyance: a nationwide study of associated mental health and perceived stress Eur J Public Health 28 2018 1050 1055 10.1093/eurpub/cky091 29846583
42 Maschke C. Niemann H. Health effects of annoyance induced by neighbour noise Noise Control Eng J 55 2007 348 356 10.3397/1.2741308
43 Park S.H. Lee P.J. Jeong J.H. Steinborn M.B. Emotions evoked by exposure to footstep noise in residential buildings PLoS ONE 13 8 2018 e0202058 10.1371/journal.pone.0202058 30102721
44 Jensen H.A.R. Rasmussen B. Ekholm O. Neighbour noise annoyance is associated with various mental and physical health symptoms: results from a nationwide study among individuals living in multi-storey housing BMC Public Health 19 2019 1 10 10.1186/s12889-019-7893-8 30606151
45 Rautio N. Filatova S. Lehtiniemi H. Miettunen J. Living environment and its relationship to depressive mood: A systematic review Int J Soc Psychiatry 64 1 2018 92 103 10.1177/0020764017744582 29212385
46 World Health Organization. WHO Housing and health guidelines. 2018.
47 Bonnefoy X. Inadequate housing and health: an overview Int. J. Environ. Pollut. 30 2007 411 429 10.1504/IJEP.2007.014819
48 Solari C.D. RoD M. Housing crowding effects on children’s wellbeing Soc Sci Res 41 2012 464 476 10.1016/j.ssresearch.2011.09.012.HOUSING 23017764
49 Amerio A. Brambilla A. Morganti A. Aguglia A. Bianchi D. Santi F. Covid-19 lockdown: housing built environment’s effects on mental health Int J Environ Res Public Health 17 16 2020 5973 10.3390/ijerph17165973 32824594
50 Kaplan S. The restorative benefits of nature: toward an integrative framework J Environ Psychol 15 3 1995 169 182 10.1016/0272-4944(95)90001-2
51 Kaplan R. The nature of the view from home: psychological benefits. vol. 33. 2001. doi: 10.1177/00139160121973115.
52 Grinde B. Patil G.G. Biophilia: Does visual contact with nature impact on health and well-being? Int J Environ Res Public Health 6 2009 2332 2343 10.3390/ijerph6092332 19826546
53 Mohamed M.A.E. Dokmeci Yorukoglu P.N. Indoor soundscape perception in residential spaces: a cross-cultural analysis in Ankara, Turkey Build Acoust 27 1 2020 35 46 10.1177/1351010X19885030
| 0 | PMC9746876 | NO-CC CODE | 2022-12-15 23:21:58 | no | Appl Acoust. 2022 Jan 1; 185:108379 | utf-8 | Appl Acoust | 2,021 | 10.1016/j.apacoust.2021.108379 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00155-5
10.1016/j.biocon.2021.109103
109103
Policy Analysis
Reduced human activity in shallow reefs during the COVID-19 pandemic increases fish evenness
China Victor ab⁎
Zvuloni Assaf c
Roll Uri d
Belmaker Jonathan ef
a The Jacob Blaustein Center for Scientific Cooperation, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Midreshet Ben-Gurion 8499000, Israel
b Interuniversity Institute for Marine Sciences, Eilat, Israel
c Israel Nature & Parks Authority, Eilat, Israel
d Mitrani Department of Desert Ecology, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Midreshet Ben-Gurion 8499000, Israel
e School of Zoology, Tel Aviv University, Tel Aviv 6997801, Israel
f Steinhardt Museum of Natural History, Tel Aviv University, Tel Aviv 6997801, Israel
⁎ Corresponding author at: The Jacob Blaustein Center for Scientific Cooperation, The Jacob Blaustein Institutes for Desert Research, Ben-Gurion University of the Negev, Midreshet Ben-Gurion 8499000, Israel.
1 4 2021
5 2021
1 4 2021
257 109103109103
20 9 2020
8 3 2021
24 3 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
The COVID-19 pandemic provides a rare opportunity to examine effects of people on natural systems and processes. Here, we collected fish diversity data from coral reefs at the Israeli Gulf of Aqaba during and after the COVID-19 lockdown. We examined beach entrances to the reef, nearby shallow reefs and deeper areas exposed mostly to divers. We found that the lockdown elicited a behavioral response that resulted in elevated species richness at designated reef entrances, predominantly influenced by increased evenness without changes to total abundances. This effect was observed both at the local scale and when several beach entrances were aggregated together. Consequently, non-extractive human activities may have substantial short-term impacts on fish diversity. Our insights could help designate guidelines to manage visitor impacts on coral reefs and aid in their prolonged persistence.
Keywords
Anthropocene
Coral reefs
Fishes
Divers
Human disturbance
Human-nature interactions
==== Body
pmc1 Introduction
Measures to control the spread of the COVID-19 pandemic precluded, or significantly reduced, human activity in nature (Rice et al., 2020). These circumstances offered a unique opportunity to test how organisms react to the absence of humans (Bates et al., 2020; Rutz et al., 2020). Several studies provide examples of reduced human pressures on natural ecosystems, cleaner air, and cleaner water (Corlett et al., 2020; Zambrano-Monserrate et al., 2020). Covid-19 lockdowns and travel restrictions have been noted to cause changes to animal behavior, with some animals reclaiming areas that they usually have been precluded from or becoming more diurnal (Manenti et al., 2020; Vardi et al., 2021; Zellmer et al., 2020).
Coral reefs are some of the most diverse ecosystems globally and face multiple threats, including climate change, overfishing, pollution, and physical destruction (Hughes et al., 2017; Munday, 2004; Riegl et al., 2009). Extractive human activities have large effects on fish abundance, diversity and evenness, as seen by the change in fish communities with MPAs compared to fished areas (Blowes et al., 2020; Claudet et al., 2006). Non-extractive recreational activities in coral reefs, such as swimming, snorkeling or scuba diving, promote human-nature interactions, but may also adversely affect biodiversity (Davenport and Davenport, 2006; Hawkins and Roberts, 1993; Zakai and Chadwick-Furman, 2002). To date, effects of such activities have predominantly focused on corals, and research on the impacts of swimmers and divers on fishes are uncommon. Nevertheless, Medeiros et al. (2007) compared two coastal reefs and found indications that recreational activities lead to lower fish assemblage evenness, driven by changes in the abundance of a single species (Abudefduf saxatilis). Other effects of recreational activities on fishes include reduced cleaning by cleaner (Titus et al., 2015); shorter latency periods and escape distances (Valerio et al., 2019); increased use of refuge during presence of divers and snorkelers (Benevides et al., 2019). On the other hand, fish feeding based tourism has be shown to boost fish diversity (Brunnschweiler and Earle, 2006; Feitosa et al., 2012) and to expand trophic niches (Drew and McKeon, 2019). These human-induced changes may have cascading effects on other fish species (Milazzo et al., 2006). Nevertheless, it remains unclear whether recreational activity impacts are chronic or acute and whether they remain local or span spatial scales.
Species richness is one of the most widely-used biodiversity metrics, but is dependent on several distinct mechanisms (McGlinn et al., 2019). To better understand variation in species richness and the mechanisms contributing to it, it should be decomposed to (1) the size of the species pool, (2) the relative abundance of species, with more even communities displaying higher richness, (3) the number of individuals, with denser communities showing higher richness, (4) intraspecific aggregations, with aggregations decreasing richness within a given scale. Such decomposed richness could produce new insights on the effect of humans on fish biodiversity.
COVID-19 lockdowns may help elucidate short- and long-term impacts of humans on coral reefs (Giglio et al., 2020). For example, if human activity causes long-term damage to the habitat, we can expect local reductions in fish density within impacted sites, which is likely to result in lower richness (Wilson et al., 2008). This reduction in diversity will remain during short-term cessation of human activity such as during COVID-19 lockdowns. However, if human activity causes short-term behavioral changes (Albuquerque et al., 2014), we could see changes to richness during COVID-19 lockdowns due to changes in spatial aggregation or evenness of fishes that are either attracted to or repelled by humans. Both long-term and short-term effects may either accumulate across spatial scales or mostly manifest at the local scale.
Here, we tracked changes to coral reef fish diversity across spatial scales from the local site level to the landscape level (5 km stretch of the coral reef reserve in Israel) following the cessation of human activity because of a COVID-19 lockdown. We measured fish diversity at three different coral reef habitats: (1) shallow areas designated as ‘entrances’ to the reefs with high levels of activity by swimmers, snorkelers, and divers; (2) shallow areas ‘near-entrances’ with low levels of human activity, and (3) 3–6 m deep ‘knolls’ which are mostly visited by SCUBA divers and are also exposed to low levels of human activity. Comparing changes to fish richness during and after a lockdown allowed us to understand the mechanisms underpinning potential human impacts on coral reef fishes.
2 Methods
2.1 Study site
We sampled three different coral reef habitats (1) ‘entrances’, (2) ‘near-entrances’, and (3) ‘knolls’ at nine sites across a ~5 km fringing reef along the northern Gulf of Aqaba, in Eilat, Israel (Fig. 1 ; Note, that not all habitats were available at each site). Due to ease of access, coral reefs in Eilat are among the most heavily visited with ~6 million tourists annually with >300,000 dives per year along a 12 km coastline (Tynyakov et al., 2017). The entire reef area sampled is within a well-enforced marine nature reserve with no fishing. While coral reefs are easily accessible from the shore, entrance to the sea is only allowed at specific marked entrances 4-8 m wide delineated by ropes (there are a total of 19 entrance points along the beach of the nature reserve).Fig. 1 Map of study area and sampling scheme. Each point represents a site sampled; colors represent habitat type. Insert shows the study area within the larger region. Representative underwater photos of each habitat type are to the left of study area map. Detailed sampling scheme at the bottom of the figure depicts which sites and habitats were sampled during each of the three sampling sessions.
Fig. 1
Post lockdown human activity varied between the different habitats we sampled. The highest activity levels were measured at ‘entrances’ where swimmers, snorkelers and divers enter and exit the water, and bathers are usually abundant. At the ‘near entrances’ swimmers, snorkelers and divers can swim above the coral reef, but are not allowed to stand or make other contact with the reef. ‘knolls’ are deeper and access to them is limited mostly to divers, (see Fig. S1 for objective measurements of human activities in each habitat type). We found no significant difference in human activity levels post-lockdown between different entrance sites (Tukey-HSD adjusted p-value>0.074).
2.2 Sampling periods during and after the first COVID-19 lockdown
The first COVID-19 lockdown in Israel (19/03/2020–20/5/2020) was very tight and access to beaches or entering to the sea was prohibited and strictly enforced (we found zero human activity in our lockdown samples; Fig. S1). Touristic activity resumed after the lockdown to levels similar to those measured during 2019 and peaked during July–August (Fig. S2). Beaches were relatively crowded, and diving activity resumed (until a second lockdown started during September 2020). We conducted three sampling sessions. In the first two sessions (during and post lockdown) we tested the effect of the lockdown on ‘entrance’ and ‘knoll’ habitats. The third sampling session was added to test for possible temporal changes and test potential movements of fish to adjacent habitats by also examining the ‘near-entrances’ habitat (see Table 1 and Fig. 1 for all comparisons made and the spatial distribution of sites).Table 1 Comparisons made in this study, their corresponding hypotheses, summary of their main findings, and significance.
Table 1Comparison between Hypothesis tested Summary of findings Significance
‘Knolls’ during (14 April–5 July) and after lockdown (16 June −17 July) Caseation of human activity will result in increased species richness We found no change in species richness Lesser visited and deep habitats do not show short term effects of cessation of activity on diversity
‘Entrances’ during (14 April–5 July) and after lockdown (16 June −17 July) Caseation of human activity will result in increased species richness Species richness during lockdown was higher due to increased evenness Local changes in human activity manifest in changes to diversity
‘Entrances’ and ‘near entrances’ post lockdown (8–27 august) Reduced human activity ‘near entrances’ relative to ‘entrances’ will result in increased species richness Species richness in near entrances was higher due to increased evenness Local changes in human activity manifest in changes to diversity
‘Entrances’ during lockdown (14 April–5 July)and ‘near entrances’ post lockdown (8–27 august) Habitats with similar low human activity levels will have similar species richness We found no differences in species richness Low levels of human activity show similar effects on diversity to complete cessation of human activity
2.3 Sampling method
Data were collected via video surveys. GoPro-8 cameras were deployed randomly within the pre-selected sites to maximize spatial coverage along the nature reserve (see Fig. 1 for sampling design). Cameras were placed without baits in natural light between 11:00 and 14:00 each day and recordings were activated for 45 min (at 4K/24P). From each video we recorded fish species identity and abundance (using “maxN”: maximum number of individuals per frame for a given species) (Bacheler and Shertzer, 2015; Campbell et al., 2015). We analyzed 20 min from the middle of each recording (minutes 10–29) to minimize the effect of deployment and collection. A total of 5144 individuals belonging to 111 different fish species were recorded (mean 17.75 species and 58.45 individuals per video). We further quantified human activity levels from these videos (for details see “Human activity” in Supplement; Fig. S2).
2.4 Analyses
The different comparisons we made are summarized in Table 1. For each comparison, we first analyzed changes to local species richness (α-diversity), total species richness across all sites (γ-diversity) and species turnover between sites (β-diversity, the ratio of total to local richness). We then examined changes in species richness across spatial scales using rarefaction curves and thus aggregating an increasing number of sites. We then used the ‘Measurement of Biodiversity (MoB)’ approach in the R package ‘mobr’ (McGlinn et al., 2019) to decompose these changes in richness into components attributed to (1) the relative abundance of species (evenness), (2) the number of individuals, (3) spatial autocorrelation due to intraspecific aggregation (see Supplement for details - “The MoB approach”).
To gain a deeper understanding of the changes in diversity, we also performed a cross-species analysis. We calculated for each species the bias-corrected log response ratio (‘SinglecaseES’ R package; James Pustejovsky, 2019), as the logged abundance post-lockdown divided by the abundance during lockdown (Hedges et al., 1999). We further explored the potential effects of fish traits in differentiating between species that increased or decreased their abundance during the lockdown. We obtained the following ecological traits per species (following Belmaker et al., 2013): fish family, diet, home-range, schooling, height in the water column, body size, and trophic level (see Supplement for full details - “Description of the predictors”). To relate these traits to the log response ratio, we adopted an exploratory approach using linear mixed models framework with fish traits as predictors, ‘fish family’ as a random effect, and the inverse of the variance of the log response ratios as weight (lme4 R package; Bates et al., 2015). All analyses were conducted in R (R Core Team, 2020).
3 Results
The COVID-19 pandemic and consequent lockdown precluded human activity from coral reefs in the Israeli Gulf of Aqaba (Figs. S4, S5).
3.1 Two scale analysis
We did not find significant changes in ‘knolls’ richness during versus post lockdown, at any scale (Fig. S3, p-value > 0.724, based on Monte Carlo simulations). However, at ‘entrances’ sites, richness during the lockdown was consistently higher relative to post lockdown (Fig. 2 ). We found these differences were manifested both at the sample level (Fig. 2A, α-diversity, p-value = 0.041 based on Monte Carlo permutations), and across all samples (Fig. 2B, γ-diversity, p-value = 0.022 based on Monte Carlo permutations). Similarly, post-lockdown richness at the ‘entrance’ sites was consistently lower than the ‘near-entrance’ sites post-lockdown (α–diversity: Fig. 2B, p-value = 0.047; γ-diversity: Fig. 2B, p-value = 0.05). In contrast, we found no significant change in richness when we compared ‘near entrance’ sites post-lockdown to ‘entrance’ sites during the lockdown nor when we compared ‘entrance’ sites during the two post-lockdown sessions (Fig. S4 panels C and D, p-value > 0.114).Fig. 2 Comparison of species richness. Panels A, B, C represent a two-scale analysis of species richness; (A) α-diversity sample-level, (B) γ-diversity, (C) β-diversity. p-Values are based on a Monte Carlo permutation procedure applied by the MoB algorithm. Test statistic D¯ is average absolute difference in richness between pairwise group comparisons. Dots in figure C represent value two times higher than upper quartile. Panel D¯ presents individual-based rarefaction curves for each habitat. Species richness (y-axis) is plotted as a function of the number of individuals (x-axis). Panels E and F represent two pairwise comparisons of a multi-scale analysis comparing pairs of habitats. Panel E compares the expected change in species richness due to differences in species abundance distribution between the ‘entrances’ during and post lockdown (green line), while the blue shaded area represents 95% acceptance intervals (p-value = 0.029, DCLF significance test). Panel F represents multi scale comparison between ‘near entrances’ and ‘entrances’ post lockdown, once again differences are significant (p-value = 0.003, DCLF significance test). Habitats are color coded purple - ‘near entrances’ post lockdown; blue – ‘entrances’ during lockdown; green – ‘entrances’ post lockdown. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 2
3.2 Measurement of Biodiversity
Next, we used the MobR approach to decompose changes in richness in the ‘entrances’ as a consequence of the lockdown. We found that differences in richness of ‘entrances’ were mainly driven by higher evenness in the species abundance distributions during the lockdown (Fig. 2A and E, p-value = 0.029 based on Diggle-Cressie-Loosmore-Ford – DCLF test). Similarly, we found that differences in richness between ‘entrances’ and ‘near entrances’ were mainly driven by higher evenness in the species abundance distribution of the ‘near entrances’ habitat (Fig. 2A and F, p-value = 0.003 based on DCLF test). The number of individuals and intraspecific aggregation did not have a significant effect on species richness (Fig. S5, p-value > 0.216).
3.3 Species level changes
We found that species showed variable changes to the removal of the lockdown (Fig. 3 ) Siganus argentus, Scolopsis ghanam and Anampses lineatus showed the highest increase in abundance during lockdown, while Crenimugil crenilabis, Abudefduf sexfasiatus and Cheilio inermis increased in abundance post-lockdown. SIMPER test (Clarke, 1993) based on Bray-Curtis dissimilarity showed that Abudefduf vaigiensis, Siganus rivulatus, Acanthurus nigrofuscus, Fistularia commersonii and Diplodus noct together contribute about 50% of the dissimilarity between ‘entrances’ sites during and post lockdown. However, a linear mixed model could not detect traits that explained the change in species abundances due to the lockdown (Table S1).Fig. 3 Changes in species abundance during and after lockdown. The figure shows species level log-ratio of mean abundance during lockdown relative to mean abundance after the lockdown in the ‘entrance’ sites. Green bars represent species which their abundance was higher during lockdown, red bars represent species which their abundance was lower during lockdown relative to their abundance post lockdown. Error bars represent 95% confidence intervals. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 3
4 Discussion
The COVID-19 lockdown was associated with a significant rise in reef fish diversity in the habitat that receives most human activity – reef entrances - but not at deeper or less disturbed habitats (Figs. 2, S3 and S4). These differences were associated with an increase in species evenness during lockdown, a phenomenon that was also apparent at ‘near-entrances’ sites post the COVID-19 lockdown (Fig. 2). While we show an effect of non-extractive human activity on fishes in the reef entrances, these effects are reversible temporally and restricted spatially. Thus, restricting human activities to few designated entrance points can minimize the total effects of human disturbances on reef fishes.
We found that the reduction in human activity during lockdown at the heavily visited ‘entrance’ habitats had similar effects to restricting human activity year-round at the ‘near entrance’ habitats (Fig. 2). These findings demonstrate that short-term temporal cessation in human activity has similar effects to long-term spatial restrictions. As the lockdown effect was short-term, we show that at least some of the human impacts to the reef are reversible and are likely related to fish behavior and not habitat alternation. For example, human disturbances at ‘entrances’ may drive some species to nearby habitats, may shift activity times to periods with less human disturbance (Gaynor et al., 2018), or cause some species to hide or inhibit their movement (Benevides et al., 2019; Côté et al., 2014). However, short term behavioral shifts between high and low disturbance sites may not be enough to mitigate the effect of human disturbance (Albuquerque et al., 2014). Changes to fish assemblages associated with intensive recreational activities lack long term monitoring, and should be better integrated into scientific research and management (Giglio et al., 2020).
In the deeper ‘knoll’ habitat, accessible mostly to divers, we did not find an effect of the lockdown on species richness (Fig. S3). This can be due to two opposing, yet non-exclusive, mechanisms: (1) divers' impacts on these deeper habitat are mostly associated with physical breakage of the coral (Albuquerque et al., 2014; Zakai and Chadwick-Furman, 2002) cause long –term habitat degradation and thus fish diversity is non-responsive to changes in human activity. (2) Alternatively, fish may perceive divers as a low threat and hence their impact may be lower than those of bathers in shallow reefs. Our data cannot currently differentiate between these two potential mechanisms.
The main driver of changes to species richness in our study was differences in evenness. At the ‘entrance’ sites communities had more even relative abundances during lockdown (Figs. 2 and S5). The effect of the lockdown on evenness was not limited to the local site scale. Changes in evenness seem to also drive differences in richness across all the spatial scales measured. Hence, while human recreational impacts may be reversible temporally, they may at the same time accumulate across all spatial scales and substantially change diversity. Our findings are consistent with other studies that found declines in evenness due to non-extractive human activities (Medeiros et al., 2007) and are further consistent with the long-term effect of fishing, as MPAs are characterized by higher evenness even when no change in fish abundance is detected (Blowes et al., 2020). It is interesting that the short term behavioral changes associated with the lockdown are similar to the long term impacts of MPAs and may suggest that, at the community level, evenness is more sensitive to human disturbance than richness or total abundance that are more commonly reported.
Interestingly, we did not find an effect of the lockdown on density (number of individuals) or intraspecific aggregation (Fig. S5). In addition, we could not detect traits that explain which species increased or decreased their abundance (Table S1). The similar density values we found could arise from a substitution of individuals of some species by individuals of others (probably occurring at local scales with no consistent spatial pattern). It is less clear why we did not detect the lockdown effect on aggregation, but we believe this could reflect the habitat patchiness across sites which was not affected by the lockdown.
While we found some changes to species diversity, we did not explore many other potential human effects on coral reef fish species. These include the long-term impact of human activity on the permanent degradation of the habitat, and changes to the demography and diet of fishes. We also did not measure direct behavioral changes or the effect of seasonal changes in touristic activity. Future studies could also benefit from exploring stress levels of fish and how they are related to human presence as this may have nontrivial cascading effects.
Although we found that human interference at ‘entrance’ sites is intensive, their total areas is small (there are a total of only 19 such entry points). Therefore, fishes that are adversely impacted by humans may simply utilize nearby habitats. This suggests that limiting recreational activity from the beach to entrance points is an effective management policy to minimize total recreational human impacts on coral reefs.
Declaration of competing interest
None.
Appendix A Supplementary data
Supplementary material
Image 1
Acknowledgments
VC is supported by the 10.13039/501100007621 Jacob Blaustein Center for Scientific Cooperation (BCSC) and the Interuniversity Institute for Marine Sciences, Eilat (IUI).
Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.biocon.2021.109103.
==== Refs
References
Albuquerque T. Loiola M. José de Anchieta C.C. Reis-Filho J.A. Sampaio C.L. Leduc A.O. In situ effects of human disturbances on coral reef-fish assemblage structure: temporary and persisting changes are reflected as a result of intensive tourism Mar. Freshw. Res. 66 1 2014 23 32
Bacheler N.M. Shertzer K.W. Estimating relative abundance and species richness from video surveys of reef fishes Fish. Bull. 113 2015 15 26
Bates D. Mächler M. Bolker B. Walker S. Fitting linear mixed-effects models using lme4 Journal of Statistical Software 67 2015 1 48
Bates A.E. Primack R.B. Moraga P. Duarte C.M. COVID-19 pandemic and associated lockdown as a “Global Human Confinement Experiment” to investigate biodiversity conservation Biol. Conserv. 248 2020 108665
Belmaker J. Parravicini V. Kulbicki M. Ecological traits and environmental affinity explain Red Sea fish introduction into the Mediterranean Glob. Chang. Biol. 19 2013 1373 1382 23505033
Benevides L.J. Cardozo-Ferreira G.C. Ferreira C.E.L. Pereira P.H.C. Pinto T.K. Sampaio C.L.S. Fear-induced behavioural modifications in damselfishes can be diver-triggered J. Exp. Mar. Biol. Ecol. 514 2019 34 40
Blowes S.A. Chase J.M. Di Franco A. Frid O. Gotelli N.J. Guidetti P. Knight T.M. May F. McGlinn D.J. Micheli F. Sala E. Mediterranean marine protected areas have higher biodiversity via increased evenness, not abundance J. Appl. Ecol. 57 3 2020 578 589
Brunnschweiler J.M. Earle J.L. A contribution to marine life conservation efforts in the South Pacific: the Shark Reef Marine Reserve, Fiji Cybium 30 4 2006 133 139
Campbell M.D. Pollack A.G. Gledhill C.T. Switzer T.S. DeVries D.A. Comparison of relative abundance indices calculated from two methods of generating video count data Fish. Res. 170 2015 125 133
Clarke K.R. Non-parametric multivariate analyses of changes in community structure Aust. J. Ecol. 18 1 1993 117 143
Claudet J. Pelletier D. Jouvenel J.Y. Bachet F. Galzin R. Assessing the effects of marine protected area (MPA) on a reef fish assemblage in a northwestern Mediterranean marine reserve: identifying community-based indicators Biol. Conserv. 130 3 2006 349 369
Corlett R.T. Primack R.B. Devictor V. Maas B. Goswami V.R. Bates A.E. Koh L.P. Regan T.J. Loyola R. Pakeman R.J. Cumming G.S. Pidgeon A. Johns D. Roth R. Impacts of the coronavirus pandemic on biodiversity conservation Biol. Conserv. 246 2020 108571 32292203
Côté I.M. Darling E.S. Malpica-Cruz L. Smith N.S. Green S.J. Curtis-Quick J. Layman C. What doesn’t kill you makes you wary? Effect of repeated culling on the behaviour of an invasive predator PLoS One 9 4 2014 e94248
Davenport J. Davenport J.L. The impact of tourism and personal leisure transport on coastal environments: a review Estuar. Coast. Shelf Sci. 67 2006 280 292
Drew J.A. McKeon M. Shark-based tourism presents opportunities for facultative dietary shift in coral reef fish PLoS One 14 8 2019 e0221781
Feitosa C.V. Chaves L.D.C.T. Ferreira B.P. De Araujo M.E. Recreational fish feeding inside Brazilian MPAs: impacts on reef fish community structure J Mar. Biol. Assoc. UK 92 2012 1525
Gaynor K.M. Hojnowski C.E. Carter N.H. Brashares J.S. The influence of human disturbance on wildlife nocturnality Science 360 2018 1232 1235 29903973
Giglio V.J. Luiz O.J. Ferreira C.E. Ecological impacts and management strategies for recreational diving: a review J. Environ. Manag. 256 2020 109949
Hawkins J.P. Roberts C.M. Effects of recreational scuba diving on coral reefs: trampling on reef-flat communities J. Appl. Ecol. 30 1993 25 30
Hedges L.V. Gurevitch J. Curtis P.S. The meta-analysis of response ratios in experimental ecology Ecology 80 1999 1150 1156
Hughes T.P. Barnes M.L. Bellwood D.R. Cinner J.E. Cumming G.S. Jackson J.B.C. Kleypas J. Van De Leemput I.A. Lough J.M. Morrison T.H. Palumbi S.R. Van Nes E.H. Scheffer M. Coral reefs in the Anthropocene Nature 546 2017 82 90 28569801
James Pustejovsky M.E. Package “SingleCaseES” Type Package Title A Calculator for Single-case Effect Sizes 2019 10.1080/00273171.2018.1466681
Manenti R. Mori E. Di Canio V. Mercurio S. Picone M. Caffi M. Brambilla M. Ficetola G.F. Rubolini D. The good, the bad and the ugly of COVID-19 lockdown effects on wildlife conservation: insights from the first European locked down country Biol. Conserv. 249 2020 108728 32863391
McGlinn D.J. Xiao X. May F. Gotelli N.J. Engel T. Blowes S.A. Knight T.M. Purschke O. Chase J.M. McGill B.J. Measurement of Biodiversity (MoB): a method to separate the scale-dependent effects of species abundance distribution, density, and aggregation on diversity change Methods Ecol. Evol. 10 2019 258 269
Medeiros P.R. Grempel R.G. Souza A.T. Ilarri M.I. Cláudio Sampaio L.S. Effects of recreational activities on the fish assemblage structure in a northeastern Brazilian reef Pan. Am. J. Aquat. Sci. 2 2007 288 300
Milazzo M. Anastasi I. Willis T.J. Recreational fish feeding affects coastal fish behavior and increases frequency of predation on damselfish Chromis chromis nests Mar. Ecol. Prog. Ser. 310 2006 165 172
Munday P.L. Habitat loss, resource specialization, and extinction on coral reefs Glob. Chang. Biol. 10 2004 1642 1647
Rice, W.L., Mateer, T.J., Reigner, N., Newman, P., Lawhon, B., Taff, B.D., 2020. Changes in recreational behaviors of outdoor enthusiasts during the COVID-19 pandemic: analysis across urban and rural communities. J. Urban Ecol. 6, juaa020.
Riegl B. Bruckner A. Coles S.L. Renaud P. Dodge R.E. Coral reefs: threats and conservation in an era of global change Ann. N. Y. Acad. Sci. 1162 2009 136 186 19432648
Rutz C. Loretto M.-C. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R.B. Ropert-Coudert Y. Tucker M.A. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evol. 4 2020 1156 1159 32572222
Titus B.M. Daly M. Exton D.A. Do reef fish habituate to diver presence? Evidence from two reef sites with contrasting historical levels of SCUBA intensity in the Bay Islands, Honduras PLoS One 10 2015 e0119645
Tynyakov J. Rousseau M. Chen M. Figus O. Belhassen Y. Shashar N. Artificial reefs as a means of spreading diving pressure in a coral reef environment Ocean Coast. Manag. 149 2017 159 164
Valerio M. Mann O. Shashar N. “Boo! Did we scare you?”: behavioral responses of reef-associated fish, prawn gobies (Amblyeleotris steinitzi and Amblyeleotris sungami) to anthropogenic diver disturbance Mar. Biol. 166 1 2019 1 9
Vardi R. Berger-Tal O. Roll U. iNaturalist insights illuminate COVID-19 effects on large mammals in urban centers Biol. Conserv. 254 2021 108953
Wilson S.K. Fisher R. Pratchett M.S. Graham N.A.J. Dulvy N.K. Turner R.A. …Rushton S.P. Exploitation and habitat degradation as agents of change within coral reef fish communities Glob. Change. Biol. 14 12 2008 2796 2809
Zakai D. Chadwick-Furman N.E. Impacts of intensive recreational diving on reef corals at Eilat, northern Red Sea Biol. Conserv. 105 2002 179 187
Zambrano-Monserrate M.A. Ruano M.A. Sanchez-Alcalde L. Indirect effects of COVID-19 on the environment Sci. Total Environ. 728 2020 138813
Zellmer A.J. Wood E.M. Surasinghe T. Putman B.J. Pauly G.B. Magle S.B. Lewis J.S. Kay C.A.M. Fidino M. What can we learn from wildlife sightings during the COVID-19 global shutdown? Ecosphere 11 2020 e03215 32834907
| 0 | PMC9746877 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 May 1; 257:109103 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.109103 | oa_other |
==== Front
Appl Acoust
Appl Acoust
Applied Acoustics. Acoustique Applique. Angewandte Akustik
0003-682X
1872-910X
Elsevier Ltd.
S0003-682X(21)00143-2
10.1016/j.apacoust.2021.108050
108050
Article
The effect of room sound absorption on a teleconference system and the differences in subjective assessments between elderly and young people
Hara Rikiya a⁎
Shimizu Takafumi b
a Architectural Design Course, Graduate School of Natural Science and Technology, Shimane University, Japan
b Architectural Design Course, Institute of Environmental Systems Science, Academic Assembly, Shimane University, Dr. Eng, Japan
⁎ Corresponding author.
1 4 2021
8 2021
1 4 2021
179 108050108050
25 10 2020
7 2 2021
11 3 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
In recent years, the rapid development of information and communication technology (ICT) and the influence of the novel coronavirus (COVID-19) have affected our lives and work in various fields such as medical and welfare, construction and manufacturing and education, etc. With this global background, teleconference systems have received attention and become a new trend. However, the acoustics of rooms using teleconference system often overlap the acoustic characteristics from multiple rooms on both the speaker and listener sides. Therefore, it can sometimes be difficult to listen to each other. A prior study suggested that the installation of sound-absorbing panels improves intelligibility and reduces the listening difficulty for young people. However, elderly people must be included in the target owing to the effects of aging. This study aimed to clarify improvements in the subjective assessments of elderly people in a room where a teleconference system is used. In addition, the differences in subjective assessments between young people and elderly people were also investigated. The results of an experiment indicate that, first, a room using a teleconference system demonstrated a greater improvement in subjective assessments after the acoustic improvements compared to the same room where face-to-face meetings. Second, the subjective assessments and improvements of them for elderly people differed greatly since older user had listening habits and experiences that varied from those of young people.
Keywords
Teleconference system
STI
Clarity(C50)
Reverberation time
Intelligibility
Listening difficulty
==== Body
pmc1 Introduction
In recent years, digital devices such as personal computers and tablet PCs have become indispensable in our lives. In addition, the Internet has become an important lifeline for several people. Thus, our lives are changing rapidly into a society of information and communication technology (ICT). ICT has a new relationship with human society and it will become the center of economic activity [1]. For example, various fields such as smart construction work, agriculture, factory control, medical and surgical support are using big data. There are also certain solutions that use ICT in the Sustainable Development Goals (SDGs) adopted at the 2015 United Nations Summit.
In particular, the number of people who have used a teleconference system is rapidly increasing owing to the influence of the novel coronavirus (COVID-19), which has spread worldwide in the past several months. As shown in Fig. 1 , Microsoft’s Teams, a typical service used for remote collaboration, had more than 115 million daily users by October 2020. By contrast, in the middle of March, the number of users was only 44 million. Thus, a rapid increase of 240% was witnessed in just 7 months. According to the Seed Planning survey, the video conferencing market in Japan was 50.3 billion yen in 2018; this, increased by 32% compared to 38.1 billion yen in 2012 [2].Fig. 1 Number of Microsoft Teams daily active users.
From the above, teleconference will become a common technology in our daily lives and work. Moreover, teleconference not only requires work meetings but also access to education and welfare sites. However, unlike traditional face-to-face meetings in the same room, a teleconference system requires multiple rooms to communicate video and sound. Therefore, the challenges in listening to each other often occur owing to the influence of reverberation because the acoustic characteristics in the rooms overlap.
In a previous study by Shimizu et al. [3], the goal was to architecturally enhance the sound intelligibility and reduce the listening difficulty for participants by using a teleconference system. Shimizu’s study suggested that the physical acoustic quantities of the conference rooms were improved, and the sound intelligibility and listening difficulty [4], [5], [6] were also improved, after installing sound-absorbing panels in the conference rooms..
However, the subjects of the prior study were young people around the age of 20, and they were considered to have relatively good listening capability. In addition, it has been generally confirmed that elderly people have weaker listening capability than young people.
It is necessary to mention the problem of a super-aging society (a society whose proportion of the population aged 65 and over exceeds 21%). The most developed countries around the world become super-aging societies owing to factors such as advances in medical technology and declining birth rates. According to Moriizumi at al. [7], most developed countries will probably become super-aging societies by 2060. In addition, according to the United Nations, as shown in Fig. 2 , the percentage of the population aged 65 and over in the world will reach 21% in the 2070 s at the earliest [8].Fig. 2 Percentage of population aged 65 years and over.
The present study aims to clarify improvements in the subjective assessments of elderly people by improving the sound in rooms where a teleconference system is used. Furthermore, the present study examines the difference in subjective assessments between elderly and young people in an acoustic environment when using a teleconference system.
2 Measurements of acoustic physical quantities
2.1 Acoustic physical quantities
2.1.1 STI
The speech transmission index (STI) objectively represents changes in a voice waveform owing to sound reverberation and noise. It takes a value from 0 to 1, where 0.6 or higher is ideal. The higher the value, the less negative sound effects (such as sound reverberation and noise). Accordingly, the clarity is high. STI is calculated from the modulation transfer function (MTF) and the coefficients for each frequency, that is, the contribution rate to the intelligibility [9], [10], [11].
2.1.2 Clarity (C50)
Clarity is an index showing the intelligibility of a voice. The higher the value, the higher is the intelligibility that can be realized. It is represented by the intensity ratios of the direct sound and the individual early reflection sound that arrive after the direct sound in a logarithm. The time to separate the direct sound from the early reflection sound is 50 ms from the moment the direct sound arrives.(1) C50=10log10∫050p2tdt∫50∞p2tdt
In a survey on conversations took place in a living room by Sato [12], C50 are proposed as design target values that 3.1 dB or more in listening situation and 12.6 dB or more in utterance situation.
2.1.3 Reverberation time (RT)
Sound reverberation is a phenomenon in which the generated sound remains in a room. It increases to a certain time and reaches a steady state. When the sound source is consequently stopped, the sound gradually weakens and becomes inaudible. The Reverberation Time (RT) is the time from when the sound source is stopped to the energy density in the room decays 60 dB from a steady value of the sound reverberation phenomenon.
RT is often calculated by the integrated impulse response method (Schroeder integration method) proposed by Schroeder [13]. The reverberation time can be obtained by finding the gradients of the reverberation energy from the reverberation decay curve equation described below [14]:(2) Et=∫t∞p2τdτ∫0∞p2τdτ
where E (t) is the reverberation decay curve (Schroeder’s decay curve), and p (τ) is the instantaneous amplitude of the measured impulse response.
2.2 Methods of measurement
In a prior study, researchers used two rooms with a teleconference system (HDX 7000 manufactured by POLYCOM Co., Ltd.) to measure acoustic physical quantities. As shown in Fig. 3 , two rooms (Room A and Room B) were connected to each other by the teleconference system using the Internet. A microphone of the teleconference system (HDX microphone array manufactured by POLYCOM Co., Ltd.) was located in Room A, and a loudspeaker with a liquid–crystal display (LCD: LC-40 AE7 manufactured by SHARP Co., Ltd.) in Room B. Moreover, there was a microphone with a sound level meter (NL-31 manufactured by RION Co., Ltd., microphone: UC-53A) and a loudspeaker (MSP7 manufactured by Yamaha Corporation) in each room. Needless to say, this system was able to record the sound transmitted from Room A with the microphone in Room A. In addition, it was possible to record the sound in Room A with the microphone in Room B via the Internet of the teleconference system. Therefore, the sounds recorded by the microphones in Room A and B were able to reproduce the sounds in single room face-to-face and two rooms using the teleconference system.Fig. 3 Block diagrams for measurements of acoustic physical quantities.
The acoustic physical quantities were measured in four discrete sound absorption levels from 0 to 3 by adjusting the number of sound-absorbing panels installed in both rooms as shown in Table 1 . The sound-absorbing panels were paper honeycomb structures sandwiched between polyester nonwoven fabrics. The frequency characteristics of the sound absorption coefficient of the sound-absorbing panels are listed in Fig. 4 ; this was measured in accordance with JIS A 1409 [15]. Notably, the size of the sound-absorbing panels was 1.08 m2 (600 × 900 mm × 2 pieces), and the sound absorption coefficient exceeded 1. Table 2 lists the experimental conditions. The measuring method was the swept-sine method (time-stretched pulse method) using the sound source as the chirp signal [16], [17], [18].Table 1 Sound absorption levels and number of sound absorption panels.
Sound absorption level Room A Room B
Level 0 (Non panel) – –
Level 1 2.94 m2
(0.49 m2 × 6) 1.96 m2
(0.49 m2 × 4)
Level 2 5.04 m2
(0.49 m2 × 8) 3.92 m2
(0.49 m2 × 8)
Level 3 7.60 m2
(0.49 m2 × 8 + 0.54 m2 × 7) 6.02 m2
(0.49 m2 × 2 + 0.63 m2 × 8)
Fig. 4 Sound absorption coefficient of sound-absorbing panel.
Table 2 Experimental conditions.
No. Utterance side Receiving side Sound-absorption level Abbreviation
1 Room A Speaker → Room A Microphone Level0 (Non) A0-A0
2 Room A Speaker → Room A Microphone Level 1 A1-A1
3 Room A Speaker → Room A Microphone Level 2 A2-A2
4 Room A Speaker → Room A Microphone Level 3 A3-A3
5 Room A Speaker → Room B Microphone Level 0 (Non) A0-B0
6 Room A Speaker → Room B Microphone Level 1 A1-B1
7 Room A Speaker → Room B Microphone Level 2 A2-B2
8 Room A Speaker → Room B Microphone Level 3 A3-B3
Fig. 5, Fig. 6 show the impulse responses recorded in Room A and Room B, respectively. The measurement results of STI, Clarity, and RT are shown in Fig. 7, Fig. 8, Fig. 9 , respectively. Clarity and reverberation time were measured in three octave bands from 500 Hz to 2 k Hz since they are approximate to the frequency of the human voice.Fig. 5 Impulse responses in Room A.
Fig. 6 Impulse responses from Room A to B.
Fig. 7 Measurement results for STI.
Fig. 8 Measurement results for Clarity.
Fig. 9 Measurement results for RT.
3 Methods
In intelligibility tests, the participants were presented with a test signal via headphones (SENNHEISER: HD 280 pro). The test signal presented to the participants were convolved with the impulse response of each sound field in the speech corpus used for the Japanese word intelligibility test [19]. The speech corpus for the intelligibility test was used four-mora Japanese words with low word-familiarity (1.0-2.5). The same impulse response convolved with the speech corpus used in the prior study was presented to the participants in this research. In this test, the participants were 20 elderly people. This was done in order to clarify the difference in subjective assessments between the elderly people and the young people conducted in the prior study. The participants were over the age of 60.
The test signal was presented to both ears of the participants. The test signal was the equivalent noise level measured a slow peak average of 65 dBA tuned to A characteristic between both loudspeakers of the headphones. The participants listened to the words of the presented test signal and answered words. The intelligibility was evaluated by the percentage of correct answers given by the participants for the presented words.
After taking the intelligibility test, participants were presented with a listening difficulty test in which they evaluated whether the presented test signal was difficult to hear. As shown in Table 3 , the evaluation of difficulty in listening was divided into four levels: “not difficult,” “slightly difficult,” “moderately difficult,” and “Very difficult.” The listening difficulty was the percentage at which the presented test signals were evaluated as “difficult.” In other words, it was the ratio of the total of the evaluations other than “not difficult” in Table 3. Both the intelligibility and listening difficulty tests were conducted for each participant.Table 3 Listening difficulty ratings.
No. Category
0 Not difficult
1 Slightly difficult
2 Moderately difficult
3 Very difficult
4 Results
4.1 Intelligibility test
Table 4 shows the results of the intelligibility test. The results for the young participants were gathered in a prior study, and the results for the elderly participants were gathered in the present test. Fig. 10 summarizes the results of the intelligibility test for both the young and elderly participants.Table 4 Result of intelligibility test.
Experimental condition Intelligibility
Young Elderly
A0-A0 84.6% 69.4%
A1-A1 88.2% 73.0%
A2-A2 91.0% 70.4%
A3-A3 93.0% 72.7%
A0-B0 67.8% 36.6%
A1-B1 72.8% 43.1%
A2-B2 73.4% 46.3%
A3-B3 81.4% 53.0%
Fig. 10 Results of intelligibility test.
For the young participants, the correct answer rate was 84.6% in the same room when the acoustic physical quantity was not improved and 93.0% when the acoustic physical quantity was improved. In contrast, in the room using the teleconference system, the correct answer rate was 67.8%, but the ratio rose to 81.4% when the acoustic physical quantity was improved. In terms of the elderly participants, the correct answer rate was 69.4% in the same room when the acoustic physical quantity was not improved and 72.7% when the acoustic physical quantity was improved. In the room using the teleconference system, the correct answer rate was 36.6%, which improved to 53.0% when the acoustic physical quantity was improved.
Moreover, Fig. 11 shows the correlation between the intelligibility and the acoustic physical quantities such as STI, Clarity and Reverberation time. As shown in Fig. 11 (a), the determination coefficients of intelligibility and STI were 0.961 for young participants and 0.918 for elderly participants. As shown in Fig. 11 (b), the coefficients of determination of the intelligibility and Clarity were also very high at 0.958 for young participants, and 0.941 for elderly participants.Fig. 11 Correlation between intelligibility and acoustic physical quantities.
4.2 Listening difficulty test
Fig. 12 shows the results of the listening difficulty test. (a) shows the result for young participants, which is the result of the prior study, and (b) shows the result for elderly participants, which is the result of the present test. This represents the ratio of the evaluation of the listening difficulty for sound under each experimental condition. In addition, Table 5 shows the value of the listening difficulty, that is, the ratio of the evaluation results excluding “not difficult.” Fig. 13 summarizes the results of the listening difficulty test for both the young and elderly participants.Fig. 12 Answer results of listening difficulty test.
Table 5 Results of listening difficulty test.
Experimental condition Listening difficulty
Young Elderly
A0-A0 44% 23%
A1-A1 44% 22%
A2-A2 20% 20%
A3-A3 24% 19%
A0-B0 100% 43%
A1-B1 88% 37%
A2-B2 88% 30%
A3-B3 88% 32%
Fig. 13 Results of listening difficulty test.
First, there is a difference in intelligibility between the sound in the same room and the sound in the room using the teleconference system. Furthermore, more than half the young participants evaluated the sound using a teleconference system as “difficult.” Secondly, when the sound-absorbing panels are installed, the more the sound is improved, the lower the listening difficulty becomes. This applies to both young and elderly participants.
In terms of the young participants, the listening difficulty was 44% in the same room when the acoustic physical quantity was not improved and this improved to 20% at absorption level 2 and 24% at absorption level 3. By contrast, in the room using the teleconference system, the listening difficulty was 100%, and 88% when the acoustic physical quantity was improved. In terms of the elderly participants, the listening difficulty was 23% in the same room when the acoustic physical quantity was not improved and 19% when the acoustic physical quantity was improved. In contrast, in the room using the teleconference system, the listening difficulty was 43%, which improved to 30% at absorption level 2 and 32% at absorption level 3.
In addition, as shown in Fig. 12 (a), in the room using the teleconference system, the percentage of young participants who evaluated as “very difficult” or “moderately difficult” decreased with an increase in the sound absorption level. However, as shown in Fig. 12 (b), for the elderly participants, the percentage of those who evaluated “very difficult” or “moderately difficult” were almost 10% under all conditions.
Fig. 14 shows the correlation between the listening difficulty and the three acoustic physical quantities. As shown in Fig. 14 (a), the determination coefficient between the listening difficulty and STI was 0.859 for young participants and 0.959 for elderly participants. As shown in Fig. 14 (b), the determination coefficient between the listening difficulty and Clarity was 0.902 for young participants and 0.980 for elderly participants, which was extremely high.Fig. 14 Correlation between listening difficulty and acoustic physical quantities.
5 Discussions
The intelligibility results for the elderly in the room using the teleconference system were extremely low at 36.6%. However, the intelligibility of elderly people was improved by 16.4% at sound absorption levels 0 to 3 in the room using the teleconference system. When using the teleconference system, it was observed that the intelligibility of young people also improved by 13.6%. However, in the same room where the sound absorption levels were 0 to 3, the intelligibility of the elderly improved only by 3.3%, and the intelligibility of the young improved only by 8.4%. This indicates that the intelligibility of the room using the teleconference system was highly improved by the sound-absorbing panels.
The intelligibility results for the room using the teleconference system were improved by the sound-absorbing panels, but not as highly as when directly speaking to each other. The intelligibility of young people in the improved room using the teleconference system was 81.4% compared to 84.6% in the same room when the teleconference system was not used. Furthermore, the intelligibility of elderly people in the improved room using the teleconference system was 53.0% compared to 69.4% in the same room.
As shown in Fig. 11 (a) and (b), it was found that STI and Clarity have an extremely high correlation with the intelligibility of young and elderly people. Furthermore, it is generally considered that an STI of 0.75 or higher is ideal. However, even if the STI was 0.75 or more, the intelligibility of the elderly people did not reach 50% or more. Therefore, more than half of the elderly people may not understand speech even if the STI was 0.75.
Accordingly, it can be observed that the intelligibility in the room using the teleconference system can be improved by sound-absorbing panels. However, the intelligibility in the room using the teleconference system still needs to be improved.
The listening difficulty in the same room improved by 20% for young people and 4% for elderly people at sound absorption levels 0 to 3. On the other hand, the listening difficulty in the room using the teleconference system improved by 12% for young people and 11% for elderly people at sound absorption levels 0 to 3. The listening difficulty in the room using the teleconference system was greatly improved by the sound-absorbing panels. However, the listening difficulty of sound absorption level 3 did not provide significant improvement over levels 1 and 2. This may be owing to the reason that the listening difficulty for improvement of the acoustic physical quantities has a certain threshold.
In Fig. 14 (a) and (b), it was found that STI and Clarity have an extremely high correlation with the listening difficulty of young and elderly people. Furthermore, even if the STI is 0.75 or higher, the listening difficulty of young people is not improved to less than 50%. Therefore, more than half the young people found it difficult to listen when the STI was 0.75. However, it is known that participants tend to judge on a scale for the entire experiment in the case of subjective assessments such as the listening difficulty. Therefore, it is possible that the narrow range of STI used in this study result in higher the listening difficulty than the wide range of STI containing less than 0.7. In addition, as mentioned above, according to Sato et al., it is suggested that the listener will not be irked if the Clarity is 3.1 dB or higher in the living room [12]. According to the experimental results, even if the Clarity was 3.1 dB (which was the designated target value) or higher, the listening difficulty of young people did not improve to less than 50%. It can be said that more than half the young people do not evaluate as “difficult to listen” when the Clarity is ≥9.4 dB.
A difference was found in the evaluation of the listening difficulty between elderly people and young people. Young people found it more difficult to listen than elderly people. The listening difficulty of elderly people in the same room were 0 to 22% lower than that of young people. Furthermore, it is interesting that the listening difficulty of elderly people in the room using the teleconference system were 51 to 58% lower than that of young people. According to a study by Uchida et al. [20], elderly people tend to underestimate the listening difficulty compared to young people if their hearing thresholds are comparable. It is speculated that elderly people do not find themselves difficult to listen. More specifically, elderly people may mistakenly think that they can listen accurately regardless of whether they can understand what they heard, so they will not pay attention to listening. Additionally, the effect was stronger in the room using the teleconference system than in the same room. Therefore, elderly people should be more careful when listening, and the acoustics of rooms using teleconference system still need to be improved.
6 Conclusions
In this study, we investigated the effects of acoustics in the same room and the room using a teleconference system on subjective assessments such as intelligibility and listening difficulty. In addition, the influence of the improvement of the acoustic physical quantities using the sound-absorbing panels on the subjective assessments was also investigated. This research aims to clarify the difference in the subjectivity between young people and elderly people, including the viewpoint of the aging society, which has become a challenge in developed countries in recent years. Moreover, we examined the improvement of the subjective assessments of young and elderly people by improving the acoustics of the room using the teleconference system.
An analysis of the data obtained from this research led to the following findings. Firstly, it was found that the improvements in acoustic quantities by the sound-absorbing panels increased the subjective assessments of both young and elderly people in the same room and the room using the teleconference system. In particular, the acoustic improvement by the sound-absorbing panels in the room using the teleconference system greatly affected the subjective assessment of the listener. Furthermore, it is possible to satisfy the design target value of the acoustic physical quantities even with a small number of sound-absorbing panels. However, the acoustics of a room using a teleconference system may not always be applicable to the existing design target values. Even if the acoustic physical quantities are improved to the design target value, this does not imply that the listener can understand what is being transmitted.
Second, elderly people have different listening abilities than young people, and thus the strength of the effect on intelligibility also differs. It is considered that elderly people tend not to admit to having listening difficulties. It is important to note that elderly people may unknowingly overlook or mishear certain sounds. Therefore, it is necessary to assume users of various ages when organizing a conference or meeting, or when a designing room using a teleconference system. This is because use of teleconference systems will become commonplace among the general public, and there will definitely be more opportunities for the elderly as well as the young to use these systems.
In this study, the sound field of a relatively small and easy-to-hear room was used for the experiment. It is also necessary to cover the sound fields of larger conference rooms and venues because not only meetings and lectures for small groups but also performances, classes, and lectures for large numbers of people will need to be accommodated.
CRediT authorship contribution statement
Rikiya Hara: Conceptualization, Formal analysis, Data curation, Writing - original draft. Takafumi Shimizu: Conceptualization, Methodology, Investigation, Writing - review & editing, Resources, Supervision.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
I would like to express my deepest gratitude to Yasutomi Matushima, Tooru Ochiai (Kuraraykuraflex Co., Ltd., Japan), Dr. Hiroshi Onaga (former professor at Kindai University), Shinji Morita, Hiromi Ishimura (former Kindai University undergraduate students) and Masaru Koike for their cooperation in the experiments and for preparing the sound-absorbing panels used in this study.
==== Refs
References
1 White Paper Information and Communications in Japan. Information and Communication Bureau, Ministry of Internal Affairs and Communications (2019), https://www.soumu.go.jp/johotsusintokei/whitepaper/eng/WP2019/2019-index.html [accessed 10 September 2020].
2 Seed Planning, Inc., 2020 Video Conference / Web Conference Latest Market and Video Communication Equipment / Service Trends (2020.3) [in Japanese].
3 Shimizu T. Onaga H. Study on acoustic improvement by sound-absorbing panels and acoustical quality assessment of teleconference systems Appl Acoust 139 2018 101 112 10.1016/j.apacoust.2018.04.021
4 Morimoto M. Sato H. Kobayashi M. Listening difficulty as a subjective measure for evaluation of speech transmission performance in public spaces J Acoust Soc America 116 3 2004 1607 1613 10.1121/1.1775276
5 Sato H. Bradley J.S. Morimoto M. Using listening difficulty ratings of conditions for speech communication in rooms J Acoust Soc America 117 3 2005 1157 1167 10.1121/1.1849936
6 H. Sato. Measurement method of listening difficulty ratings. J INCE/J 2011; 35(1): pp. 69–74. Institute of noise control engineering [in Japanese].
7 R. Moriizumi, K. Kamata. International Comparison of the National Population Projections in Industrialized Countries. Journal of Population Problems. 69-3 (2013.9) pp. 27–47 [in Japanese].
8 United Nations. World Population Prospects: The 2019 Revision, “https://population.un.org/wpp/” [accessed 10 September 2020].
9 Satoh F. Measurement techniques of room impulse response J Acoust Soc Japan 58–10 2002 669 676 [in Japanese]
10 Houtgast T. Steeneken H.J.M. The Modulation Transfer Function in Room Acoustics as a Predictor of Speech Intelligibility Acustica 28 1973 66 73
11 ISO 9921:2003 Ergonomics-Assessment of speech communication
12 Sato H. Morimoto M. Kobuki K. The effect of reverberation sound on conversation in living rooms J Acoust Soc Japan 66–11 2010 541 551 [in Japanese]
13 Schroeder M.R. New method of measuring reverberation time J Acoust Soc America 37 6 1965 1187 1188 10.1121/1.1939454
14 Hanyu T. Calculation of room acoustic parameters from impulse response J Acoust Soc Japan 69–6 2013 291 296 [in Japanese]
15 JIS A 1409. Method for measurement of sound absorption coefficients in a reverberation room. Japanese Industrial Standard Committee (1998) [in Japanese].
16 Aoshima N. Computer-generated pulse signal applied for sound measurement J Acoust Soc America 69 5 1981 1484 1488 10.1121/1.385782
17 Tachibana H. Yano H. Measurement of environmental noise and architectural acoustics (Sound technology series 8) 2004 CORONA Publishing Co., LTD. Acoustical Society of Japan [in Japanese]
18 ISO 18233:2006 Acoustics-Application of new methods in building and room acoustics.
19 [dataset] T. Kondo, S. Amano, S. Sakamoto, Y. Suzuki. Development of familiarity-controlled word-lists (FW07). Institute of Electronics. J Inform Commun Eng 2008; 107(432): pp. 43–48. Institute of Electronics, Information and Communication Engineering [in Japanese].
20 Uchida Y. Nakashima T. Ando F. Niino N. Shimokata H. Prevalence of self-perceived auditory problems and their relation to audiometric thresholds in a middle-aged to elderly population Acta Otolaryngol 123 5 2003 618 626 10.1080/00016480310001448 12875585
| 0 | PMC9746878 | NO-CC CODE | 2022-12-15 23:21:58 | no | Appl Acoust. 2021 Aug 1; 179:108050 | utf-8 | Appl Acoust | 2,021 | 10.1016/j.apacoust.2021.108050 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00256-1
10.1016/j.biocon.2021.109204
109204
Policy Analysis
The conservation and ecological impacts of the COVID-19 pandemic
Primack Richard B. a⁎
Bates Amanda E. b
Duarte Carlos M. c
a Department of Biology, Boston University, 5 Cummington Mall, Boston, MA 02215, USA
b Department of Ocean Sciences, Memorial University of Newfoundland, St. John's, NL A1C 5S7, Canada
c Red Sea Research Centre (RSRC) and Computational Bioscience Research Center (CBRC), King Abdullah University of Science and Technology, Thuwal 23955, Saudi Arabia
⁎ Corresponding author.
1 6 2021
8 2021
1 6 2021
260 109204109204
26 5 2021
28 5 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Keywords
COVID
==== Body
pmc1 Introduction
The outbreak of the COVID-19 pandemic at the start of 2020 led to dramatic reduction and alteration of human activity. People were confined to their homes, and international travel essentially ceased. National parks and other protected areas either closed or limited visitation, though some parks, particularly in urban areas, showed dramatically increased use. Universities and schools closed or shifted to online teaching, and field work was suspended or disrupted. Businesses closed and economic activity was reduced or changed in substantial ways.
All of these societal shifts have implications for the conservation of biological diversity, and the functioning of ecological communities. At the start of the pandemic lockdown in March and April of 2020, we, along with our colleagues, wrote a series of papers describing how the pandemic might affect conservation in the broadest sense, including education, research, ecotourism, citizen science programs, and wildlife protection (Corlett et al., 2020; Bates et al., 2020; Rutz et al., 2020). We pointed out that even though the pandemic is tragedy of enormous significance, it is also an opportunity to study the relationship between humans and the environment. What happens when tourism to natural areas rapidly declines? What happens when management actions are abruptly suspended? What happens when people's options for recreation are limited to exploring their local communities? How does wildlife respond to reduction of traffic, both on land and in the water, and associated noise? These types of questions both describe the effects of the pandemic, and illuminate the wider impacts we have on biodiversity through management and consumption of natural resources. It is also important to understand how the disruption of university courses, research activities, and employment will impact on research programs and the careers of students and recent graduates, as well as how public opinion and policy towards wildlife will change as a result of the pandemic.
To accomplish this goal, in the middle of 2020, we invited researchers from around the world to contribute articles to a Special Issue of Biological Conservation focusing on the conservation impacts of the pandemic lockdown. We wanted to encourage researchers to take an early look at the conservation impacts, with a goal towards describing new methods that could potentially be applied more generally. We recognize that many of the papers presented here are still preliminary and somewhat limited in scope and design, but we felt that it was important to take advantage of this unprecedented opportunity and give researchers a chance to publish early results. This compilation of studies from around the world allows a first look at the variability and scope of wildlife and conservation programs responses that will help formulate better informed hypotheses and improve subsequent research designs.
For all of the papers submitted to this Special Issue, we required the authors to provide a valid control for comparison. In many cases the control would be a natural system that had been monitored for one or more years prior to the pandemic. In the best case, a system would have been monitored for several years prior to the pandemic, during the pandemic, and then again after the pandemic lockdown had been eased. Alternatively, a system could be examined during the pandemic and then followed as it returned to normal after the lockdown.
While we had this requirement for controls, we also recognized that the pandemic is an unplanned experiment that happened suddenly and without warning. Therefore, some of the papers in this Special Issue have smaller sample sizes and less formal designs than we expect from articles submitted in normal times. For the sake of presenting the best available research on this topic as quickly as possible, we were willing to accept articles with these limitations but required that authors interpret their results in the context of these limitations.
This Special Issue contains over 30 articles, making this probably the largest Special Issue ever published by Biological Conservation. It is striking how many countries and continents are presented in these articles, and the range of topics covered. We have also gathered together additional examples and anecdotes from around the world to create a broad synthesis of the ecological and conservation impacts of the pandemic (Bates et al., 2021).
Our hope is that conservation biologists reading these articles will see opportunities for using these same approaches in their own research and managers will consider possibilities for adjusting their management plans.
2 Changes in visitation and protected areas
Given that the COVID-19 pandemic heavily impacted people's ability and willingness to travel, it is no surprise that several articles in this issue focus on the effects of changes in ecotourism during the pandemic. A study by China et al. (2021) surveyed coral reefs in the Gulf of Aqaba in Israel, and found that the species richness of fish was greater during the lockdown, with tourism suspended, than after the lockdown, when tourism resumed. Similarly, Soto et al. (2021) monitored 29 beaches in Latin America during and after the lockdown. They discovered that in the absence of people, the beach vegetation began to recover, and local animals such as ghost crabs increased in abundance. There was also notably less litter and noise. Quesada-Rodríguez et al. (2021) studied a different aspect of the beach system: the nesting of leatherback sea turtles (Dermochelys coriacea) in Pacuare Reserve, Costa Rica. Despite the lack of income from tourists and educational programs, with the absence of humans and increased staff patrolling, the overall percentage of nesting turtles was higher than normal and hatching success of the eggs was at an appropriate level. These studies both demonstrate the influential role tourism has on ecosystems, and show that recovery is possible even for heavily impacted beach and ocean ecosystems.
However, not all of the effects from decreased tourism have been positive. In an innovative study, Souza et al. (2021) investigated changes in searches of information about national parks using Google Search. They found that during the pandemic lockdown, there were large declines in searches for information about national parks, especially international parks that depend on foreign travel. This slump in online interest was reflected in real life consequences for many parks. Smith et al. (2021) outline how revenue from international ecotourism in South Africa declined by 90% during the pandemic lockdown as international travelers were unable and unwilling to visit South Africa's parks. Likewise, Miller-Rushing et al. (2021) provide an extensive survey of USA parks, including many iconic parks such as Yellowstone and Great Smokies, and describe the extent of changes and severe reductions in visitation, management, research, staffing and education. Most educational and visitor programs moved online, and much of the research and management activities of the parks were simply cancelled.
Some wildlife populations have been negatively affected as well. In their article, Hentati-Sundberg et al. (2021) describe the effects on the common murre (Uria aalge) population of a coastal Swedish island. The absence of tourists to the island during the pandemic in 2020 led to increased visitation to the island of white tailed eagles (Haliaeetus albicilla), which normally avoid people. The eagles caused substantial disturbance to common murre colonies and a lower breeding success due to egg predation by gulls and crows during disturbance episodes.
The inability to travel has also affected urban ecosystems. Because traveling to distant locations and using indoor exercise facilities was not possible during the pandemic, many urban dwellers instead crowded into nearby parks to relax and exercise. In a study of a park in the Boston suburbs, Primack and Terry (2021) found that in the first two months of the pandemic, new social trails increased the network of trails by 36%, approximately the same length of new trails that had been created in the previous 48 years.
The COVID-19 pandemic has served to further illuminate the complex relationship between conservation and ecotourism. These studies demonstrate how tourism can both contribute to or impede the restoration of ecosystems, depending on the situation. New approaches need to be developed which allow tourism to have positive impacts on nature, at the same time as flagging those activities with harmful impacts.
3 Urban impacts
At the start of the pandemic, there were many anecdotal reports of wildlife entering cities and becoming more active. However, it was always uncertain if wildlife patterns really changing or if people were just outside more and making more wildlife observations. This topic was examined by Vardi et al. (2021) using an analysis of iNaturalist observations. They found that reports of large mammals venturing into cities during the pandemic appear to be exaggerated. The one exception is mountain lions (Puma concolor) which appear to have become more common in cities during the pandemic.
While the influx of wildlife to cities may have been exaggerated in some cases, there have been several instances of wildlife leaving cities. Gilby et al. (2021) examined the regional impact of the pandemic lockdown on coastal and urban ecosystems in Eastern Australia. The most notable change was Torresian crows (Corvus orru) departing from urban areas due to lack of food and foraging on nearby beaches, where they had surprisingly large ecological impacts due to outcompeting native scavengers, feeding on insects and invertebrates, and predating on the eggs of native birds. Soh et al. (2021) also examined the effects of the pandemic on bird communities, in this case in Singapore. They found that pigeons and other bird species shifted their foraging areas quite substantially after the pandemic lockdown was implemented as the birds could no longer readily find food discarded by people. Pigeons also spent more time foraging and less time resting due to the presumed difficulty of finding food.
In addition to the animal populations, the environmental characteristics of cities have been impacted. At a site in Colombia, Ulloa et al. (2021) measured noise levels during and after the pandemic lockdown. They found that noise levels were lower during the pandemic due to less human activity, and people noticed the sounds of wildlife more when there was less noise pollution.
4 Wildlife and technology
The pandemic highlighted the importance of using technology to continue monitoring wildlife and ecosystems even when people could not be in the field. Huveneers et al. (2021) describes the role played by an extensive acoustic monitoring system in the waters off Australia's coast. This system allows researchers to monitor changes in behavior and range of marine species, especially sharks and other fish species, during the pandemic when shark tourism was halted. It was found that species differed in their responses when they were no longer being fed at the site. Camera traps are another technology that will likely be a large part of future wildlife research. Blount et al. (2021) describe the increasing importance of camera traps in monitoring wildlife at night and documenting illegal activities, even before the pandemic. Traps took on greater importance during COVID-19 when in-person wildlife monitoring had to be suspended.
As people are spending more time outside, citizen science networks such as iNaturalist, eBird, and the National Phenology Network have also played a large role during the pandemic; however, increases in reported observations have not been uniform. Crimmins et al. (2021) reports that major citizen science platforms in the USA mostly experienced increases in observations during the pandemic, but most of the growth was in the eastern USA and more in urban areas. This geographic shift in reporting needs to be considered in any analysis investigating changes in species ranges and phenology. The findings by Crimmins et al. (2021) are confirmed by a more detailed analysis of eBird data in the USA by Hochachka et al. (2021). They show that observers generally shifted towards more urban habitat and away from rare wetland habitat. Also, changes varied considerably among regions, suggesting that generalizations of shifting bias are difficult to make.
Analyses of citizen science networks have been conducted for other countries as well. Sánchez-Clavijo et al. (2021) found that observations of birds in Colombia using eBird and iNaturalist were high during the lockdown but were more concentrated in urban areas. There were also fewer total bird species reported and fewer observations of rare birds in more natural areas. Basile et al. (2021) found similar results in Italy, Spain, and the U.K.; they report an increase in citizen scientist activity in urban areas and a decrease in non-urban areas. These studies suggest that people are making more observations near their urban homes due to the lockdown.
5 Impacts of roads
One of the most serious threats to wildlife in general, and rare and endangered species in particular, is the growing network of roads. For many wildlife species, collisions with vehicles represent the greatest threat to their populations. To reduce this threat, measures have been implemented such as posting signs warning motorists about wildlife crossing, constructing under- and over-passes, and erecting fencing along roads. Wildlife biologists have also been monitoring highways to document the number, location, and species of animals killed by collisions. These baseline studies have allowed wildlife biologists to examine how the pandemic lockdown affected the pattern of wildlife being killed along highways.
Shilling et al. (2021) reports on an extensive study of wildlife mortality along roads in various USA states. This study found that with reduced traffic on highways during the pandemic lockdown, the mortality declined by 34% overall. There was a 58% decline in mortality during the pandemic for the mountain lion (Puma concolor), an apex predator of special conservation interest. In a comparable study from the Australian island of Tasmania, Driessen (2021) reports that roadkills decreased by 46% during the time of the pandemic. Bíl et al. (2021) also studied wildlife mortality along roads in 11 European countries. In four countries, Spain, Israel, Estonia, and Czechia, a reduction of road traffic during the pandemic resulted in a 40% decline in wildlife mortality. This situation contrasted with Sweden where there was no major lockdown and correspondingly no reduction in wildlife mortality.
Another major impact from roads is noise pollution. In a study of noise pollution in Boston protected areas, Terry et al. (2021) found that as traffic volumes and other human activities declined with the start of the pandemic, sound levels at two parks decreased as expected. However, at a third park, sound levels actually increased; even though there was less traffic, the vehicles were going much faster and making more noise.
These studies demonstrate that when there is rigorous, quantitative protocol for monitoring the environment near roads, the effects of the pandemic can be clearly demonstrated. The key question is whether the insights gained by this research can lead to changes in management of protected areas, such as the reducing the number of vehicles or speed limits, to reduce the chance of wildlife-vehicle collisions and the level of noise pollution.
6 Wildlife management
The pandemic has caused changes in many management and harvesting practices, with direct impacts on wildlife populations. LeTourneux et al. (2021) describe how reduced hunting activity during the pandemic allowed snow geese (Anser caerulescens), an over-abundant species, to feed more effectively on their spring migration grounds. As a consequence, the geese were in better body condition than previous years, leading to higher breeding success. This article emphasizes the importance of hunting for species management. Human impacts on animal populations are also shown by Coll et al. (2021) reporting on a rebound in shrimp populations during the lockdown, due to a decrease in harvesting by the Spanish fishing fleet. However, this effect was short lived once fishing resumed, suggesting that a sustained reduction in fishing is needed for marine ecosystems to recover.
Sumasgutner et al. (2021) propose establishing a Global Anthropause Raptor Research Network to target how this group of iconic, keystone species responds to changing levels of human disturbance and activity. This project has the potential to provide considerable insight as many raptor species avoid human presence, and so may change their distribution and behavior during the COVID-19 pandemic. This project may also engage the public in conservation, as hawks, eagles, and other raptors hold special interest for many people.
A review by Cooke et al. (2021) also investigates aquatic ecosystems, considering the positive and negative impacts that the pandemic could have on freshwater fish populations. This overview includes demand for food, monitoring, research, compliance, and management interventions. The authors argue that the pandemic provides insights into how this resource could be better managed. Hopefully, fish biologists throughout the world will heed this call to action, and search for the data needed to provide greater insight.
The lack of wildlife management and protection during the time of the pandemic also occurred at a time when rural people were often out of work. Aditya et al. (2021) provides an example of how illegally harvesting of pangolins in India increased during the time of the pandemic, as indicated by wildlife seizures by government officials.
The likelihood that the COVID-19 virus originally spread from wildlife to people has sparked recommendations for management practices aimed at preventing future pandemics. Dobson et al. (2020) suggest that a reduction in the wildlife trade and the handling and eating of wildlife by people is an effective strategy for preventing another pandemic. However, while the most likely source of COVID-19 is mammals, there is also a need to address the problems associated with other groups of animals. Borzée et al. (2021) argue that concerns about the spread of disease in East Asian countries should also extend to amphibians, with increased regulation of amphibian farming and the amphibian pet trade to reduce the chance of disease spread.
The pandemic is also affecting the careers of young conservation biologists who are learning the skills needed to manage protected areas and preserve biodiversity. A survey by Ramvilas et al. (2021) of Indian early-career conservation researchers found that their fieldwork, travel, and funding was highly disrupted by the pandemic. They hope that stakeholders will have a greater role in conservation priorities in the post-pandemic world.
7 Policy and attitudes
Several countries have used the pandemic to enact changes in their conservation policy. A paper by Huang et al. (2021) describes revisions to the Wildlife Protection Law in China which are intended to restrict the use of wild animals for food and traditional medicine. Combined with improved management of protected areas, these changes may provide benefits for wildlife populations and reduce the spread of diseases between humans and wild animals. However, while China and other Asian countries are moving towards increased protection of national parks, endangered species, and wildlife, many countries do not share these priorities. As described by Vale et al. (2021), the Brazilian government is taking advantage of the pandemic to weaken environmental legislation and enforcement in a misguided attempt to stimulate economic activity. As a result, the ability of the Amazon rain forest to protect biodiversity and sequester carbon is being reduced.
Public engagement has also suffered somewhat during the pandemic, although there are strong initiatives at work to get people involved in protecting wildlife and reducing the spread of disease. Special biodiversity events such as City Nature Challenges or Biodiversity Days help increase interest in community-engaged science, though these types of events have faced difficulties during the pandemic. Kishimoto and Kobori (2021) report that volunteer participation in the Tokyo City Nature Challenge declined by around 60% during the pandemic, though the number of species observed was about the same. The spatial pattern of observations also changed from clustered to scattered. Lack of interest is also sometimes accompanied by negative attitudes towards wildlife. Using a survey, Lu et al. (2021) found that a large proportion of people in China have misconceptions of the role of bats in spreading disease to humans. Such misconceptions can lead to destruction of bats and their habitats, and a loss of the ecosystem services that they provide. After people watched a bat conservation lecture their attitudes towards and knowledge about bats improved. These findings indicates that improved public education about conservation and nature could go far to bolster public engagement and promote positve sentiments towards wildlife.
8 Conclusion
The purpose of this special issue has been to present an early group of studies that have measured the conservation and ecological impacts of the pandemic and associated lockdown. These studies primarily compare systems before and during the pandemic, and as such represent valuable initial studies of natural systems. Many of these systems will continue to be monitored as the pandemic ends, and follow-up studies will document recovery back to pre-pandemic conditions. It is also true that some systems will take years to return to normal, and some systems will not ever return to their pre-pandemic state. Prominent conservation biologist Lovejoy (2021) expresses a nuanced perspective and cautious optimism as he points out the connections between the COVID-19 pandemic, climate change, and the protection of biodiversity. He reminds us that nature is resilient and can recover if we give it a chance.
We plan to organize a second Special Issue later in 2021/2020 of papers that take a more thorough and long-term perspective of the conservation impacts of the pandemic lockdown, including impacts on biodiversity itself as well as management, monitoring, education, training, community science networks, and ecotourism.
Declaration of competing interest
There is no conflict of interest regarding this paper.
Acknowledgments
10.13039/501100001804 Canada Research Chairs program.
==== Refs
References
Aditya V. Goswami R. Mendis A. Roopa R. Scale of the issue: mapping the impact of the COVID-19 lockdown on pangolin trade across India Biol. Conserv. 257 2021 109136 10.1016/j.biocon.2021.109136
Basile M. Russo L.F. Russo V.G. Senese A. Bernardo N. Birds seen and not seen during the COVID-19 pandemic: the impact of lockdown measures on citizen science bird observations Biol. Conserv. 256 2021 109079 34580546
Bates A.E. Primack R.B. Moraga P. Duarte C.M. COVID-19 pandemic and associated lockdown as a “global human confinement experiment” to investigate biodiversity conservation Biol. Conserv. 248 2020 108665 32549587
Bates A.E. Primack R.B. Duarte C.M. PAN-Environment Working Group Global COVID-19 lockdown highlights humans as both threats and custodians of the environment Biol. Conserv. 2021 109175 10.1016/j.biocon.2021.109175
Bíl M. Andrášik R. Cícha V. Arnon A. Kruuse M. Langbein J. Náhlik A. Niemi M. Pokorny B. Colino-Rabanal J. Rolandsen C.M. Seiler A. COVID-19 related travel restrictions prevented numerous wildlife deaths on roads: a comparative analysis of results from 11 countries Biol. Conserv. 256 2021 109076 34580545
Blount D. Chynoweth M.W. Green A.M. Sekercioğlu Ç.H. Review: COVID-19 highlights the importance of camera traps for wildlife conservation research and management Biol. Conserv. 256 2021 108984 10.1016/j.biocon.2021.108984
Borzée A. Kielgast J. Wren S. Angulo A. Chen S. Magellan K. Messenger K.R. Hansen-Hendrikx C.M. Baker A. Santos M.M.D. Kusrini M. Jiang J. Maslova I.V. Das I. Park D. Bickford D. Murphy R.W. Che J. Van Do T. Nguyen T.Q. Chuang M.F. Bishop P.J. Using the 2020 global pandemic as a springboard to highlight the need for amphibian conservation in eastern Asia Biol. Conserv. 255 2021 108973 35125500
China V. Zvuloni A. Roll U. Belmaker J. Reduced human activity in shallow reefs during the COVID-19 pandemic increases fish evenness Biol. Conserv. 257 2021 109103 10.1016/j.biocon.2021.109103
Coll M. Ortega-Cerdà M. Mascarell-Rocher Y. Ecological and economic effects of COVID-19 in marine fisheries from the Northwestern Mediterranean Sea Biol. Conserv. 255 2021 108997 33558776
Cooke S.J. Twardek W.M. Lynch A.J. Cowx I.G. Olden J.D. Funge-Smith S. Lorenzen K. Arlinghaus R. Chen Y. Weyl O.L.F. Nyboer E.A. Pompeu P.S. Carlson S.M. Koehn J.D. Pinder A.C. Raghavan R. Phang S. Koning A.A. Taylor W.W. Bartley D. Britton J.R. A global perspective on the influence of the COVID-19 pandemic on freshwater fish biodiversity Biol. Conserv. 253 2021 108932
Corlett R.A. Primack R.B. Devictor V. Maas B. Goswami V.R. Bates A.E. Koh L.P. Regan T.J. Loyola R. Pakeman R.J. Cumming G.S. Pidgeon A. Johns D. Roth R. Impacts of the coronavirus pandemic on biodiversity conservation Biol. Conserv. 246 2020 108571 32292203
Crimmins T.M. Posthumus E. Schaffer S. Prudic K.L. COVID-19 impacts on participation in large scale biodiversity-themed community science projects in the United States Biol. Conserv. 256 2021 109017
Dobson A.P. Pimm S.L. Hannah L. Kaufman L. Ahumada J.A. Ando A.W. Bernstein A. Busch J. Daszak P. Engelmann J. Kinnaird M.F. Li B.V. Loch-Temzelides T. Lovejoy T. Nowak L. Roehrkranz P.R. Vale M.M. Ecology and economics for pandemic prevention Science 369 2020 379 381 32703868
Driessen M.M. COVID-19 restrictions provide a brief respite from the wildlife roadkill toll Biol. Conserv. 256 2021 109012 33612848
Gilby B.L. Henderson C.J. Olds A.D. Ballantyne J.A. Bingham E.L. Elliott B.B. Jones T.R. Kimber O. Mosman J.D. Schlacher T.A. Potentially negative ecological consequences of animal redistribution on beaches during COVID-19 lockdown Biol. Conserv. 253 2021 108926
Hentati-Sundberg J. Berglund P.A. Hedjströmb A. Olsson O. COVID-19 lockdown reveals tourists as seabird guardians Biol. Conserv. 254 2021 e108950
Hochachka W.M. Alonso H. Gutiérrez-Expósito C. Miller E. Johnston A. Regional variation in the impacts of the COVID-19 pandemic on the quantity and quality of data collected by the project eBird Biol. Conserv. 254 2021 108974 34629475
Huang Q. Wang F. Yang H. Valitutto M. Songer M. Will the COVID-19 outbreak be a turning point for China’s wildlife protection: new developments and challenges of wildlife conservation in China Biol. Conserv. 254 2021 108937 33518771
Huveneers C. Jaine F.R.A. Barnett A. Butcher P.A. Clarke T.M. Currey-Randal L.M. Dwyer R.G. Ferreira L.C. Gleiss A.C. Hoenner X. Ierodiaconou D. Lédée E.J.I. Meekan M.G. Pederson H. Rizzari J.R. van Ruth P.D. Semmens J.M. Taylor M.D. Udyawer V. Walsh P. Heupel M.R. Harcourt R. The power of national acoustic tracking networks to assess the impacts of human activity on marine organisms during the COVID-19 pandemic Biol. Conserv. 256 2021 108995 34580542
Kishimoto K. Kobori H. COVID-19 pandemic drives changes in participation in citizen science project “city nature challenge” in Tokyo Biol. Conserv. 255 2021 109001 34565806
LeTourneux F. Grandmont T. Dulude-de Broin F. Martin M.C. Lefebvre J. Kato A. Bêty J. Gauthier G. Legagneux P. COVID19-induced reduction in human disturbance enhances fattening of an overabundant goose species Biol. Conserv. 255 2021 e108968
Lovejoy T.E. Nature, COVID-19, disease prevention, and climate change Biol. Conserv. 2021 (in press)
Lu M. Wang X. Ye H. Wang H. Qui S. Zhang H. Liu Y. Luo J. Feng J. Does public fear that bats spread COVID-19 jeopardize bat conservation? Biol. Conserv. 254 2021 108952 33518772
Miller-Rushing A.J. Athearn N. Blackford T. Brigham C. Cohen L. Cole-Will R. Edgar T. Ellwood E.R. Fisichelli N. Pritz C.F. Gallinat A.S. Gibson A. Hubbard A. McLane S. Nydick K. Primack R.B. Sachs S. Super P.E. COVID-19 pandemic impacts on conservation research, management, and public engagement in US national parks Biol. Conserv. 2021 (in press)
Primack R.B. Terry C. New social trails made during the pandemic increase fragmentation of an urban protected area Biol. Conserv. 255 2021 108993
Quesada-Rodríguez C. Orientale C. Diaz-Orozco J. Sellés-Rios B. Impact of 2020 COVID-19 lockdown on environmental education and leatherback sea turtle (Dermochelys coriacea) nesting monitoring in Pacuare Reserve, Costa Rica Biol. Conserv. 255 2021 10898
Ramvilas G. Dhyani S. Kumar B. Sinha N. Raghavan R. Selvaraj G. Dikavar N. Anoop V.K. Shalu K. Sinha A. Kulkarni A. Das S. Molur S. Insights on COVID-19 impacts, challenges, and opportunities for India’s biodiversity research: from complexity to building adaptations Biol. Conserv. 255 2021 109003 35125501
Rutz C. Loretto M.-C. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R.B. Ropert-Coudert Y. Tucker M.A. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evol. 4 2020 1156 1159 32572222
Sánchez-Clavijo L.M. Martínez-Callejas S.J. Acevedo-Charry O. Diaz-Pulido A. Gómez-Valencia B. Ocampo-Peñuela N. Ocampo D. Olaya-Rodríguez M.H. Rey-Velasco J.C. Soto-Vargas C. Ochoa-Quintero J.M. Differential reporting of biodiversity in two citizen science platforms during COVID-19 lockdown in Colombia Biol. Conserv. 256 2021 109077 35702146
Shilling F. Nguyen T. Saleh M. Khant Kyaw M. Tapia K. Trujillo G. Bejarano M. Waetjen D. Peterson J. Kalisz G. Sejour R. Croston S. Ham E. A reprieve from US wildlife mortality on roads during the COVID-19 pandemic Biol. Conserv. 256 2021 109013 34580543
Smith M.K.S. Smit I.P.J. Swemmer L.K. Mokhatla M.M. Freitag S. Roux D.J. Dziba L. Sustainability of protected areas: vulnerabilities and opportunities as revealed by COVID-19 in a national park management agency Biol. Conserv. 255 2021 108985
Soh M.C.K. Pang R.Y.T. Ng B.X.K. Lee B.P.Y.-H. Loo A.H.B. Er K.B.H. Restricted human activities shift the foraging strategies of feral pigeons (Columba livia) and three other commensal bird species Biol. Conserv. 253 2021 108927
Soto E.H. Botero C.M. Milanés C.B. Rodríguez-Santiago A. Palacios-Moreno M. Díaz-Ferguson E. Velázquez Y.R. Abbehusen A. Guerra-Castro E. Simoes N. Muciño-Reyes M. Souza Filho J.R. How does the beach ecosystem change without tourists during COVID-19? Biol. Conserv. 255 2021 108972
Souza C.N. Rodrigues A.C. Correia R.A. Normande I.C. Costa H.C.M. Guedes-Santos J. Malhado A.C.M. Carvalho A.R. Ladle R.J. No visit, no interest: how COVID-19 has affected public interest in world’s national parks Biol. Conserv. 256 2021 109015
Sumasgutner P. Buij R. McClure C.J.W. Shaw P. Dykstra C.R. Kumar N. Rutz C. Raptor research during the COVID-19 pandemic provides invaluable opportunities for conservation biology Biol. Conserv. 2021 (in press)
Terry C. Rothendler M. Zipf L. Dietze M.C. Primack R.B. Effects of the COVID-19 pandemic on noise pollution in three protected areas in metropolitan Boston (USA) Biol. Conserv. 256 2021 109039 34580544
Ulloa J.S. Hernández-Palma A. Acevedo-Charry O. Gómez-Valencia B. Cruz-Rodríguez C. Herrera-Varón Y. Roa M. Rodríguez-Buriticá S. Ochoa-Quintero J.M. Listening to cities during the COVID-19 lockdown: how do human activities and urbanization impact soundscapes in Colombia? Biol. Conserv. 255 2021 108996
Vale M.M. Berenguer E. Argollo de Menezes M. Viveiros de Castro E.B. Pugliese de Sigueira L. Portela R.C.Q. The COVID-19 pademic as an opportunity to weaken environmental protection in Brazil Biol. Conserv. 255 2021 108994 33568834
Vardi R. Berger-Tal O. Roll U. iNaturalist insights illuminate COVID-19 effects on large mammals in urban centers Biol. Conserv. 254 2021 108953 33424027
| 0 | PMC9746885 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 Aug 1; 260:109204 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.109204 | oa_other |
==== Front
Appl Acoust
Appl Acoust
Applied Acoustics. Acoustique Applique. Angewandte Akustik
0003-682X
1872-910X
The Author(s). Published by Elsevier Ltd.
S0003-682X(21)00399-6
10.1016/j.apacoust.2021.108305
108305
Article
Indoor soundscapes at home during the COVID-19 lockdown in London – Part I: Associations between the perception of the acoustic environment, occupantś activity and well-being
Torresin Simone ab⁎
Albatici Rossano a
Aletta Francesco c
Babich Francesco b
Oberman Tin c
Stawinoga Agnieszka Elzbieta d
Kang Jian c
a Department of Civil Environmental and Mechanical Engineering, University of Trento, Italy
b Institute for Renewable Energy, Eurac Research, Bozen/Bolzano, Italy
c UCL Institute for Environmental Design and Engineering, The Bartlett, University College London, London, UK
d Management and Committees, Eurac Research, Bozen/Bolzano, Italy
⁎ Corresponding author at: Department of Civil Environmental and Mechanical Engineering, University of Trento, Italy.
20 7 2021
1 12 2021
20 7 2021
183 108305108305
21 4 2021
8 6 2021
11 7 2021
© 2021 The Author(s)
2021
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Since the outbreak of the COVID-19 pandemic, as a result of the adoption of worldwide lockdown measures, the home environment has become the place where all the daily activities are taking place for many people. In these changed social and acoustical contexts, we wanted to evaluate the perception of the indoor acoustic environment in relation to traditional and new activities performed at home, i.e., relaxation, and working from home (WFH). Taking London as a case study, the present paper presents the results of an online survey administered to 464 home workers in January 2021. The survey utilized a previously developed model for the assessment of indoor soundscapes to describe the affective responses to the acoustic environments in a perceptual space defined by comfort (i.e. how comfortable or annoying the environment was judged) and content (i.e., how saturated the environment is with events and sounds) dimensions. A mixed-method approach was adopted to reinforce result validity by triangulating data from questionnaires and spontaneous descriptions given by participants. In this first part of the study, the main objectives were: (1) evaluating differences in soundscape evaluation, in terms of comfort and content dimensions, based on the activity performed at home, (2) identifying appropriate conditions for WFH and relaxation, and (3) investigating associations between psychological well-being and indoor soundscapes. The results showed that the environments were perceived as more comfortable and slightly fuller of content when rated in relation to relaxation than for WFH, thus suggesting a stricter evaluation of the acoustic environment in the latter case. As regards the second objective, spaces that were more appropriate for relaxation had high comfort, whereas spaces appropriate for WFH resulted more private and under control, i.e. with high comfort and low content scores. Lastly, better psychological well-being was associated with more comfortable soundscapes, both for WFH (rs = 0.346, p < .0005), and relaxation (rs = 0.353, p < .0005), and with lower content while WFH (rs = −0.133, p = .004). The discussion points out the need of considering the implications of changed working patterns to rethink the design of soundscapes in residential buildings, also in relation to potential well-being outcomes that will be further investigated in the Part II of the study.
Keywords
Indoor soundscape
Indoor environmental quality
Acoustic design
Well-being
COVID-19
WFH
==== Body
pmc1 Introduction
Since the global outbreak of the SARS-CoV-2 virus (COVID-19 disease), severe lockdown measures have been adopted by governments all around the world to prevent the spread of the virus. Stay-at-home mandates have transformed houses into places where to spend the entire day while working, home-schooling, taking care of families, nourishing, training, socializing, and finally resting.
Several studies have been reporting on the changed outdoor and indoor acoustic contexts during the COVID-19 pandemic and on the impacts of this unprecedented acoustic scenario on people. A general reduction of outdoor noise levels has been observed due to contingency measures in several countries [1], [2], [3], [4], [5], [6], [7], [8]. In terms of noise annoyance, studies have showed a persistence [9] or reduction [10], [11], [12] in annoyance due to outdoor noise, a reduction [11], persistence [10] or increase [9], [12] in annoyance towards neighbourś noise, and an increase in annoyance due to noise from onés own dwelling [9], [10] during the lockdown compared to before.
In this sudden acceleration towards remote working and schooling, houses had to host new social functions and to face a new set of challenges, making even more evident now than before the importance of housing to the physical and mental health of building occupants [13]. The indoor soundscape literature has recently stressed the opportunity given by acoustic design to improve health, well-being, and quality of life of building occupants through a perception-based approach, thus making a step forward compared to a mere noise annoyance reduction [14]. Being the perception of the acoustic environment (i.e. the soundscape [15]) highly context dependent, the question is thus how the acoustic conditions at home can support home activities in this changed pandemic context, while ensuring the well-being of the occupants.
With the purpose to address this question, an online survey has been conducted targeting home workers living in UK (London) and Italy. Differently from other studies that have focused on the noise – annoyance binomial during the pandemic (i.e., how much were you annoyed by these noise sources?), the present survey has been designed from a soundscape perspective in order to explore both positive and negative effects of sounds and noises, depending on the specific task performed at home during the lockdown. Questions about working typology, housing features, person-related traits, living and urban contexts have been complemented with questions from ISO soundscape standards [16], tailored to address the peculiarities of indoor residential spaces [17]. In particular, the survey evaluated the perceived affective quality of the acoustic environment through an indoor soundscape assessment model developed by the authors [18], building on the existing outdoor soundscape model by Axelsson et al. [19]. The model allows to describe the emotional response to indoor acoustic environments in a two-dimensional perceptual space (cfr. Fig. 1 ) where the orthogonal axes are comfort (an annoying – comfortable continuum) and content (an empty - full of content continuum). In the previous laboratory investigation [18] the two dimensions were found to explain together 83% of the total variance in the assessment made by test participants of 20 indoor acoustic environments on 97 attribute rating scales. According to this perceptual space, engaging indoor soundscapes are both comfortable and full of content, detached soundscapes are both annoying and empty, intrusive and uncontrolled soundscapes are annoying and full of content, and finally private and controlled indoor soundscapes are both comfortable and empty.Fig. 1 Model of affective response to indoor residential acoustic environments from [18]
The present survey constitutes the first field application of the model to evaluate the emotional reaction to the residential acoustic environment in relation to two specific activities performed at home during the lockdown, namely working from home (hereinafter WFH) and relaxing. Results are presented in two intertwined papers. In Part I, the main goals are to explore differences in the comfort/content rating of the built environment based on the performed activity, and to identify comfort/content combinations characterizing an acoustic environment perceived as appropriate for WFH and relaxing.
Moreover, we wanted to investigate the relationship between the perceived acoustic environment at home, described through the coordinates on the two comfort and content dimensions, and the psychological well-being of building occupants. Evidence on effects of sound exposure on quality of life, well-being and mental health is far from being conclusive due to a lack of longitudinal and interventional studies [20], [21]. Previous research in outdoor soundscape literature suggested a positive association between positively perceived soundscapes and faster stress-recovery processes, better self-reported health conditions and higher self-reported well-being [21], [22], [23], [24]. Poor noise conditions at home can result in increased prevalence of mental health symptoms [25] and impede the acquisition of psycho-social benefits from home [26]. For instance, traffic noise was found to cause emotional disorders and hyperactivity in children [20]. Spending the lockdown in houses with poor indoor environmental conditions (including acoustic discomfort) was found to be associated with a higher risk of moderate–severe and severe depressive symptoms [27]. In indoor spaces, natural sounds coming from the outdoor environment and unmasked by the drop in anthropogenic noise in cities during the lockdown [28], [29], [30] might provide improved cognitive performance and stress recovery [31], despites findings of restorative effects of natural sounds being still inconclusive [32]. Listening to music can be a coping strategy for dealing with stress, can involve emotion regulation and provide psychotherapeutic effects [33]. Indoor human sounds from family members can also have a restorative potential and activate cerebral functioning [34], but can also result highly disrupting when WFH during a pandemic [9]. In a study with university students during the COVID-19 lockdown, higher exposure to mechanical sounds was found to be associated with a lower restorative quality of home and a worse self-rated health, whereas nature sounds were positively associated with restorative quality, and in turn with better self-rated health [35]. Positive and negative outcomes from the acoustic environment might depend on the type of activity carried out at home [17] and on the degree of interference with the task with which people are engaged [36]. As such, associations between the two soundscape dimensions based on the perceived affective quality responses (comfort and content) and the subjective psychological well-being are investigated in relation to the two considered home activities, i.e. working and relaxing at home.
In the following, the analysis will focus on data gathered from the London sample. Research questions addressed in this paper include:1. RQ1. Is there a difference in the evaluation of a space depending on the activity in which the occupant is engaged?
2. RQ2. What are the comfort-content combinations, if any, that describe an environment being appropriate for WFH and relaxing at home?
3. RQ3. Are comfort and content related to occupant́s psychological well-being?
2 Methods
2.1 Participants
An online survey was administered to adult participants via Prolific participant pool [37], [38] on 18 and 19 January 2021, while London was in a lockdown condition [39]. Potential participants were filtered through the following prescreening criteria available in the platform: age (18 – 65 years old), no self-reported hearing difficulties, indicating London (UK) as area of residence, and WFH during the COVID-19 lockdown. After excluding 9 participants that failed an attention check included in the survey (cf. Q26 – Appendix A), 464 participants (181 males, 282 females, 1 other; mean age: 32.2 years; SD: 9.1 years) were considered for the data analysis. The survey took on average 29 min to complete and participants were offered a small monetary compensation as a token of appreciation for their time. The study was approved via the UCL IEDE Ethics departmental low-risk procedure on November 26th, 2020.
2.2 Questionnaire design
Study data from the online survey were collected and managed using REDCap electronic data capture tools hosted at University College London (UCL) [40], [41]. An excerpt of the questionnaire used for the online survey is provided in Appendix A where only the questions that are relevant to the present study (Part I and Part II) have been reported. The questionnaire included both closed- and open-ended questions. Given the complex and multi-faceted nature of soundscape investigations, a mixed-method approach was adopted [42] by complementing quantitative data from closed-ended questions with data from open-ended questions, according to the principle of triangulation commonly applied in behavioral and social sciences [43]. As depicted in Fig. 2 , the questionnaire was made of an introductory section including the information sheet and consent form and five more sections focusing on: (1) the WFH activity; (2) leisure activities performed at home; (3) housing features; (4) the urban context; and (5) person-related characteristics.Fig. 2 Schematic representation of the main sections composing the questionnaire. Questions are listed in Appendix A.
In the first section, participants were asked to provide information about the type of work performed at home (Q1), by rating the importance of several activities (e.g., online meetings, reading, thinking/creative thinking). The frequency of headphone use when WFH was assessed in Q2. Participants had then to indicate a room that could be relevant for their WFH activity (Q3), and to describe the dominance of several categories of sounds as perceived in this room (Q4). The question was adapted from the ISO/TS 12913-2 (Method A) [16], by including the following sound sources relevant for indoor soundscapes: traffic noise, other noise from outside (e.g. sirens, construction, industry, loading of goods), natural sounds, human beings outside, other human beings present at home, neighbours, building services at home, building services of neighbours and common areas, and music or TV played by participants themselves. The type of window view from the room where WFH was assessed in Q5, by rating the dominance of vegetation, sky and other buildings when looking outside. A spontaneous description of positive and negative effects of sound exposure while WFH was collected through an open-ended question (Q6). This was done in order not to constraint answers into pre-defined categories and to collect information about aspects that might have not been covered by the researchers in the rest of the questionnaire. The appropriateness of the surrounding sound environment for WFH was evaluated in Q7, by adapting the corresponding question included in the ISO/TS 12913-2 (Method A) [16]. Perceived affective quality responses were collected in Q8. This part was adapted from the ISO/TS 12913–2 (Method A) [16], by using the eight attributes derived in [18] (i.e., Comfortable; Intrusive, uncontrolled; Engaging; Empty; Private, controlled; Annoying; Full of content; Detached).
Similarly, in the second section, participants were asked to indicate a room relevant for relaxation activities (Q9), which could therefore be different from the one used for WFH, and to answer consequently about sounds heard (Q10), the components of view from window (Q11), the positive and negative effects on relaxation (Q12), the appropriateness of the sound environment (Q13), and the affective response to the acoustic environment while relaxing (Q14). Open-ended questions investigated the impact on several leisure activities. In the present paper, only responses related to watching TV, reading, and listening to music will be considered.
In the third section, information about the housing context were collected and specifically on: the ownership status (Q15), the house size (Q16), the house typology (Q17, i.e. detached single family, semi-detached or terraced house, apartment block), room exposure to quiet or noisy areas (Q18), other people at home (Q19, e.g., roommates, children), the number of people living at home (Q20), and the building service typologies for ventilation (Q21), heating (Q22), and cooling (Q23).
The fourth section was related to the urban context. Participants were asked to provide their postcode (Q24) and to describe the urban area where they live (Q25, i.e. urban, suburban, or rural).
In the last section, person-related information was collected. Noise sensitivity was assessed through a reduced number of items (Q26) extracted from the Weinstein’s Noise Sensitivity Scale [44], which is consistent in providing a user profile similar to that of the full scale [45]. Subjective psychological well-being was evaluated through the WHO-5 (Q27) well-being index [46]. The WHO-5 is based on five questions having as time frame the previous two weeks and it has been found to have adequate validity in screening for depression [46]. Lastly, demographic information about age (Q28) and gender (Q29) were collected.
2.3 Data analysis
Statistical analyses were run in IBM SPSS Statistics 26 [47] and in R [48], while qualitative analyses have been conducted in NVivo 12 software.
2.3.1 Projection of the perceptual attribute dimensions onto the circumplex model
The scores derived from the assessment of indoor soundscapes on the eight attributes for the two investigated activities, WFH (Q8) and relaxation (Q14), were processed to be represented as points into circumplex models with coordinates for the two dimensions comfort and content, following the procedure described in ISO/TS 12913–3 [43]: while no standard or technical specifications currently exist for indoor soundscape [49], we felt it would be sensible to process the data by analogy as per the recommendations of Part 3 of the ISO 12913 series [43]. Every data point represents the assessment by one participant, in relation to the activity being investigated. The coordinates for comfort and content are calculated as:Comfort=c-a+cos45°∙pc-iu+cos45°∙en-d
Content=f-em+cos45°∙iu-pc+cos45°∙en-d
where a is annoying, c is comfortable, d is detached, em is empty, en is engaging, f is full of content, iu is intrusive - uncontrolled, and pc is private, controlled. The coordinates are divided by (4+√ 32) to scale the resulting values between −1 and +1.
2.3.2 Assessing associations and differences between groups
As assessed by Shapiro-Wilk's test, content scores resulted normally distributed, (p > .05), while well-being values from the WHO-5 index and comfort scores failed to meet normality assumptions (p < .05). As such, non-parametric tests were used to analyze data. Associations were assessed through Spearman's rank-order correlation. Kruskal-Wallis tests were run to evaluate differences between independent groups while the Wilcoxon signed-rank test was used to determine whether there was a median difference between paired observations (i.e., responses given by same participants). Please notice that for a Wilcoxon signed-rank test, the median difference is obtained as the median of the differences between the paired values and not as the difference of the medians of the two groups. The statistical significance threshold was set at 0.05.
2.3.3 Qualitative analysis
The material collected by the open-ended questions was coded in NVivo 12 software, organizing the excerpts according to patterns of semantic content via constant comparisons of data [50]. Only verbal descriptors with more than five occurrences have been retained.
3 Results
Frequency distributions were computed to explore categorical and ordinal variables (cf. Fig. 3 ). The type of work carried out at home was mainly individual, desk based, focused work, whereas online meetings, telephone conversations, reading, and creative thinking were reported as the most relevant activities. Working from home was to a lesser extent characterized by individual focused work away from the desk, by the use of technical equipment and the reception of visitors, clients, or customers (Fig. 3 a). The majority of respondents (73.9%) reported using headphones at least sometimes while WFH (Fig. 3 b). Bedrooms (41.6%) and living rooms (41.6%, also considering open spaces kitchen-living rooms) were the most used spaces where WFH (Fig. 3 c). As shown in Fig. 3 d, relaxation took place in living rooms (62.5%, if also considering open spaces kitchen-living rooms), followed by bedrooms (33.2%). Having in mind the ambiguity of the term ‘studio’, there is a possibility that while answering to questions Q3 and Q9 both participants living in a ‘studio flat’ and having a ‘separate study space’ in their house have answered in this category (cf. Fig. 3 c and d). However, it should be noticed that those data have only be used to derive information about the exposure to a quiet or noisy side in combination with data from Q18, as reported in the following, and therefore this possible misunderstanding had no impact on the following analyses.Fig. 3 Absolute and percentage values of responses (N = 464) to: (a) relevance of different activities to WFH (Q1.1 – Q1.8); (b) frequency of headphone use while WFH (Q2); room used for WFH (c - Q3) and for relaxation (d - Q9); perceived dominance of sounds while WFH (e – Q4) and while relaxing (f – Q10); perceived dominance of different components from the window view in the room chosen for WFH (g – Q5) and for relaxation (h – Q11); perceived appropriateness of the sound environment for WFH (i – Q7) and for relaxation (j – Q13); (k) ownership status (Q15); (l) house size (Q16); (m) housing typology (Q17); (n) people living with (recoded from Q19); (o) number of people at home (Q20); (p) type of urban area (Q25). Labels for categories having less than 10 occurrences have been omitted.
The sound environment while WFH and during relaxation was reported to be dominated by music or TV played by the respondents themselves, by sound generated by other human beings present at home, followed by outdoor sounds and neighbours, as detailed in Fig. 3 e-f. The view from windows in rooms employed for working and relaxing was most often dominated by the view of sky and other buildings, and to a lesser extent by vegetation (Fig. 3 g-h). As regards soundscape appropriateness, the sound environment was evaluated as very and perfectly appropriate for working (56.5%) and relaxing (62.2%) by the majority of respondents.
Most participants do not own the house they live in (55.5%, Fig. 3 k) and almost half of the dwellings (47.4%) have a surface area ranging between 40 and 80 m2 (Fig. 3 l). Semi-detached or terraced houses (42.5%) and apartment blocks (42.5%) were the most common housing typologies (Fig. 3 m). The spaces where people worked and relaxed at home overlooked urban areas described as quiet (respectively, 52.6 and 51.1%) or as noisy (45.9 and 48.3%, 1.5 and 0.6% missing). The majority of respondents lived with someone else but without children (65.7%), followed by those living also with children (22.0%) and by those living alone (12.3%, cf. Fig. 3 n). In most cases the household consisted of two (37.3%) or three (23.7%) people (cf. Fig. 3 o).
As regards the building services, 13.8%, 5.6%, and 3.2% of respondentś dwellings were equipped respectively with air-systems for ventilation, cooling and heating (e.g., HVAC systems, mechanical ventilation, air conditioners).
Houses were mainly located in areas described as urban (68.1%, Fig. 3 p). The map of the city of London depicting the residential areas of respondents is reported in Fig. 4 and it shows a good coverage of the city by the survey.Fig. 4 Area of residence of respondents in London. 3% of postcodes were missing.
The WHO-5 well-being index averaged 53.7 ± 19.24 (Mean ± SD), with 100 representing the best quality of life. Noise sensitivity index scored on average 64.19 ± 19.25 (Mean ± SD), with higher scores denoting higher sensitivity to noise.
In the following, results related to the three research questions are presented (sections 3.1 – 3.3), with reference to quantitative and qualitative analyses.
3.1 Difference in soundscape evaluation based on the activity
3.1.1 Quantitative analysis
A Wilcoxon signed-rank test was conducted to determine whether there was a median difference in comfort, content, and perceived soundscape appropriateness when spaces were considered in relation to the two different activities (i.e., working or relaxing). Data in this section are medians unless otherwise stated. There was a significant increase in comfort, (median difference: 0.07), z = 5.895, p < .0005, and content, (median difference: 0.06), z = 6.259, p < .0005, when evaluated for relaxation compared to WFH, while the evaluation of soundscape appropriateness was not significantly different between the two activities, z = 1.658, p = .097. However, it must be noted that the respondents rated the soundscapes for the rooms most often employed to perform those activities and that the chosen rooms might have been different. As such, the dissimilar characteristics of the rooms (and related acoustic features) employed for working and relaxing might have confounded the observed differences in comfort and content evaluations.
Therefore, a further test was run on the subgroup of respondents that chose the same room for both WFH and relaxation (N = 212). Results confirm a significant increase in comfort (median difference: 0.07), z = 4.401, p < .0005, and content (median difference: 0.03), z = 3.093, p = .002, of the same space when evaluated for relaxation compared to when evaluated for WFH (cf. Fig. 5 ). Of the 212 participants that selected the same room for both WFH and relaxation, 118 evaluated the environment as more comfortable for relaxing, 67 evaluated the environment as more comfortable for WFH, while 27 expressed no difference in comfort evaluation. As regards content scores, 110 respondents expressed higher content while relaxing, 77 expressed higher content while WFH, while 25 respondents expressed no difference. In terms of soundscapes appropriateness, 96 participants expressed no difference, 75 evaluated the sound environment as better for relaxing, whereas 41 respondents evaluated the sound environment as better for working.Fig. 5 Boxplots of comfort and content scores by type of the activity performed: working from home (WFH) and relaxation (REL). Data refer to those respondents (N: 212) that chose the same room for both WFH and relaxation. Inside the boxes, the central line is the median value, n.s.: not significant, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001.
3.1.2 Qualitative analysis
Table 1 presents the main themes extracted from the analysis of the questions “In your view, how is the sound environment currently (positively and negatively) affecting your working activity from home?” and “In your view, how is the sound environment currently (positively and negatively) affecting your leisure activities at home? (While watching TV, reading, listening to music)”. While the evaluation of specific sound sources is reported in Part II of the study, here the focus is on general themes that might help confirming and understanding the differences in soundscape evaluation based on the performed activity.Table 1 Main themes extracted from open-ended questions on the impacts of the sound environment on WFH (Q6) and relaxation (Q12).
Working from home (N = 274) Relaxing at home (N = 282)
Distracting, disrupting (41%) Music, TV, (noise cancelling) headphones mask other noises (28%)
Hearing sounds is beneficial while working (23%) Distracting, disrupting (20%)
Music, TV, (noise cancelling) headphones mask other noises (16%) Noises are present but provide no impact (11%)
Habituation to the sound environment (9%) Annoying when reading (12%)
Block out noise while concentrating (3%) Doing it when less noisy (8%)
TV, music played by themselves self-distracting (3%) Having to increase the volume (7%)
Lack of control (2%) Hearing sounds is beneficial while relaxing (6%)
Closing windows (2%) Habituation to the sound environment (6%)
Closing doors (2%) Closing windows (2%)
As regards the impact on WFH, participants reported the sound environment often being distracting and disruptive to their work (N = 112). The lack of control over urban noises or the noises from neighbours were mentioned as a cause of annoyance and frustration (N = 6). In addition to closing windows (N = 5) and doors (N = 5) because of noise from outside or inside the building, participants resort to listening to music, TV and wearing noise-cancelling headphones (N = 43) to help concentration, provide the wanted background and drown out unwanted sounds. However, podcasts, music and TV played by respondents could also provide further sources of distraction (N = 7). Some of the respondents mentioned they had become accustomed to their surrounding sound environment over time, thus indicating habituation to the acoustic conditions they were exposed to (N = 26). Others could simply “block out noise” and isolate themselves while concentrating on the activity they were engaged with (N = 8). Many respondents (N = 62) expressed the beneficial effect of listening to sounds in the background compared to having a completely silent environment, as this could help them feel less lonely, provide some contact with the outside world, relax and feel comforted by the sounds of the family.
Acoustic conditions can be detrimental also for leisure activities performed at home (N = 56). Respondents reported the need to turn up the volume of TV or music because of a noisy environment (N = 20) or to close the windows because of noise from outside (N = 5). Poor acoustic conditions can be problematic when reading, as this activity does not provide any opportunity for masking the background noise (N = 34). However, participants generally reported being less affected while relaxing than during work, because the acoustic environment is usually overpowered by the sound of TV and music (N = 79). Leisure activities are often carried out in the evening when the road traffic is reduced and construction works have stopped, thus resulting in a quieter environment (N = 22). For some of the respondents listening to noise is not an issue while relaxing (N = 32), or they simply got used to it over time (N = 16). The acoustic stimuli can even be beneficial (N = 18), as the sound environment can provide sources of distraction that can be conductive to relaxation, to feeling connected, comforted and less alone, as rendered in the following excerpts:“If I have too much quiet then it gives me too much opportunity to think. So, in order to relax, I need my brain to be occupied with something else.”
“The sounds of the street are comforting to me when I relax, I feel at home.”
“I like hearing outside noises to keep me feeling connected”
“The noises make you feel like you aren't alone.”
3.2 Comfort-content combinations in relation to soundscape appropriateness
Affective responses to the indoor acoustic environments have been represented in the perceptual space defined by comfort and content dimensions according to the procedure described in par. 2.3.1. The result is given in Fig. 6 , where each data point represents the soundscape assessment by one participant in the room employed for WFH (Fig. 6 a) and for relaxing (Fig. 6 b). In the scatter plot, data points have been grouped by the perceived appropriateness of the acoustic environment to working and relaxing at home (3 categories: not at all & slightly; moderately; very & perfectly). Indoor soundscapes rated as more appropriate for WFH (Kruskal-Wallis, comfort: χ2(2) = 195.844, p < .0005; content: χ2(2) = 86.827, p < .0005) and for relaxation (Kruskal-Wallis, comfort: χ2(2) = 168.699, p < .0005; content: χ2(2) = 47.824, p < .0005) were characterized by significantly higher comfort and significantly lower content scores than those judged as inappropriate for the two activities. Differences in soundscape evaluation across the two activities are investigated in the next section.Fig. 6 Projection of the affective responses to the indoor acoustic environment onto the bidimensional circumplex model defined by the comfort – content dimensions. Points are grouped by soundscape appropriateness to (a) working and (b) relaxing at home. Crosses depict the centroids of the different groups.
3.3 Association between comfort, content and well-being
A Spearman's rank-order correlation was run to assess the relationship between comfort scores, content scores and psychological well-being. Results showed a statistically significant, moderate positive correlation between comfort and well-being for both WFH, rs = 0.346, p < .0005, and relaxation, rs = 0.353, p < .0005 (cf. Fig. 7 a). Comfortable acoustic environments were associated with higher psychological well-being. As regards content, there was a statistically significant, weak negative correlation between content and well-being when soundscape was evaluated for relaxation, rs = -0.133, p = .004 (cf. Fig. 7 b). Differently, the relationship between soundscape content when WFH and psychological well-being was not statistically significant.Fig. 7 Scatter plot of comfort (a) and content (b) scores by psychological well-being. Data are grouped for the two activities in relation to which the soundscape assessment was carried out, i.e. working from home (WFH) and relaxation (REL).
4 Discussion
The study presented the results of an online survey conducted in London with the purpose of exploring the relationships between indoor soundscapes, working and relaxing activities, and psychological well-being. Given the complex and multi-facet problems encountered in soundscape studies, methodological triangulation has been suggested [43] and often applied in previous research to reinforce result validity. In the following, the three main research questions underpinning the study are discussed by triangulating the results from rating scales with those from the qualitative analysis of free format responses.
4.1 RQ1. Is there a difference in the evaluation of a space depending on the activity in which the occupant is engaged?
Spaces were rated as more comfortable and more content-rich when considered for relaxing than for working from home. The result was confirmed also when selecting the subset of participants that indicated the same room both for relaxation and for WFH. The analysis of open-ended questions can help to shed light on the reasons behind the difference. Despite appropriateness ratings not being significantly different when the environment was evaluated for WFH or relaxation, the analysis of free-format answers showed that the sound environment has been reported to be generally less disruptive during relaxation activities involving listening to music and watching TV compared to WFH and reading. Results revealed that during relaxation, music and TV were reported to overpower the sound environment with sounds over which people had control. Leisure activities were often carried out at quieter times, when the road traffic was reduced, and construction works had ceased. The finding is also consistent with previous studies reporting noise annoyance to be related to the amount of task disruption by noise [17], [36], [51]. Overall, the analysis suggests that WFH was more heavily affected by the surrounding acoustic environment, which led to a more stringent assessment of the space in terms of comfort. The lack of difference in appropriateness ratings between relaxation and WFH might be due to the fact that the lockdown conditions blurred the distinction between the spaces for relaxation (typically the home environment) and for work (typically the office). During the pandemic, home became both the place where stress and restoration took place, and this might have resulted in participants not being able to distinguish the border between WFH and relaxation and to correctly report on space appropriateness in relation to those two conditions.
While several studies have addressed the impact of noise on valence-related dimensions (mostly annoyance) as a function on the activity at hand [17], [36], [51], [52], [53], [54], this was the first time the impact of acoustic conditions on a content dimension has been explored in indoor environments. Higher content scores while relaxing are likely due to the saturation of the residential space with music and TV sounds. On the other side, lower content scores while WFH might result from being focused on cognitively demanding tasks or from a higher use of headphones during remote working, in both cases leading to a higher isolation to the surrounding sound environment.
4.2 RQ2. What are the comfort-content combinations, if any, that describe an environment being appropriate for WFH and relaxing at home?
Indoor soundscapes perceived as more appropriate for WFH and for relaxation were characterized by higher comfort scores and lower content scores than those that were rated as inappropriate. By plotting the affective responses to the acoustic environments in the comfort – content space (cf. Fig. 6), it can be observed that environments perceived as more appropriate for WFH were mainly located in the quadrant of perceived privacy and control over the environment, characterized by high comfort and low content scores (cf. Fig. 1). This might be partially due to participants reporting lower content in relation to WFH than for relaxation, as observed in the previous section. Furthermore, WFH might require soundscapes that are more private and that are perceived as more controlled compared with relaxation. Indeed, indoor soundscapes appropriate to relaxation were more evenly positioned in the half-plane characterized by positive comfort scores, thus being either perceived as engaging or as private and under control.
Overall, results suggest that soundscapes characterized by positive valence might be adequate for relaxation, but not necessarily supportive for working if not coupled with low content.
4.3 RQ3. Are comfort and content related to occupantś psychological well-being?
The results showed a significant association between soundscapes characterized by positive valence in relation to home working and relaxation (i.e., comfortable, pleasant), and the psychological well-being of respondents (cf. Fig. 7 a), in line with the trends highlighted in the (outdoor) soundscape literature [21], [22], [23], [24]. As reported in the literature, access to a high-quality acoustic environment might elicit positive mental states in building occupants, thus fostering psychological resilience and reducing the risks of mental health problems [55]. Moreover, spaces with positive soundscapes can be beneficial for health and well-being by providing psychophysiological recovery from stressors [56]. Conversely, psychological issues might make people more susceptible to acoustic conditions, thus resulting in a more negative perception of the environment and a stronger need for high-quality acoustic conditions [57].
As regards content, a weak negative correlation was found between content scores and psychological well-being in relation to WFH (cf. Fig. 7 b). The association was not significant when considered for relaxation. Due to the observed need for a private and controlled soundscape for home working, high content can result in perceived disturbance and frustration that might induce mental health issues.
4.4 Limitations
Results presented in this study need to be interpreted considering some limitations. Firstly, due to the cross-sectional nature of the study we cannot draw causal claims about the observed associations. The study focused on two out of many activities that are performed at home (i.e., working and relaxing), not addressing scenarios involving people that are external to the house (e.g., temporary visitors), due to the lockdown situation. In a post-pandemic scenario, other factors might affect the perception of the acoustic environment when, for instance, hosting friends at home. While a private and under control soundscape was relevant to WFH, it is likely that a more engaging soundscape would be appropriate for a more convivial situation. Next, the study relied on self-reporting questionnaires, that can result in respondents misunderstanding or not correctly estimating and reporting the objects of investigation. This is particularly true when assessing the emotional response to the acoustic environments by self-reports, as affective qualities might not be accessible by individuals [58]. Furthermore, most of the topics included in the survey have been investigated with structured but not validated questionnaires due to the scant evidence published in the existing literature. Lastly, due to a lack of reference data, it cannot be assumed that the sample is statistically representative of Londoner homeworkers population. However, recruiting participants through an online research platform allowed to avoid some of the limitations of snowball sampling, such as collecting results from participants sharing the same background (e.g., researchers in the acoustic field). Moreover, it can be assumed that people working from home are digitally connected similarly to those engaged with online platforms, in order to being able to perform office work remotely.
5 Conclusions
The paper reported on the results of an online survey conducted on 464 home workers in London in January 2021 during the COVID-19 lockdown. The study constituted a first application of the indoor soundscape model [18] for the assessment of the acoustic environment in relation to two main activities performed at home during the pandemic, i.e. relaxing and home working. Evidence extracted from the analysis of data collected from rating scales and open-ended questions have been combined to increase results validity through methodological triangulation. The main findings are as follows:(1) Spaces were rated as more comfortable and more content-rich when considered for relaxation than for WFH. Despite the non-significant difference in soundscape appropriateness between relaxation and WFH, the more stringent assessment of the same space in terms of comfort and the analysis of free format responses suggest that WFH is more affected by the acoustic environment compared to relaxation.
(2) Indoor soundscapes perceived as more appropriate for WFH and for relaxation were characterized by higher comfort scores and lower content scores than those that were rated as inappropriate. Soundscapes that are appropriate for relaxation are characterized by (positive) comfort conditions, and can be both full of content (i.e., engaging) or empty (i.e., private and under control). Differently, spaces that are more appropriate to home working are comfortable but also tend to be poor in content; i.e., they are perceived as private and under control.
(3) Psychological well-being was positively associated with comfortable soundscapes both in relation to WFH, rs = 0.346, p < .0005, and relaxation, rs = 0.353, p < .0005. As regards content, a weak negative correlation was found between content scores and psychological well-being in relation to WFH, rs = -0.133, p = .004, but not for relaxation.
Evaluating the affective response to the indoor acoustic environment in the comfort – content space helped identifying conditions that were appropriate to home working, compared to leisure activities, reaching a more in-depth knowledge compared to appraisals based on annoyance evaluation. Notably, the new functions dwellings have been called to host since the COVID-19 outbreak, and that will likely last in the post-pandemic era, can make building occupants differently vulnerable to the acoustic conditions at home, and more demanding of high-quality acoustic environments.
In the second part of the study, the influence of acoustical, building, urban, and person-related factors on indoor soundscape perception and well-being will be investigated, thus allowing to explain part of the identified associations between perceived acoustic conditions and occupantś well-being.
Funding
This work was funded by the Chartered Institution of Building Services Engineers (CIBSE) within the project ‘Home as a place of rest and work: the ideal indoor soundscape during the Covid-19 pandemic and beyond’. This work was supported by the Programma di cooperazione Interreg V-A Italia-Svizzera 2014–2020, project QAES [ID no. 613474]; and the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme [grant agreement No. 740696].
CRediT authorship contribution statement
Simone Torresin: Conceptualization, Methodology, Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Visualization. Rossano Albatici: Conceptualization, Writing - review & editing. Francesco Aletta: Conceptualization, Methodology, Writing - review & editing. Francesco Babich: Conceptualization, Writing - review & editing. Tin Oberman: Conceptualization, Writing - review & editing. Agnieszka Elzbieta Stawinoga: Formal analysis, Writing - review & editing. Jian Kang: Conceptualization, Supervision, Funding acquisition, Writing - review & editing.
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A – Questionnaire excerpt
ID Question Scale Label
Working from home
Q1 “Please indicate how much each of these activities is relevant to your work from home.”
(Q1.1 Online meetings; Q1.2 Telephone conversations; Q1.3 Reading; Q1.4 Thinking/creative thinking; Q1.5 Individual focused work, desk based; Q1.6 Individual focused work away from your desk; Q1.7 Using technical/specialist equipment or material; Q1.8 Hosting visitors, clients or customers) Likert Not at all (1) – Really important (5); Not applicable (6)
Q2 “How often do you use headphones while working from home?” Likert Never (1) – Always (5)
Q3 “Now please focus on one room that is relevant for your working activity at home:” – Studio; Kitchen; Living room; Kitchen – living room, Bedroom
Q4 “To what extent do you hear the following types of sounds while working from home in your [piping: Q3]?”
(Q4.1 Traffic noise from outside - e.g. cars, buses, trains, airplanes; Q4.2 Other noise from outside - e.g. sirens, construction, industry, loading of goods; Q4.3 Natural sounds from outside - e.g., singing birds, flowing water, wind in vegetation; Q4.4 Sounds from human beings from outside - e.g. conversation, laughter, children at play, footsteps; Q4.5 Sounds from other human beings present in your house - e.g. conversation, music, TV, laughter, children at play, footsteps; Q4.6 Sounds from neighbors - e.g. conversation, music, TV, laughter, children at play, footsteps; Q4.7 Sounds from building services of your house - e.g. heating, cooling, ventilation systems, toilet flushes; Q4.8 Sounds from building services of your neighbours / common areas - e.g. heating, cooling, ventilation systems, let flushes, lift; Q4.9 Music or TV played by you - through headphones or loudspeakers) Likert Not at all (1) – Dominates completely (5); Not applicable (6)
Q5 “To what extent do you see the following elements from windows, if any, present in your [piping: Q3]?”
(Q5.1 Vegetation; Q5.2 Sky; Q5.3 Other buildings) Likert Not at all (1) – Dominates completely (5); Not applicable (no window) (6)
Q6 “In your view, how is the sound environment currently (positively and negatively) affecting your working activity from home? - e.g. heard noises and sounds, building characteristics, urban environment” – Text field
Q7 “To what extent is your present surrounding sound environment appropriate to working from home?” Likert Not at all (1) – Perfectly (5)
Q8 “For each of the 8 scales below, to what extend do you agree or disagree that the present surrounding sound environment while you are working from home is:”
(Q8.1 Comfortable; Q8.2 Intrusive, uncontrolled; Q8.3 Engaging; Q8.4 Empty; Q8.5 Private, controlled; Q8.6 Annoying; Q8.7 Full of content; Q8.8 Detached) Likert Strongly agree (5) – Strongly disagree (1)
Relaxing at home
Q9 “Now please focus on one room that is relevant for your relaxing activities at home:” – Studio; Kitchen; Living room; Kitchen – living room; Bedroom; Bathroom
Q10 “To what extent do you hear the following types of sounds while relaxing in your [piping: Q9]?”
(Q10.1 Traffic noise from outside - e.g. cars, buses, trains, airplanes; Q10.2 Other noise from outside - e.g. sirens, construction, industry, loading of goods; Q10.3 Natural sounds from outside - e.g., singing birds, flowing water, wind in vegetation; Q10.4 Sounds from human beings from outside - e.g. conversation, laughter, children at play, footsteps; Q10.5 Sounds from other human beings present in your house - e.g. conversation, music, TV, laughter, children at play, footsteps; Q10.6 Sounds from neighbors - e.g. conversation, music, TV, laughter, children at play, footsteps; Q10.7 Sounds from building services of your house - e.g. heating, cooling, ventilation systems, toilet flushes; Q10.8 Sounds from building services of your neighbours / common areas - e.g. heating, cooling, ventilation systems, let flushes, lift; Q10.9 Music or TV played by you - through headphones or loudspeakers) Likert Not at all (1) – Dominates completely (5); Not applicable (6)
Q11 “To what extent do you see the following elements from windows, if any, present in your [piping: Q9]?”
(Q11.1 Vegetation; Q11.2 Sky; Q11.3 Other buildings) Likert Not at all (1) – Dominates completely (5); Not applicable (no window) (6)
Q12 “In your view, how is the sound environment currently (positively and negatively) affecting your leisure activities at home? - e.g. heard noises and sounds, building characteristics, urban environment”
(While watching TV, reading, listening to music) – Text field
Q13 “To what extent is your present surrounding sound environment appropriate to relax at home?” Likert Not at all (1) – Perfectly (5)
Q14 “For each of the 8 scales below, to what extend do you agree or disagree that the present surrounding sound environment while you are working from home is:”
(Q14.1 Comfortable; Q14.2 Intrusive, uncontrolled; Q14.3 Engaging; Q14.4 Empty; Q14.5 Private, controlled; Q14.6 Annoying; Q14.7 Full of content; Q14.8 Detached) Likert Strongly agree (5) – Strongly disagree (1)
The house in which you live
Q15 “As regards your house, what is your ownership status?” – Rent – not owned, Owned; Other
Q16 “What is the size of your house?” – Floor area ≤ 40 m2; 40 m2 < Floor area ≤ 80 m2; 80 m2 < Floor area ≤ 110 m2; Floor area > 110 m2
Q17 “What type of house do you live in?” – Detached single family; Semi-detached or terraced house; Apartment block; Other*
Q18 “Please indicate whether the following spaces are present in your house and whether they face a noisy side (e.g. facing a major road, a railway, a busy pedestrian street) or a quiet side (e.g. facing an internal courtyard, a garden, a small street) or whether they are windowless. Please note that it is possible to have multiple noisy or quiet sides”
(Studio; Kitchen; Living room; Kitchen - Living room / open plan; Bedroom; Bathroom) – It faces a noisy side; It faces a quiet side; It is a windowless room; Room not present
Q19 “Who are you currently living with?” – Alone; With roommate(s); With a spouse/partner; With child(ren); With parent(s) or other family members
Q20 “Including yourself, how many people live in your home?” – 1; 2; 3; 4; 5+
Q21 “How do you ventilate your house? Select all that apply” Multiple choice I open the windows; I have mechanical ventilation
Q22 “How do you heat your house? Select all that apply” Multiple choice Radiators; Radiant floor; Electric heaters; Fireplace; Stove; Air systems; Other*
Q23 “How do you cool your house? [Select all that apply]” Multiple choice I have no cooling systems; Radiant systems (e.g. floor, ceiling, etc.); Full air systems (e.g. air conditioners); Air movement devices (e.g. ceiling or desktop fans); By opening windows
The urban context where you live in
Q24 Where do you live in London? Please provide your postcode – Text field
Q25 How would you describe the area where you live? – Urban; Suburban; Rural
Finally, something about you
Q26 “Please state to what extent you disagree/agree with the following sentences:”
(Q26.1 I am sensitive to noise; Q26.2 I find it difficult to relax in a place that́s noisy; Q26.3 I get mad at people who make noise that keeps me from falling asleep or getting work done; Q26.4 I get annoyed when my neighbours are noisy; Q26.5 I get used to most noises without much difficulty) Likert Slider: Totally disagree (0) – Totally agree (1 0 0)
Q27 “Please indicate for each of the five statements which is closest to how you have been feeling over the last two weeks. Notice that higher numbers mean better well-being.”
(Q27.1 I have felt cheerful and in good spirits; Q27.2 I have felt calm and relaxed; Q27.3 I have felt active and vigorous; Q27.4 I woke up feeling fresh and rested; Q27.5 My daily life has been filled with things that interest me; Q27.6 It is important that you pay attention to this study. Please select: 'All of the time') Likert All of the time (5) – At no time (0)
Q28 “How old are you?” – Text field
Q29 “How would you describe your gender?” – Male - including transgender men; Female - including transgender women; Other - e.g. non-binary, gender-fluid, agender*; Prefer not to say
* The “Other” option was followed by a text field in which participants were asked to specify their answer.
Acknowledgments
The authors thank the Department of Innovation, Research and University of the Autonomous Province of Bozen/Bolzano for covering the Open Access publication costs.
==== Refs
References
1 Aletta F. Oberman T. Mitchell A. Tong H. Kang J. Assessing the changing urban sound environment during the COVID-19 lockdown period using short-term acoustic measurements Noise Mapp 7 2020 1 12 10.1515/noise-2020-0011
2 Basu B. Murphy E. Molter A. Sarkar Basu A. Sannigrahi S. Belmonte M. Investigating changes in noise pollution due to the COVID-19 lockdown: The case of Dublin, Ireland Sustain Cities Soc 65 2021 102597 10.1016/j.scs.2020.102597
3 Zambrano-Monserrate M.A. Ruano M.A. Sanchez-Alcalde L. Indirect effects of COVID-19 on the environment Sci Total Environ 728 2020 138813 10.1016/j.scitotenv.2020.138813 32334159
4 Asensio C. Pavón I. de Arcas G. Changes in noise levels in the city of Madrid during COVID-19 lockdown in 2020 J Acoust Soc Am 148 3 2020 1748 1755 10.1121/10.0002008 33003833
5 Aletta F, Osborn D, Osborn D. The COVID-19 global challenge and its implications for the environment – what we are learning. UCL Open Environ 2020;2:8–10. 10.14324/111.444/ucloe.000008.
6 Sakagami K. How did the ‘state of emergency’ declaration in Japan due to the COVID-19 pandemic affect the acoustic environment in a rather quiet residential area? UCL Open Environ 2020;2:1–9. 10.14324/111.444/ucloe.000009.
7 Manzano JV, Pastor JAA, Quesada RG, Aletta F, Oberman T, Mitchell A, et al. The “sound of silence” in Granada during the COVID-19 lockdown. Noise Mapp 2021;8:16–31. 10.1515/noise-2021-0002.
8 Redel-Macías M.D. Aparicio-Martinez P. Pinzi S. Arezes P. Cubero-Atienza A.J. Monitoring sound and its perception during the lockdown and de-escalation of COVID-19 pandemic: A Spanish study Int J Environ Res Public Health 18 7 2021 3392 10.3390/ijerph18073392 33805936
9 Andargie MS, Touchie M, O’Brien W. Case study: A survey of perceived noise in Canadian multi-unit residential buildings to study long-term implications for widespread teleworking. Build Acoust 2021:1351010X2199374. 10.1177/1351010x21993742.
10 Şentop Dümen A. Şaher K. Noise annoyance during COVID-19 lockdown: A research of public opinion before and during the pandemic J Acoust Soc Am 148 6 2020 3489 3496 10.1121/10.0002667 33379907
11 Bartalucci C. Bellomini R. Luzzi S. Pulella P. Torelli G. A survey on the soundscape perception before and during the COVID-19 pandemic in Italy Noise Mapp 8 2021 65 88 10.1515/noise-2021-0005
12 Lee P.J. Jeong J.H. Attitudes towards outdoor and neighbour noise during the COVID-19 lockdown: A case study in London Sustain Cities Soc 67 2021 102768 10.1016/j.scs.2021.102768 33585168
13 Awada M. Becerik-Gerber B. Hoque S. O'Neill Z. Pedrielli G. Wen J. Ten questions concerning occupant health in buildings during normal operations and extreme events including the COVID-19 pandemic Build Environ 188 2021 107480 10.1016/j.buildenv.2020.107480
14 Torresin S. Aletta F. Babich F. Bourdeau E. Acoustics for Supportive and Healthy Buildings : Emerging Themes on Indoor sustainability Sustain 6054 2020 10.3390/su12156054
15 ISO 12913-1:2014 - Acoustics - Soundscape Part 1: Definition and conceptual framework 2014.
16 ISO TS 12913-2:2018 - Acoustics - Soundscape part 2: Data collection and reporting requirements 2018.
17 Torresin S. Albatici R. Aletta F. Babich F. Kang J. Assessment methods and factors determining positive indoor soundscapes in residential buildings: A systematic review Sustain 11 2019 5290 10.3390/su11195290
18 Torresin S. Albatici R. Aletta F. Babich F. Oberman T. Siboni S. Indoor soundscape assessment: A principal components model of acoustic perception in residential buildings Build Environ 182 2020 107152 10.1016/j.buildenv.2020.107152
19 Axelsson Ö. Nilsson M.E. Berglund B. A principal components model of soundscape perception J Acoust Soc Am 128 5 2010 2836 2846 10.1121/1.3493436 21110579
20 Clark C. Paunovic K. WHO Environmental Noise Guidelines for the European Region: A Systematic Review on Environmental Noise and Quality of Life, Wellbeing and Mental Health Int J Environ Res Public Health 15 2018 2400 10.3390/ijerph15112400 30380665
21 Erfanian M. Mitchell A.J. Kang J. Aletta F. The psychophysiological implications of soundscape: A systematic review of empirical literature and a research agenda Int J Environ Res Public Health 16 19 2019 3533 10.3390/ijerph16193533 31546577
22 Aletta F. Oberman T. Mitchell A. Erfanian M. Lionello M. Kachlicka M. Associations between soundscape experience and self-reported wellbeing in open public urban spaces: a field study Lancet 394 2019 S17 10.1016/S0140-6736(19)32814-4
23 Aletta F, Molinero L, Astolfi A, Di Blasio S, Shtrepi L, Oberman T, et al. Exploring associations between soundscape assessment, perceived safety and well-being: A pilot field study in Granary Square, London. Proc Int Congr Acoust 2019;2019-Septe:7946–53. 10.18154/RWTH-CONV-238876.
24 Aletta F. Oberman T. Kang J. Associations between positive health-related effects and soundscapes perceptual constructs: A systematic review Int J Environ Res Public Health 15 2018 2392 10.3390/ijerph15112392 30380601
25 Ormandy D. Housing and health in Europe: the WHO LARES project 2009 Routledge
26 Kearns A. Hiscock R. Ellaway A. MaCintyre S. “Beyond four walls”. The psycho-social benefits of home: Evidence from West Central Scotland Hous Stud 15 3 2000 387 410 10.1080/02673030050009249
27 Amerio A. Brambilla A. Morganti A. Aguglia A. Bianchi D. Santi F. Covid-19 lockdown: Housing built environment’s effects on mental health Int J Environ Res Public Health 17 16 2020 5973 10.3390/ijerph17165973 32824594
28 Derryberry E.P. Phillips J.N. Derryberry G.E. Blum M.J. Luther D. Singing in a silent spring: Birds respond to a half-century soundscape reversion during the COVID-19 shutdown Science (80-) 370 6516 2020 575 579 10.1126/science:abd5777
29 Lenzi S, Sádaba J, Lindborg P. Soundscape in Times of Change : Case Study of a City Neighbourhood During the COVID-19 Lockdown 2021;12. 10.3389/fpsyg.2021.570741.
30 Maggi A.L. Muratore J. Gaetán S. Zalazar-Jaime M.F. Evin D. Pérez Villalobo J. Perception of the acoustic environment during COVID-19 lockdown in Argentina J Acoust Soc Am 149 6 2021 3902 3909 10.1121/10.0005131 34241433
31 Ratcliffe E. Sound and soundscape in restorative natural environments: A narrative literature review Front Psychol 2021 10.3389/fpsyg.2021.570563
32 Hedblom M. Gunnarsson B. Schaefer M. Knez I. Thorsson P. Lundström J.N. Sounds of nature in the city: No evidence of bird song improving stress recovery Int J Environ Res Public Health 16 8 2019 1390 10.3390/ijerph16081390 30999690
33 MacDonald RAR. Music, health, and well-being: A review. Int J Qual Stud Health Well-Being 2013;8. 10.3402/qhw.v8i0.20635.
34 Tanaka Y.L. Kudo Y. Effects of familiar voices on brain activity Int J Nurs Pract 18 2012 38 44 10.1111/j.1440-172X.2012.02027.x 22776531
35 Dzhambov A.M. Lercher P. Stoyanov D. Petrova N. Novakov S. Dimitrova D.D. University students’ self-rated health in relation to perceived acoustic environment during the covid-19 home quarantine Int J Environ Res Public Health 18 2021 1 21 10.3390/ijerph18052538
36 Zimmer K. Ghani J. Ellermeier W. The role of task interference and exposure duration in judging noise annoyance J Sound Vib 311 3-5 2008 1039 1051 10.1016/j.jsv.2007.10.002
37 Peer E. Brandimarte L. Samat S. Acquisti A. Beyond the Turk: Alternative platforms for crowdsourcing behavioral research J Exp Soc Psychol 70 2017 153 163 10.1016/j.jesp.2017.01.006
38 Palan S. Schitter C. Prolific.ac—A subject pool for online experiments J Behav Exp Financ 17 2018 22 27 10.1016/j.jbef.2017.12.004
39 The Health Protection (Coronavirus, Restrictions) (No. 3) and (All Tiers) (England) (Amendment) Regulations 2021 n.d. https://www.legislation.gov.uk/uksi/2021/8/introduction/made (accessed February 16, 2021).
40 Harris P.A. Taylor R. Thielke R. Payne J. Gonzalez N. Conde J.G. Research electronic data capture (REDCap) - a metadata-driven methodology and workflow process for providing translational research informatics support J Biomed Inform 42 2 2009 377 381 18929686
41 Harris P.A. Taylor R. Minor B.L. Elliott V. Fernandez M. O'Neal L. The REDCap consortium: Building an international community of software platform partners J Biomed Inform 95 2019 103208 10.1016/j.jbi.2019.103208 31078660
42 Johnson R.B. Onwuegbuzie A.J. Mixed methods research: A research paradigm whose time has come Educ Res 33 7 2004 14 26
43 ISO TS 12913-3:2019 - Acoustics - Soundscape part 3: Data analysis 2019.
44 Weinstein N.D. Individual differences in critical tendencies and noise annoyance J Sound Vib 68 2 1980 241 248
45 Benfield J.A. Nurse G.A. Jakubowski R. Gibson A.W. Taff B.D. Newman P. Testing Noise in the Field: A Brief Measure of Individual Noise Sensitivity Environ Behav 46 3 2014 353 372 10.1177/0013916512454430
46 Topp C.W. Østergaard S.D. Søndergaard S. Bech P. The WHO-5 well-being index: A systematic review of the literature Psychother Psychosom 84 3 2015 167 176 10.1159/000376585 25831962
47 Corp I.B.M. IBM SPSS Statistics for Windows Version 26 2019
48 Team RC, others. R: A language and environment for statistical computing 2013.
49 Ercakmak UB, Dokmeci Yorukoglu PN. Comparing Turkish and European Noise Management and Soundscape Policies : A Proposal of Indoor Soundscape Integration to Architectural Design and Application. Acoustics 2019;1:847–65. 10.3390/acoustics1040051.
50 Hallberg L.-M. The “core category” of grounded theory: Making constant comparisons Int J Qual Stud Health Well-Being 1 3 2006 141 148 10.1080/17482620600858399
51 OUIS D. Annoyance from road traffic noise: A review J Environ Psychol 21 1 2001 101 120 10.1006/jevp.2000.0187
52 Park S.H. Lee P.J. Yang K.S. Kim K.W. Relationships between non-acoustic factors and subjective reactions to floor impact noise in apartment buildings J Acoust Soc Am 139 3 2016 1158 1167 10.1121/1.4944034 27036252
53 Marquis-Favre C. Premat E. Aubrédue D. Noise and its effects - A review on qualitative aspects of sound. Part II: Noise and annoyance Acta Acust United with Acust 91 2005 626 642
54 Guski R. Personal and social variables as co-determinants of noise annoyance Noise Health 1 1999 45 56 12689499
55 Rohde L. Larsen T.S. Jensen R.L. Larsen O.K. Framing holistic indoor environment: Definitions of comfort, health and well-being Indoor Built Environ 29 8 2020 1118 1136 10.1177/1420326X19875795
56 Van Kamp I, Klaeboe R, Brown AL, Lercher P. Soundscapes, human restoration and quality of life. In: Kang J, Schulte-Fortkamp B, editors. Soundscape Built Environ., Boca Raton, FL, USA: CRC Press; 2016, p. 43–68. 10.1201/b19145-4.
57 Lercher P. Van Kamp I. Von Lindern E. Botteldooren D. Perceived soundscapes and health-related quality of life, context, restoration, and personal characteristics. Soundscape Built Environ 89 2016
58 Fiebig A. Jordan P. Moshona C.C. Assessments of Acoustic Environments by Emotions – The Application of Emotion Theory in Soundscape Front Psychol 11 2020 10.3389/fpsyg.2020.573041
| 0 | PMC9746886 | NO-CC CODE | 2022-12-15 23:21:58 | no | Appl Acoust. 2021 Dec 1; 183:108305 | utf-8 | Appl Acoust | 2,021 | 10.1016/j.apacoust.2021.108305 | oa_other |
==== Front
Ann Med Psychol (Paris)
Ann Med Psychol (Paris)
Annales Medico-Psychologiques
0003-4487
0003-4487
Elsevier Masson SAS.
S0003-4487(22)00137-8
10.1016/j.amp.2022.05.006
Original Article
Examining the relationship between the thinking styles and the motivation aspects of the individuals working in the health sector in Turkey during the COVID-19 pandemic: The case of hospital staff
Examiner la relation entre les styles de pensée et les aspects de motivation des personnes travaillant dans le secteur de la santé en Turquie pendant la pandémie de COVID-19 : le cas du personnel hospitalierKıroğlu Arslan Işıl
Department of Health Management, Faculty of Health Sciences, Ardahan University, Ardahan, Turkey
13 5 2022
13 5 2022
15 9 2021
5 5 2022
© 2022 Elsevier Masson SAS. All rights reserved.
2022
Elsevier Masson SAS
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Aim
The aim of this study is to evaluate the relationship between the thinking styles of hospital staff and their motivation tools during the pandemic in Turkey.
Materials and methods
Top 100 hospitals in Turkey that have the highest number of examinations on the list, constitute the population and 1220 participants from all over Turkey took part in the study. The data set was analyzed with descriptive statistics, independent groups t-test, ANOVA and regression analysis.
Results
Most of the sample were women (66.4%), undergraduate (49%), working in a medical position (72%) and nurses (50.6%). T-test conclusion was showed that intrinsic motivation values are higher than extrinsic motivation values. There was no significant difference between administrative/medical position status and the mean score of any of the scales. As the results of regression analysis, it was determined that thinking style significantly predicted motivation.
Conclusions
The results of the present study suggest that hospital staff used the experiential thinking style more than rational thinking and behaved with intrinsic motivation more than extrinsic motivation during COVID-19 pandemic.
Résumé
Objectif
L’objectif de cette étude est d’évaluer la relation entre les styles de pensée du personnel hospitalier et leurs outils de motivation pendant la pandémie en Turquie.
Matériels et méthodes
Les 100 meilleurs hôpitaux en Turquie, qui ont le plus grand nombre d’examens sur la liste, constituent la population, et 1 220 participants de toute la Turquie ont participé à l’étude. L’ensemble de données a été analysé avec des statistiques descriptives, des groupes indépendants t-test, ANOVA et une analyse de régression.
Résultats
L’échantillon était composé pour la plupart de femmes (66,4 %), de premier cycle (49 %), travaillant dans un poste médical (72 %) et des infirmières (50,6 %). La conclusion du test-T a montré que les valeurs de motivation intrinsèques sont plus élevées que les valeurs de motivation extrinsèques. Il n’y a pas de différence significative entre le statut du poste administratif/médical et le score moyen de l’une des échelles. Selon les résultats de l’analyse de régression, il a été déterminé que le style de pensée prédisait de manière significative la motivation.
Conclusions
Les résultats de la présente étude suggèrent que le personnel hospitalier a utilisé le style de pensée expérientielle plus que la pensée rationnelle, et s’est comporté avec une motivation intrinsèque plus qu’une motivation extrinsèque pendant la pandémie de COVID-19.
Keywords
COVID-19
Hospital staff
Motivation
Pandemic
Thinking
Mots clés
COVID-19
Motivation
Pandémie
Pensée
Personnel hospitalier
==== Body
pmc1 Introduction
Changing living conditions appear in the fields of work with the developments in the information age. Especially with the effect of the COVID 2019 pandemic, the conditions of many working areas have changed radically. Healthcare workers experience high levels of work stress even under normal conditions. The crisis of coronavirus disease 2019 (COVID-19) is putting additional pressure on healthcare personnel [[1], [2]]. On the other hand, it is of great importance to determine which factors affect the professional attitudes of healthcare professionals in such events, so that hospitals can continue to function even in such difficult conditions [3]. The sudden outbreak of the pandemic has resulted in the need for urgent response measures to identify and diagnose patients, treat them, and combat contamination intensively. In this process, healthcare professionals who take the risk of infection for themselves and their relatives, were at the forefront of the fight against the disease [4]. Hence, there is a need for individuals (healthcare professionals) who can solve the problems they encounter in working environments, are open to change and development, support change, and make an effort for this.
The education system that trains healthcare professionals in line with the expectations and preferences of organizations in the information age, expects students to learn and use various ways of thinking, such as processing the information they receive, problem-solving, handling the problems creatively and critically, and having the ability to implement them effectively in the learning process [5]. However, the purpose of hospitals that carry out service activities is to transform unhealthy input into a healthy output. At this point, the most important task of managers should be carefully direct all attention of the staff to the human input process. Because the smallest mistake to be made has a high risk of adversely affecting human health and human life [[6], [7]]. Moreover, healthcare services are becoming more complex and branched out to meet needs. Knowledge and technology in the medical field are expanding at an incredible rate, making it difficult for healthcare personnel to keep up with the developing knowledge. Patients’ needs are also changing. It is being switched from the diagnosis and treatment of a single acute problem to the long-term follow-up of multiple and related chronic conditions [8]. The basic condition for healthcare professionals serving human health to be able to offer this service in the best way is to be healthy. There are some basic conditions and possibilities of being physically and mentally healthy for them. Unless these conditions and opportunities, which can be summarized as regulating the working conditions of healthcare workers and gaining their democratic rights and thus improving their living conditions, are not provided, healthcare workers cannot be expected to provide satisfying health services [9]. As the main input is human and human life in hospitals, which have a complicated organizational structure, the quality of the output at each stage of the system is vital. Service execution that customers/patients will perceive as the quality is only possible with motivated and well-guided personnel. Hence, health institution managers must first motivate their internal customers, namely their employees, and increase their satisfaction to satisfy their external customers. Making conscious motivation management by using various material and non-pecuniary motivation tools and methods are the paths followed by systematic organizations [10]. In the context of motivation theories, various studies have been conducted from past to present in order to determine the factors that motivate individuals in organizations [[11], [12], [13], [14], [15], [16], [17], [18]], This study is based on the “intrinsic motivation” and “extrinsic motivation” offered in Self-Determination Theory [19]. Self-Determination Theory analyzes the natural developmental tendencies and innate psychological needs of people that are the basis for self-motivation and personality integration. The main difference between intrinsic and extrinsic motivation; while extrinsic incentives, pressures, and material rewards form the basis for people to act with extrinsic motivation, the intrinsic motivation enables the person to take action focused on fun or struggle [20]. Intrinsic motivation occurs naturally by doing things that one loves or that attracts one's attention [21]. The behaviors of living beings that they perform with internal motivation rely on positive experiences based on exercising for self-improvement and increasing capacity rather than for a material gain [20].
Extrinsic motivation, contrary to intrinsic motivation, is associated with an end goal and belongs to many behaviors that are more than just for their own good. Being extrinsically motivated involves behaviors with the intention of achieving some distinguishable outcome, such as achieving a reward, getting rid of guilt, or gaining someone else's approval [11]. Extrinsic motivation elements are performance-based incentives that fall to the share of employees out of the job itself. Although not as effective as intrinsic motivation, expectations for extrinsic rewards can provide positive changes in employees’ attitudes [11]. From a holistic view, determining the factors that motivate individuals in the work environment is important in terms of creating an organizational environment that will maximize the motivation of the employees.
In the business environment, as in all areas of life, people use their reasoning skills before they are motivated to take any action. According to Semerci [[22], [23]], thinking people can produce, use, and evaluate knowledge. The most basic feature that distinguishes humans from other living beings is their cognition. Thinking is the activity of using existing information, combining information at necessary points, separating, classifying and evaluating new information, as well as learning new information [[24], [25]] According to Nickerson [26], thinking includes logical reasoning, critical thinking, creative thinking, problem-solving, and decision making.
If individuals exhibit different characteristics from each other in terms of their appearance, physical abilities, interests, tastes, and knowledge, they also differ in their mental development (mental schemas), learning, and thinking activities (styles). These differences result in people using their abilities in different ways, reacting differently to events, and making different choices [[24], [27]]. The concept of “style”, one of the variables of the individual difference that educational psychologists have focused on in recent years, in general terms, “is a link between talent and personality; the individual's combination of choice in using their abilities is how they choose to apply their knowledge and skills” [28]. The thinking styles, defined by the concept of style, show the works and preferences that the individual enjoys doing and indicate the power to benefit from intelligence rather than intelligence itself and maximizing intelligence [29]. Thinking styles, which can also be expressed as cognitive styles, are defined as “consistent individual differences in preferred methods to organize and to process information and experiences” by Allinson and Hayes [30]. The intellectual development of an individual needs to discover and develop thinking styles in which creative thinking, decision making, problem-solving, evaluation, and reasoning skills are effective [31].
Looking at the studies on thinking styles, it is seen that three comprehensive approaches are frequently emphasized. These are (a) Cognitive-Experiential Self-Theory of Personality (CEST); (b) Myer-Briggs Type Indicator (Thinking Style Classifications); and (c) Theory of Mental Self-Government. At the same time, new approaches that try to explain the thinking and decision-making processes argue that emotion plays an important role in cognitive-based decisions. In fact, CEST is one of the approaches suggesting that logic/cognition-based explanations alone are not sufficient for decision making or information processing, and that includes processes based on emotion/intuition into the decision-making process [[32], [33]] . According to CEST, there are two systems that do not depend on each other but operate in parallel in processing information. One of these, the logical/cognitive system, is an inferential system based on culturally transmitted rules of reasoning. It works consciously, partly slowly, analytically, mainly verbally and relatively emotion-free. The experiential system, on the other hand, is a learning system and belongs to the pre-conscious level rather than consciousness. It works fast, automatic, holistic, mostly non-verbal, and linked to emotions [34]. Cognitive and experiential systems work in parallel and interactively, and behavior toward finding solution is determined by the joint effect of these two systems. The system to be used more predominantly reflects an individual tendency [35].
The need for individuals, working in a service environment where workflows occur rapidly, to find a quick solution to a problem distinguishes these individuals from those working in other business environments (sectors). Since thinking is especially associated with learning and teaching activities, there are many studies specifically for the Education Sector. Dinçer and Saracaloğlu [24] conducted a study to examine the thinking styles of teacher candidates in terms of various variables. According to the research findings; statistically significant differences were found between the thinking styles of the teachers candidates and the learning program, class, gender, age, graduated area, and the perception of socioeconomic level. Similarly, Berkant and Tüzer [22] conducted a study to explore the thinking styles of classroom teachers and to associate the styles with some demographic variables (gender, marital status, professional seniority, etc.). Çubukçu [31] focused on revealing the thinking styles that affect the learning of teacher candidates and also revealed that the styles are closely related to the individual situations (age, gender) and socio-economic status (features such as hobbies, leadership experience, work experience) of teacher candidates. Fer [28] examined whether the thinking styles of teacher candidates changed depending on the variables of gender, university resources and program resources and obtained significant findings that they did. Golian [36] did a research whether there is a difference in thinking styles between senior library managers working in the public and technical service fields in libraries with institutional memberships to the Association of Research Libraries (ARL). Also, a few studies are found to explore the styles of health personnel candidate students [[5], [37]]. However, there is no such research has been found for those working in the Health Sector. Karadağ, Alparslan and İşeri [5] did a research on the critical thinking tendencies and learning styles of midwifery and nursing students. While there was a significant gender difference in some sub-dimensions of the California Critical Thinking Disposition Inventory, no significant difference was found according to the department and class. Zhang's study [37] concerns the contingent nature of the relationships of student–teacher style match (or mismatch) to students’ academic achievement. Kanbay et al. [38] studied the critical thinking and problem solving skills of undergraduate nursing students and made comparisons between classes. Although there is a difference in critical thinking mean scores between classes, this difference is not statistically significant.
There are many studies related motivation applied on hospital staff, but any study is not found on the relationship of motivation and thinking styles. The aim of this study is to contribute to the literature by exploring the motivation factors and thinking styles of health personnel in Turkey and revealing the relationship between motivation factors and thinking styles during COVID-19 pandemic.
2 Material and method
2.1 Study design, setting and sample
The research focuses on defining the effects of thinking and decision-making styles of individuals working in hospitals, which are service-producing organizations of the health sector, on their motivational aspects (internal-external). The research unit is hospital staff who performs difficult tasks to solve the health problems of people.
In this study, descriptive (exploratory) research model is used. The purpose of the descriptive model is to define the research problem, the variables related to this problem and the relationships between the variables. Thus, it will be possible to make forward-looking predictions and to develop suggestions [39].
The hypotheses put forward for the purpose of the study are as follows:• H1: thinking/decision making has an effect on motivation;
• H2: individuals with experiential thinking style are motivated by intrinsic motivation tools;
• H3: individuals with experiential thinking style are motivated by extrinsic motivation tools;
• H4: individuals with rational thinking style are motivated by intrinsic motivation tools;
• H5: individuals with rational thinking style are motivated by extrinsic motivation tools.
The study was based on the “2017 Period Clinical Examination Numbers on the Basis of Hospitals” obtained from the 2017 January-October Period Polyclinic, Hospitalization, Intensive Care and Emergency Service Statistics report published by the Ministry of Health, General Directorate of Public Hospitals. There are top 100 hospitals that have the highest number of examinations in Turkey on the list and make up the population of the research. Hospitals were invited to participate in the research through official permission channels, and the principle of voluntary participation, which is accepted as one of the ethical principles in scientific research, was observed [40]. All hospitals with a positive response were visited until a sample enabling the representation of the population was obtained within the specified period. 25 of these 100 hospitals make up the sample of the hospital. Hospitals that accepted to participate in research (25 hospitals in 15 provinces), are seen showing a wide distribution in Turkey (Fig. 1 ).Fig. 1 The provinces included in the sample on the map of Turkey.
It was planned to contact the hospital management units in order to ensure the participation of hospital personnel in the data collection process. In this direction, an application was made to the local health authority to which each hospital is affiliated, through institutional correspondence. In this process, the local health authority first consulted the hospital and received the approval that the research did not pose an obstacle for them. Afterwards, it was reported through official channels that the necessary permissions were obtained if deemed appropriate by the directorate. Hospitals were visited on common dates set for hospital staff and researcher, and data were collected from administrative staff and medical units via a questionnaire.
2.2 Instruments
The Interview/Demographics Form, the Intrinsic/Extrinsic Motivation Tools Scale, and the Rational-Experiential Inventory were the data collection tools of this study. The interview form consisted of questions including sociodemographic characteristics of the hospital staff such as the age, gender, education level, marital status, department, job position, and information about their experiences in making decisions such as factors that led to a successful result.
Motivation questionnaire (the Intrinsic/Extrinsic Motivation Tools Scale) items were developed based on Mottaz [41], Brislin et al. [42], and Mahaney and Lederer's [43] motivation works to measure the direction (intrinsic or extrinsic) of motivation of hospital staff [[41], [42], [43]]. It included 24 items containing two dimensions and was adapted to Turkish by Dündar, Özutku and Taşpınar in 2007 [12]. Intrinsic motivation was measured by 9 items and extrinsic motivation was measured by 15 items. Reliability of the scale created by Dündar, Özutku and Taşpınar [12] was determined using the “Cronbach alpha” criterion based on the “internal consistency” method. Cronbach alpha values of the scale of intrinsic and extrinsic motivation tools were calculated as 0.83 and 0.84, respectively. As the answer to the question “To what extent do the following situations about your job encourage and motivate you?”; there are five options such as “1 – does not encourage at all”, “2 – it encourages a little”, “3 – it encourages moderately”, “4 – it highly encourages”, “5 – it highly encourages”.
In order to determine the thinking style, the Rational-Experiential Inventory, which was developed by Pacini and Epstein [34] and adapted to Turkish by Türk [44], was preferred. A Likert type metric with five intervals was used for the answers of the expressions in the scales. The original form of the scale with 40 items consists of four subscales. These are rational ability, rational engagement, experiential ability, and experiential engagement. In the rational skill subscale, there are expressions indicating a high level skill in analytical-rational thinking; in the subscale of rational thinking engagement, it includes statements about having confidence in thinking in an analytical-rational way and enjoying it. The experiential skills subscale includes expressions reporting skills related to the individual's intuitive perceptions, and the experiential engagement subscale includes expressions reporting trusting on emotions and intuitions while making decisions. The reliability of the original scale was found as 0.90 for rational thinking style and 0.87 for experiential thinking style. In this section, “Please mark the following expressions about your feelings, beliefs and behaviors by evaluating the option (between 1 and 5) appropriate for you. Try to reflect the first effect you have while answering.” directive and 5-point Likert type measurement graded between “absolutely wrong” and “absolutely true” were preferred.
2.3 Data Collection
Data were collected via hospital visits. Each hospital that declared that they would participate in the study was contacted and a common date was determined for the visit. All departments in the hospital were visited in company with a nurse or an officer determined by the hospital administration to assist the study. The employees in the departments were informed about the study and each employee who wanted to take part in the study was asked to fill in a questionnaire. After all departments were visited in this way, they returned to the starting section and the answered forms were collected by checking. If there were missing parts in the form, the participant was supported to complete it.
Data collection took approximately a year due to the responses to the acceptance of participation from hospitals where the task intensity is high and decision mechanisms are always urgent. Consequently, about 1,400 forms were collected from 20 hospitals, but 1,220 of them could be used in the research.
2.4 Ethical Considerations
Ethics report was received from Ardahan University Social and Human Sciences Ethics Committee (date: 8 July 2019) for “Ethics Committee Approval”. All participants were informed, and consents were obtained before data collection.
2.5 Data Analysis
Statistical analyses were carried out using SPSS 20.0 software. The number and percentage of participants in the categorical variables were expressed as (n), and (%), respectively, and as mean ± standard deviation (X ± SD) for the numeric variables. The comparisons between scale mean and sample characteristics, data were evaluated via Independent Samples Test and analysis of variance. Relationships between scales/sub-scales were evaluated using linear regression analysis.
3 Results
3.1 Sample characteristics (Independent Samples Test, Analysis of Variance)
A total of 1,220 individuals, including 340 administrative and 880 medical staff, were included in the study. The mean (±SD) age of all participants was 33.70 ± 8.39 (range: 17–65) years, 810 of them were female (66.4%) and 830 of them were married (68%). Most of the participants were undergraduate (49%) and work in nursing staff (50.6%) (Table 1 ).Table 1 Demographic characteristics of the sample.
Table 1Characteristics n %
Staff type
Administrative 340 27.9
Medical 880 72.1
Gender
Female 810 66.4
Male 410 33.6
Marital status
Single 390 32.0
Married 830 68.0
Education level
High school 252 20.7
Associate degree 285 23.4
Undergraduate 598 49.0
Graduate and above 85 7.0
Staff position
Administrative unit manager 26 2.1
Biologist 2 0.2
Civil servant 99 8.1
Clinical support staff 9 0.7
Computing/information system staff 6 0.5
Data preparation and control operator 82 6.7
Deputy chief physician 4 0.3
Elderly care 1 0.1
Engineer 1 0.1
First and emergency aid 15 1.2
Health officer 27 2.2
Healthcare technician 22 1.8
Hospital manager 1 0.1
Laboratory 33 2.7
Medical doctor 51 4.2
Medical secretary 64 5.2
Medical technician 55 4.5
Medical unit manager 5 0.4
Midwife 33 2.7
Nurse 617 50.6
Nutritionist 6 0.5
Occupational health and safety specialist 6 0.5
Patient admission/registration 22 1.8
Pharmacist 1 0.1
Physiotherapist 8 0.7
Psychologist 3 0.2
Security guard 1 0.1
Social worker 6 0.5
Speech and language therapist 1 0.1
Technician 6 0.5
Training officer 7 0.6
Mean scores of intrinsic motivation scale, extrinsic motivation scale, rational and experiential thinking styles and some characteristics of the participants (gender, education and administrative/medical position status) were compared with Independent Samples Test and ANOVA. T-test conclusion was showed (Table 2 ) that the intrinsic motivation scores of male staff (Mean = 4,05, SD = 0.73) were significantly [t(1218) = 3.192, P = 0.001] lower than female staff (Mean = 4.19, SD = 0.69). There was no statistically significant difference between mean scores of extrinsic motivation scale and gender (Table 3 ) [t(1218) = 1.502, P = 0.133], rational thinking style and gender (Table 4 ) [t(1218) = 0.654, P = 0.513] and experiential thinking style and gender (Table 5 ) [t(1218) = 1.502, P = 0.133]. When the two tables (Table 2, Table 3) are examined together in terms of gender, it is noteworthy that intrinsic motivation values (Mean for females: 4.188, for males: 4.050) are higher than extrinsic motivation values (Mean for females: 3.175, for males: 3.133) in both males and females. Therefore, it can be concluded that hospital staff are more motivated with intrinsic motivation tools than extrinsic motivation tools.Table 2 Results of T-test analysis for intrinsic motivation.
Table 2Group n Mean SD t df P
Female 810 4.188 0.698 3.192 1218 0.001
Male 410 4.050 0.729
Medical 880 4.151 0.697 −0.758 1218 0.449
Administrative 340 4.116 0.750
Table 3 Results of T-test analysis for extrinsic motivation.
Table 3Group n Mean SD t df P
Female 810 3.175 0.465 1.502 1218 0.133
Male 410 3.133 0.467
Medical 880 3.968 0.801 −1.196 1218 0.232
Administrative 340 3.906 0.840
Table 4 Results of T-test analysis for rational thinking style.
Table 4Group n Mean SD t df P
Female 810 3.099 0.431 0.654 1218 0.513
Male 410 3.081 0.464
Medical 880 3.101 0.801 −1.089 577.8 0.277
Administrative 340 3.070 0.840
Table 5 Results of T-test analysis for experiential thinking style.
Table 5Group n Mean SD t df P
Female 810 3.175 0.465 1.502 1218 0.133
Male 410 3.133 0.467
Medical 880 3.170 0.452 −0.986 1218 0.324
Administrative 340 3.140 0.502
Also, there was no significant difference between administrative/medical position status and the mean score of intrinsic [t(1218) = −0,758, P = 0.449] and extrinsic [t(1218) = −1,196, P = 0.232] motivation scales, rational [t(577,8) = −1,089, P = 0.277] and experiential [t(1218) = −0,986, P = 0.324] thinking styles (Table 2, Table 3, Table 4, Table 5, respectively).
A one-way ANOVA was performed among the participants to compare their high school, associate degree, undergraduate and graduate and above degree graduation levels whether the intrinsic and extrinsic motivation, and rational and experiential inventory rates had a different effect in terms of education level or not. According to the homogeneity test of variances, it was observed that group variances were not homogeneous for intrinsic motivation, rational and experiential thinking styles. In this case, Welch or Brown-Forsythe tests had to be performed, since the classical F test could not be performed. According to the level of education, the intrinsic motivation (P = 0.176), rational thinking style (P = 0.790) and experiential thinking style (P = 0.256) usage rates of the staff did not differ at the 5% significance level (Table 6, Table 7, Table 8 , respectively). Additionally, there was not a significant effect of education level differences on extrinsic motivation (Table 9 ) used at work at P < 0.05 [F (3, 1216) = 1,562, P = 0.197].Table 6 Results of variance analysis for intrinsic motivation.
Table 6Group n Mean SD Source SS df MS F P
High sch. 252 4.194 0.645 Between g. 2.134 3 0.711 1.405 0.240
Associate deg. 285 4.108 0.775 Within g. 615,947 1216 0.507
Undergraduate 598 4.121 0.721 Total 618,081 1219
Graduate and + 85 4.242 0.599
Table 7 Results of variance analysis for rational thinking style.
Table 7Group n Mean SD Source SS df MS F P
High sch. 252 3.130 0.521 Between g. 1.393 3 0.464 2.396 0.067
Associate deg. 285 3.049 0.423 Within g. 235,687 1216 0.194
Undergraduate 598 3.107 0.407 Total 237,080 1219
Graduate and + 85 3.028 0.456
Table 8 Results of variance analysis for experiential thinking style.
Table 8Group n Mean SD Source SS df MS F P
High sch. 252 3.184 0.539 Between g. 0.922 3 0.307 1.414 0.237
Associate deg. 285 3.127 0.471 Within g. 264,312 1216 0.217
Undergraduate 598 3.177 0.426 Total 265,234 1219
Graduate and + 85 3.100 0.483
Table 9 Results of variance analysis for extrinsic motivation.
Table 9Group n Mean SD Source SS df MS F P
High sch. 252 4.049 0.768 Between g. 3.089 3 1.030 1.562 0.197
Associate deg. 285 3.927 0.840 Within g. 801,711 1216 0.659
Undergraduate 598 3.923 0.819 Total 804,800 1219
Graduate and + 85 3.936 0.790
3.2 Regression Analysis
When the preferences of the participants regarding the sub-dimensions of motivation are examined, the average scores are 4.14 ± 0.71 for intrinsic motivation and 3.95 ± 0.81 for extrinsic motivation. The mean score of the experiential thinking style sub-dimension of the thinking styles is 3.16 ± 0.46 and the average score of the logical thinking style, the other sub-dimension, is 3.09 ± 0.44.
Correlation analysis was used to investigate whether there is a relationship between the variables. As a result of the Pearson Correlation Analysis conducted to determine the relationship between the scores obtained from the motivation sub-dimensions and the sub-dimensions of thinking styles, a statistically significant relationship was found between the scores at P < 0.01 level as seen in Table 10 .Table 10 Results of correlation analysis.
Table 10Variables 1 2 3 4
1. Experiential thinking style 1 0.765a 0.120a 0.093a
2. Rational thinking style 0.765a 1 0.090a 0.093a
3. Intrinsic motivation 0.120a 0.090a 1 0.660a
4. Extrinsic motivation 0.093a 0.093a 0.660a 1
a Correlation is significant at the 0.01 level.
When it is examined whether the motivation sub-dimensions and thinking styles sub-dimensions of the employees participating in the research differ or not, firstly a strongly positive correlation was found between experiential thinking style and rational thinking style as being two parts of thinking style inventory. There was a weakly positive correlation between experiential thinking style and two sub-dimensions of motivation (intrinsic and extrinsic). Similarly a weakly positive correlation was found between rational thinking style and intrinsic and extrinsic motivation (Table 10).
Linear regression analysis was used to evaluate the predictive power of the thinking/decision making variable on the motivation variable. As a result of the analysis, it was determined that thinking/decision making significantly predicted motivation [F (1, 1218) = 15.552, P < 0.001] and that thinking explained 1,3% of the variance of motivation variable. The findings showed that the independent variable explained only a small part of the variance in motivation. The predictive effect of the experiential thinking style on intrinsic motivation sub-dimension was tested by linear regression analysis, and it was found that the results were statistically significant [F (1, 1218) = 17.865, P < 0.001] and it was seen that the experiential thinking style explained the variance of intrinsic motivation at the level of 1.4%. According to this, when the experiential thinking activities of health workers increase, their motivation in the dimension of internality increases. Besides, the predictive effect of experiential thinking style on extrinsic motivation was also tested, and it was found that the results were statistically significant [F (1, 1218) = 10.529, P < 0.01] and experiential thinking style explained the variance of extrinsic motivation at 1% level. Findings show that experiential thinking style explains only a small part of the variance in extrinsic motivation.
The results of the linear regression analysis performed to evaluate the power of rational thinking sub-dimension level to predict intrinsic motivation were statistically significant [F (1, 1218) = 9.899, P < 0.01] and it was determined that the rational thinking sub-dimension explained about 1% of the change in the intrinsic motivation level (Table 11 ). Accordingly, it was seen that the act of rational thinking explained very little of the variance in intrinsic motivation. The predictive power of rational thinking style on extrinsic motivation was also examined, and the results of the linear regression analysis performed were found to be statistically significant [F (1, 1218) = 10.547, P < 0.01]. It was determined that the rational thinking sub-dimension explained 1% of the total variance of extrinsic motivation. The findings show that rational thinking explains a very low part of the variance in extrinsic motivation. Also, it can be said that the predictive power of the experiential thinking sub-dimension on intrinsic motivation is higher than the predictive power of the rational thinking sub-dimension.Table 11 Results of Linear Regression analysis.
Table 11Independent Variable B β t P R2 F
1. Thinking 0.188 0.112 3.944 0.000 0.013 15.552**
2. Experiential thinking 0.184 0.120 4.227 0.000 0.014 17.865**
3. Experiential thinking 0.161 0.093 3.245 0.001 0.009 10.529*
4. Rational thinking 0.145 0.090 3.146 0.002 0.008 9.899*
5. Rational thinking 0.171 0.093 3.248 0.001 0.009 10.547*
Note 1. *: P < 0.01, **: P < 0.001. Dependent Variable: 1. Motivation, 2. Intrinsic motivation, 3. Extrinsic motivation, 4. Intrinsic motivation, 5. Extrinsic motivation.
4 Discussion
The current study was conducted to evaluate the relationship between the motivational aspect of healthcare personnel and administrative personnel working in hospitals and their thinking styles. The most basic finding obtained is that there is a positive linear relationship between motivation and thinking, and thinking scores are significantly predicted motivation. The expected result before the study was in this direction and it was achieved. Although there are few studies [[45], [46], [47], [48], [49]] in the literature focusing on the relationship between thinking styles and motivation, they support our finding in this study. For example, in their studies Belousova and Mochalova [45] focused on measuring the relationship between thinking styles and motivational characteristics, starting from the idea that the motivational characteristics of managers determine their personal professional success and effectiveness in organizational activities. According to the results of the research, it was suggested that the relationship between thinking style and need for achievement was characteristic for managers with different professional orientations. This confirmed that the need for achievement and motivation to achieve are one of the determinants of the thinking style of managers. In another study on thinking styles, psychological needs and motivation toward education of instructional psychology students, Doménech-Betoret and Gómez-Artiga [46] found that thinking styles have a significant and positive impact on student psychological need satisfaction. In turn, psychological need satisfaction has a significant and positive impact on student intrinsic motivation. In particular, such a study has not been found in the health sector.
Also, the second main finding is that in general hospital staff, acting with intrinsic motivation tools is more intense than it is motivated by extrinsic motivation tools. Similar results have also been obtained in previous studies hold on healthcare personnel in Turkey on motivation supports the results of this study. In these studies [[50], [51], [52], [53]], which focus more on discovering the motivational aspect and the relationship between the organizational motivation tools used and the job satisfaction of the staff, the importance of intrinsic motivation or Herzberg's motivational tools for healthcare personnel is mentioned.
Besides, it was observed that females were more motivated with intrinsic motivation tools than the male group. There is no significant difference between the groups in terms of extrinsic motivation. It was observed that motivation tools did not make a significant difference between medical personnel and administrative personnel in terms of neither intrinsic motivation nor extrinsic motivation. In addition, when the education levels (high school, associate degree, undergraduate, graduate and above) are considered, no differences were found in motivation with intrinsic and extrinsic motivation tools.
As a result of the analysis, no meaningful finding was found that any of the logical and experiential thinking styles were used more by hospital staff, and it was found that the usage rates were close to each other. No significant difference was observed between both gender (male/female) and position (medical/administrative) groups on the basis of logical and experiential thinking styles. As a result of the ANOVA test, no statistically significant differences were found between education groups on the basis of logical and experiential thinking styles. Although it has been stated in previous studies that education has an effect on thinking style [[24], [31]] there is a decision making process for practice in the workplace environment. Especially in the health sector, rapid decision making requires practical thinking. The difference in experience rather than education is likely to make a difference in thinking style. It is suggested that the difference in experience neglected in this study should be included in future studies.
The hypotheses of the study were tested by linear regression analysis. The expectation that “thinking/decision making has an effect on motivation” put forward by the first hypothesis was met (P < 0.001). As stated in the study, the health sector includes critical processes in that the slightest mistake will lead to significant results. Therefore, it is necessary to act very carefully. Pre-motivational thinking/decision-making for implementation has an exceptional importance in the health sector. The reality of this situation is supported in the first hypothesis. In other hypotheses, the relationships between the sub-dimensions of the variables were discussed. As well as the hypothesis “individuals with experiential thinking style are motivated by intrinsic motivation tools” is supported, the highest variance explained in all hypothesis tests emerged in this analysis (1.4%). It is expected that experiential thinking, which indicates the practical thinking gained by the experience of hospital staff, who is mostly motivated by intrinsic motivation tools, has a high relationship with intrinsic motivation [F (1.1218) = 17.865, P < 0.001]. The predictive power on “motivation of individuals with experiential thinking style with extrinsic motivation tools” expected by the third hypothesis was tested and accepted statistically [F (1.1218) = 10.529, P < 0.01].
On the other hand, the last two hypotheses are intended to test the predictive power of rational thinking. The fourth hypothesis questioned “individuals with rational thinking style being motivated by intrinsic motivation tools” and the fifth hypothesis questioned “individuals with rational thinking style being motivated by extrinsic motivation tools” were supported at P < 0.01 level as the results of linear regression analysis.
5 Conclusion and recommendations
All hypotheses were supported according to the findings of the study. The first hypothesis and the most basic finding was “thinking/decision making has an effect on motivation”. Employees in the health sector, which has a vital importance in the service sectors, take urgent and rapid decisions in an environment where they need to act rapidly, and the relationship between motivation sources while implementing these decisions is important. Motivation is the effort and determination of people to achieve a certain goal. The aim of hospital staff is to improve human health and even save lives. In line with this valuable purpose, it will create quality hospital services with the motivation and guidance of the hospital staff. The quality of services supply is also among the issues of importance in the public sphere. In order to solve the tools (internal/external) that employees are motivated to, revealing the ways of thinking and making decisions is an important factor. Because, as in every point of life, individuals use their reasoning skills before they are motivated to take any action in the business environment. The ways that people prefer while performing their daily activities or using their skills in the workplace are referred to as “thinking style”.
With the increase and diversification of health needs, the services offered in the health sector have also increased and created an intense competitive environment. One of the basic conditions for making a difference in service quality and patient satisfaction for the organizations in the sector is to increase the service delivery quality of the hospital staff. In this direction, in order to increase competitiveness, provide quality service and increase patient satisfaction; hospital staff working under intense workload and stress need quick thinking and high motivation [[51], [53], [54], [55]]. Hospital management can take action to motivate their staff, especially by focusing her/his motivation style (intrinsic/extrinsic). Dündar, Özutku and Taşpınar [12] suggested that managers can gain competitive advantage with higher individual and organizational performance levels by discovering the motivation tools that are more effective on the motivation of their staff. As internal motivation tools, employees should be given the opportunity to show their creativity in their work, the importance and dignity of the work should be adopted by the employees, and the employees should be made to feel that they are important. In order to provide motivation with external motivation tools, additional payments to be made depending on success, providing training and promotion opportunities, and establishing a good communication between management and employees can affect motivation positively.
The second main finding is that in most of the hospital staff, acting with intrinsic motivation tools is more likely to be motivated by extrinsic motivation tools. It is expected that healthcare workers, who show the highest self-sacrifice and act altruistically, especially during the COVID-19 period, are motivated by intrinsic motivation tools. Intrinsic motivation tools are defined as tools related to the content of the job (to achieve a job, a challenging job, to be independent at work, to be given responsibility in the workplace, personal and professional development, the importance of the contribution made by the employee, etc.) in Herzberg's Two Factor Theory [43]. Based on this, it can be concluded that healthcare professionals love their jobs and are committed to their duties. The importance of the individual contribution they make in the business environment motivates the employees. At this point, it is recommended that managers give their employeesa a chance to own their work and prove themselves, and finally to appreciate the employees. Another important point reached was on the relationships between the sub-dimensions of motivation (internal/external) and the sub-dimensions of thinking style (experiential/logical). The highest explained variance was between experiential thinking style and intrinsic motivation with a rate of 1.4% among sub-dimensions. This situation can be interpreted as “experiential thinking, which expresses the practical thinking gained from the experiences of the hospital staff, causes them to act with intrinsic motivation”. As Pacini and Epstein [34] pointed out, the experiential system is a non-conscious learning system. It works quickly, automatically and in connection with emotions.
The COVID-19 pandemic has put healthcare professionals around the world in an unprecedented situation, forcing them to make challenging decisions and work under extreme pressure. These decisions include how to allocate scarce resources to patients who are all equally in need, how to meet their own physical and mental health needs alongside those of patients, how to carry out their wishes and duties towards patients in harmony with family and friends, and how to provide care for all severely unwell patients with constrained or inadequate resources. This difficult and fast decision-making obligation can lead to moral injury or mental health problems in healthcare professionals [2]. Making health personnel feel safe during the pandemic is the most effective factor in their performing of their duties, and this effect occurs mostly when it comes from their organizations [3]. In addition, some motivational factors suggested in this process are getting support from colleagues, relatives and other segments of the society; to work and live with appreciation and gratitude; and to strengthen one's ability to self-reflect especially on one's will, potential, and courage [4].
It is important to apply similar studies on motivation and thinking styles to health personnel, where personal contribution is intense. With the results that will emerge, organizations and managers will have the opportunity to create motivation maps for their personnel. Besides, it is recommended that similar studies be repeated a while after the pandemic and comparisons should be made.
Informed Consent
All participants were informed on the questionnaire form and their consent was received. In addition, verbal information was given. Also, this study was approved by the Ardahan University Social and Human Sciences Ethics Committee (date: 8 July 2019) for “Ethics Committee Approval”, which is among the acceptance criteria of the Ministry of Health.
Disclosure of interest
The author declares that he has no competing interest.
Acknowledgements
I wish to thank the Ardahan University Coordinatorship of Scientific Research Projects for supporting my project. Also, I sincerely thank the health workers who participated in this study and gave support, and my students who made great efforts to collect data. My special thanks go to Yusuf Sevgiler, Feyziye Tombak Dizili, Kenan Demirbağ for editing and Murat Kizilkaya for analysis support.
==== Refs
References
1 Galbraith N. Boyda D. McFeeters D. Hassan T. The mental health of doctors during the COVID-19 pandemic BJ Psych Bull 45 2021 93 97
2 Greenberg N. Docherty M. Gnanapragasam S. Wessely S. Managing mental health challenges faced by healthcare workers during COVID-19 pandemic BMJ 368 2020 m1211 32217624
3 Malesza M. Factors informing healthcare workers’ willingness to work during the COVID-19 pandemic medRxiv 2021 1 19
4 Sperling D. Ethical dilemmas, perceived risk, and motivation among nurses during the COVID-19 pandemic Nurs Ethics 28 2021 9 22 33000673
5 Karadağ M. Alparslan Ö. İşeri Ö.P. Ebelik ve hemşirelik öğrencilerinin eleştirel düşünme eğilimleri ve öğrenme stilleri Çukurova Med J 43 2018 13 21
6 Dalkıran M. TSK asker hastanelerinde görev yapan sağlık personelinin daha verimli çalışmasına etki eden motive edici faktörlerin belirlenmesi (Yayınlanmamış yüksek lisans tezi) 1993 Gülhane Askeri Tıp Akademisi Ankara
7 Soykenar M. Sağlık işletmelerinde personelin motivasyonunu etkileyen faktörler: Dokuz Eylül Üniversitesi Hastanesinde örnek bir uygulama (Yayınlanmamış doktora tezi) 2008 Dokuz Eylül Universitesi İzmir
8 Ralston J.D. Larson E.B. Crossing to safety: transforming healthcare organizations for patient safety J Postgrad Med 51 2005 61 15793345
9 Özdemir M. Sağlık çalışanlarının sorunları Toplum ve Hekim 14 1996 402 411
10 Aykanat S. Sağlık işletmelerinde motivasyonun personel verimliliği üzerine etkilerine ilişkin bir araştırma (Yüksek lisans tezi) 2002 Gazi Üniversitesi Ankara
11 Deci E.L. Ryan R.M. Williams G.C. Need satisfaction and the self-regulation of learning Learn Individual Differences 8 1996 165 183
12 Dündar S. Özutku H. Taşpınar F. İçsel ve dışsal motivasyon araçlarının işgörenlerin motivasyonu üzerindeki etkisi: Ampirik bir inceleme Ticaret ve Turizm Eğitim Fakültesi Dergisi 2 2007 105 119
13 Fisher D.C. Yuan Y.X. What motivates employees? A comparison of US and Chinese responses Int J Hum Resource Manag 9 1998 516 528
14 Kovach A.K. What motivates employees? Workers and supervisors give different answers Business Horizons 30 1987 58 66
15 Leete L. Wage equity and employee motivation in non-profit and for-profit organizations J Econ Behav Organ 43 2000 423 446
16 Mak B.L. Sockel H.Y. A confirmatory factor analysis of employee motivation and retention Inf Manag 38 2006 265 276
17 Ross D.L. Attitudes and work motivation of subgroups of seasonal hotel workers Serv Ind J 15 1995 295 313
18 Wiley C. What motivates employees according to over 40 years of motivation surveys? Int J Manpower 18 1997 263 280
19 Deci E.L. Ryan R.M. Handbook of self-determination research 2004 University of Rochester Press New York
20 Ryan R.M. Deci E.L. Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being Am Psychol 55 2000 68 78 11392867
21 Wong S.C. Ladkin A. Exploring the relationship between employee creativity and job-related motivators in the Hong Kong hotel industry Int J Hosp Manag 27 2008 426 437
22 Berkant H.G. Tüzer L. Sınıf öğretmenlerinin düşünme stillerinin çeşitli değişkenler açısından incelenmesi Uluslararası Sosyal Araştırmalar Dergisi 11 2018 477 488
23 Semerci N. Öğretmenin görevi: Düşünmeyi geliştirmek Fırat Üniversitesi Sosyal Bilimler Dergisi 9 1999 209 216
24 Dinçer B. Saracaloğlu A.S. Öğretmen adaylarının düşünme stilleri profillerinin çeşitli değişkenler açısından incelenmesi Türk Eğitim Bilimleri Dergisi 9 2011 701 744
25 MEB Düşünme eğitimi dersi 6,7,8. sınıf öğretmen kılavuz kitabı 2007 MEB Yayınları Ankara
26 Nickerson R.S. Why teach thinking? Baron J.B. Sternberg R.J. Series of books in psychology. Teaching thinking skills: theory and practice 1987 W H Freeman/Times Books/Henry Holt & Co NY 27 37
27 Sternberg R.J. Thinking styles 1997 Cambridge University Press NY
28 Fer S. Aday öğretmenlerin düşünme stilleri nedir? Proceedings of the XIV. Ulusal Eğitim Bilimleri Kongresi 2005 Pamukkale Üniversitesi Eğitim Bilimleri Fakültesi Denizli
29 Cano Garcia F. Hughes E.H. Learning and thinking styles: an analysis of their interrelationship and influence on academic achievement Educ Psychol 20 2000 413 430
30 Allinson C.W. Hayes J. The cognitive style index: a measure of intuition analysis for organizational research J Manag Stud 33 1996 119 135
31 Çubukçu Z. Proceedings of the XIII. Ulusal Eğitim Bilimleri Kurultayı Öğretmen adaylarının düşünme stillerinin öğrenme biçimlerini tercih etmelerindeki etkisi 2004 İnönü Üniversitesi Malatya
32 Epstein S. Cognitive-experiential self-theory of personality Millon T. Lerner M.J. Comprehensive Handbook of Psychology, Vol. 5: Personality and Social Psychology 2003 Wiley & Sons Hoboken, NJ 159 184
33 Evans J. Dual processing accounts of reasoning, judgement and social cognition Ann Rev Psychol 59 2008 255 278 18154502
34 Pacini R. Epstein S. The relation of rational and experiential information processing styles to personality, basic beliefs, and the ratio-bias phenomenon J Pers Soc Psychol 76 1999 972 987 10402681
35 Türk E.G. Artar M. Mantıksal deneyimsel düşünme ölçeğinin uyarlanması: Geçerlik ve güvenirlik çalışması Ankara Üniversitesi Eğitim Bilimleri Fakültesi Dergisi 47 2014 1 18
36 Golian L.M. Thinking style preferences among academic librarians: practical tips for effective work relationships 1999 ACRL Ninth National Conference Detroit, Michigan
37 Zhang L. Does student–teacher thinking style match/mismatch matter in students’ achievement? Educ Psychol 26 2006 395 409
38 Kanbay Y. Aslan Ö. Işık E. Kılıç N. Hemşirelik lisans öğrencilerinin problem çözme ve eleştirel düşünme becerileri Yüksekögretim ve Bilim Dergisi 3 2013 244 251
39 Kurtuluş K. Pazarlama araştırmaları 7th ed. 2004 Literatür Yayıncılık Istanbul
40 Aydın İ. Sosyal bilimlerde araştırmadan yayına etik değerler</CT> Karakütük K. Sosyal Bilimlerde Süreli Yayıncılık 1. Ulusal Kurultay Bildirileri 2006 TÜBİTAK ULAKBİM Sosyal Bilimler Veri Tabanı Ankara [s. 71-80]
41 Mottaz J.C. The relative importance of intrinsic and extrinsic rewards as determinants of wok satisfaction Sociol Q 26 1985 365 385
42 Brislin R.W. Kabigting F. Macnab B. Zukis B. Worthley R. Evolving perceptions of Japanese workplace motivation Int J Cross Cult Manag 5 2005 87 103
43 Mahaney CR, Lederer AL. The effect of intrinsic and extrinsic rewards for developers on information systems project success. Project Management Journal 37(4):42-54.
44 Türk E.G. Ergenlerin düşünme biçemlerini yordayan faktörler: Anne baba, üstbiliş ve epistemolojik inançlar (Yayınlanmamış Ddktora tezi) 2011 Ankara Universitesi Ankara
45 Belousova A. Mochalova Y. The relationship of thinking style and motivation features of sales and advertising managers Behav Sci 10 2020 68 32183430
46 Doménech-Betoret F. Gómez-Artiga A. The relationship among students’ and teachers’ thinking styles, psychological needs and motivation Learn Individual Differences 2 2014 89 97
47 Fan W. Zhang L.F. Are achievement motivation and thinking styles related? A visit among Chinese university students Learn Individual differences 19 2009 299 303
48 Rafat A.H. Novinfar A.E. Hasanloo H.O. Hedayati A. Study of the relationship between thinking styles and achievement motivation of students Contemp Educ Res J 1 2013 16 21
49 Varisoglu B. Motivation to learn Turkish in foreign students with different thinking styles in terms of functional and formal aspects Educ Res Rev 13 2018 328 335
50 Ağırbaş İ. Çelik Y. Büyükkayıkçı H. Motivation means and job satisfaction: a study on deptuy of head physicians working at the hospitals of department of social insurance organization Hacettepe Sağlık İdaresi Dergisi 8 2005 326 350
51 Doğanlı B. Demirci Ç. Sağlık kuruluşu çalışanlarının (hemşire) motivasyonlarını belirleyici faktörler üzerine bir araştırma Celal Bayar Üniversitesi İktisadi ve İdari Bilimler Fakültesi Dergisi 21 2014 47 60
52 Kıdak L. Aksaraylı M. Sağlık hizmetlerinde motivasyon faktörleri Manisa Celal Bayar Üniversitesi Sosyal Bilimler Dergisi 7 2009 75 94
53 Orhaner E. Mutlu S. Sağlık personelinin iş tatmininin motivasyon üzerine etkisi Uluslararası Sağlık Yönetimi ve Stratejileri Araştırma Dergisi 4 2018 74 93
54 Karakaya A. Ay F.A. Çalışanların motivasyonunu etkileyen faktörler: Sağlık çalışanlarına yönelik bir araştırma Cumhuriyet Üniversitesi Sosyal Bilimler Dergisi 31 2007 55 67
55 Korkmaz S. Hastanelerde doktor, hemşire ve ebelerin motivasyonunu etkileyen faktörler: Bir uygulama (Yayımlanmamış yüksek lisans tezi) 2008 Çağ Üniversitesi Sosyal Bilimler Enstitüsü Mersin
| 0 | PMC9746888 | NO-CC CODE | 2022-12-15 23:21:58 | no | Ann Med Psychol (Paris). 2022 May 13; doi: 10.1016/j.amp.2022.05.006 | utf-8 | Ann Med Psychol (Paris) | 2,022 | 10.1016/j.amp.2022.05.006 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Published by Elsevier Ltd.
S0006-3207(22)00150-1
10.1016/j.biocon.2022.109597
109597
Article
Ruling the roost: Avian species reclaim urban habitat during India's COVID-19 lockdown
Madhok Raahil ab
Gulati Sumeet ab⁎
a Food and Resource Economics Group, Faculty of Land and Food Systems, University of British Columbia, 2357 Main Mall, Vancouver, BC V6T 1Z4, Canada
b The Wildlife and Conservation Economics Lab, University of British Columbia, 2357 Main Mall, Vancouver, BC V6T 1Z4, Canada
⁎ Corresponding author at: 341-2357 Main Mall, University of British Columbia, Vancouver, BC V6T 1Z4, Canada.
19 5 2022
7 2022
19 5 2022
271 109597109597
6 2 2021
2 5 2022
14 5 2022
Crown Copyright © 2022 Published by Elsevier Ltd. All rights reserved.
2022
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
As we retreated to our dwellings in the “anthropause” of spring 2020, were the wildlife sightings in previously crowded spaces a reclamation of habitat, or a mere increase in detection? We leverage an increase in balcony birdwatching, a million eBird entries, and difference-in-difference techniques to test if urban avian species richness rose during India's COVID-19 lockdown. Controlling for effort, birdwatchers in the 20 most populous cities observed a 16% increase in the number of species during lockdown. While human activity stopped overnight, and noise and visual pollution decreased soon after, increased species diversity was observed 1–2 weeks later; evidence that gradual population recovery, not better detection, underlay our results. We find at-risk, and rare, species among those reclaiming cities, implying that reducing human disturbance in urban areas can protect threatened species. Increased species diversity likely derives from a reduction in noise and air pollution associated with the lockdown, implying that urban planners should consider conservation co-benefits of urban policies when designing sustainable cities.
Keywords
COVID-19
Biodiversity
Conservation
India
Wildlife
Coexistence
==== Body
pmc1 Introduction
On March 27th, 2020, three days after India's abrupt cessation of human activity, a reporter posted a video on Twitter of a Sambar deer crossing a road in Chandigarh—among India's most population dense cities (Ghazali, 2020). Two days later, in the same city, a leopard was seen and tranquilized (Tribune News, 2020). Similar sightings were reported elsewhere; a Nilgai walked along a usually bustling street in a Delhi suburb, a critically endangered Malabar Civet wandered alongside traffic in Kozikhode, Kerala, and flamingos returned in unusually large numbers to the Mumbai wetlands (Guy and Gupta, 2020). These sightings were not unique to India. As countries reined in human activity to fight the spread of COVID-19, animals were reported in urban locations across the globe, including pumas in Santiago, Chile, dolphins in the Bosphorous, mountain goats in Wales, and deer in Nara, Japan, and Romford, UK. Were these chance sightings, or a systematic increase in animal presence in previously forbidding urban areas? The cascade of global COVID-19 lockdowns provides a unique opportunity to shed light on this (Bates et al., 2020; Rutz et al., 2020).
Here, we use citizen science data from before and during India's lockdown to evaluate how the “anthropause” (a term coined by Rutz et al. (2020)) affects urban avian diversity. We also investigate if changes in avian diversity derive from a change in abundance, or from improved detection. Finally, we employ a species level analysis to provide a deeper understanding of the avifauna repopulating the cities in our sample. Together, our findings have important management lessons for optimizing the design of sustainable cities.
We study birds because they are a known proxy for ecological health (Morrison, 1986; Xu et al., 2018), are easily observable, and are among the few animals documented with unparalleled spatiotemporal coverage. We use over one million bird sightings from eBird, the world's largest citizen science platform (Cornell Lab of Ornithology, 2020; Sullivan et al., 2009), to test whether observed species richness in India's 20 most populous cities rose during the COVID-19 lockdown. As wildlife surveys stalled under lockdown (Corlett et al., 2020), eBird usage soared (Hochachka et al., 2021). This allows us to present a robust and comprehensive analysis of our question.
India is among the best settings to learn about human-wildlife interactions. It juxtaposes a backdrop of unrivaled biodiversity, home to 12% of the world's bird species (Jayadevan et al., 2016), against dense urban centres that are particularly forbidding for wildlife. Indeed, 21 out of the world's 30 most polluted cities (IQAir, 2019), and 4 out of the 10 most traffic congested cities are in India (TomTom International, 2020). These cities typically have noise levels deemed by the World Health Organization as unhealthy for humans (Dantewadia et al., 2020), and featured a deep and sudden reduction in human activity during spring 2020 (see section S1.1 in Supplementary Materials for timeline).
Our methodology addresses two main empirical challenges. First, we establish a credible counterfactual to ensure that eBird entries before and after March 25th 2020 are statistically similar except for the imposition of lockdown. The crux of this strategy is the use of the difference-in-difference (DD) technique (Angrist and Pischke, 2008), with the same 48-day period in 2019 as the basis for our counterfactual. Second, we ascertain if changes in observed species diversity represent changes in animal behaviour (abundance) or human behaviour (detection ability). Several studies discuss this in detail (Manenti et al., 2020; Simons et al., 2007; Williams et al., 2002), yet few show how to disentangle these competing mechanisms. Here, we introduce a way to separate the abundance and detection mechanisms by estimating impacts over time. As changes in human activity, and associated reductions in noise and pollution, occur almost immediately after lockdown, any sizeable lag suggests that changes in avian presence rather than improved detection underlie our results.
2 Methods
2.1 Selecting a counterfactual
To estimate the causal impact of India's lockdown on avian diversity, we need to compare observed species richness during lockdown with a counterfactual when lockdown never occurred. The weeks prior to lockdown are inappropriate as the available species pool differed. Migratory species arrive during India's warm winter and depart in Spring, coinciding with the lockdown period (Veen et al., 2005). A simple pre-post comparison thus conflates lockdown with species migration.
We choose the same period in 2019 as our counterfactual. The assumption being, had lockdown not occurred, species richness would have evolved as it did during the same 48-day window in 2019 (the parallel trend assumption). The difference in species richness before and after the fourth Wednesday in March 2019 captures species migration, and we subtract this from the pre-post difference in 2020. Any remaining “difference in differences” (Angrist and Pischke, 2008) is devoid of the bias from migration (see S1.6 for a mathematical derivation).
For this method to work, we first need to establish that species diversity during March and April 2019 was not systematically different from other non-lockdown years. To verify the robustness of our counterfactual, we compare the daily mean species richness per trip on a given day in 2019 to the same day in 2018. Since there was no lockdown in either year, we expect no difference in daily species richness. Fig. S5B confirms this, showing no systematic difference between 2019 and 2018 and thereby validating the robustness of our choice of 2019 as the counterfactual period. See Section S1.7.2 for more details on this placebo check.
2.2 Mitigating selection bias
Whereas our counterfactual accounts for species migration, it does not account for the changing nature of birdwatching after March 25th 2020 (see Section S1.4). We use a DD design that eliminates biases from these contemporaneous changes by taking advantage of eBird's requirement to enter a trip protocol (e.g., stationary or travelling; Table S1) and other trip characteristics alongside a species checklist (see Section S1.2). Immediately following lockdown, there is a near-quadrupling of stationary users across India (Fig. 1A) and the number of trips they report (Fig. 1B), consistent with reports of surges in balcony birding in isolation (Fortin, 2020). Correspondingly, there is an unsurprising reduction in users taking travelling trips. To exclude new sign-ups, who likely have different characteristics than their pre-lockdown counterparts, we impose a participation constraint to compare checklists from a constant user base (Section S1.4, Table S3). In our preferred sample, only users recording at least two trips in the 24 days before and after lockdown—called consistent users—are selected.Fig. 1 Daily birdwatching activity in India from eBird trips reported between March 1st and April 17th, 2020. Data is pre-processed as described in Section S1.3. Panel A) plots the number of unique user ID's active each day; panel B) plots the total number of daily trips; panel C) plots collective, daily species richness. The vertical dashed line denotes lockdown (March 25th, 2020).
Fig. 1
Before lockdown, collective species richness was substantially higher on travelling trips compared to stationary ones, because users could venture deeper into bird habitats (Fig. 1C). After lockdown, this trend reverses as travelling becomes restricted, resulting in a species richness decline when protocols are pooled (Fig. S1). Within stationary trips, however, species richness generally increases post-lockdown. Thus, we make all pre-post comparisons within trips of the same type to remove the protocol bias.
Additional biases arise from the changing schedule of birdwatching during lockdown (Fig. 2 ) and rural-urban differences (discussed in Section S1.4). We use hour-of-day fixed effects to control for the former and restrict our analysis to the top 20 densest cities (Table S2) to control for the latter.Fig. 2 Hour-of-day distribution of eBird trips reported in 2020 in the top 20 cities by population density. Data covers consistent users who recorded at least two trips in each of the 24 days before and after lockdown. Janta Curfew (March 22nd) is dropped.
Fig. 2
Lastly, we control for a range of climatic and behavioural variables—including weather and trip duration—that change during our study window and also impact species diversity (Section S1.5). For example, mean temperature was 2° warmer during lockdown compared to the 24 days prior (Table S4B). Equivalent warming is also observed across the same period in 2019. In contrast, rainfall was 0.2 mm higher during lockdown compared to before, whereas in 2019 rainfall decreased over the same period. Our climatic controls account for such weather aberrations relative to the counterfactual so that species observations before and after March 25th 2020 are statistically similar except for the imposition of lockdown, delivering precise estimates of the ecological impact of lockdown in India's urban core.
2.3 Difference-in-difference model
Combining our counterfactual with the procedure for mitigating selection bias, we proceed to estimate the DD specification (Angrist and Pischke, 2008). The goal is to compare changes in outcomes in a treatment group before and after a policy date (the first difference) with changes in outcomes over the same period in a counterfactual where the policy was never implemented (second difference). We estimate:(1) SRijdyt=α+δTreatmenty×Tt+γTreatmenty+λTt+Xijdyt+ηp+μd+θt+εijdyt
where SR ijdyt is species richness observed by user i on trip j in district d during year y and day-of-year t. Treatment y is a dummy for 2020. T t is a time dummy for the post period t ∈ [Policy t , Policy t + 24], where Policy t is the 4th Wednesday of March—the “policy date”. The pre-policy period in 2020 is March 1st until Policy t, and the post-policy period spans 24 days afterwards. X ijdyt is a vector of weather and trip covariates (Section S1.5). ηp is a protocol fixed effect that ensures all pre-post comparisons are made among trips of the same type. μ d is a district fixed effect and ensures comparisons are made across trips within the same district (Table S2). θ t is a set of temporal controls including hour-of-day fixed effects and weekend fixed effects. δ is our parameter of interest, denoting the causal impact of India's COVID-19 lockdown on species richness. Section S1.6 provides a detailed mathematical derivation.
2.4 Dynamic difference in differences
The term δ in Eq. (1) captures an average impact over 24 days of lockdown. We decompose this to parameterize marginal effects over time in order to investigate whether our estimates represent actual changes in species presence, or whether birds are just easier to observe in the absence of human activity. A reduction in noise pollution associated with the reduction in human activity—as found by Mishra et al. (2021) during India's lockdown—may lead to an increase in species detection. Indeed, Simons et al. (2007) find that detection probability is 42% lower with 10 dB of white noise. Since human activity stopped overnight, a lag between lockdown and higher species diversity suggests a gradual recovery of species. We estimate:(2) SRijdyt=α+∑k=−4K=4δkTreatmenty×Tk+γTreatmenty+∑k=−4K=4λkTk+Xijdyt+υd×t+ηp+μd+θt+εijdyt
where δk is the DD estimate in time bin k. We use 8 bins with 6 days each so that every bin has the same number of days (48 days/8 bins = 6 days per bin). k = (−6, 0] is omitted so that all estimates are relative to the six days before (and including the day of) lockdown. We also include a district level time trend, d × t, to control for a linear district-specific trend in species richness that may be observed even in the absence of lockdown. All other parameters are defined and interpreted in the same way as Eq. (1).
2.5 Marginal species identification
SRijdyt measures changes in the number of species, but not which ones are seen more or less often during lockdown. Thus, we conduct a species-level analysis to complement our DD estimates. We use the same sample used in our DD regression and calculate frequency distributions for individual species in 2019 and 2020 in Bangalore and Delhi—two cities with the most eBird activity. We then identify 28 “marginal” species (out of 211 total) in Bangalore and 30 (out of 147) in Delhi according to three criteria: a) the average daily proportion of checklists reporting the species after the policy date is higher in 2020 than 2019; b) the average daily proportion of checklists reporting the species pre-lockdown in 2020 is no more than in 2019, thus excluding species consistently more common in 2020; c) the species is observed on at least seven days in the post-lockdown period in 2020.
Having established which species are observed more often during lockdown, we classify the rarity of these marginal species at a global and local level. For the global classification, we list the IUCN Red List category. Since this may misrepresent the local threat level, we classify a species as locally rare if their 2019 reporting frequency is in the bottom 25th percentile, a common threshold in the literature (Gaston, 1994).
Note that δ in Eq. (1) reveals the net change in species diversity during lockdown. Whereas the marginal species analysis identifies those observed more often, there may also be species that retreated away during lockdown. For completeness, we investigate this opposing scenario by reversing the criteria above. This procedure identifies species observed less frequently during India's lockdown, which we call retreating species.
3 Results
3.1 Lockdown increases species diversity in cities
Before turning to the formal regression results, we first illustrate our research design, without controls or fixed effects. Fig. 3 shows mean species richness per trip on a given day in 2020 relative to that on the same day in 2019. Species richness on travelling trips drops after the fourth Wednesday in March (the policy date), but gradually increases among stationary trips, compared to 2019. This corroborates a story of recovering species abundances (or diversity, see Discussion in Section 4). Importantly, in the absence of lockdown (left of vertical dashed line), there was no systematic difference in species richness between 2020 and 2019, bolstering our choice of 2019 as the counterfactual. Fig. S2 shows the same illustration for number of users and trips. Stationary activity spikes and travelling activity drops after the 2020 policy date compared to 2019 due to mobility restrictions in the former but not latter.Fig. 3 Daily species richness in 2020 relative to 2019 in India. Selected users meet 2-trip participation constraint in both years. The policy date (dashed vertical line) is the 4th Wednesday of March, the date of lockdown announcement in 2020. Solid lines describe mean species richness per trip across users on a given day in 2020 minus the value from the same day in 2019.
Fig. 3
The formal DD estimates (Eq. (1)) show robust evidence that a reduction in human activity increases observed avian diversity (Fig. 4A). The first coefficient—our preferred specification—shows that India's COVID-19 lockdown increases species richness by 2.27 species per trip (p < 0.01) in the top 20 cities, equivalent to 16% of the pre-lockdown mean. The other two coefficients show that tightening the participation constraint to five and ten trips yields point estimates of 2.22 (16% increase) and 1.99 (14% increase), respectively. Since more experienced users are selected when the participation constraint tightens, the number of selected users drops and sample size reduces each time.Fig. 4 Difference in difference results. White circles in panel A) describe the main DD estimate from Eq. (1). Panel B) shows the dynamic estimates from Eq. (2). The x-axis denotes 6-day time bins and negative values denote days before lockdown. In both panels, bars show 95% confidence intervals, the estimation sample is pre-processed, covers the top 20 cities, and data from March 22nd, 2020 (the Janta Curfew) are dropped. All regressions include district, hour-of-day, and protocol fixed effects as well as controls for trip duration, rain, temperature, number of observers, distance to nearest birding hotspot, and a weekend dummy. Standard errors are robust to heteroskedasticity.
Fig. 4
Next, we decompose the DD estimates into weekly bins (Eq. (2)) to investigate whether our estimates represent changes in animal or human behaviour. The increase in species richness is detected after a one-week lag (p < 0.01), and persists through the second week of lockdown (Fig. 4B). The district-time trend in Eq. (2) ensures that these weekly estimates reflect deviations from a linear trend, thereby identifying the impact of lockdown and not a trend in species richness that would have occurred regardless. Pre-lockdown, there is no statistical difference in week-to-week species richness between 2020 and 2019, providing formal support for the parallel trend assumption.
Our results are robust to a variety of alternative specifications (see Section S1.7). Including all cities in India, instead of only the top 20 most population dense ones, reduces the point estimate nearly three times. This is due to the inclusion of rural areas where human activity was less affected by the lockdown (Fig. S3, column 2). Despite the participation constraint and effort covariates, we also try a specification with user fixed effects to capture differences in ability between users. However, since very few users are observed during the pre- and post-period in both 2019 and 2020, we cannot add a user fixed effect directly in Eq. (1). Instead, we compute the pre-post difference in species richness in each year separately with a user fixed effect, and then manually subtract the coefficients (this is mechanically equivalent to our DD specification). We continue to find increased species richness during lockdown, but precision decreases because there is substantially less variation within a user and therefore larger standard errors (Fig. S3, columns 3 and 4).
We also estimate our dynamic specification under the five-trip participation constraint. The increase in species richness in the first and second week of lockdown is once again observed (Fig. S4). The magnitude and precision are similar to that in Fig. 4B, and there is also no pre-trend leading up to the policy date.
Lastly, our results are also robust to the level of aggregation (Section S1.7.3). We re-estimate our main specification at the city-level using mean species richness per trip across all trips recorded in a city-day, weighted by the number of trips in the day. The coefficients are largely similar to our main results at the trip level (Fig. S7). See Section S1.7 for more details as well as additional sensitivity checks.
3.2 Both common and rare species repopulate cities
Fig. 5 illustrates frequency distributions for four marginal species in Bangalore and Delhi (see Fig. S5 for full set). In Bangalore, the Black-rumped Flameback Woodpecker is never reported in 2019, and never reported in the pre-period of 2020, but we find several reports of the species 1–3 weeks into lockdown. In Delhi, the Black-rumped Flameback is reported by a larger proportion of checklists in the pre-period in both years. In 2019, however, it is no longer observed during the post-period, but in 2020, continues to be reported throughout lockdown. A similar pattern is seen for the Large-billed Crow in Delhi: reported by a similar proportion of eBird checklists in the pre-period of 2019 and 2020, but mainly observed in the post period in 2020.Fig. 5 Marginal species distributions of four example species in Delhi (panel A) and Bangalore (panel B). The x-axis is the number of days relative to the 4th Wednesday of March, the lockdown date in 2020. The y-axis is the share of checklists reporting the species on a given day. The sample of checklists consists of stationary trips from the main regression sample under the 2-trip participation constraint. Distributions for remaining marginal species are shown in Figs. S9 and S10 in the Supplementary materials.
Fig. 5
Among marginal species in Delhi and Bangalore, a handful are rare and the majority are common species detected more frequently. The IUCN classification yields two globally near-threatened species: the Black-headed Ibis in Bangalore and the Alexandrine Parakeet in Delhi (distributions in Fig. 5). The remaining marginal species in both cities are Least Concern. Our local rarity criteria classify seven (out of 28) marginal species in Bangalore (Table S5) as locally rare and six (out of 30) in Delhi (Table S6).
In stark contrast to the 28 marginal species in Bangalore and 30 in Delhi, we find five retreating species in Bangalore and only one in Delhi. This indicates that the increase in abundance and diversity of species during lockdown more than compensates for the few species simultaneously exiting urban spaces. Figs. S9 and S10 show frequency distributions for these species in Bangalore and Delhi, respectively. In Delhi, the Rosy Starling is never reported in the pre-period of 2019 or 2020. In the post-period, however, it is observed less frequently on average in 2020 compared to 2019. In Bangalore, most retreating species are observed at similar daily frequencies in both years, but slightly less often in the 2020 post-period. Unlike “true” marginal species—such as the Black-rumped Flameback in Bangalore—there is no equivalent for retreating species in either city (always seen in 2019 and only seen in 2020 pre-period). In terms of rarity, all six species are listed as Least Concern by the IUCN (Table S7).
4 Discussion
Our analysis presents a measurable change in the viewing of avian species. The cessation in human activity on March 25th 2020 was abrupt and strongly enforced. If the additional 2.27 species from our DD estimate were always present, but undetected because of distractions from human activity, then users should detect additional species immediately following lockdown (when human activity ceased). We find that balcony bird-watching soars the next day (Fig. 1A, B), but the fact that it takes up to two weeks to detect additional species suggests that the abundance of incumbent species, or the emergence of species previously absent, gradually grew until the probability of detection was high enough two weeks after lockdown (Fig. 4B). On the other hand, the viewing of these additional species seems to return to pre-lockdown levels by the end of the last week of lockdown. We do not have a verifiable hypothesis for why this occurs.
Our species-level analysis confirms that our findings are indeed based on both an increase in the abundance of incumbent species as well as the emergence of previously absent species (Fig. 5). Some species are seen frequently in Delhi and Bangalore before and after the 4th Wednesday of March in both years, but a greater proportion of checklists report them in the post lockdown period in 2020. Some are never seen in 2019 altogether, and only seen in 2020 during lockdown, exhibiting a “true” marginal observation.
Among the marginal species, the Black-headed Ibis in Bangalore, and the Alexandrine Parakeet in Delhi, are globally near-threatened according to the IUCN. At a local level, we classify 25% and 20% of species seen more frequently during lockdown as locally rare in Bangalore and Delhi, respectively. Rare species face greater extinction risk and are more sensitive to environmental changes (Gaston, 1994), making our study useful for allocating scarce conservation budgets. Our results suggest that investment in making our cities more wildlife friendly can also protect some at-risk species, not just the urban specialists.
At least two environmental mechanisms can explain the increase in avian diversity and abundance found in this study. The first is related to noise pollution. The abundance and occupancy of avian species are negatively impacted by noise pollution (McClure et al., 2013; Shannon et al., 2016). Specifically, elevated noise levels mask mating signals and defense mechanisms (Slabbekoorn, 2013). A February 2020 report in LiveMint, an e-paper in India, reports that average noise levels recorded by monitors in residential areas of Mumbai, Delhi, Bengaluru, Kolkata, Chennai and Hyderabad were 10 dB higher than the maximum recommended by the Central Pollution Control Board. For the city of Kanpur, Uttar Pradesh, Mishra et al. (2021) find average sound levels in the range of 42–87 dB before lockdown, and a reduction to 38–66 dB during lockdown (Dantewadia et al., 2020). Since the lockdown reduced traffic and other anthropogenic noise, certain species may reoccupy the landscape in larger numbers. However, lower noise pollution also increases the ability of observers to hear bird calls, but our dynamic DD results (Fig. 4B) show evidence ruling out this alternative explanation.
The second possibility is a drop in air pollution. 21 of the 30 most polluted global cities are in India (IQAir, 2019). This includes 9 cities in our sample. During lockdown, these cities experienced unprecedented reductions in PM2.5 and other pollutants (Mahato et al., 2020; Sharma et al., 2020). Exposure to particulates can reduce species diversity (Liang et al., 2020; Sanderfoot and Holloway, 2017). Therefore, air quality improvements may underlie the higher species diversity we observe. However, lower air pollution, especially particulate pollution, also improves visibility, another important factor for species detection. Again, our dynamic results rule out this mechanism.
Taken together, this paper makes three important contributions to our understanding of anthropogenic pressures on avian diversity. First, we estimate the causal impact of reducing human activity on avian diversity in urban settings (Fig. 4A). We share this contribution with a nascent literature studying the impact of COVID-19 lockdowns on wildlife elsewhere (Vardi et al., 2021). Studies based on pre-COVID data typically evaluate species diversity along an urban gradient (see Chace and Walsh (2006) for a review). Such analyses confound changes in human activity with changes in habitat (Verma and Murmu, 2015; Xu et al., 2018). Instead, we compare species diversity as human activity varies within the same urban habitat.
Second, we present a method separating abundance from detection in observational surveys, something particularly important for studies analyzing avian populations. While technology-based methods capture animal presence, observational surveys typically capture species detection. Terrestrial species are increasingly monitored with autonomous cameras (Silveira et al., 2003), GPS collars, and radio collars (Cagnacci et al., 2010). In avian surveys, however, human observation remains the primary method to collect data, whether via systematic surveys or citizen science. Despite this, only 5% of 224 ornithology papers reviewed by Rosenstock et al. (2002) address the issue of abundance vs. observation. We formally investigate whether changes in species diversity represent changes in presence, or mere improvements in observer cognition from reduced noise and visual pollution.
Third, we present robust estimates derived from a very large sample collected over a geographically and culturally diverse region, extending external validity to other developing countries where urbanization is accelerating and large swathes of species are imperiled (Jenkins et al., 2013; Myers et al., 2000; Newbold et al., 2016). Many studies estimating the impact of COVID-19 lockdowns on animal presence or behaviour use small samples in distinct study sites in developed countries. For example, Manenti et al. (2020) survey water birds at an artificial lake during Italy's COVID-19 lockdown and find higher species richness compared to 2019. Derryberry et al. (2020) find that birds increase their acoustic distance during San Francisco's lockdown, and possibly improve breeding success (Manenti et al., 2020). In contrast, our estimates represent a statistically robust average impact of reduced human activity on avian diversity for 20 large cities across India.
In addition to the empirical contributions, our results also have important conservation policy implications. Our findings imply that policies influencing human activity, and consequently noise and air pollution, in urban centres also have conservation co-benefits. The most prominent examples of such policies are those managing traffic: congestion pricing, odd-even license plate restrictions, road closures, etc. However, most research evaluating urban traffic policies focus on congestion, air pollution, and human health effects (Farda and Balijepalli, 2018; Kumar et al., 2017; Simeonova et al., 2019; Zhou et al., 2010). Our results imply that researchers and policymakers should also consider their effect on avian (and other species) diversity. By not including these impacts they are likely underestimating policy benefits.
It is also important to note that there are important benefits from bird watching and engaging with nature in its own right (Bratman et al., 2019; Maldonado et al., 2018; Shanahan et al., 2016). A change in the viewing of birds documented in this paper (Fig. 1) represents a change in the well-being for bird-watchers during an unprecedented time. However, it behooves us to recognize that only those with the ability and resources to engage in leisure, in the face of the economic and humanitarian crisis precipitated by the pandemic, may be able to engage in this possibility.
CRediT authorship contribution statement
Raahil Madhok: Conceptualization, methodology, software, formal analysis, data curation, writing original draft, reviewing and editing draft, visualization. Sumeet Gulati: Methodology, resources, writing original draft, reviewing and editing draft, supervision, project administration.
Data and materials availability
Further details on data supporting the findings of this study are available in the Supplementary Materials. Raw datasets and code for reproducing summary tables, figures, and empirical estimates are deposited in a github repository: https://github.com/rmadhok/ebird-lockdown.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supplementary data
Supplementary material
Image 1
Acknowledgements
We are grateful to Nicola Koper and the rest of the C19-Wild Group, Patrick Baylis, and three anonymous referees for constructive suggestions. We also thank participants of the Canadian Economics Association, the UBC WCEL seminar, and the BIOECON conference for helpful comments. This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors. Raahil Madhok received doctoral research funding from the SSHRC CGS Doctoral Fellowship.
Appendix A Supplementary material for this article can be found online at https://doi.org/10.1016/j.biocon.2022.109597.
==== Refs
References
Angrist J.D. Pischke J.S. Mostly harmless econometrics: an empiricist’s companion Mostly Harmless Econometrics: An Empiricist’s Companion 2008 Princeton University Press Princeton, NJ 10.1111/j.1475-4932.2011.00742.x
Bates A.E. Primack R.B. Moraga P. Duarte C.M. COVID-19 pandemic and associated lockdown as a “Global human confinement experiment” to investigate biodiversity conservation Biol. Conserv. 248 2020 108665 10.1016/J.BIOCON.2020.108665
Bratman G.N. Anderson C.B. Berman M.G. Cochran B. de Vries S. Flanders J. Folke C. Frumkin H. Gross J.J. Hartig T. Kahn P.H. Kuo M. Lawler J.J. Levin P.S. Lindahl T. Meyer-Lindenberg A. Mitchell R. Ouyang Z. Roe J. Scarlett L. Smith J.R. van den Bosch M. Wheeler B.W. White M.P. Zheng H. Daily G.C. Nature and mental health: an ecosystem service perspective Sci. Adv. 5 2019 10.1126/SCIADV.AAX0903/SUPPL_FILE/AAX0903_SM.PDF
Cagnacci F. Boitani L. Powell R.A. Boyce M.S. Animal ecology meets GPS-based radiotelemetry: a perfect storm of opportunities and challenges Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 365 2010 2157 2162 10.1098/RSTB.2010.0107 20566493
Chace J.F. Walsh J.J. Urban effects on native avifauna: a review Landsc. Urban Plan. 74 2006 46 69 10.1016/J.LANDURBPLAN.2004.08.007
Corlett R.T. Primack R.B. Devictor V. Maas B. Goswami V.R. Bates A.E. Koh L.P. Regan T.J. Loyola R. Pakeman R.J. Cumming G.S. Pidgeon A. Johns D. Roth R. Impacts of the coronavirus pandemic on biodiversity conservation Biol. Conserv. 246 2020 108571 10.1016/J.BIOCON.2020.108571
Cornell Lab of Ornithology eBird Basic Dataset Version: EBD_relApr-2020 2020
Dantewadia P. Devulpalli S. Padmanabhan V. Which is India’s noisiest city? URL 2020 LiveMint (accessed 6.23.20) https://www.livemint.com/news/india/which-is-india-s-noisiest-city-11581527059879.html
Derryberry E.P. Phillips J.N. Derryberry G.E. Blum M.J. Luther D. Singing in a silent spring: birds respond to a half-century soundscape reversion during the COVID-19 shutdown Science 370 2020 575 579 10.1126/SCIENCE.ABD5777 32972991
Farda M. Balijepalli C. Exploring the effectiveness of demand management policy in reducing traffic congestion and environmental pollution: car-free day and odd-even plate measures for Bandung city in Indonesia Case Stud. Transp. Policy 6 2018 577 590 10.1016/J.CSTP.2018.07.008
Fortin J. The birds are not on lockdown, and more people are watching them URL 2020 New York Times (accessed 6.23.20) https://www.nytimes.com/2020/05/29/science/bird-watching-coronavirus.html
Gaston K.J. What is rarity? Rarity 1–21 1994 10.1007/978-94-011-0701-3_1
Ghazali M. Sec 9 - 10 dividing road Chandigarh, nature re-claming it’s territory URL 2020 Twitter (accessed 6.23.20) https://twitter.com/ghazalimohammad/status/1243472152579342336?s=20
Guy J. Gupta S. Flamingos flock to a locked-down Mumbai URL 2020 CNN Travel (accessed 6.23.20) https://www.cnn.com/travel/article/flamingos-mumbai-lockdown-scli-intl/index.html
Hochachka W.M. Alonso H. Gutiérrez-Expósito C. Miller E. Johnston A. Regional variation in the impacts of the COVID-19 pandemic on the quantity and quality of data collected by the project eBird Biol. Conserv. 254 2021 108974 10.1016/J.BIOCON.2021.108974
IQAir 2019 world air quality report URL https://www.iqair.com/world-most-polluted-cities/world-air-quality-report-2019-en.pdf 2019 (accessed 6.23.20)
Jayadevan P. Jayapal R. Pittie A. A checklist of the birds of India Indian BIRDS 11 2016 113 172 10.2307/4511115
Jenkins C.N. Pimm S.L. Joppa L.N. Global patterns of terrestrial vertebrate diversity and conservation Proc. Natl. Acad. Sci. U. S. A. 110 2013 E2603 E2610 10.1073/PNAS.1302251110/SUPPL_FILE/PNAS.201302251SI.PDF
Kumar P. Gulia S. Harrison R.M. Khare M. The influence of odd–even car trial on fine and coarse particles in Delhi Environ. Pollut. 225 2017 20 30 10.1016/J.ENVPOL.2017.03.017 28343101
Liang Y. Rudik I. Zou E.Y. Johnston A. Rodewald A.D. Kling C.L. Conservation cobenefits from air pollution regulation: evidence from birds Proc. Natl. Acad. Sci. 117 2020 30900 30906 10.1073/pnas.2013568117 33234568
Mahato S. Pal S. Ghosh K.G. Effect of lockdown amid COVID-19 pandemic on air quality of the megacity Delhi, India Sci. Total Environ. 730 2020 139086 10.1016/J.SCITOTENV.2020.139086
Maldonado J.H. Moreno-Sánchez R.del P. Espinoza S. Bruner A. Garzón N. Myers J. Peace is much more than doves: the economic benefits of bird-based tourism as a result of the peace treaty in Colombia World Dev. 106 2018 78 86 10.1016/J.WORLDDEV.2018.01.015
Manenti R. Mori E. Di Canio V. Mercurio S. Picone M. Caffi M. Brambilla M. Ficetola G.F. Rubolini D. The good, the bad and the ugly of COVID-19 lockdown effects on wildlife conservation: insights from the first european locked down country Biol. Conserv. 249 2020 108728 10.1016/J.BIOCON.2020.108728
McClure C.J.W. Ware H.E. Carlisle J. Kaltenecker G. Barber J.R. An experimental investigation into the effects of traffic noise on distributions of birds: avoiding the phantom road Proc. R. Soc. B Biol. Sci. 280 2013 10.1098/RSPB.2013.2290
Mishra A. Das S. Singh D. Maurya A.K. Effect of COVID-19 lockdown on noise pollution levels in an indian city: a case study of Kanpur Environ. Sci. Pollut. Res. Int. 28 2021 46007 46019 10.1007/S11356-021-13872-Z 33884552
Morrison M.L. Bird populations as indicators of environmental change Curr. Ornithol. 3 1986 429 451 10.1007/978-1-4615-6784-4_10
Myers N. Mittermeler R.A. Mittermeler C.G. Da Fonseca G.A.B. Kent J. Biodiversity hotspots for conservation priorities Nature 4036772 403 2000 853 858 10.1038/35002501
Newbold T. Hudson L.N. Arnell A.P. Contu S. De Palma A. Ferrier S. Hill S.L.L. Hoskins A.J. Lysenko I. Phillips H.R.P. Burton V.J. Chng C.W.T. Emerson S. Gao D. Hale G.P. Hutton J. Jung M. Sanchez-Ortiz K. Simmons B.I. Whitmee S. Zhang H. Scharlemann J.P.W. Purvis A. Has land use pushed terrestrial biodiversity beyond the planetary boundary? A global assessment Science (80-.) 353 2016 10.1126/SCIENCE.AAF2201/SUPPL_FILE/NEWBOLD-SM.PDF 291–288
Rosenstock S.S. Anderson D.R. Giesen K.M. Leukering T. Carter M.F. Landbird counting techniques: current practices and an alternative Auk 119 2002 46 53 10.2307/4090011
Rutz C. Loretto M.C. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R.B. Ropert-Coudert Y. Tucker M.A. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evol. 49 4 2020 1156 1159 10.1038/s41559-020-1237-z 2020
Sanderfoot O.V. Holloway T. Air pollution impacts on avian species via inhalation exposure and associated outcomes Environ. Res. Lett. 12 2017 10.1088/1748-9326/AA8051
Shanahan D.F. Bush R. Gaston K.J. Lin B.B. Dean J. Barber E. Fuller R.A. Health benefits from nature experiences depend on dose Sci. Rep. 61 6 2016 1 10 10.1038/srep28551 2016
Shannon G. McKenna M.F. Angeloni L.M. Crooks K.R. Fristrup K.M. Brown E. Warner K.A. Nelson M.D. White C. Briggs J. McFarland S. Wittemyer G. A synthesis of two decades of research documenting the effects of noise on wildlife Biol. Rev. 91 2016 982 1005 10.1111/BRV.12207 26118691
Sharma S. Zhang M. Anshika Gao J. Zhang H. Kota S.H. Effect of restricted emissions during COVID-19 on air quality in India Sci. Total Environ. 728 2020 138878 10.1016/J.SCITOTENV.2020.138878
Silveira L. Jácomo A.T.A. Diniz-Filho J.A.F. Camera trap, line transect census and track surveys: a comparative evaluation Biol. Conserv. 114 2003 351 355 10.1016/S0006-3207(03)00063-6
Simeonova E. Currie J. Nilsson P. Walker R. Congestion pricing, air pollution, and children’s health J. Hum. Resour. 56 2019 10.3368/JHR.56.4.0218-9363R2 0218–9363R2
Simons T.R. Alldredge M.W. Pollock K.H. Wettroth J.M. Experimental analysis of the auditory detection process on avian point counts Auk 124 2007 986 999 10.1093/AUK/124.3.986
Slabbekoorn H. Songs of the city: noise-dependent spectral plasticity in the acoustic phenotype of urban birds Anim. Behav. 85 2013 1089 1099 10.1016/J.ANBEHAV.2013.01.021
Sullivan B.L. Wood C.L. Iliff M.J. Bonney R.E. Fink D. Kelling S. eBird: a citizen-based bird observation network in the biological sciences Biol. Conserv. 142 2009 2282 2292 10.1016/j.biocon.2009.05.006
TomTom International TomTom traffic index ranking 2020 TomTom Int BV https://www.tomtom.com/en_gb/traffic-index/ranking/ (accessed 4.17.22)
Tribune News Leopard spotted in Chandigarh’s Sector 5, tranquillised : The Tribune India URL 2020 Trib. News (accessed 6.23.20) https://www.tribuneindia.com/news/chandigarh/leopard-spotted-in-chandigarhs-sector-5-63171
Vardi R. Berger-Tal O. Roll U. iNaturalist insights illuminate COVID-19 effects on large mammals in urban centers Biol. Conserv. 254 2021 108953 10.1016/J.BIOCON.2021.108953
Veen J. Yurlov A. Delany S. AI M. Selivanova M. GC B. An Atlas of Movements of Southwest Siberian Waterbirds 2005
Verma S.K. Murmu T.Das Impact of environmental and disturbance variables on avian community structure along a gradient of urbanization in Jamshedpur, India PLoS One 10 2015 e0133383 10.1371/JOURNAL.PONE.0133383
Williams B.K. Nichols J.D. Conroy M.J. Analysis and Management of Animal Populations : Modeling, Estimation, and Decision Making 2002 Academic Press
Xu X. Xie Y. Qi K. Luo Z. Wang X. Detecting the response of bird communities and biodiversity to habitat loss and fragmentation due to urbanization Sci. Total Environ. 624 2018 1561 1576 10.1016/J.SCITOTENV.2017.12.143 29929265
Zhou Y. Wu Y. Yang L. Fu L. He K. Wang S. Hao J. Chen J. Li C. The impact of transportation control measures on emission reductions during the 2008 Olympic Games in Beijing, China Atmos. Environ. 44 2010 285 293 10.1016/J.ATMOSENV.2009.10.040
| 0 | PMC9746919 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2022 Jul 19; 271:109597 | utf-8 | Biol Conserv | 2,022 | 10.1016/j.biocon.2022.109597 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00069-0
10.1016/j.biocon.2021.109017
109017
Article
COVID-19 impacts on participation in large scale biodiversity-themed community science projects in the United States
Crimmins Theresa M. ab⁎
Posthumus Erin ab
Schaffer Sara ab
Prudic Kathleen L. a⁎⁎
a School of Natural Resources and the Environment, University of Arizona, 1604 E. Lowell St., Tucson, AZ 85721, USA
b USA National Phenology Network, 1311 E. 4th St., Ste. 325, Tucson, AZ 85721, USA
⁎ Correspondence to: T M. Crimmins, School of Natural Resources and the Environment, University of Arizona, 1604 E. Lowell St., Tucson, AZ 85721, USA.
⁎⁎ Corresponding author.
4 3 2021
4 2021
4 3 2021
256 109017109017
21 9 2020
29 1 2021
9 2 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Shutdowns associated with the COVID-19 pandemic have had extensive impacts on professional and volunteer-based biodiversity and conservation efforts. We evaluated the impact of the widespread pandemic-related closures in the spring of 2020 on participation patterns and rates on a national and a state-by-state basis in the United States in four biodiversity-themed community science programs: eBird, eButterfly, iNaturalist, and Nature's Notebook. We compared the number of participants, observations submitted, and proportion of observations collected in urban environments in spring 2020 to the expected values for these metrics based on activity in the previous five years (2015–2019), which in many cases exhibited underlying growth.
At the national scale, eButterfly and Nature's Notebook exhibited declines in the number of participants and number of observations submitted during the spring of 2020 and iNaturalist and eBird showed growth in both measures. On a state-by-state basis, the patterns varied geographically and by program. The more popular programs – iNaturalist and eBird – exhibited increases in the Eastern U.S. in both the number of observations and participants and slight declines in the West. Further, there was a widespread increase in observations originating from urban areas, particularly in iNaturalist and eBird. Understanding the impacts of lockdowns on participation patterns in these programs is crucial for proper interpretation of the data. The data generated by these programs are highly valuable for documenting impacts of pandemic-related closures on wildlife and plants and may suggest patterns seen in other community science programs and in other countries.
Keywords
Citizen science
Conservation
Coronavirus
eBird
eButterfly
iNaturalist
Nature's Notebook
Pandemic
==== Body
pmc1 Introduction
The COVID-19 pandemic has had extensive impact on all facets of human society (Bates et al., 2020; Diffenbaugh et al., 2020). To limit virus transmission, swift closures of public spaces including college campuses, K-12 schools, theaters, sports venues, and parks and recreation facilities swept through the United States in March 2020 and remained in place for variable durations across states through subsequent months. Consequently, tourism, recreation behaviors, and other forms of human activity patterns have been dramatically impacted (Bakar and Rosbi, 2020; Nicola et al., 2020). The dramatic shifts in human activities have had clear effects on wildlife and biodiversity; anecdotes suggest some wildlife may be moving into new areas or changing their behavior, while others may be at risk of increased exploitation or disturbance (Corlett et al., 2020; Rutz et al., 2020).
Community science – also referred to as citizen science, volunteer science, and public participation in scientific research – provides significant value to conservation efforts in both urban and non-urban areas (Cooper et al., 2007; Devictor et al., 2010; McKinley et al., 2017; Sullivan et al., 2017). Community science programs are characterized as scientific research conducted at least in part by amateur or volunteer scientists (Bonney et al., 2009; Dickinson et al., 2012). Designed to engage non-professionals in the act of science and data, these programs frequently yield data at spatial and temporal scales far beyond what professional scientists can achieve when working alone. Community science programs lead to increases in science literacy and an understanding of the process of “doing science”, a deepened sense of place, and a greater understanding and appreciation for the plants and animals they are observing (Dickinson et al., 2012; Evans et al., 2020). As such, community science programs were widely advertised during early weeks of the shutdown in the U.S. as stimulating and meaningful activities for children and adults alike during school and office closures (Bowman and Gibson, 2020; Crimmins, 2020; Piñon, 2020) as well as an alternative approach for data collection that might mitigate the shutdown of formal research and monitoring activities (Bowser et al., 2020; Kornfeld, 2020; Zellmer et al., 2020).
In the weeks immediately following the issuance of COVID-related shutdown orders in the U.S., several community science programs reported spikes in participation. Zooniverse, City Nature Challenge, and Stall Catchers – all large scale community science programs or platforms – reported an increase in participation in March and April 2020 (Bowser et al., 2020; Kubis, 2020; Dinneen, 2020; Young, 2020). Several of these programs, which can be undertaken by individuals on personal computers at home, reported an increase in participation of three to five times the rate of previous years during the same time period (Bowser et al., 2020; Kubis, 2020). Zooniverse participants completed classifications of galaxies, animal photos, and more at three times the rate of previous years as of April 3, 2020 (Bowser et al., 2020), and participants in the Stall Catchers project assisted with Alzheimer's research at levels 38% higher than in 2019 (personal communication, P. Michelucci, July 24, 2020). SciStarter, which connects participants with thousands of community science programs, reported increased interest in projects focused on environmental health and identifying and observing birds during the shutdown (Kornfeld, 2020).
Whether the boost in community science project participation documented among some programs early in the shutdown extended to all types of community science programs remains unknown. Here, we explore the impact of the shutdown on participation in four biodiversity-themed community science programs in the U.S.: eButterfly (e-butterfly.org), iNaturalist (inaturalist.org), Nature's Notebook (naturesnotebook.org), and eBird (ebird.org). Each of these programs exists to document and share biodiversity observations to support science and conservation. Because participants in these programs typically step outside to identify and assess plants and animals, we anticipate these programs may show different patterns in participation and data submissions from those reported by community science programs that are undertaken completely online. An understanding of the impacts of the pandemic on participation in biodiversity-themed programs is necessary for analysts exploring these data in future studies, as shifts in the intensity or geographic scope of participation may necessitate statistical techniques that account for consequential irregularities in the datasets. The rich and geographically extensive volunteer-contributed reports of plants and animals originating from these programs have the potential to provide important insight into wildlife responses to pandemic-related closures, provided that data interpretation accounts for the impacts of lockdown on data collection. Further, a clearer understanding of changes in program participants' contributions during lockdown is valuable to program staff aiming to support participants as fully as possible. Finally, the findings specific to these four programs in the U.S. may point to what might be expected regarding patterns in participation and consequent impacts on resultant data in other community science programs and in other countries.
We predicted that the shutdown would lead to a drop in the number of participants contributing to the four biodiversity-themed community science programs as well in the amount of observations submitted, due to the increased demands in other parts of participants' lives during this period. Second, we expected the locations where participants collected observations to change during the shutdown, due to the closure of parks and reserves, natural spaces, and facilities such as nature centers and arboreta. Specifically, we expected to see a greater proportion of observations submitted from urban areas than prior to closures, due to stay-at-home orders limiting participants' movement. Finally, we hypothesized that the number of active participants, the amount of data submitted, and the proportion of observations submitted from urban areas in each state would all be affected proportionally by the amount of time a state was formally under lockdown.
2 Materials and methods
2.1 Community science programs
The data evaluated in this study represent four popular biodiversity-themed community science programs in the U.S. The programs vary in their aims, complexity in participating, and levels of standardization, though all contribute critical data and information for documenting and tracking status and trends in biodiversity (Kelling et al., 2019). Data from all four programs are frequently utilized by scientists, conservation organizations, and land management agencies to understand distributions and trends in species and to inform decisions (Cooper et al., 2007; Ellwood et al., 2017).
eButterfly engages participants in documenting checklists of butterflies across North America (Prudic et al., 2017). Participants submit their observations for a new or existing location on a web browser; all locations are stored to encourage repeated observations from established locations. Similar to eBird, participants choose from one of four types of sampling protocols and are presented with a checklist of butterfly species known to occur in the state or province; participants are invited to report presence or absence for all species on the list. Participants are encouraged to submit photos of their observations so that species identification can be verified by other participants in the community. Over 1000 species of butterflies and moths have been contributed to eButterfly to-date (eButterfly, 2020).
iNaturalist engages participants across the globe to photo document plants, animals, fungi, and algae (Seltzer, 2019). Photos are uploaded through a web browser or mobile application to an online community where other participants verify the species identification (Nugent, 2018; Unger et al., 2020). Species identification is also facilitated by a machine learning algorithm which evaluates the submitted photo and makes suggestions on species identification to the participant (Van Horn et al., 2018). Since the program's launch, over 300,000 species have been documented worldwide through iNaturalist (Loarie, 2020). Projects and events can also be created within the platform, such as bioblitz and City Nature Challenge events in which participants survey the biodiversity of a specific area during a defined time period. Dozens of such events took place across the U.S. in spring 2020, despite pandemic lockdowns.
Nature's Notebook, coordinated by the USA National Phenology Network (USA-NPN), engages individuals and groups of participants observing collectively in documenting plant and animal phenology across the U.S. (Denny et al., 2014). Participants first register one or more locations (sites) at which they make repeated observations, then register individual plants and/or a checklist of animal species to observe at each site. Participants collect observations of the status of seasonal growth and development (conditions such as presence of leaves, open flowers, or ripe fruits in plants and presence of individuals, mating, courtship calling, or egg laying in animals) via a web browser or mobile application. Participants are encouraged to make observations 2–3 times per week during the season when plants and animals are active and indicate the presence or absence of each phenological stage at each visit (Rosemartin et al., 2014). Protocols are currently available for participants to track the phenology of over 1000 species of plants and nearly 400 species of insects, fish, amphibians, reptiles, birds, and mammals (USA National Phenology Network, 2020a).
eBird engages a global network of participants who submit observations of birds to a central data repository via a web browser or the eBird Mobile application (Sullivan et al., 2014). Participants report bird species identity, occurrence, and relative abundance at either pre-defined birding hotspots or observer-specified locations; locations can be saved and returned to for repeat observations. Participants choose from one of four types of sampling protocols and are presented with a checklist of bird species most likely to be observed at their selected location; participants are invited to report presence or absence and number of individuals for all species on the list. Some participants report only occasionally; others complete daily checklists (Sullivan et al., 2009). As of 2019, eBird boasted 10,721 bird species in the program's taxonomy (Team eBird, 2019).
Citizen science programs generally have shown growth in recognition and participation over the past decade (McKinley et al., 2017). Three of the four programs examined – iNaturalist, Nature's Notebook, and eBird – similarly experienced either steady or exponential growth in participation in recent years (Fig. 1a, b).Fig. 1 Long-term patterns in participation among four biodiversity-themed community science programs. a) Number of participants, b) observations submitted, and c) percentage of observations originating from urban areas contributed to eButterfly, iNaturalist, Nature's Notebook, and eBird in the U.S., March–June 2015–2020. In a) and b), eButterfly and Nature's Notebook are plotted on the primary y-axis and iNaturalist and eBird are plotted on the secondary y-axis.
Fig. 1
2.2 Data preparation
We downloaded the prepackaged eBird “basic sampling event dataset” from the eBird website on August 15, 2020 (eBird Basic Dataset, 2020). This dataset includes all validated observations and unique participants from checklists entered into eBird as well as covariates entered into the checklists regarding location and effort, but not species (Sullivan et al., 2014).
We accessed iNaturalist “research grade” observations through the Global Biodiversity Information Facility filtering by state, month, year, and unique participant (GBIF, 2020). Research grade observations are observations with a date, latitude/longitude coordinates, and a consistent species identification made by at least two reviewers (Ueda, 2020), which is analogous to the internal vetting processes of eBird and eButterfly. We accessed eButterfly data through the eButterfly database. All records for observations within the United States were retained.
For Nature's Notebook, we downloaded all “status and intensity” records collected 2015–2020 from the USA-NPN National Phenology Database using the rnpn package (USA National Phenology Network, 2020b). Status and intensity records reflect each time an observer recorded data on an individual plant or an animal at location over the course of the season (Rosemartin et al., 2018). We excluded data contributed by the National Ecological Observatory Network (NEON) and records contributed at locations outside of the U.S. We treated each instance of observing a single organism on a single date as an “observation,” consistent with the definition of an observation in the other community science programs in this study.
For each program-specific dataset, we excluded all records collected in months other than March, April, May, and June and we removed all observations falling outside of the United States. Next, we intersected observation locations with a shapefile representing the boundaries of urban areas (U.S. Census Bureau, 2017) and assigned a binary value of urban/non-urban to each observation based on its latitude/longitude reported location. Finally, we tallied the number of observations and the number of unique participants for each program in each year, and then again by state in each year. Similarly, for each program, we calculated the percentage of observations within each year that fell within urban areas as well as the percentage of observations within urban areas in each state in each year.
2.3 Statistical analyses
To determine the impact of the shutdowns on participation in community science programs, we examined the number of individuals contributing observations and the number of observations submitted. Because several of the variables under examination exhibit growth over the past five years (Fig. 1, Table A.1), we performed a likelihood ratio test to select between linear and polynomial models for each program. Residuals were normally distributed as determined by a visual inspection of a QQ plot. We tested homogeneity of variance by plotting fitted values versus residuals. The final models selected appear in Table A.2. We then constructed a model between 2015 and 2019 and used this model to create an expected 2020 value with a 95% prediction interval for 2020 (Knowles and Frederick, 2016). We then compared the predicted 2020 value to the observed 2020 value, calculated the percent difference between the two, and then assessed whether the observed fell outside of the predicted 95% interval as our measure of significance (Knowles and Frederick, 2016). We evaluated both the number of unique participants contributing to the program and the number of observations submitted in each of the programs (eButterfly, iNaturalist, Nature's Notebook, and eBird) for the entire U.S. as well as for each state in the U.S. For the state-by-state analyses, iNaturalist and eBird data were log transformed, and Nature's Notebook and eButterfly data were square root-transformed. We also used this approach to evaluate whether a larger proportion of records originated from within urban areas in the spring of 2020.
For all three metrics (number of observations, number of unique participants, percent urban observations), we evaluated the effect of stay at home orders on the percent change between the observed and expected 2020 values in each of the programs (eButterfly, iNaturalist, Nature's Notebook, and eBird) for the entire U.S. as well as for each state. Number of stay at home days by state were acquired from the National Academy for State Health Policy (2020).
All analyses were performed in Rv3.5.3 with RStudio v1.2.5001 as the integrated development environment. Both data and R code are archived in Zenodo (DOI: https://doi.org/10.5281/zenodo.4430966).
3 Results
Spring (March-Jun) participation rates vary dramatically across the four programs evaluated in this study (Fig. 1, Table A.1). iNaturalist and eBird engage tens to hundreds of thousands of participants each spring - far more than Nature's Notebook, which engages thousands, and eButterfly, which engages hundreds of individuals each spring. Accordingly, the quantities of incoming observations also vary among the programs: eButterfly participants report thousands of observations each spring, where eBird participants report millions of observations. Participants in iNaturalist and Nature's Notebook contribute hundreds of thousands of observations each spring. Nature's Notebook boasts the highest rate of observations originating from urban areas; eButterfly's observations are submitted primarily from non-urban areas.
In 2020, two of the four programs, eButterfly and Nature's Notebook, experienced fewer participants than expected, and Nature's Notebook saw significantly fewer observations than expected (Fig. 2 , Table A.1). In contrast, both iNaturalist and eBird show sustained activity or increases in these variables across the nation, though gains over what was predicted were non-significant (Fig. 2a, b). All programs but eButterfly experienced more observations originating in urban areas in 2020 than expected, and this proportion was significantly greater than expected for iNaturalist and eBird (Fig. 2c).Fig. 2 Difference between predicted and observed values in a) the number of participants, b) observations submitted, and c) percentage of observations originating from urban areas contributed to eButterfly, iNaturalist, Nature's Notebook, and eBird in the U.S., March–June 2015–2020.
Fig. 2
The number of participants and amount of data coming into each program is markedly greater in certain states (Table A.4). California is among the top five states in all four programs in terms of participants and observations contributed 2015–2019, and Texas and New York are in the top five states for both metrics in three of the four programs during the pre-COVID springs. The extent to which the number of participants and amount of incoming data from these states was impacted in spring of 2020 was not consistent among programs. For example, the levels of participation in California and Texas declined noticeably across programs in 2020, though the measures changed little for New York.
3.1 Contributing participants
State-by-state analyses revealed widespread decreases in participation across all four programs, though spatial patterns in changes varied by program. eButterfly exhibited significant drops in participation in Alaska, Hawai'i, and through the Great Plains states and also showed sharp increases in participation in other states, though the increase over expected levels of participation were only significant in Utah (Fig. 3a, Table A.4). iNaturalist demonstrated decreases in participation in 2020 over expected numbers nearly nationwide, with significant decreases in many western states as well as decreases in states that contribute the largest proportions of observations and participants (Fig. 3b, Table A.4). Changes in participation in Nature's Notebook were spatially patchy (Fig. 3c). California, a top-contributing state in Nature's Notebook pre-COVID, saw a significant decline in participation in 2020, though other top-observing states, including Massachusetts and New York, remained steady in 2020 (Table A.4). Similar to iNaturalist, eBird showed a significant decrease in participation over what was expected based on previous years in many western states as well as significant decreases in Eastern Seaboard states (Fig. 3d).Fig. 3 Percent difference in the observed number of participants in March–June 2020 from the expected number of participants in March–June 2020 based on participation patterns in March–June 2015–2019 in four biodiversity community science programs: a) eButterfly, b) iNaturalist, c) Nature's Notebook, and d) eBird. Blue tones indicate fewer participants than expected in 2020; red tones indicate more participants than expected in 2020; hatching indicates a significant difference between predicted and observed number of participants in 2020 (p < 0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 3
3.2 Observation activity
Overall patterns of change in observations in 2020 paralleled the patterns seen in participants. For all states combined, Nature's Notebook participants contributed significantly fewer observations in 2020 (98,256 observations) compared to what was expected (95% prediction interval: 105,980–170,849; Table A.3). eButterfly, iNaturalist, and eBird each exhibited a non-significant increase in the number of participants over what was expected based on 2015–2019 patterns (Table A.3).
Spatial patterns of change in observations submitted to the eButterfly program (Fig. 3a, Table A.4) paralleled changes observed in participants (Fig. 3a). Changes in observations contributed to iNaturalist and eBird both exhibited a fairly clear east-west gradient, where western states generally showed decreases in observations and states east of the hundredth meridian tended to show increases in observations (Fig. 4b, d). Finally, most states exhibited a decrease in the number of observations reported to Nature's Notebook in 2020 (Fig. 4c).Fig. 4 Percent difference in observed observations submitted in March–June 2020 from the expected number of observations in March–June 2020 based on participation patterns in March–June 2015–2019 in four biodiversity community science programs: a) eButterfly, b) iNaturalist, c) Nature's Notebook, and d) eBird. Blue tones indicate fewer observations than expected in 2020; red tones indicate more observations than expected in 2020; hatching indicates a significant difference between predicted and observed number of observations submitted in 2020 (p < 0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 4
3.3 Shift in geography of observations
When all states were combined, the percent of observations submitted from within urban areas significantly increased in 2020 over what was expected for both iNaturalist and eBird (Fig. 2c). In 2020, 45% of iNaturalist and 46% of eBird observations originated in urban areas (iNaturalist 95% prediction interval: 41–44%; eBird 95% prediction interval: 37–42%; Table A.3). The percentage of observations submitted from within urban areas decreased non-significantly in both eButterfly and Nature's Notebook over what was expected based on 2015–2019 patterns (Table A.3).
State-specific results varied appreciably by program in the shift of observations submitted from urban and non-urban areas. Across much of the western U.S. and the Ohio Valley, the proportion of observations submitted from within urban areas dropped sharply in 2020 in the eButterfly program, though none of these decreases were significant (Fig. 5a, Table A.4). In contrast, iNaturalist and eBird both exhibited increases in the proportion of observations reported from within urban areas in 2020 across the majority of states, and the shifts toward more urban observations were significant for many states in the eBird program (Fig. 5b, d). Patterns apparent in Nature's Notebook were mixed, with large increases in the proportion of observations reported from within urban areas increasing in states in the Southeast, Northeast, and West, and decreasing in many Great Plains states (Fig. 5c).Fig. 5 Percent difference in the proportion of observations submitted from within an urban area in March–June 2020 from the expected proportion of observations submitted from within an urban area in March–June 2020 based on participation patterns in March–June 2015–2019 in four biodiversity community science programs: a) eButterfly, b) iNaturalist, c) Nature's Notebook, and d) eBird. Blue tones indicate a smaller proportion of observations submitted from within urban areas than expected in 2020; red tones indicate a larger proportion of observations submitted from within urban areas than expected in 2020; hatching indicates a significant difference between predicted and observed percent of records submitted from within urban areas in 2020 (p < 0.05). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 5
3.4 Influence of length of lockdown on participants, observations, and percent urban observations
There was a suggestive but inconclusive positive relationship between the number of days states were in lockdown and the number of participants contributing data to eButterfly by state (p = 0.103, adj r2 = 0.03; Table A.5), such that the longer a state was in lockdown, the greater the number of participants contributing in 2020. There were similarly significantly positive relationships between the number of days in lockdown and the percentage of observations submitted from urban areas to both eButterfly (p = 0.032, adj r2 = 0.07) and Nature's Notebook (p = 0.066, adj r2 = 0.05), such that states experiencing longer periods of lockdown were associated with a higher proportion of observations submitted from urban areas. The number of days in lockdown did not show a relationship with the number of participants in iNaturalist, Nature's Notebook, or eBird; in the proportion of observations submitted from within urban areas to iNaturalist or eBird; or with the number of observations contributed to any of the programs (Table A.5).
4 Discussion
This study evaluated impacts of COVID-related stay-at-home orders and widespread closures on the participation and activity level in four biodiversity-themed community science programs in the United States. The four studies evaluated here vary by orders of magnitude in terms of the numbers of participants and observations submitted (Fig. 1). Further, within programs, data contributions vary by geography, with certain states accounting for a large proportion of the participation. This is important because even small changes in participation in states that account for a large proportion of participation can translate to substantial impacts to overall participation numbers for a program.
Overall, the results of this evaluation revealed variable patterns in activity among the programs and across geography - inconsistent with our expectations that all programs would show uniform drops in participation as a consequence of the pandemic. The two programs exhibiting the greatest participation, iNaturalist and eBird, showed similarities in their patterns of change.
4.1 Changes in participant activity varied by program and geography
We had predicted that both the number of participants and the amount of observations submitted across the U.S. in spring 2020 would be fewer than what would have been expected had COVID not occurred. Though we see a clear overall decrease in participants and observations submitted to Nature's Notebook, these patterns did not hold for the other three programs. Further, patterns of change varied dramatically among states and programs.
The patterns exhibited in participants and incoming observations across the U.S. in iNaturalist and eBird follow an interesting pattern oriented along a longitudinal gradient. The largest decreases in both metrics were observed in western states and increases were generally observed in eastern states, and a more in-depth assessment should be undertaken to fully evaluate the reasons for this pattern. One explanation for the increases documented in iNaturalist, especially in the Northeast, may be that iNaturalist continued to encourage participation in local and regional BioBlitz events and other community biodiversity projects throughout spring 2020 (City Nature Challenge, 2020). The City Nature Challenge, an event that takes place in cities worldwide and utilizes the iNaturalist platform, occurred late in April in 2020 (City Nature Challenge, 2020). In 2020, 244 cities participated in the City Nature Challenge, a substantial increase over 2019, when 159 cities participated (Young, 2020). Many of the U.S. in the City Nature Challenge in 2020 were concentrated in the eastern portion of the country. In addition, iNaturalist featured instructions on how to participate in the program safely during the pandemic on their homepage from April to June of 2020 (Iwane, 2020); this may also account for increased participation in the program. eBird similarly experienced intense activity in May because of an annual springtime event. Global Big Day, occurring annually in the spring, engages birders worldwide in documenting and celebrating birds. Global Big Day took place on May 9, 2020 and broke records for participation, yielding a larger than 30% increase in participants over 2019 (Team eBird, 2020). Finally, social justice movements such as #BlackBirdersWeek and #BlackInNature that took place in the spring 2020 (Mock, 2020) may also account for the upticks observed in these two programs.
The patterns we see in eButterfly participation for 2020 across states and for the U.S. as a whole is complicated by two factors outside of COVID. First, the program released a new version of the web platform with associated messaging to the community in mid-May; the need to adjust to a new interface may have slowed users' contributions to some extent. Second, reports of butterflies are typically low in spring (March–June) in the U.S. due to their phenology. Patterns in eButterfly participation may be driven by the comparatively low sample sizes in this program.
Nature's Notebook exhibited highly variable patterns of increases and decreases in participants and incoming observations in 2020. The dramatic increases in participation seen in several states, including Indiana, Oklahoma, Louisiana, and Colorado are likely due to the establishment of several new groups of individuals tracking phenology in these states. A unique aspect of Nature's Notebook is that monitoring can be undertaken by individuals as well as by community or regionally-organized groups referred to as Local Phenology Programs (LPPs). Organizations such as nature centers, arboreta, land conservancies, and National Wildlife Refuges use Nature's Notebook to meet a diversity of outcomes, including asking and answering scientific questions about the impact of environmental change, informing natural resource management and decision-making, and educating and engaging the public. Several new LPPs were established in early 2020 in the states depicting the largest increases in participants; one of these states was the focus of a data collection campaign in late 2019 and early 2020. Newly established LPPs are also the likely reason for the increase in observations seen in several states in 2020, including Indiana and New Jersey. The large increase in participants in Texas is likely the result of the launch of a new campaign focused on tracking juniper pollen in this state in late 2019. The clear decrease in participation and incoming observations observed in California, Tennessee, New York, and other states are likely attributable to closures of public spaces such as parks, nature centers, natural areas, and schools where many active Nature's Notebook LPP sites exist.
The mixed patterns we see in participation and incoming data in these four programs in the spring of 2020 are partially in conflict with the reports of record-breaking participation in other community science projects (Bowser et al., 2020; Kubis, 2020; Dinneen, 2020). One reason for such differences may be the way in which volunteers participate: in many of the programs boasting large increases, volunteers participate completely online using a computer or other device. In contrast, the programs evaluated in this study focus on outdoor phenomena, and participants typically step outside to identify or evaluate individual organisms. Many parts of the country were still experiencing inclement weather in March, April, and even into May, which may have encouraged participation in computer-based programs and discouraged participation in programs requiring time spent outside.
We expect that we also see decreases in participation in Nature's Notebook and eButterfly because many formerly active participants no longer had time available to dedicate to the efforts during a period characterized by major upheaval and change in both personal and professional lives. Click rates reported by Constant Contact for Nature's Notebook newsletters - which remained constant from 2019 to 2020 - support the notion that participants continued to care for the program despite a decline in their participation during spring of 2020. This bodes well for the future of these community science programs, suggesting that once participants feel settled in their lives again, they may reengage.
4.2 Shift toward urban observation locations in more popular programs
We had predicted that participants would log a larger proportion of observations from urban locations in 2020 as a result of the stay-at-home orders issued across the country over the spring period. eBird and iNaturalist exhibited the clearest and most widespread shifts toward increased urban-based observations contributed in 2020. iNaturalist exhibited a clear increase in all three measures, suggesting enthusiastic involvement in this program in urban areas, likely resulting at least in part from major growth in City Nature Challenge events. eBird also showed growth the number of incoming observations, though not in the number of participants, suggesting increased participation, especially in urban areas, by approximately the same number of participants as in spring 2019. A shift toward urban participation during COVID lockdown has been reported for iNaturalist in Europe as well BIOCON-20-00460, this issue.
Findings for eButterfly and Nature's Notebook were more mixed. We observed a significant increase in the percent urban observations in New York. We suspect many participants who live in urban areas such as New York City and travel to more butterfly biodiversity spring locations such as the southwest and California switched their behavior to local environs, but more in- depth analysis is needed. Many other states show drops in the proportion of observations submitted from within urban areas in eButterfly; the states showing shifts away from urban areas are also those exhibiting decreases in overall participation (Figs. 3a and 5a).
Patterns of shifts in Nature's Notebook show large increases in urban participation in many states, which is likely the result of the USA-NPN's concerted efforts to encourage participants to register new sites and continue monitoring close to home if the facilities where they had previously been collecting observations were closed. Recognizing the potential for significant drops in Nature's Notebook activity due to such closures, USA-NPN staff sent email newsletters and social media messages throughout spring 2020 encouraging participants to establish new sites in their yards or nearby, accessible locations to offset the loss of incoming data from sites no longer accessible. The positive relationship between the proportion of observations originating from urban areas and the length of lockdown in both eButterfly and Nature's Notebook suggests that participants responded and reoriented their activities to locations closer to their homes. Incidentally, visitation to urban, peri-urban, and other natural areas dramatically increased during stay-at-home lockdowns (Fisher et al., 2020; Goodier and Rayman, 2020), consistent with the large-scale shift toward urban observations in the community science programs evaluated in this study. The increases in urban observations might reflect either increased usage of urban greenspaces or a shift to greater observation activity closer to urban dwellings, or both.
4.3 Conservation implications
Several federal and state agencies and other conservation organizations rely on data from programs such as those evaluated in this study to inform management decision making. For example, data contributed to Nature's Notebook have been used to develop phenological indicators of wildfire danger (Nathan et al., 2019); a sudden drop in incoming observations on these indicator species could negatively impact managers assessing wildfire danger in public lands. Similarly, the California Department of Fish and Wildlife leveraged iNaturalist and eBird observations to develop a connectivity plan and identify key land acquisitions to grow and maintain corridors (Jennings et al., 2019). The results of this study demonstrate that pandemic-related shutdowns can have serious consequences on the availability of volunteer-contributed data necessary to support these sorts of management and planning activities. This is especially true for states where community science is more widely adopted and data contribution is high, such as California, which experienced a drop in incoming data in spring 2020 over what was expected based on previous years in all four programs evaluated.
A long-recognized benefit of community science programs is that they contribute valuable insights that are otherwise not possible to achieve. That community science programs fill in gaps in knowledge and understanding is particularly true during pandemic-related closures, when many other forms of monitoring have been shuttered (Pennisi, 2020). One way in which observations contributed through community science programs might prove especially useful is in documenting the changes in wildlife, such as increases in species richness, higher breeding success, and reduced road-killing that have occurred as a result of reduced traffic and other changes associated with pandemic-related closures (Manenti et al., 2020). The results of this study indicate that participation in these volunteer programs have been affected as well; even so, the incoming data stand to provide one of the best approaches for documenting wildlife responses to COVID-related shutdowns. The findings specific to the four programs evaluated here may point to what might be expected regarding patterns in participation and consequent impacts on resultant data in other community science programs and in other countries.
The results of this study also underscore the value of greenspaces and urban and peri-urban parks. The importance of urban greenspaces to support biodiversity as well as mental health during lockdown and closures has rapidly been documented (Kleinschroth and Kowarik, 2020; Slater et al., 2020). We see clear evidence that people appreciate these spaces as opportunities to document wildlife, plants, progression of phenological events like leaf-out and flowering over the course of the season. The closure of many parks and public facilities where participants in Nature's Notebook in particular had regularly observed prior to the COVID shutdowns resulted in a clear drop in incoming data in the spring of 2020. Second, it seems highly likely that the greater proportion of observations originating from urban locales during shutdowns is being collected at greenspaces that have remained open, including city parks or open lots. An increased understanding of the importance of greenspaces for the biodiversity they support as well as in maintaining mental health will help city planners manage them as ecosystems (Plummer et al., 2020).
The findings of this analysis offer insights for staff managing biodiversity-themed community science programs. Program staff may use the changes documented here to encourage adaptations to participation that better suit participants' limited options during closures or to emphasize particular activities that better match their current tendencies in participation. For example, the Maryland Department of Natural Resources invited participants to create their own ‘State Park’ in their local private backyards and share their creations and wildlife observations with others on social media and iNaturalist. Similarly, the California Academy of Sciences, the home of iNaturalist, modified their City Nature Challenge in San Francisco during spring 2020 to accommodate social distancing and travel restrictions (California Academy of Sciences, 2020). The findings of this study may also provide insight for staff to most effectively reinvigorate participants once it is possible to return to pre-shutdown levels of activity.
As well, the pandemic-related changes in program participation documented in this study are important for data users to consider. The clear geographic shifts documented here may result in otherwise inexplicable changes in the composition, abundance, or range of species reported. Likewise, decreases in species reports during the spring of 2020 may be directly traceable to declines in participation in these programs and therefore may necessitate careful use of statistical techniques BIOCON-20-00460, this issue.
5 Conclusions
In this study, we evaluated the impact of the shutdown on participation patterns and rates in four national-scale biodiversity-themed community science programs: eBird, eButterfly, iNaturalist, and Nature's Notebook. We had predicted a decline in the number of participants and observations contributed to the four programs as a result of COVID-related lockdowns, but found that patterns were not as clear or stark as we had feared. Overall, Nature's Notebook exhibited the largest declines in participants and observations compared to what was expected for spring 2020, and iNaturalist showed large increases over what was expected in both metrics. Further, as predicted, both iNaturalist and eBird experienced significant increases in the proportion of records coming from urban areas. Patterns varied by state and by program. Finally, we anticipated changes in participation to be driven by the length of lockdown; these patterns were weak.
Our findings suggest that participation in the community science programs evaluated had adapted as a result of lifestyle changes imposed by pandemic-related closures. Participants have generally continued their activity, albeit in different locations than previously. Though the numbers of participants generally decreased in some programs compared to what was expected for 2020, the amount of incoming data appears to be impacted to a lesser degree, offering a sense of hope for the future of these programs and the incoming data. That participants in these programs are persevering is encouraging, as the rich and geographically extensive volunteer-contributed reports of plants and animals originating from these programs have the potential to provide important insight into wildlife responses to pandemic-related closures and yield data to offset losses due to the shuttering of formal plant and animal monitoring efforts.
Declaration of competing interest
The authors declare no conflicts of interest.
Appendix A Table A.1 Total number of participants, observations submitted, and percentage of observations originating from urban areas contributed to eButterfly, iNaturalist, Nature's Notebook, and eBird in the U.S., March–June 2015–2020.
Table A.1Program Year Participants Observations %Urban observations
eButterfly 2015 318 10,373 28%
2016 299 8066 22%
2017 281 10,391 23%
2018 223 6361 20%
2019 204 5722 23%
2020 184 5547 19%
iNaturalist 2015 9963 185,519 36%
2016 17,745 351,788 37%
2017 21,242 589,864 38%
2018 32,876 902,758 40%
2019 75,578 1,441,358 41%
2020 110,023 1,945,420 45%
Nature's Notebook 2015 1411 107,850 30%
2016 1582 106,068 44%
2017 1922 123,691 46%
2018 2188 132,627 44%
2019 1937 126,387 47%
2020 1744 98,256 51%
eBird 2015 66,846 1265,152 38%
2016 79,622 1,464,060 37%
2017 91,016 1744,873 38%
2018 107,925 2,144,422 39%
2019 130,385 2,486,899 39%
2020 128,225 2,948,944 46%
Table A.2 Model selection.
Table A.2Program y Model selected
eButterfly Observations Linear
Participants Linear
%Urban Linear
iNaturalist Observations Polynomial
Participants Polynomial
%Urban Linear
Nature's Notebook Observations Linear
Participants Linear
%Urban Polynomial
eBird Observations Linear
Participants Linear
%Urban Polynomial
Table A.3 Predicted 2020 counts, observed 2020 counts, 95% predicted 2020 interval, and percent change between predicted and observed participants, contributed observations, and percent of observations originating from within urban areas, March–June 2020, for four community science programs. *Denotes 2020 actual value falls outside of 95% prediction interval.
Table A.3Program Observed 2020 participants Predicted 2020 participants 95% prediction interval Percent change (observed vs. predicted 2020 participants)
Nature's Notebook 1744 2328 1460–3195 −25
eButterfly 184 174 116–231 6
iNaturalist 110,023 75,389 12,331–138,448 46
eBird 128,225 141,773 123,001–160,545 −10
Program Observed 2020 observations Predicted 2020 observations 95% prediction interval Percent change (observed vs. predicted 2020 observations)
Nature's Notebook 98,256 138,415 105,980–170,849 −29*
eButterfly 5547 4880 0–11,898 13
iNaturalist 1,945,420 1,613,212 1014,473–2,211,951 21
eBird 2,948,944 2,758,238 2,439,344–3,077,132 7
Program Observed 2020 %urban observations Predicted 2020 %urban observations 95% prediction interval Percent change (observed vs. predicted 2020 %urban observations)
Nature's Notebook 51% 56% 30–75% −3%
eButterfly 19% 20% 8–32% −5%
iNaturalist 45% 43% 41–44% 6%*
eBird 46% 40% 37–42% 16%*
Table A.4 Predicted 2020 counts, observed 2020 counts, 95% predicted 2020 interval, and percent change between predicted and observed participants, contributed observations, and percent of observations originating from within urban areas by state, March–June 2020, for four community science programs. *Denotes 2020 actual value falls outside of 95% prediction interval. Tables are sorted by number of observations, participants, or %urban observations reported in 2020.
Table A.4Table A.4.a. eButterfly predicted and observed counts of observations.
State 2020 observations Predicted 2020 observations 95% prediction interval Percent difference
South Carolina 1108 748 336–1305 48
Virginia 646 303 71–643 113*
Vermont 637 567 230–1062 12
Arizona 396 317 77–682 25
Massachusetts 294 223 42–553 32
Texas 254 579 240–1045 −56
North Carolina 246 230 50–596 7
California 242 624 267–1158 −61*
Arkansas 219 85 0–311 158
Idaho 196 2 0–89 11278*
New Jersey 191 64 0–270 200
Florida 183 295 69–689 −38
Michigan 140 227 35–513 −38
Georgia 112 84 0–329 34
Rhode Island 83 46 0–228 79
Maryland 73 308 79–684 −76*
Pennsylvania 72 12 0–142 504
Maine 70 76 0–319 −8
Indiana 67 22 0–161 201
New Mexico 58 72 0–290 −19
Washington 51 9 0–123 443
Utah 34 1 0–56 6445
Oregon 33 23 0–176 41
Wisconsin 30 19 0–168 60
Iowa 23 63 0–288 −63
Colorado 21 11 0–137 83
Minnesota 18 1 0–90 1135
Connecticut 12 47 0–241 −74
New York 12 74 0–298 −84
New Hampshire 11 30 0–196 −63
Ohio 8 189 31–471 −96*
Delaware 2 0 0–61 2413
Illinois 2 0 0–70 7111
Nevada 2 6 0–131 −67
Wyoming 1 0 0–83 223
Alabama 0 7 0–125 −100
Alaska 0 5 0–128 −100
District of Columbia 0 0 0–67 −100
Hawaii 0 0 0–89 −100
Kansas 0 1 0–83 −100
Kentucky 0 0 0–65 −100
Louisiana 0 1 0–88 −100
Mississippi 0 2 0–46 −100
Missouri 0 18 0–178 −100
Montana 0 1 0–89 −100
Nebraska 0 1 0–55 −100
North Dakota 0 4 0–43 −100
Oklahoma 0 0 0–89 −100
South Dakota 0 3 0–49 −100
Tennessee 0 2 0–106 −100
West Virginia 0 8 0–131 −100
Table A.4.b. eButterfly predicted and observed counts of participants.
State 2020 participants Predicted 2020 participants 95% prediction interval Percent difference
Vermont 20 18 9–30 14
Virginia 19 14 5–25 40
Arizona 12 8 2–18 56
Massachusetts 12 5 1–13 143
California 11 17 7–29 −35
South Carolina 10 8 2–18 23
Michigan 9 9 3–19 0
North Carolina 9 12 4–23 −22
Washington 7 3 0–10 130
Florida 5 13 5–25 −62*
Maine 5 4 0–11 35
Maryland 5 6 1–15 −16
New Jersey 5 4 0–10 39
New Mexico 5 2 0–8 108
Ohio 5 5 1–13 −1
Connecticut 4 2 0–8 80
Pennsylvania 4 4 1–11 −8
Texas 4 11 4–22 −64
Georgia 3 6 1–14 −46
Iowa 3 1 0–6 114
New Hampshire 3 4 0–11 −19
Rhode Island 3 1 0–6 213
Utah 3 0 1–3 2373*
Arkansas 2 1 0–7 52
Colorado 2 2 0–8 −12
Minnesota 2 1 0–5 203
New York 2 6 1–15 −67
Oregon 2 3 0–10 −40
Wisconsin 2 2 0–7 32
Delaware 1 0 1–4 175
Idaho 1 1 0–6 1
Illinois 1 0 0–4 109
Indiana 1 2 0–8 −55
Nevada 1 1 0–5 −6
Wyoming 1 0 1–3 925
Alabama 0 1 0–6 −100*
Alaska 0 3 0–9 −100*
District of Columbia 0 0 0–4 −100*
Hawaii 0 0 1–2 −100*
Kansas 0 0 1–3 −100*
Kentucky 0 0 1–3 −100*
Louisiana 0 1 0–5 −100*
Mississippi 0 0 1–2 −100*
Missouri 0 2 0–7 −100*
Montana 0 0 0–4 −100*
Nebraska 0 0 1–2 −100*
North Dakota 0 0 2–1 −100*
Oklahoma 0 0 0–4 −100*
South Dakota 0 0 1–2 −100*
Tennessee 0 1 0–6 −100*
West Virginia 0 1 0–5 −100*
Table A.4.c. eButterfly predicted and observed percent of observations originating from urban areas. Prediction intervals >100% are reported to indicate the size of the interval, even though >100% is not possible.
State 2020 %urban observations Predicted 2020 %urban observations 95% prediction interval Percent difference 2020
New York 67 13 −36–63 394*
Ohio 63 32 −20–80 95
Georgia 62 24 −24–75 156
New Hampshire 55 7 −42–58 640
Arizona 45 38 −23–77 17
Maryland 45 26 −7–88 71
Massachusetts 45 27 −20–78 68
Indiana 43 32 −19–85 37
Florida 37 27 −20–78 38
Wisconsin 37 29 −21–78 28
New Jersey 36 25 −21–78 43
Virginia 33 16 −35–65 108
Washington 18 17 −33–65 3
California 17 16 −33–66 9
Pennsylvania 14 12 −39–61 19
Maine 13 9 −40–59 36
Utah 12 5 −47–53 144
Rhode Island 11 31 −15–83 −66
Texas 10 23 −26–74 −55
Iowa 9 44 −19–81 −80
New Mexico 9 16 −5–90 −47
Oregon 9 30 −34–68 −70
Connecticut 8 19 −32–68 −55
Vermont 7 15 −34–65 −54
Minnesota 6 6 −24–76 −7
North Carolina 6 27 −44–54 −79
South Carolina 5 9 −40–59 −44
Michigan 4 14 −34–64 −69
Arkansas 2 4 −45–53 −51
Alabama 0 6 −42–56 −100
Alaska 0 19 −29–68 −100
Colorado 0 8 −40–59 −100
Delaware 0 23 −23–73 −100
District of Columbia 0 57 7–104 −100*
Hawaii 0 20 −27–72 −100
Idaho 0 16 −32–67 −100
Illinois 0 28 −21–80 −100
Kansas 0 6 −44–59 −100
Kentucky 0 17 −31–64 −100
Louisiana 0 35 −14–82 −100
Mississippi 0 17 −29–61 −100
Missouri 0 2 −50–51 −100
Montana 0 4 −48–57 −100
Nebraska 0 4 −46–51 −100
Nevada 0 9 −38–56 −100
North Dakota 0 4 −48–55 −100
Oklahoma 0 11 −40–64 −100
South Dakota 0 4 −47–51 −100
Tennessee 0 4 −43–54 −100
West Virginia 0 3 −44–57 −100
Wyoming 0 3 −43–54 −100
Table A.4.d. iNaturalist predicted and observed counts of observations.
State 2020 observations Predicted 2020 observations 95% prediction interval Percent difference
California 421,217 646,163 298,846–1,371,409 −35
Texas 324,382 417,169 191,859–847,278 −22
Florida 102,353 89,616 38,673–196,973 14
New York 69,034 55,263 24,233–125,965 25
Virginia 64,914 48,953 21,810–111,707 33
Massachusetts 62,134 34,308 15,718–79,581 81
North Carolina 57,913 54,998 23,821–128,904 5
Ohio 56,446 63,488 28,495–148,840 −11
Pennsylvania 53,803 39,138 18,068–90,312 37
New Jersey 51,865 45,634 20,711–102,604 14
Maryland 48,320 44,035 19,535–99,190 10
Illinois 46,632 51,279 22,674–121,561 −9
Washington 34,030 39,698 18,419–91,715 −14
Oregon 33,731 37,700 16,826–83,479 −11
Arizona 32,992 60,004 26,386–135,104 −45
Vermont 31,067 62,015 27,096–140,588 −50
Minnesota 30,584 25,307 10,970–57,865 21
Wisconsin 27,845 26,248 12,412–59,896 6
Tennessee 27,436 24,027 10,716–54,993 14
Georgia 25,908 15,146 6783–34,783 71
Alabama 25,809 27,035 12,479–57,656 −5
Michigan 25,430 27,326 11,988–63,880 −7
Colorado 23,825 25,432 11,158–58,757 −6
Louisiana 21,120 16,983 7511–40,129 24
New Mexico 20,597 15,388 6991–33,828 34
Arkansas 18,297 15,936 7017–34,138 15
Oklahoma 17,626 17,294 7542–38,104 2
Indiana 14,546 7444 3376–17,452 95
Missouri 13,430 11,929 5068–26,152 13
South Carolina 13,363 16,100 6898–35,862 −17
Connecticut 13,021 12,398 5551–25,433 5
Utah 12,369 12,579 5704–29,990 −2
New Hampshire 11,881 7538 3498–16,869 58
Mississippi 11,638 7725 3365–18,211 51
Nevada 11,224 16,394 7787–40,065 −32
Kentucky 10,675 6795 3021–15,530 57
Idaho 8260 9394 4382–20,943 −12
Nebraska 8220 3214 1359–7286 156*
Maine 7877 11,197 4725–26,598 −30
Kansas 7470 8675 3833–18,720 −14
Alaska 7418 16,885 6830–36,057 −56
West Virginia 5925 5862 2613–13,397 1
Hawaii 5594 17,054 7854–38,337 −67*
Iowa 4845 4106 1760–9393 18
Rhode Island 4416 1667 701–3778 165*
Montana 4204 4268 1824–9863 −1
District of Columbia 3897 7608 3223–17,827 −49
Delaware 3573 4704 2054–10,395 −24
South Dakota 2879 2690 1207–6258 7
Wyoming 2573 3980 1830–8782 −35
North Dakota 812 1338 592–2835 −39
Table A.4.e. iNaturalist predicted and observed counts of participants.
State 2020 participants Predicted 2020 participants 95% prediction interval Percent difference
California 17,307 30,102 17,997–51,653 −43*
Texas 11,120 16,064 9115–27,829 −31
Florida 7534 8086 4815–14,115 −7
New York 4437 4953 2772–8669 −10
North Carolina 4371 4547 2657–7939 −4
Pennsylvania 3882 3467 2037–5890 12
Virginia 3755 4587 2715–8345 −18
Massachusetts 3653 3846 2239–6782 −5
Ohio 3195 4254 2525–7687 −25
Maryland 2865 3020 1665–5466 −5
Washington 2864 3528 2126–6215 −19
Georgia 2628 2128 1280–3624 24
Illinois 2375 2815 1656–4924 −16
New Jersey 2173 2136 1226–3626 2
Oregon 2166 3486 2030–6194 −38
Minnesota 2110 2197 1306–3891 −4
Tennessee 2102 2216 1332–3886 −5
Arizona 2087 3898 2178–6472 −46*
Colorado 1984 3243 1853–5443 −39
Michigan 1971 2044 1238–3664 −4
Wisconsin 1670 1904 1079–3178 −12
Missouri 1612 1393 843–2416 16
Connecticut 1456 1265 730–2201 15
Alabama 1407 1667 947–2963 −16
Vermont 1372 2366 1389–4116 −42*
Utah 1342 1836 1099–3300 −27
Indiana 1335 1201 701–2060 11
South Carolina 1300 1615 917–2826 −19
Louisiana 1149 1431 836–2488 −20
New Hampshire 1054 1013 594–1726 4
Oklahoma 1030 1138 666–1980 −10
New Mexico 963 1527 874–2658 −37
Arkansas 929 1082 641–1941 −14
Kentucky 884 1113 634–1874 −21
Maine 818 1131 664–1999 −28
Nebraska 788 594 336–1014 33
Hawaii 592 1727 1022–2941 −66*
Nevada 580 1355 820–2313 −57*
Idaho 579 1017 579–1796 −43
Iowa 555 645 385–1112 −14
Kansas 524 649 390–1240 −19
Mississippi 519 756 420–1291 −31
West Virginia 495 690 401–1206 −28
District of Columbia 467 1079 619–1868 −57*
Montana 448 845 465–1428 −47*
Rhode Island 399 320 180–547 24
Delaware 344 493 285–862 −30
Wyoming 304 752 421–1318 −60*
Alaska 242 1084 617–1870 −78*
South Dakota 213 363 215–621 −41*
North Dakota 74 196 112–343 −62*
Table A.4.f. iNaturalist predicted and observed percent of observations originating from urban areas. Prediction intervals >100% are reported to indicate the size of the interval, even though >100% is not possible.
State 2020 %urban observations Predicted 2020 %urban observations 95% prediction interval Percent difference
District of Columbia 100 102 88–117 −2
New Jersey 64 63 47–78 2
New York 62 57 43–72 8
Illinois 58 58 42–73 0
Massachusetts 57 73 57–87 −21
Virginia 57 55 40–70 4
Georgia 55 41 41–70 33
Maryland 55 56 25–56 −2
Pennsylvania 55 41 28–56 32
Connecticut 54 60 34–65 −10
Rhode Island 54 49 45–75 9
Texas 51 48 34–63 5
Florida 48 44 29–58 9
Louisiana 46 44 24–53 4
Missouri 46 39 29–60 20
Washington 46 40 25–55 15
Indiana 45 54 38–69 −16
North Carolina 45 52 24–53 −14
South Carolina 45 39 35–65 16
Minnesota 44 40 25–55 11
Tennessee 44 31 17–46 44
California 43 42 27–57 4
Nebraska 42 40 26–56 5
Michigan 41 26 11–41 57*
Ohio 41 40 26–55 2
Kansas 39 27 12–42 43
Utah 38 25 10–39 53
Oklahoma 37 37 21–52 −1
Colorado 35 32 11–41 12
Delaware 35 26 17–47 38
Hawaii 34 34 18–49 1
Oregon 34 31 17–48 9
Wisconsin 34 24 10–39 40
Iowa 33 34 19–49 −4
Alabama 30 45 30–59 −33
Arizona 26 27 8–37 −2
Arkansas 26 23 12–42 13
Nevada 26 26 9–41 0
Maine 23 19 4–34 23
Mississippi 23 29 14–44 −21
New Mexico 22 20 7–36 10
Kentucky 21 20 4–35 7
West Virginia 21 31 17–47 −32
Alaska 20 20 5–36 −2
Idaho 18 23 1–32 −23
North Dakota 18 16 8–39 10
Montana 17 19 5–34 −12
New Hampshire 16 19 3–33 −17
South Dakota 15 13 6–34 13
Vermont 15 20 −1–28 −26
Wyoming 10 8 −7–22 27
Table A.4.g. Nature's Notebook predicted and observed counts of observations.
State 2020 observations Predicted 2020 observations 95% prediction interval Percent difference
Massachusetts 12,238 8075 4687–12,627 52
New York 9712 15,005 9730–20,572 −35*
Minnesota 9385 13,879 8997–19,570 −32
Arizona 7720 6290 3218–10,248 23
Michigan 7137 6783 3616–11,565 5
Tennessee 6072 11,694 7169–17,093 −48*
California 5699 15,462 10,506–21,709 −63*
Maine 4565 4633 2157–8290 −1
Indiana 3515 802 37–2522 338*
North Carolina 3463 4214 1895–7625 −18
New Hampshire 2722 4223 1747–7424 −36
Colorado 2209 4520 1864–7998 −51
Ohio 2161 1454 260–3463 49
Oregon 1842 1980 449–4466 −7
Illinois 1800 2580 808–5483 −30
Pennsylvania 1726 2401 689–4940 −28
New Jersey 1665 144 0–1146 1053*
Louisiana 1477 492 0–1951 200
Maryland 1452 1348 200–3448 8
New Mexico 1085 1865 413–4157 −42
Texas 1030 2311 536–4942 −55
Washington 856 1559 304–3729 −45
Wisconsin 834 915 67–2789 −9
Georgia 795 368 0–1859 116
Virginia 790 2424 723–5119 −67
Florida 709 1906 493–4379 −63
Mississippi 707 392 0–1786 80
Iowa 599 256 0–1453 134
Utah 483 386 0–1746 25
Kentucky 436 678 16–2286 −36
West Virginia 381 740 40–2548 −48
Arkansas 372 230 0–1397 62
Kansas 356 708 17–2319 −50
South Dakota 343 901 57–2884 −62
Vermont 325 155 0–1248 109
Wyoming 320 358 0–1727 −11
Alabama 277 503 0–2061 −45
Alaska 222 226 0–1503 −2
Missouri 205 794 27–2689 −74
District of Columbia 138 246 0–1478 −44
Delaware 130 50 0–814 158
South Carolina 99 970 71–2746 −90
Oklahoma 65 105 0–1085 −38
Connecticut 54 403 0–1843 −87
Montana 47 584 1–2090 −92
Rhode Island 20 66 0–982 −70
Idaho 13 355 0–1666 −96
Nebraska 4 95 0–1148 −96
Nevada 1 154 0–1218 −99
Hawaii 0 56 0–1009 −100
North Dakota 0 543 7–1953 −100*
Table A.4.h. Nature's Notebook predicted and observed counts of participants.
State 2020 participants Predicted 2020 participants 95% prediction interval Percent difference
New York 231 190 114–303 22
Texas 179 30 6–77 491*
Massachusetts 130 132 62–220 −1
California 88 250 160–364 −65*
Arizona 81 104 48–186 −22
North Carolina 76 73 25–138 4
Minnesota 66 97 39–174 −32
Colorado 65 118 56–207 −45
Michigan 64 50 13–105 28
Pennsylvania 63 53 15–113 19
Illinois 62 53 17–116 16
Maine 61 79 32–148 −22
Oregon 45 61 20–121 −27
Tennessee 35 60 21–125 −41
Washington 34 31 5–76 11
Indiana 33 18 1–56 85
Maryland 31 51 15–106 −39
New Hampshire 31 30 4–74 4
Wisconsin 31 30 5–73 4
New Mexico 30 35 8–88 −15
Ohio 30 25 2–70 22
Virginia 30 49 14–106 −39
Oklahoma 27 6 0–34 362
Louisiana 24 11 0–43 110
Wyoming 19 11 0–45 66
Kentucky 18 112 53–193 −84*
South Dakota 15 17 0–57 −10
Utah 15 15 0–52 −2
District of Columbia 13 16 0–51 −18
New Jersey 12 13 0–47 −6
Kansas 11 21 2–59 −47
Mississippi 11 8 0–38 30
Florida 10 31 5–82 −68
Missouri 10 21 1–63 −52
West Virginia 9 24 3–68 −62
Vermont 8 10 0–42 −17
Connecticut 7 10 0–44 −31
Georgia 7 15 0–51 −54
Iowa 6 12 0–45 −51
Montana 5 9 0–39 −41
Arkansas 4 8 0–36 −49
Alaska 3 7 0–34 −59
Idaho 3 23 2–73 −87
Nebraska 3 6 0–35 −52
Delaware 2 3 0–26 −22
Rhode Island 2 4 0–27 −45
South Carolina 2 7 0–34 −69
Alabama 1 9 0–39 −89
Nevada 1 6 0–34 −85
Hawaii 0 6 0–32 −100
North Dakota 0 16 0–55 −100*
Table A.4.i. Nature's Notebook predicted and observed percent of observations originating from urban areas. Prediction intervals >100% are reported to indicate the size of the interval, even though >100% is not possible.
State 2020 %urban observations Predicted 2020 %urban observations 95% prediction interval Percent difference
District of Columbia 100 86 29–141 17
Nevada 100 74 19–129 35
South Carolina 100 69 15–127 44
Rhode Island 100 50 −7–105 99
Idaho 100 44 −10–98 127*
Delaware 100 17 −34–71 483*
Kentucky 100 80 24–134 25
Georgia 99 29 −26–85 240*
Florida 99 58 5–113 70
Oklahoma 95 47 −8–104 103
Michigan 95 66 13–128 44
Arkansas 92 61 9–116 50
Maryland 91 34 −18–87 168*
Indiana 84 51 −3–107 64
Illinois 79 66 10–124 20
Oregon 78 27 −27–83 190
West Virginia 72 33 −21–90 117
Washington 71 55 2–110 30
Massachusetts 71 47 −4–99 50
Connecticut 70 28 −23–84 152
Virginia 67 57 2–110 18
Iowa 59 51 −4–102 15
Texas 59 81 29–132 −27
Mississippi 58 17 −40–73 234
North Carolina 56 22 −30–74 157
Wyoming 51 56 4–111 −9
Pennsylvania 50 25 −28–83 104
Ohio 50 63 2–119 −20
Nebraska 50 20 −38–72 148
Minnesota 45 31 −27–89 45
Arizona 43 43 −10–96 0
New Mexico 43 36 −20–92 20
Wisconsin 42 62 6–116 −33
Maine 39 32 −23–85 22
Louisiana 36 5 −55–61 638
New York 34 46 −5–101 −25
California 32 21 −39–81 54
Colorado 28 49 −7–100 −43
Utah 26 56 −1–110 −54
Missouri 21 31 −25–81 −32
New Hampshire 17 1 −53–58 1508
Tennessee 12 4 −53–58 213
Vermont 10 27 −26–84 −63
Alaska 8 24 −35–76 −68
South Dakota 7 59 6–112 −89
New Jersey 5 71 19–127 −93*
Kansas 2 12 −40–67 −84
Alabama 0 36 −19–92 −100
Hawaii 0 19 −32–77 −100
Montana 0 11 −44–68 −100
North Dakota 0 35 −21–88 −100
Table A.4.j. eBird predicted and observed counts of observations.
State 2020 observations Predicted 2020 observations 95% prediction interval Percent difference
New York 218,659 193,087 138,498–263,557 13
California 213,295 242,914 170,640–340,753 −12
Pennsylvania 150,625 141,104 98,730–199,461 7
Texas 142,329 183,661 131,394–254,042 −23
Florida 124,574 144,507 103,823–199,484 −14
Michigan 121,443 128,107 93,063–177,884 −5
Ohio 113,267 119,427 85,514–166,551 −5
Washington 104,635 99,251 70,081–140,343 5
Massachusetts 102,812 95,707 69,494–136,349 7
Wisconsin 100,974 120,192 86,576–168,677 −16
Virginia 97,540 90,976 64,118–129,777 7
Colorado 96,765 102,331 74,187–144,880 −5
Illinois 96,251 96,219 70,344–134,582 0
Oregon 93,816 106,668 75,532–151,217 −12
Maryland 90,380 73,715 52,453–105,951 23
Minnesota 86,422 69,095 49,730–95,886 25
Arizona 71,975 95,744 67,924–133,395 −25
North Carolina 71,618 60,374 42,985–84,597 19
New Jersey 65,790 74,360 52,729–104,084 −12
Indiana 53,294 47,535 33,904–66,255 12
Maine 51,133 58,214 40,629–80,450 −12
Georgia 46,017 52,421 36,887–71,876 −12
Connecticut 45,663 48,831 36,021–68,532 −6
Tennessee 40,169 38,444 27,639–55,040 4
Missouri 39,218 33,731 23,977–46,735 16
Vermont 38,630 42,852 30,649–61,221 −10
New Mexico 32,585 34,544 24,331–49,447 −6
Montana 32,076 40,364 28,503–55,613 −21
South Carolina 30,587 29,147 20,263–40,799 5
New Hampshire 30,463 28,789 20,374–40,543 6
Kansas 29,057 34,048 24,441–48,024 −15
Utah 28,953 34,712 24,509–50,626 −17
Idaho 28,661 25,750 18,519–36,328 11
Alaska 22,279 39,909 28,803–55,014 −44*
Kentucky 21,007 18,452 12,969–25,513 14
Iowa 20,324 20,069 14,159–28,504 1
Louisiana 18,950 21,666 15,859–30,553 −13
Alabama 18,849 19,737 14,349–27,406 −5
Nebraska 17,612 16,371 11,662–22,938 8
Wyoming 15,586 16,382 11,709–23,148 −5
Oklahoma 15,286 18,772 13,428–26,233 −19
North Dakota 14,601 15,578 11,287–21,891 −6
Arkansas 14,477 13,724 9760–19,093 5
West Virginia 12,870 14,848 10,603–21,057 −13
Delaware 12,505 16,711 11,794–23,176 −25
Mississippi 10,798 10,663 7745–14,983 1
Rhode Island 10,436 9299 6558–13,094 12
South Dakota 10,324 11,724 8293–16,479 −12
Nevada 9945 13,610 9758–19,056 −27
District of Columbia 8446 7376 5246–10,219 15
Hawaii 4973 10,543 7518–14,804 −53*
Table A.4.k. eBird predicted and observed counts of participants.
State 2020 participants Predicted 2020 participants 95% prediction interval Percent difference
California 9385 12,104 9355–15,317 −22
New York 8039 7965 6227–10,180 1
Texas 6012 8267 6580–10,418 −27*
Florida 5879 8233 6567–10,253 −29*
Pennsylvania 5872 5871 4602–7406 0
Ohio 4559 5817 4558–7320 −22*
Virginia 4456 4796 3734–6148 −7
Massachusetts 4391 4703 3687–5968 −7
Washington 4376 4754 3763–6079 −8
Michigan 4369 5039 4019–6369 −13
Illinois 4004 4326 3389–5533 −7
North Carolina 3935 4101 3232–5258 −4
Wisconsin 3842 4539 3561–5677 −15
Colorado 3777 4439 3497–5632 −15
Maryland 3470 3641 2854–4589 −5
Arizona 3204 5087 3969–6421 −37*
New Jersey 3188 4126 3267–5162 −23*
Oregon 3097 3637 2858–4619 −15
Minnesota 2793 3141 2510–3937 −11
Georgia 2705 3157 2447–4036 −14
Indiana 2551 2598 2058–3293 −2
Tennessee 2085 2378 1880–3041 −12
Connecticut 2015 2139 1695–2674 −6
Missouri 2007 2153 1718–2744 −7
South Carolina 1999 2542 1996–3174 −21
Maine 1878 2814 2237–3525 −33*
Utah 1654 2281 1798–2875 −27*
New Mexico 1511 2276 1786–2896 −34*
New Hampshire 1416 1830 1449–2301 −23*
Vermont 1374 1779 1387–2258 −23*
Montana 1344 1644 1289–2098 −18
Idaho 1314 1397 1090–1765 −6
Kentucky 1139 1361 1055–1733 −16
Iowa 1130 1259 987–1600 −10
Kansas 1123 1463 1156–1880 −23*
Alabama 1114 1402 1090–1761 −21
Louisiana 1033 1537 1222–1953 −33*
Oklahoma 934 1300 1019–1639 −28*
Delaware 924 1513 1196–1915 −39*
West Virginia 920 1204 925–1541 −24*
Wyoming 903 1348 1065–1706 −33*
Nebraska 895 1060 841–1339 −16
Arkansas 857 1041 819–1315 −18
Rhode Island 740 712 560–904 4
Nevada 735 1356 1078–1711 −46*
Alaska 725 1694 1328–2162 −57*
District of Columbia 651 951 741–1213 −32*
Mississippi 634 842 661–1075 −25*
South Dakota 510 655 521–823 −22*
North Dakota 413 597 460–761 −31*
Hawaii 344 847 669–1073 −59*
Table A.4.l. eBird predicted and observed percent of observations originating from urban areas. Prediction intervals >100% are reported to indicate the size of the interval, even though >100% is not possible.
State Observed 2020 %urban observations Predicted 2020 %urban observations 95% prediction interval Percent difference 2020 observations (observed – predicted / predicted ∗ 100)
District of Columbia 100 103 97–110 −3
Illinois 63 63 57–68 1
New Jersey 61 56 50–62 8
Florida 60 55 49–61 8
Massachusetts 60 60 54–65 0
Connecticut 59 64 58–70 −8
Georgia 58 52 46–58 11*
California 57 48 43–54 18*
Rhode Island 54 49 43–54 11
Washington 53 40 35–46 33*
Louisiana 51 37 31–42 40*
Maryland 50 49 43–55 3
Kentucky 49 43 38–49 13*
North Carolina 49 46 40–52 6
Texas 49 39 34–45 24*
Virginia 49 46 40–52 5
Ohio 47 38 32–44 25*
Colorado 46 36 30–42 26*
Tennessee 46 41 35–47 11
Minnesota 45 46 40–52 −3
South Carolina 45 41 35–47 9
Alabama 44 35 29–41 27*
Pennsylvania 44 45 39–50 −3
Indiana 43 37 32–43 17*
New York 43 46 40–52 −8
Oregon 42 36 30–41 18*
Hawaii 41 35 29–41 17*
Mississippi 41 41 36–47 0
Missouri 41 40 34–47 3
Nevada 40 38 32–44 5
New Mexico 40 35 30–41 14
Michigan 37 36 31–42 1
Wisconsin 37 31 25–36 20*
Oklahoma 35 40 34–46 −13
Utah 35 31 25–37 15
Delaware 34 34 28–39 1
Kansas 33 28 22–34 18
New Hampshire 33 34 29–40 −4
Arizona 32 29 24–35 10
Arkansas 32 34 28–40 −6
Alaska 30 23 18–29 29*
Nebraska 29 26 20–31 15
Idaho 26 25 19–31 5
Iowa 26 34 28–39 −21*
Maine 24 25 20–31 −7
West Virginia 22 25 19–31 −13
Montana 21 18 12–24 14
North Dakota 19 20 15–26 −5
Wyoming 19 20 14–26 −8
Vermont 18 17 11–23 8
South Dakota 13 16 10–22 −16
Table A.5 Correlation between the length of stay-at-home orders (days) and counts of 2020 participants 2020 observations, and 2020 percent of observations originating from within urban areas, March–June 2020, for four community science programs.
Table A.5Program y x Adj r squared F1,49 statistic p value Estimate Standard error
Nature's Notebook 2020 observations Length stay at home (days) −0.01973 0.0327 0.8571
eButterfly 2020 observations Length stay at home (days) −0.01963 0.03732 0.8476
iNaturalist 2020 observations Length stay at home (days) −0.01954 0.04176 0.8389
eBird 2020 observations Length stay at home (days) −0.02041 0.0000615 0.9934
Nature's Notebook 2020 participants Length stay at home (days) −0.01969 0.03467 0.8531
eButterfly 2020 participants Length stay at home (days) 0.03397 2.758 0.1031 −3.305 1.99
iNaturalist 2020 participants Length stay at home (days) −0.01711 0.159 0.6918
eBird 2020 participants Length stay at home (days) −0.0007943 0.9603 0.3319
Nature's Notebook 2020 %urban observations Length stay at home (days) 0.04846 3.547 0.06561 2.788 1.48
eButterfly 2020 %urban observations Length stay at home (days) 0.07229 4.896 0.03161 1.5966 0.7216
iNaturalist 2020 %urban observations Length stay at home (days) −0.01922 0.05694 0.8124
eBird 2020 %urban observations Length stay at home (days) −0.008157 0.5955 0.444
Acknowledgements
We are very grateful to the thousands of dedicated participants in iNaturalist, eButterfly, Nature's Notebook, and eBird for sharing their time and talents. We are also grateful to the wonderful staff supporting the four community science programs. We thank Mike Crimmins and Jeff Oliver for statistical advice and technical support and Kent McFarland for helpful pre-review comments. We also thank two anonymous reviewers for helpful comments that improved the manuscript.
Funding information
This work was supported through a Cooperative Agreement from the 10.13039/100000203 U.S. Geological Survey [G18AC00135] and a Cooperative Agreement from the 10.13039/100000202 U.S. Fish & Wildlife Service [F19AC00168].
Data statement
The data and code used in this analysis are available at https://zenodo.org/record/4430966#.X_uQmlNKiV4.
==== Refs
References
Bakar, N.A. and S. Rosbi. 2020. Effect of Coronavirus disease (COVID-19) to tourism industry. International Journal of Advanced Engineering Research and Science 7: 189-193. doi:10.22161/ijaers.74.23.
Bates A.E. Primack R.B. Moraga P. Duarte C.M. COVID-19 pandemic and associated lockdown as a “Global Human Confinement Experiment” to investigate biodiversity conservation Biol. Conserv. 248 2020 108665 32549587
BIOCON-20-00460 Birds seen and lost at the time of COVID-19: impact of lockdown measures on bird observations from citizen science Biol. Conserv. 2020 this issue
Bonney R. Cooper C.B. Dickinson J. Kelling S. Phillips T. Rosenberg K.V. Shirk J. Citizen science: a developing tool for expanding science knowledge and scientific literacy BioScience 59 2009 977 984 10.1525/bio.2009.59.11.9
Bowman S. Gibson L. Here are 50 ways to celebrate 50 years of Earth Day (from your home during the pandemic) Indianapolis Star, Apr 17, 2020 https://www.indystar.com/story/news/environment/2020/04/17/earth-day-celebrations-coronavirus-stay-at-home-order/5137118002/?fbclid=IwAR1KbB0zDBLhc6LCd3KNncrRhFjYCZgDCZGg-kg68NPHxxIc_RxjjiyuCu4
Bowser A. Parker A. Long A. Citizen science and COVID-19: the power of the (distanced) crowd Wilson Center Blog, Jun 22, 2020 https://www.wilsoncenter.org/blog-post/citizen-science-and-covid-19-power-distanced-crowd 2020
California Academy of Sciences. 2020. City Nature Challenge. https://www.calacademy.org/citizen-science/city-nature-challenge. Accessed Dec 1, 2020.
City Nature Challenge. 2020. How it got started. https://citynaturechallenge.org/about/. Accessed Dec 1, 2020.
Cooper C.B. Dickinson J. Phillips T. Bonney R. Citizen science as a tool for conservation in residential systems Ecol. Soc. 12 2007 11 http://www.ecologyandsociety.org/vol12/iss2/art11/
Corlett R.T. Primack R.B. Devictor V. Maas B. Goswami V.R. Bates A.E. Koh L.P. Regan T.J. Loyola R. Pakeman R.J. Cumming G.S. Pigeon A. Johns D. Roth R. Impact of the coronavirus pandemic on biodiversity conservation Biol. Conserv. 246 2020 108751
Crimmins, T. 2020. Citizen science programs offer great options for families affected by recent school closures. Arizona Daily Star, 26 Mar, 2020.
Denny E.G. Gerst K.L. Miller-Rushing A.J. Tierney G.L. Crimmins T.M. Enquist C.A.F. Guertin P. Rosemartin A.H. Schwartz M.D. Thomas K.A. Weltzin J.F. Standardized phenology monitoring methods to track plant and animal activity for science and resource management applications Int. J. Biometeorol. 58 2014 591 601 10.1007/s00484-014-0789-5 24458770
Devictor V. Whittaker R.J. Beltrame C. Beyond scarcity: citizen science programmes as useful tools for conservation biology Divers. Distrib. 16 2010 354 362 10.1111/j.1472-4642.2009.00615.x
Dickinson J.L. Shirk J. Bonter D. Bonney R. Crain R.L. Martin J. Phillips T. Purcell K. The current state of citizen science as a tool for ecological research and public engagement Front. Ecol. Environ. 10 2012 291 297 10.1890/110236
Dinneen, J. 2020. Covid-19 can't stop citizen science. Undark Magazine, 17 Apr, 2020. https://undark.org/2020/04/17/covid-19-citizen-science/.
eBird Basic Dataset eBird 2020 Ithaca, New York http://www.ebird.org
eButterfly. 2020. Explore data. http://www.e-butterfly.org/ebapp/en/observations/explore?view=species&subview=grid. Accessed Dec 3, 2020.
Ellwood E. Crimmins T.M. Miller-Rushing A.J. Citizen science and conservation: recommendations for a rapidly moving field Biol. Conserv. 208 2017 1 4 10.1016/j.biocon.2016.10.014
Evans K.L. Ewen J.G. Guillera-Arroita G. Johnson J.A. Penteriani V. Ryan S.J. Sollmann R. Gordon I.J. Conservation in the maelstrom of Covid-19 – a call to action to solve the challenges, exploit opportunities and prepare for the next pandemic Anim. Conserv. 23 2020 235 238 10.1111/acv.12601 32837242
Fisher, B., W. Corcoran, C. Hill-James, B. Langton, H. Sommer, and N. Grima. 2020. The importance of urban natural areas and urban ecosystem services during the COVID-19 pandemic. SocArXiv 10.31235/osf.io/sd3h6 .
GBIF.org. GBIF Occurrence Download doi:10.15468/dl.wqjr5j. Accessed Aug 16, 2020.
Goodier M. Rayman J. Covid-19 is highlighting cities' unequal access to green space CityMetric, Jun 3, 2020 https://www.citymetric.com/fabric/covid-19-highlighting-cities-unequal-access-green-space-5168 2020
Iwane, T. 2020. Exploring nature when you're stuck at home. https://www.inaturalist.org/blog/31664-exploring-nature-when-you-re-stuck-at-home. Accessed Dec 1, 2020.
Jennings M. Conlisk E. Haeuser E. Foote D. Lewison R. Climate Resilient Connectivity for the South Coast Ecoregion of California Final Report prepared for California Department of Fish and Wildlife http://www.conservationecologylab.com/uploads/1/9/7/6/19763887/sdsu_climateresilientconnectivity_finalreport_48p.pdf 2019
Kelling S Johnston A Bonn A Fink D Ruiz-Gutierrez V Bonney R Fernandez M Hochachka W M Julliard R Kraemer R Guralnick R Using semistructured surveys to improve citizen science data for monitoring biodiversity BioScience 69 2019 170 179 10.1093/biosci/biz010 30905970
Kleinschroth F. Kowarik I. COVID-19 crisis demonstrates the urgent need for urban greenspaces Front. Ecol. Environ. 18 2020 318 319 32834788
Knowles, J.E. and C. Frederick. 2016. merTools: Tools for Analyzing Mixed Effect Regression Models. R package version 0.3.0.
Kornfeld M. As scientists struggle with rollbacks, stay at home orders and funding cuts, citizens fill the gap Inside Climate News, Jun 23, 2020 https://insideclimatenews.org/news/22062020/citizen-science-coronavirus 2020
Kubis A.D. Volunteers flock to sign up as citizen scientists Here & Now, WBUR. Apr 14, 2020 https://www.wbur.org/hereandnow/2020/04/14/volunteers-citizen-scientists 2020
Loarie, S. 2020. We passed 300,000 species observed on iNaturalist!! https://www.inaturalist.org/blog/42626-we-passed-300-000-species-observed-on-inaturalist#:~:text=In%20fact%2C%20even%20though%20the,total%20plant%20diversity%20(33%25). Accessed Dec 3, 2020.
Manenti R. Mori E. Di Canio V. Mercurio S. Picone M. Caffi M. Brambilla M. Ficetola G.F. Rubolini D. The good, the bad and the ugly of COVID-19 lockdown effects on wildlife conservation: insights from the first European locked down country Biol. Conserv. 249 2020 108728 10.1016/j.biocon.2020.108728 32863391
McKinley D.C. Miller-Rushing A.J. Ballard H.L. Bonney R. Brown H. Cook-Patton S.C. Evans D.M. French R.A. Parrish J.K. Phillips T.B. Ryan S.F. Shanley L.A. Shirk J.L. Stepenuck K.F. Weltzin J.F. Wiggins A. Boyle O.D. Briggs R.D. Chapin S.F. III Hewitt D.A. Preuss P.W. Soukup M.A. Citizen science can improve conservation science, natural resource management, and environmental protection Biol. Conserv. 208 2017 15 28 10.1016/j.biocon.2016.05.015
Mock J. ‘Black Birders Week’ promotes diversity and takes on racism in the outdoors Audubon Magazine, Jun 1 https://www.audubon.org/news/black-birders-week-promotes-diversity-and-takes-racism-outdoors 2020
Nathan E. Roth K. Pivovaroff A.L. Flowering phenology indicates plant flammability in a dominant shrub species Ecol. Indic. 109 2019 105745 10.1016/j.ecolind.2019.105745
National Academy for State Health Policy. 2020. Chart: Each state's COVID-19 reopening and reclosing plans and mask requirements. https://www.nashp.org/governors-prioritize-health-for-all/. Accessed Aug 7, 2020.
Nicola M. Alsafi Z. Sohrabi C. Kerwan A. Al-Jabir A. Iosifidis C. Agha M. Agha R. The socio-economic implications of the coronavirus pandemic (COVID-19): a review Int. J. Surg. 2020 10.1016/j.ijsu.2020.04.018
Nugent, J. iNaturalist: Citizen science for 21st-century naturalists. Science Scope, vol. 41, no. 7, Mar. 2018, p. 12+. Gale Academic OneFile, https://link-gale-com.ezproxy1.library.arizona.edu/apps/doc/A531045587/AONE?u=uarizona_main&sid=AONE&xid=205f9e36. Accessed Aug 20, 2020.
Pennisi E. Pandemic Robs Field Scientists of ‘Once-in-a-lifetime’ Moments Science Magazine Apr 15 2020 10.1126/science.abc2655
Piñon N. Summer camp canceled? Here are some citizen science projects to do at home Mashable, Jul 11 https://mashable.com/article/citizen-science-for-kids-summer/ 2020
Plummer R. McGrath D. Sivarajah S. How cities can add accessible green space in a post-coronavirus world The Conversation, June 11 https://theconversation.com/how-cities-can-add-accessible-green-space-in-a-post-coronavirus-world-139194 2020
Prudic K.L. McFarland K.P. Oliver J.C. Hutchinson R.A. Long E.C. Kerr J.T. Larrivée M. eButterfly: leveraging massive online citizen science for butterfly conservation Insects 8 2017 53 28524117
Rosemartin A.H. Crimmins T.M. Enquist C.A.F. Gerst K.L. Kellerman J.L. Posthumus E.E. Weltzin J.F. Denny E.G. Guertin P. Marsh L.R. Organizing phenological data resources to inform natural resource conservation Biol. Conserv. 173 2014 90 97 10.1016/j.biocon.2013.07.003
Rosemartin, A.H., E.G. Denny, K.L. Gerst, R. L. Marsh, T.M. Crimmins, and J.F. Weltzin. 2018. USA National Phenology Network Observational Data Documentation. U.S. Geological Survey Open-File Report 2018–1060. doi: 10.3133/ofr20181060.
Rutz C. Loretto M.-C. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R.B. Ropert-Coudert Y. Tucker M.A. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evol. 4 2020 1156 1159 10.1038/s41559-020-1237-z 32572222
Seltzer C. Making biodiversity data social, shareable, and scalable: reflections on iNaturalist & citizen science Biodivers. Inf. Sci. Stand. 2019 10.3897/biss.3.46670
Slater S.J. Christiana R.W. Gustat J. Recommendations for keeping parks and green space accessible for mental and physical health during COVID-19 and other pandemics Prev. Chronic Dis. 17 2020 E59
Sullivan B.L. Wood C.L. Iliff M.J. Bonney R.E. Fink D. Kelling S. eBird: a citizen-based bird observation network in the biological sciences Biol. Conserv. 142 2009 2282 2292
Sullivan B.L. Aycrigg J.L. Barry J.H. Bonney R.E. Bruns N. Cooper C.B. Damoulas T. Dhondt A.A. Dietterich T. Farnsworth A. Fink D. The eBird enterprise: an integrated approach to development and application of citizen science Biol. Conserv. 169 2014 1 40 10.1016/j.biocon.2013.11.003
Sullivan B.L. Phillips T. Dayer A.A. Wood C.L. Farnsworth A. Iliff M.J. Davies I.J. Wiggins A. Fink D. Hochachka W.M. Rodewald A.D. Rosenberg K.V. Bonney R. Kelling S. Using open access observational data for conservation action: a case study for birds Biol. Conserv. 208 2017 5 14 10.1016/j.biocon.2016.04.031
Team eBird. 2019. Team eBird taxonomy update—complete! https://ebird.org/news/2019-ebird-taxonomy-update. Accessed Dec 3, 2020.
Team eBird. 2020. Global big day 2020: birding's biggest team. https://ebird.org/news/global-big-day-2020-birdings-biggest-team. Accessed Dec 1, 2020.
U.S. Census Bureau. 2017. TIGER/Line Shapefile, 2017, 2010 nation, U.S., 2010 Census Urban Area National. https://catalog.data.gov/dataset/tiger-line-shapefile-2017-2010-nation-u-s-2010-census-urban-area-national.
Ueda, K. 2020. iNaturalist Research-grade Observations. iNaturalist.org. Occurrence dataset 10.15468/ab3s5x. Accessed via GBIF.org Dec 1, 2020.
Unger S. Rollins M. Tietz A. Dumais H. iNaturalist as an engaging tool for identifying organisms in outdoor activities J. Biol. Educ. 2020 10.1080/00219266.2020.1739114
USA National Phenology Network. 2020. Data Dashboard. https://www.usanpn.org/data/dashboard. Accessed Dec 1, 2020.
USA National Phenology Network. 2020b. Plant and Animal Phenology Data. Data type: Status Records. 01/01/2010-8/07/2020 for the United States. USA-NPN, Tucson, Arizona, USA. Data set accessed 08/07/2020 at 10.5066/F78S4N1.
Van Horn, G., O. Mac Aodha, Y. Song, Y. Cui, C. Sun, A. Shepard, H. Adam, P. Perona, and S. Belongie. 2018. The iNaturalist species classification and detection dataset. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. 8769-8778.
Young, A. City Nature Challenge RESULTS. https://www.inaturalist.org/projects/city-nature-challenge-2020/journal/34792-city-nature-challenge-results. Accessed Nov 30, 2020.
Zellmer A.J. Wood E.M. Surasinghe T. Putman B.J. Pauly G.B. Magle S.B. Lewis J.S. Kay C.A.M. Fidino M. What can we learn from wildlife sightings during the COVID-19 global shutdown? Ecosphere 11 2020 e03215 10.1002/ecs2.3215
| 0 | PMC9746923 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 Apr 4; 256:109017 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.109017 | oa_other |
==== Front
Appl Ergon
Appl Ergon
Applied Ergonomics
0003-6870
1872-9126
Elsevier Ltd.
S0003-6870(21)00198-8
10.1016/j.apergo.2021.103551
103551
Article
Validity of the occupational sitting and physical activity questionnaire (OSPAQ) for home-based office workers during the COVID-19 global pandemic: A secondary analysis
Dillon Kirsten a∗
Hiemstra Madison a
Mitchell Marc a
Bartmann Nina b
Rollo Scott cd
Gardiner Paul A. ae
Prapavessis Harry a
a School of Kinesiology, Faculty of Health Sciences, The University of Western Ontario, London, ON, Canada
b Center for Advanced Hindsight, Duke University, Durham, NC, USA
c Healthy Active Living and Obesity Research Group, Children's Hospital of Eastern Ontario Research Institute, Ottawa, ON, K1H 8L1, Canada
d School of Epidemiology and Public Health, Faculty of Medicine, University of Ottawa, Ottawa, ON, K1H 8M5, Canada
e School of Health & Wellbeing, Faculty of Health, Engineering and Sciences, University of Southern Queensland, Queensland, Australia
∗ Corresponding author. Faculty of Health Sciences, The University of Western Ontario, London, Ontario, N6A 3K7, Canada.
12 8 2021
11 2021
12 8 2021
97 103551103551
25 4 2021
14 7 2021
6 8 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
High levels of occupational sitting is an emerging health concern. As working from home has become a common practice as a result of COVID-19, it is imperative to validate an appropriate self-report measure to assess sitting in this setting. This secondary analysis study aimed to validate the occupational sitting and physical activity questionnaire (OSPAQ) against an activPAL4™ in full-time home-based ‘office’ workers (n = 148; mean age = 44.90). Participants completed a modified version of the OSPAQ and wore an activPAL4™ for a full work week. The findings suggest that the modified OSPAQ has fair levels of validity in terms of correlation for sitting and standing (ρ = 0.35–0.43, all p < 0.05) and agreement (bias = 2–12%) at the group level; however, estimates were poor at an individual level, as suggested by wide limits of agreement (±22–30%). Overall, the OSPAQ showed to be an easily administered and valid questionnaire to measure group level sitting and standing in this sample of adults.
Keywords
OSPAQ
activPAL™
Measurement-of-agreement
==== Body
pmc1 Introduction
Sedentary behaviour is defined as any waking behaviour in a seated, lying or reclining posture while expending less than or equal to 1.5 metabolic equivalents (Tremblay et al., 2017). Increased time spent sedentary has been associated with a higher risk of type 2 diabetes, cardiovascular disease and all-cause mortality (Katzmarzyk et al., 2019), independent from physical activity levels (Owen et al., 2010). Office-working adults have been shown to spend up to 77% of their workday sitting (Thorp et al., 2012), and therefore represent an at-risk population for high levels of sedentary time. The health concerns associated with high amounts of sedentary behaviour are increasingly being recognized with, for example, the recent launch of the Canadian 24-Hour Movement Guidelines for Adults, which provide time-specific recommendations for limiting daily sedentary time (Ross et al., 2020). Accordingly, it is not surprising that numerous interventions have targeted sedentary behaviour reduction in the workplace (Blackburn et al., 2020).
In response to the SARS-CoV-2 (COVID-19) global pandemic, many desk-based workers have pivoted from working at the office to working from home. For instance, findings from a survey targeting American full- or part-time office-workers (n = 5858), found that only 20% reported working from home almost all the time or full-time pre-COVID; however, since the onset of COVID-19, these proportions have risen to 71% working from home most of the time or full-time (Pew Research Center, 2020). This rapid, unplanned and unequipped transition from office-to home-based settings for many office workers, in addition to social distancing and lockdown measures, has been linked to significant decreases in overall physical activity as well as significant increases in total daily sitting time (Ammar et al., 2020; Fitbit, 2020; Tison et al., 2020). Evidence suggests that these decreases in physical activity may also be having a negative impact on mental health outcomes, including increased depression, loneliness, stress and decreased positive overall mental health (Meyer et al., 2020).
This work-from-home trend seems likely to outlive the global COVID-19 pandemic, as this pivotal transition seems to have already changed the environment of future office work. For instance, many workers have reported both preference and employer granted options for in-office/work-from-home flexibility indefinitely (Anderson et al., 2021; Pew Research Center, 2020). Hence, it is important that interventions begin to target this new growing population's sedentary behaviour patterns. Currently, evidence is sparse regarding interventions aimed at reducing sedentary behaviours in office workers who work from home. In addition to effective interventions, accurate measures to capture sedentary behaviour in this new “office environment” need to be tested and validated. Ideally, a measure that is less expensive and can be easily and quickly distributed, such as a self-report questionnaire, is urgently needed to advance sedentary research in this new segment of the working population.
While the preferred method of sedentary behaviour measurement is with a device that can differentiate sitting from standing (i.e., activPAL™), this type of device-based measurement is usually expensive (e.g., costs associated with purchasing each device, delivery to participants, dressings needed) and relatively invasive to ask participants to wear. Due to this cost barrier and added participant burden, there are a number of self-report questionnaires that have been developed and used in the literature. Questionnaires that have been previously used to assess sedentary behaviours in office working adults include, but are not limited to, the International Physical Activity Questionnaire (IPAQ; Ekelund et al., 2006), the Workforce Sitting Questionnaire (Aittasalo et al., 2017), the Sedentary Behaviour Questionnaire (SBQ; Rosenberg et al., 2010) and the Occupational Sitting and Physical Activity Questionnaire (OSPAQ; Chau et al., 2012). Amongst these, the OSPAQ has been used in an array of populations such as university students (Dillon et al., 2021), university employees (Headley et al., 2018), sedentary obese individuals (Lohana and Yadav, 2020), health professionals (Zafiropoulos et al., 2019) and of relevance, office workers (Nelson-Wong et al., 2020; Rollo and Prapavessis, 2020; Urda et al., 2017), to measure time spent sitting, standing, walking and perfoming heavy labour tasks during work hours. The OSPAQ measures sitting and standing as separate behaviours, thus, making it an ideal self-report tool to properly classify sedentary behaviour separate from physical inactivity. It is also very easy to implement as it only consists of three questions, minimizing participant burden. Validation studies using the OSPAQ have previously been conducted in various populations and demonstrated mixed levels of agreement and reliability depending on the occupation (i.e., sedentary versus non-sedentary) and device-based measure used (i.e., Actigraph versus activPAL™) (Chau et al., 2012; Jancey et al., 2014; Maes et al., 2020; van Nassau et al., 2015). Whether these findings can be replicated among traditional office-workers now working from home warrants investigation. It is imperative to establish the ‘construct validity’ of this questionnaire in an at-home ‘office’ worker population to allow future research to confidently assess sedentary behaviour within this setting, without the need for costly device-based measures. It is also important to note that these previous validation studies carry several limitations and pose risk of bias due to inadequate sample size (van Nassau et al., 2015) or use a device that cannot accurately differentiate sedentary behaviour (i.e., sitting) from physical inactivity (i.e., standing) (Chau et al., 2012; Jancey et al., 2014).
Validity evidence is lacking towards a questionnaire that can be administered to home-based office workers. Hence, a secondary analysis of data from an unpublished randomized controlled trial (NCT04488796) was undertaken to examine the measurement of agreement between the OSPAQ and the activPAL4™ inclinometer for estimating percentage of time spent sitting, standing and moving (i.e., walking) during work hours in office-working adults who had transitioned to working from home due to the COVID-19 pandemic.
2 Material and methods
2.1 Study design & population
We performed a secondary analysis on data from an unpublished pre-registered randomized controlled trial (NCT04488796) that aimed to decrease and break up time spent sedentary among home-based office workers. Data were collected from September to December 2020. Participants were full-time, home-based office workers living in London, Ontario or the surrounding area. Individuals were eligible to participate if they were 18 years or older, self-declared working full-time (i.e., employed 30+ hours/week) 5 days per week (i.e., Monday to Friday), self-declared working at least 3 days per week from their home, were able to read and write in English and had access to a computer with Internet and email. Participants were ineligible if they were planning on leaving their current employer or taking a leave of absence/vacation for more than three consecutive workdays for the duration of the study. They were also ineligible if they self-declared having a medical condition or physical limitation that prevented them from being physically active.
Participants were recruited using a number of strategies. First, contact was made via email with relevant liaisons and/or senior executives of potential businesses of interest (i.e., offices/businesses that were known to be working from home due to COVID-19). If interested, they were asked to email all full-time employees within their respective office/business inviting them to participate. Second, recruitment emails were sent directly to home-based office-working employees whose contact information was publicly available on company or institution websites (e.g., employee directories). Third, home-based office workers were recruited via recruitment posters distributed on various social media platforms (i.e., Facebook, Instagram, Twitter, LinkedIn). The recruitment emails included relevant study details (i.e., objective, eligibility criteria, brief procedures) and instructed interested individuals to contact the researcher via email if they wished to participate or wanted to receive additional details prior to making a decision. The study was approved by the institutional research ethics board.
2.2 Procedure & measures
After receiving a study invitation email, interested participants were sent a link with a unique authorization code and asked to complete an online questionnaire through a survey website called SoSci (www.soscisurvey.de). The online questionnaire consisted of a Letter of Information, informed consent and a baseline questionnaire assessing relevant demographic characteristics and outcomes of interest (i.e., primary and secondary measures including the OSPAQ). Upon completion of the first questionnaire, participants were emailed a PDF version of the Letter of Information/Informed Consent and were asked to sign the form (digitally) and send it back to research personnel, along with their address for activity monitor delivery. Participants received the activPAL4™ device via courier and were instructed to apply the device on Sunday evening and to wear the device all day for a period of 5 working days (Monday through Friday). Upon receiving the activPAL4™, participants also received a link (via email) to a detailed video outlining the proper procedures on how to apply the device. If there was any confusion, they were asked to either email or call one of the researchers. Upon finishing, they were instructed to place the device into the return envelope that was provided, and it was picked up via courier the following Saturday. Participants then underwent a 4-week intervention period, filling out the OSPAQ at the end of each workweek (i.e., Friday). During the fourth week, they again wore the activPAL™ device and this was the period used for this secondary analysis validation study.
2.3 OSPAQ-revised
The percentage of time spent sitting, standing and moving (i.e., walking) during work hours was measured using a modified version of the OSPAQ (Chau et al., 2012). Due to the sedentary ‘office’ setting, “heavy labour or physically demanding tasks” was removed from the questionnaire. This decision was made based off previous work done in the field that have reported low or zero prevalence of this behaviour in the workplace (Chau et al., 2012; Jancey et al., 2014; van Nassau et al., 2015). Participants were asked to record both the total number of days and hours they worked in the last 7 days. Participants were then asked to record the percentage of time spent sitting, standing, and moving (i.e., walking) on a typical workday in the last 7 days (i.e., “How would you describe your typical workday for the last 7 days? This involves only time spent in work-related activities and does not include what you did in your leisure time."). The sum of all percentages were to equal a total of 100% (e.g., 80% occupational sitting, 10% occupational standing and 10% walking). Time spent in each behaviour (minutes) was calculated as follows: [Minutes worked in the last 7 days/Days at work in the last 7 days] × [Percentage of the behaviour reported (i.e., sitting/standing/moving)/100].
2.4 activPAL™
The activPAL™ is currently considered the most accurate field-based measure of sitting time and sit-to-stand transitions (Kozey-Keadle et al., 2011). The activPAL4™ was the model used in the present study and is a small device worn on the midline anterior aspect of the thigh (right or left) that can differentiate between sedentary, standing and free moving activity using proprietary algorithms (Intelligent Activity Classification, PAL Technologies). Participants were instructed to wear the device for a full work week (i.e., Monday-Friday) at baseline as well as during the last intervention week. The activPAL™ monitor has been shown to be highly accurate as direct observation has shown a perfect correlation for time spent sitting/lying, standing and walking in primary school aged children (Aminian and Hinckson, 2012) and has been used in many previous validation studies involving adults (Clark et al., 2013; Júdice et al., 2015). The activPAL™ default settings were used, the validation wear time protocol was set to the ‘24-hour protocol’ (allowing 4 hours of non-wear per day), and data were downloaded in custom duration epochs (15 seconds) via activPAL™ Professional Software (version 8.11.4.61) and transferred to Microsoft Excel (version 16.44). Participants were required to have at least three valid workdays from Monday-Friday to be used in data analysis, which is consistent with previous studies (Edwardson et al., 2017). In the baseline questionnaire, participants were asked to report the start and end time of their workday (i.e., What are the hours you work in-between?). The data analyzed for each participant's workday included the data between the self-declared start time (i.e., 9:30am) up to (and including) the last 15 seconds (5:29:45pm) before the official end time (i.e., 5:30pm). Average daily sedentary time (minutes per day) was calculated [total amount of time/average number of days] using two different equations. First, as the sum of ‘sedentary’, ‘primary lying’ and ‘secondary lying’ time. Second, all the behaviours included in the first approach plus time spent in ‘seated transportation’. Time spent standing was calculated from the ‘upright time’. Time spent moving was done as two separate calculations, the first consisting solely of ‘stepping time’ and the second combining ‘stepping time’ with ‘cycling time’. Each valid day of data was totaled and then averaged for the number of valid days to calculate average daily time (minutes) for the week. The percentage of time spent sitting, standing and walking from the activPAL4™ was calculated as follows [average minutes spent in the behaviour (i.e., sitting, standing or moving): per workday/total minutes of work time (i.e., 9:30 a.m. to 5:30 p.m. = 8 hours*60)] × [100].
2.5 Statistical analysis
Statistical procedures were conducted in SPSS Statistics, Version 27 (SPSS Inc., Chicago, Illinois), GraphPad Prism version 9.0.2 (GraphPad Software Inc., San Diego, CA) and Stata Statistical Software Release 11.0 (StataCorp LP, College Station, TX) software programs. The level of significance was set at p < 0.05. Descriptive statistics were calculated for all demographic variables collected at baseline and are shown as mean (standard deviation (SD)) or number (percentage) of the sample. Univariate ANOVAs (continuous variables) and chi-square tests (categorical variables) were conducted to ensure that there were no systematic differences between participants with valid and invalid data (all p-values > 0.05). Bland and Altman (1999) do not recommend excluding outliers; however, they do suggest assessing the influence of outliers on the results. Therefore, we ran the analysis both before and after removing extreme outliers with a winsorization technique (Guttman and Smith, 1969). A total of 11 data points were imputed this way. The removal of extreme outliers did not impact the results and were therefore left in the analysis.
Spearman correlation coefficients were calculated to assess the degree of association between the activPAL4™ and modified OSPAQ. The strength of the correlation was interpreted as poor (<0.30), fair (0.30–0.50), moderately strong (0.60–0.80), or very strong (>0.80) (Chan 2003). Limits of agreement between the activPAL4™ and the modified OSPAQ were determined according to the recommendations by Bland and Altman (Bland and Altman, 1986). The difference [OSPAQ − activPAL4™] of the two paired measurements (as a percentage) was plotted against the average [(OSPAQ + activPAL4™)/2] of the two measurements (as a percentage). Percentage was deemed the most appropriate way to express the data because the OSPAQ is asked and interpreted as a percentage. The Bland-Altman plots expressed in minutes can be found in supplementary data Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5 . The mean difference, or bias, between the methods and the 95% limits of agreement intervals were calculated. Linear regression was used to determine linear bias. Significant linear bias indicates that the variability remained constant across average values while the mean difference increased significantly as average values increased. Therefore, where linear regression showed to be significant, the Bland–Altman plot presents the trend line for mean difference obtained from the regression and limits of agreement (±1.96 SD).Fig. 1 Bland–Altman plot of agreement of total self-report (OSPAQ) sitting time (including transportation time) with device-derived (activPAL4™) sitting time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (−5.63), with the dashed lines representing the limits of agreement (± 29.87).
Fig. 1
Fig. 2 Bland–Altman plot of agreement of self-report (OSPAQ) sitting time (excluding transport time) with device-derived (activPAL4™) sitting time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (−1.58), with the dashed lines representing the limits of agreement (± 28.31).
Fig. 2
Fig. 3 Bland–Altman plot of agreement of self-report (OSPAQ) standing time with device-derived (activPAL4™) sitting time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (−11.71), with the dotted lines representing the limits of agreement (± 22.34).
Fig. 3
Fig. 4 Bland–Altman plot of agreement of self-report (OSPAQ) moving time (including cycling time) with device-derived (activPAL4™) moving time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (+6.44), with the dashed lines representing the limits of agreement (± 15.36).
Fig. 4
Fig. 5 Bland–Altman plot of agreement of self-report (OSPAQ) moving time (excluding cycling time) with device-derived (activPAL4™) moving time (n = 95). The y axis is the difference between the two measures and the x axis is the average of the two, both expressed as a percentage. The bolded dashed line shows the mean difference between the two measures (+6.60), with the dashed lines representing the limits of agreement (± 14.93).
Fig. 5
2.6 Missing data
On any given variable at a single assessment point, the maximum percentage of missing data was 28% (n = 41). Of the 148 participants that filled out the baseline questionnaire, 108 of them had valid activPAL4™ data and of those, 95 had valid self-report data. Independent samples t-tests revealed that those who had valid activPAL4™ data were not different from those who did not have valid data on all demographic variables (p-values > 0.05). Taken together, all missing data were considered random. Hence, we decided to exclude missing data from the analysis.
3 Results
Descriptive statistics for the demographic variables, days worked from home and minutes worked per week are shown in Table 1 . Percentages and minutes of device based (activPAL4™) and self-reported (OSPAQ) behaviour characteristics during work hours are illustrated in Table 2 . The spearman rank correlation coefficient data between the activPAL4™ device and modified OSPAQ are displayed in Table 3 . All the spearman correlations were found to be significant (p < 0.05). The correlation of the activPAL4™ device with sitting and standing were fair (ρ = 0.35–0.43) and the correlation with moving was poor (ρ = 0.21–0.22).Table 1 Participant characteristics presented as mean (SD) or count (%) of group.
Table 1Variable Total sample (n = 148) Valid data (n = 108) Statistic (valid vs invalid) p-level
Age (years) 44.90 (SD = 11.41) 45.52 (SD = 11.38) F(1,147) = 1.181 0.279
Gender Χ2 (2) = 2.784 0.249
Male 40 (27.0%) 30 (27.8%)
Female 107 (72.3%) 78 (72.2%)
Non-Binary 1 (0.7%) 0 (0.0%)
Ethnicity Χ2 (4) = 0.688 0.953
White 126 (85.1%) 91 (84.3%)
Asian 7 (4.7%) 6 (5.6%)
Black or African American 3 (2.0%) 2 (1.9%)
Hispanic or Latino 4 (2.7%) 3 (2.8%)
Other 4 (2.7%) 3 (2.8%)
BMI (kg/m2) 27.33 (SD = 5.74) 27.42 (SD = 5.94) F(1,147) = 0.105 0.747
Level of Education Χ2 (4) = 1.741 0.783
Highschool Diploma 13 (8.8%) 9 (8.3%)
College Degree 26 (17.6%) 17 (15.7%)
University Degree 57 (38.5%) 42 (38.9%)
Masters 30 (20.3%) 22 (20.4%)
Doctorate (i.e., MD, PhD) 22 (14.9%) 18 (16.7%)
Marital Status Χ2 (4) = 4.533 0.339
Single 26 (17.6%) 22 (20.4%)
Married or equivalent 107 (72.3%) 73 (67.6%)
Separated or equivalent 7 (4.7%) 6 (5.6%)
Divorced 7 (4.7%) 6 (5.6%)
Widowed 1 (0.7%) 1 (0.9%)
Work Sector Χ2 (3) = 2.546 0.467
Private 61 (41.2%) 41 (38.0%)
Public 77 (52.0%) 58 (53.7%)
Charity 2 (1.4%) 2 (1.9%)
Other 7 (4.7%) 6 (5.6%)
Days worked from home
Five 118 (79.7%) 83 (76.9%) Χ2 (1) = 2.048 0.175
Four 19 (12.8%) 15 (13.9%) Χ2 (1) = 0.395 0.782
Three 11 (7.4%) 10 (9.3%) Χ2 (1) = 1.938 0.289
Physical Activitya Χ2 (1) = 0.107 0.852
Yes 85 (57.4%) 61 (56.5%)
No 62 (42.6%) 46 (42.6%)
Minutes Worked 494.56 (SD = 62.55) 493.71 (SD = 65.82) F(1,147) = 0.074 0.785
a In the past 3 months, have you been active for a minimum of 30 min/day on at least 3 days of the week? (i.e., jogging, biking, swimming).
Table 2 Percentages and minutes of device-based (activPAL4™) and self-reported (OSPAQ) behaviours during work hours.
Table 2Variable (activPAL4™: n = 108, OSPAQ: n = 95)
Mean (SD), Median (Range) (n = 108)
Mean (SD), Median (Range)
% Sitting
activPAL4™ a75.00 (12.48), 74.96 (27.22–94.32) b78.82 (13.34), 79.57 (27.96–130.41)
OSPAQ 71.82 (18.82), 75.00 (20.00–97.00)
Minutes Sitting
activPAL4™ a365.85 (72.57), 369.61 (136.09–502.51) b384.50 (77.61), 386.72 (144.12–625.95)
OSPAQ 339.51 (108.28), 336.00 (68.57–612.00)
% Standing
activPAL4™ 25.09 (12.46), 25.04 (5.68–72.78)
OSPAQ 13.46 (13.48), 10.00 (0.00–70.00)
Minutes Standing
activPAL4™ 122.60 (64.26), 113.81 (27.24–393.01)
OSPAQ 61.36 (56.88), 45.00 (0.00–294.00)
% Moving (i.e., walking)
activPAL4™ 8.06c (4.75), 6.96 (1.54–24.09) d8.23 (4.93), 7.17 (1.54–26.45)
OSPAQ 14.72 (10.86), 10.00 (0.00–50.00)
Minutes Moving (i.e., walking)
activPAL4™ c39.64 (25.13), 33.40 (7.40–146.82) d40.51 (26.52), 34.89 (7.40–166.61)
OSPAQ 68.58 (51.04), 54.00 (0.00–240.00)
a Without transportation time included.
b With transportation time included.
c Without cycling time included.
d With cycling time included.
Table 3 Concurrent validity of the OSPAQ with the activPAL4™.
Table 3Variable (n = 95)
ρ
% Sittinga 0.406***
% Sittingb 0.425***
% Standing 0.349**
% Movingc 0.224*
% Movingd 0.211*
*Significant at the 0.05 level.
** Significant at the 0.001 level.
*** Significant at the 0.0001 level.
a Without transportation time included.
b With transportation time included.
c Without cycling time included.
d With cycling time included.
The Bland-Altman plots for percentage of time spent sitting, standing and moving is displayed in Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5.
For total time spent sitting (Fig. 1), linear regression showed a significant positive association between the difference in the two measures (self-reported minus activPAL4™ derived sitting time) and the average of these two measures (B = 0.42, SE = 0.12, p = 0.001). Thus, the mean difference is estimated at −37.97% + 0.42 x average of the two measures. At mean levels of average self-reported/activPAL4™-derived sitting time (76.29%), the mean difference indicated self-reported sitting time was −5.63% lower than activPAL4™-derived sitting time with wide limits of agreement (±29.87%). When excluding transportation time from the device based sitting time (Fig. 2), the linear regression again showed a significant positive association between the difference in the two measures and the average of these two measures (B = 0.55, SE = 0.11, p = < 0.001). Thus, the mean difference is estimated at −42.63 percent +0.55 x average of the two measures. At mean levels of average self-reported/activPAL4™-derived sitting time (74.26%), the mean difference indicated self-reported sitting time was −1.58% lower than activPAL4™-derived sitting time with wide limits of agreement (±28.31%).
For percentage of time spent standing during the workday (Fig. 3), linear regression was not significant and log transformation did not appear to limit the spreading of the data points. Thus, the original Bland-Altman methods were used (Bland and Altman, 1999). Examination of the Bland-Altman plot showed a systematic underestimation of time spent standing with a mean difference of −11.71 (SD = 11.21) and wide 95% limits of agreement (±22.34%).
For percentage of time spent moving including device measured cycling time (Fig. 4), linear regression showed a significant positive association between the difference in the two measures (self-reported minus activPAL4™ derived moving time) and the average of these two measures (B = 1.18, SE = 0.14, p < 0.001). Thus, the mean difference is estimated at −7.04% + 1.18 x average of the two measures. At mean levels of average self-reported/activPAL4™-derived moving time (11.44%), the mean difference indicated self-reported moving time was 6.44% higher than activPAL4™-derived moving time with wide limits of agreement (±15.36%). After excluding device measured cycling time for percentage of time spent moving (Fig. 5), linear regression was still significant (B = 1.21, SE = 0.14, p < 0.001). Thus, the mean difference is estimated at −7.10% + 1.21 x average of the two measures. At mean levels of average self-reported/activPAL4™-derived moving time (11.36%), the mean difference indicated self-reported moving time was 6.60% higher than activPAL4™-derived moving time with wide limits of agreement (±14.93%).
The Bland-Altman plot for minutes of time spent sitting, standing and moving can be found in the supplementary data Fig. 1, Fig. 2, Fig. 3, Fig. 4, Fig. 5.
4 Discussion
4.1 Main finding
The aim of the current study was to assess the criterion validity and absolute measurement of agreement of the OSPAQ against the activPAL4™ device for measuring occupational sitting, standing and moving in a sample of home-based office workers. Findings indicated fair levels of validity (ρ = 0.35–0.43, all p < 0.05) and acceptable agreement (mean difference = 2–12%) when comparing self-reported sedentary and standing with the device at the group level; however, estimates were poor at an individual level, as suggested by wide limits of agreement (±22–30%). For moving time, we observed poor levels of validity (ρ = 0.21–0.22, all p < 0.05) and acceptable agreement (bias = 6–7%). Although the observed biases for moving time were reasonable the 95% limits of agreement were too large (±15%) to have confidence in recommending the self-report measure for use at the individual level. Thus, the modified OSPAQ may be appropriate for use in large-scale studies examining group-level data rather than for studies requiring estimates of an individual's sedentary behaviour profile. Beyond these general conclusions, there are some other observations that warrant commentary.
When looking at the Bland-Altman plots for sitting time (Fig. 1, Fig. 2), it seems the more people sit, the more accurate they are at recalling their sitting time, which was confirmed after performing linear regression (p < 0.001). For example, at around the 80% mark, the individual data points cluster more around the midline compared to the 50% mark, indicating better agreement at a higher sitting percentage. For standing, it is suggested through the data that participants consistently underestimate the time in this behaviour (Fig. 3). This is illustrated by the fact that most of the data points are below the midline. When looking at the moving time plots (Fig. 4, Fig. 5), we see patterns of inconsistent overestimation. That is, the less people move, the more accurate they were at recalling the behaviour, which again was confirmed by a linear regression (p < 0.001). Specifically, the individual data points of these plots are more clustered around the midline when people moved for 10% or less of their workday.
4.2 Relevant literature
With respect to previous work in this field, the first validation study conducted by Chau et al. (2012) used a convenience sample of office-workers (n = 99) and reported a moderate level of agreement between the OSPAQ and device (Actigraph) for estimating time sitting and standing. The authors reported the difference between the two measures as ‘small’, with the 95% limits of agreement for sitting ranging from −141.63 to 185.18 min (326.81-min range). Although these large ranges could be attributed to the fact that the Actigraph accelerometer cannot differentiate between sedentary behaviour (i.e., sitting) and physical inactivity (i.e., standing), the findings were similar to what we observed in our study. Specifically, when examining our Bland-Altman plots in minutes as opposed to percentages, our 95% limits of agreement ranged from −162.69 to 122.26 min (284.95-min range) for the measurement excluding transportation time and −196.89 to 115.95 min (312.84- minute range) when including transportation time. In sum, we observed similar results to the study by Chau et al. (2012) in regard to agreement for time spent sitting; however, the spearman correlations were stronger than the current study for sitting and standing time (ρ = 0.65 and 0.49 respectively) and similar for walking time (ρ = 0.29).
A later study also sought to validate the OSPAQ against a device-based measure using a sample of full-time university office workers (n = 41) (Jancey et al., 2014). Similar to Chau et al. (2012), the correlations reported were stronger than the current study for sitting standing and walking (i.e., moving) (r = 0.58, r = 0.45, r = 0.45 respectively). Contrary to Chau et al. (2012), Jancey et al. (2014) concluded a moderate level of agreement for standing and walking time, but systematic variation for sedentary time. It is important to note that the device used in this study was also an Actigraph, which makes the ability to measure posture impractical, as previously stated. These observations differ from the current study as we found systematic underestimation for standing time, while observing overall poor level of agreement for sitting and moving (i.e., walking). The 95% limits of agreement reported by Jancey et al. (2014) for sitting time were −784.7 to 733.9 min (1518.6-min rage), which is much greater than both Chau et al. (2012) (326.81) and the present findings (312.84). For standing time, they reported 95% limits of agreement of −324.6 to 269.7 min (594.3- minute range) compared to our findings of −180.80 to 49.64 min (230.44-min range). For time spent walking, their 95% limits of agreement were −269.2 to 280.8 min (550-min range), much larger than our findings of moving time with (−56.15 to 105.79 min; range of 161.94) and without (−52.74 to 104.06 min; range of 156.80) cycling time included.
The most recent validation study for the OSPAQ used a sample of both sedentary (n = 65) and non-sedentary (n = 331) workers (Maes et al., 2020). Consistent with the previous studies, the correlations reported were stronger than the current study for sitting, standing and walking (ρ = 0.53, ρ = 0.53, ρ = 0.49 respectively). The authors did not interpret the results of their Bland-Altman plots, but we examined the supplementary data file in order to make comparisons to the current study. For the purpose of relevance, we only further discuss results obtained from the sedentary worker data. The 95% limits of agreement for sitting appeared to range from approximately −45%–50% (95% range), compared to our ~60% range. For time spent standing, 95% limits of agreement appeared to be around −45%–30% (75% range) compared to our ~45%. This is the biggest discrepancy as we found systematic underestimation for standing time whereas their plot appeared randomly scattered. For time spent walking, it appears that the 95% limits of agreement ranged from about −20% to 45% (65% range), larger than our ~30%. While the Maes et al. study is an improvement in terms of the device used to measure sedentary behaviour time compared to the previously mentioned studies, it still poses the risk of misinterpretation as it is not stated how the Axivity AX3 accelerometer compares to the activPAL4™.
The only other validation study that has used an activPAL™ was conducted by van Nassau et al. (2015), using staff from a non-government health agency (n = 42) to compare the two device measures across a number of time points. In terms of correlations, they report similar findings to the present study for sitting (ρ = 0.37) and a weaker correlation for standing (ρ = 0.20). Correlations for walking or moving time were not reported. Unfortunately, this study also did not formally interpret their Bland-Altman plots, so in order to make comparisons we had to approximate numbers from the figure in their paper. Overall, the 95% limits of agreement appeared similar to those of Chau et al. (2012) and the present study. For sitting, values ranged from −120 to 210 min (~330-min). For standing, values ranged from around −75 to 75 min (~150-min).
4.2.1 Implications and future directions
Overall, the above-mentioned findings, along with the present findings, are consistent in demonstrating acceptable validity for measuring sitting and standing with the modified OSPAQ at a group level. However, the large 95% limits of agreement between the modified OSPAQ and activPAL™ or other related devices limits its use at the individual level, particularly with respect to intervention work. Previous studies targeting sedentary behaviour have only resulted in reductions of ~40 min or less (Brakenridge et al., 2018; Chu et al., 2016; Jancey et al., 2016). This reduction in sedentary behaviour unfortunately falls well within the 95% limits of agreement shown in our study and the other research discussed. Put another way, based on the lack of OSPAQ sensitivity (accuracy) evidence at the individual level, intervention studies are likely not powerful enough to show sedentary behaviour change outside the limits of agreement to be statistically significant.
There are a number of reasons as to why the questionnaire may not be performing adequately at the individual level. First, it could be the case that a one-week recall is too long. Future studies should look to validate the questionnaire when occupational behaviours are recalled at the end of each day as opposed to each week. Previous work has shown increased levels of agreement when implementing self-reported questionnaires in this fashion (Moulin et al., 2020). Second, there may be an educational piece necessary when administering this questionnaire. People may misinterpret or misunderstand the questions, unaware of whether to include certain aspects of their workday in the recall (i.e., lunch break). Thus, future studies should investigate as to whether educating or providing a quick tutorial or example beforehand would improve agreement. Third, Bland-Altman advises authors to reproduce their results (Bland and Altman, 1999). In other words, under the same circumstances, when re-administered a month or two later is the agreement level the same? Although van Nassau et al. (2015) touched upon this in their paper, their lack of interpretation of Bland-Altman plots highlights the need for future work to incorporate this kind of paradigm. Lastly, as the OSPAQ only assesses total sitting time, it is important to note that break frequency and duration also are key behaviours related to health outcomes that should be targeted. Therefore, while research continues to assess the OSPAQ and other similar questionnaires, questionnaires such as the revised SitQ-7 (Sui and Prapavessis 2016) that assesses break frequency and duration also need to be assessed and validated in this population.
4.3 Strengths and limitations
The main strength of this study was the use of the activPAL4™ device, which is the gold standard for measuring sedentary time. We were also the first to investigate the validity of the OSPAQ in a sedentary occupation working from home. While working from home was first intended to be temporary due to the COVID-19 pandemic, it seems as though this shift in work environment from the office to home could persist into the foreseeable future and beyond. Because of this shift in how office work is being conducted, it is important to evaluate how this may affect the validity of self-report sedentary questionnaires. A further strength is that the average number of valid days that were required to be included in the analysis was higher than previous studies. Lastly, our sample size and the variability of the sample is a strength, as it makes our study more generalizable given the wide array of sedentary workers that were recruited, compared to previous studies that only recruited office workers from a single company or office space.
This study also had limitations that should be taken into account when interpreting the findings. First, participants were not asked to record their start and end time of each working day while wearing the device. Therefore, participants’ self-reported start and end times might not be exact to their actual workday and thus, could be why the observed findings were not strong at the individual level. Additionally, our inclusion criteria only required a 50% or more work-from-home status; as participants did not record their workdays, we have no way of controlling for, or separating the work in office or at home days collected during the valid days. Lastly, practice effects could have impacted the results, as participants filled out the questionnaire four times prior to the assessment included in this secondary analysis. Thus, they may have improved their ability to recall their behaviour over the 4-week intervention period, leading to better levels of agreement at week 4 than what we might have seen at baseline. Alternatively, without feedback from previously self-reported sedentary behaviour, participants may have not optimally learned how to self-evaluate and thus improve the assessment of the targeted behaviour. We were unable to shed light on this issue as the sequence of measuring activPAL4™ device data and OSAPQ data was not harmonized at baseline (i.e., OSAPQ was assessed before activPAL4™).
5 Conclusion
The modified OSPAQ shows acceptable criterion validity for accurate estimates of overall sitting and standing time but not moving time in the context of at-home office working adults. The 95% limits of agreement for percentage of time spent sitting, standing and moving (i.e., walking) were large (±15–30%) indicating that the OSAPQ may not be appropriate for measuring occupational sedentary and active behaviours at the individual level in this workplace setting. With home-based office work predicted to be a permanent feature for desk-based workers (Anderson et al., 2021) and the cost and burden associated with administering devices to large populations, it is imperative to have a validated self-report measure to allow for accurate assessment of movement patterns during work hours. Although further validation is required (i.e., responsiveness to change), the modified OSPAQ is an easily administered and acceptable self-report method for measuring at-home sitting and standing time at a group level.
Funding
This research was funded by Zilveren Kruis Zorgverzekeringen N.V. and Centene Corporation.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supplementary data
The following is the supplementary data to this article:Multimedia component 1
Multimedia component 1
Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.apergo.2021.103551.
==== Refs
References
Aittasalo M. Livson M. Lusa S. Romo A. Vähä-Ypyä H. Tokola K. Sievänen H. Mänttäri A. Vasankari T. Moving to business - changes in physical activity and sedentary behavior after multilevel intervention in small and medium-size workplaces BMC Publ. Health 17 1 2017 319 10.1186/s12889-017-4229-4
Aminian S. Hinckson E.A. Examining the validity of the ActivPAL monitor in measuring posture and ambulatory movement in children Int. J. Behav. Nutr. Phys. Activ. 9 2012 119 10.1186/1479-5868-9-119
Ammar A. Brach M. Trabelsi K. Chtourou H. Boukhris O. Masmoudi L. Bouaziz B. Bentlage E. How D. Ahmed M. Müller P. Müller N. Aloui A. Hammouda O. Paineiras-Domingos L.L. Braakman-Jansen A. Wrede C. Bastoni S. Pernambuco C.S. Mataruna L. Effects of COVID-19 home confinement on eating behaviour and physical activity: results of the ECLB-COVID19 international online survey Nutrients 12 6 2020 1583 10.3390/nu12061583 32481594
Anderson Janna Rainie Lee Vogels Emily Experts say the ‘new normal’in 2025 will be far more tech-driven, presenting more big challenges 2021 Retrieved from PEW: https://www.pewresearch.org/internet/2021/02/18/experts-say-the-new-normal-in-2025-will-be-far-more-tech-driven-presenting-more-big-challenges Pew Research Center
Blackburn Nicole E. Wilson Jason J. McMullan Ilona Caserotti Paolo Giné-Garriga Maria Wirth Katharina Coll-Planas Laura Blancafort Alias Sergi Roque Marta Deidda Manuela Kunzmann Andrew T. Dallmeier Dhayana Tully Mark A. The effectiveness and complexity of interventions targeting sedentary behaviour across the lifespan: a systematic review and meta-analysis Int. J. Behav. Nutr. Phys. Activ. 17 1 2020 1 18 10.1186/s12966-020-00957-0 REV 53
Bland J.M. Altman D.G. Statistical methods for assessing agreement between two methods of clinical measurement Lancet (N. Am. Ed.) 1 8476 1986 307 310
Bland J.M. Altman D.G. Measuring agreement in method comparison studies Stat. Methods Med. Res. 8 2 1999 135 160 10.1177/096228029900800204 10501650
Brakenridge C.L. Healy G.N. Winkler E. Fjeldsoe B.S. What do workers do to reduce their sitting time? The relationships of strategy use and workplace support with desk-based workers' behavior changes in a workplace-delivered sitting-reduction and activity-promoting intervention J. Occup. Environ. Med. 60 11 2018 1026 1033 10.1097/JOM.0000000000001419 30095591
Chan Y.H. Biostatistics 104: correlational analysis Singap. Med. J. 44 12 2003 614 619
Chau J.Y. Van Der Ploeg H.P. Dunn S. Kurko J. Bauman A.E. Validity of the occupational sitting and physical activity questionnaire Med. Sci. Sports Exerc. 44 1 2012 118 125 10.1249/MSS.0b013e3182251060 21659903
Chu A.H.Y. Ng S.H. Tan C.S. Win A.M. Koh D. M€uller-Riemenschneider F. A systematic review and meta-analysis of workplace intervention strategies to reduce sedentary time in white-collar workers Obes. Rev. 17 5 2016 467 481 10.1111/obr.12388 26990220
Clark B.K. Winkler E. Healy G.N. Gardiner P.G. Dunstan D.W. Owen N. Reeves M.M. Adults' past-day recall of sedentary time: reliability, validity, and responsiveness Med. Sci. Sports Exerc. 45 6 2013 1198 1207 10.1249/MSS.0b013e3182837f57 23274615
Dillon K. Rollo S. Prapavessis H. A combined health action process approach and mHealth intervention to reduce sedentary behaviour in university students - a randomized controlled trial Psychol. Health 1–20 2021 10.1080/08870446.2021.1900574 Advance online publication
Edwardson C.L. Winkler E. Bodicoat D.H. Yates T. Davies M.J. Dunstan D.W. Healy G.N. Considerations when using the activPAL monitor in field-based research with adult populations J. Sport Health Sci. 6 2 2017 162 178 10.1016/j.jshs.2016.02.002 30356601
Ekelund U. Sepp H. Brage S. Becker W. Jakes R. Hennings M. Wareham N.J. Criterion-related validity of the last 7-day, short form of the International Physical Activity Questionnaire in Swedish adults Publ. Health Nutr. 9 2 2006 258 265 10.1079/phn2005840
Fitbit The impact of coronavirus on global activity Available online at https://blog.fitbit.com/covid-19-global-activity/ 2020 (accessed April 9, 2021)
Guttman I. Smith D.E. Investigation of rules for dealing with outliers in small samples from the normal distribution: I: estimation of the mean Technometrics 11 3 1969 527 550
Headley S. Hutchinson J. Wooley S. Dempsey K. Phan K. Spicer G. Janssen X. Laguilles J. Matthews T. Subjective and objective assessment of sedentary behavior among college employees BMC Publ. Health 18 1 2018 768 10.1186/s12889-018-5630-3
Jancey J.M. McGann S. Creagh R. Blackford K.D. Howat P. Tye M. Workplace building design and office-based workers' activity: a study of a natural experiment Aust. N. Z. J. Publ. Health 40 1 2016 78 82 10.1111/1753-6405.12464
Jancey J. Tye M. McGann S. Blackford K. Lee A.H. Application of the occupational sitting and physical activity questionnaire (OSPAQ) to office based workers BMC Publ. Health 14 2014 762 10.1186/1471-2458-14-762
Júdice P.B. Santos D.A. Hamilton M.T. Sardinha L.B. Silva A.M. Validity of GT3X and Actiheart to estimate sedentary time and breaks using ActivPAL as the reference in free-living conditions Gait Posture 41 4 2015 917 922 10.1016/j.gaitpost.2015.03.326 25852024
Katzmarzyk P.T. Powell K.E. Jakicic J.M. Troiano R.P. Piercy K. Tennant B. Sedentary behavior and health: update from the 2018 physical activity guidelines advisory committee Med. Sci. Sports Exerc. 51 6 2019 1227 1241 10.1249/MSS.0000000000001935 31095080
Kozey-Keadle S. Libertine A. Lyden K. Staudenmayer J. Freedson P.S. Validation of wearable monitors for assessing sedentary behavior Med. Sci. Sports Exerc. 43 8 2011 1561 1567 10.1249/MSS.0b013e31820ce174 21233777
Lohana S.T. Yadav T. Effect of chair aerobics on quality of life in sedentary obese individuals Central European J. Sport Sci. Med. 29 1 2020 21 27 10.18276/cej.2020.1-03
Maes I. Ketels M. Van Dyck D. Clays E. The occupational sitting and physical activity questionnaire (OSPAQ): a validation study with accelerometer-assessed measures BMC Publ. Health 20 1 2020 1072 10.1186/s12889-020-09180-9
Meyer J. McDowell C. Lansing J. Brower C. Smith L. Tully M. Herring M. Changes in physical activity and sedentary behavior in response to COVID-19 and their associations with mental health in 3052 US adults Int. J. Environ. Res. Publ. Health 17 18 2020 6469 10.3390/ijerph17186469
Moulin M.S. Lee C.J. Tucker P. Prapavessis H. Irwin J.D. Weekly recall of sedentary time: validity of 2 weekly self‐reported measures in undergraduate students Transl. Sports Med. 3 2 2020 127 133
Nelson-Wong E. Gallagher K. Johnson E. Antonioli C. Ferguson A. Harris S. Johnson H. Miller J.B. Increasing standing tolerance in office workers with standing-induced back pain Ergonomics 63 7 2020 804 817 10.1080/00140139.2020.1761034 32330093
Owen N. Healy G.N. Matthews C.E. Dunstan D.W. Too much sitting: the population health science of sedentary behavior Exerc. Sport Sci. Rev. 38 3 2010 105 113 10.1097/JES.0b013e3181e373a2 20577058
Pew Research Center How the coronavirus outbreak has – and hasn't – changed the way Americans work Available online at https://www.pewresearch.org/social-trends/2020/12/09/how-the-coronavirus-outbreak-has-and-hasnt-changed-the-way-americans-work/ December 2020 (accessed April 9, 2021)
Rosenberg D.E. Norman G.J. Wagner N. Patrick K. Calfas K.J. Sallis J.F. Reliability and validity of the sedentary behavior questionnaire (SBQ) for adults J. Phys. Activ. Health 7 6 2010 697 705 10.1123/jpah.7.6.697
Rollo S. Prapavessis H. A combined health action process approach and mHealth intervention to increase non-sedentary behaviours in office-working adults-A randomised controlled trial Appl. Psychol. Health Well-Being 12 3 2020 660 686 10.1111/aphw.12201 32342662
Ross R. Chaput J.P. Giangregorio L.M. Janssen I. Saunders T.J. Kho M.E. Poitras V.J. Tomasone J.R. El-Kotob R. McLaughlin E.C. Duggan M. Carrier J. Carson V. Chastin S.F. Latimer-Cheung A.E. Chulak-Bozzer T. Faulkner G. Flood S.M. Gazendam M.K. Healy G.N. Canadian 24-Hour Movement Guidelines for Adults aged 18-64 years and Adults aged 65 years or older: an integration of physical activity, sedentary behaviour, and sleep Appl. Physiol., Nutr. Metabol. Physiol. Appl., Nutr. Metabol. 45 Suppl. 2 2020 S57 S102 10.1139/apnm-2020-0467 10
Sui W. Prapavessis H. Testing the face validity and reliability of a modified SIT-Q 7 day recall questionnaire measuring sedentary time and break frequency & duration J. Sport Exerc. Psychol. 38 2016 S262
Thorp A.A. Healy G.N. Winkler E. Clark B.K. Gardiner P.A. Owen N. Dunstan D.W. Prolonged sedentary time and physical activity in workplace and non-work contexts: a cross-sectional study of office, customer service and call centre employees Int. J. Behav. Nutr. Phys. Activ. 9 2012 128 10.1186/1479-5868-9-128
Tremblay M.S. Aubert S. Barnes J.D. Saunders T.J. Carson V. Latimer-Cheung A.E. Chastin S. Altenburg T.M. Chinapaw M. SBRN Terminology Consensus Project Participants Sedentary behavior research network (SBRN) - terminology consensus project process and outcome Int. J. Behav. Nutr. Phys. Activ. 14 1 2017 75 10.1186/s12966-017-0525-8
Tison G.H. Avram R. Kuhar P. Abreau S. Marcus G.M. Pletcher M.J. Olgin J.E. Worldwide effect of COVID-19 on physical activity: a descriptive study Ann. Intern. Med. 173 9 2020 767 770 10.7326/M20-2665 32598162
Urda J.L. Larouere B. Verba S.D. Lynn J.S. Comparison of subjective and objective measures of office workers' sedentary time Prevent. Med. Rep. 8 2017 163 168 10.1016/j.pmedr.2017.10.004
van Nassau F. Chau J.Y. Lakerveld J. Bauman A.E. van der Ploeg H.P. Validity and responsiveness of four measures of occupational sitting and standing Int. J. Behav. Nutr. Phys. Activ. 12 2015 144 10.1186/s12966-015-0306-1
Zafiropoulos B. Alison J.A. Heard R. Physical activity levels of allied health professionals working in a large Australian metropolitan health district - an observational study J. Multidiscip. Healthc. 12 2019 51 62 10.2147/JMDH.S189513 30655673
| 34403840 | PMC9746924 | NO-CC CODE | 2022-12-15 23:21:58 | no | Appl Ergon. 2021 Nov 12; 97:103551 | utf-8 | Appl Ergon | 2,021 | 10.1016/j.apergo.2021.103551 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00036-7
10.1016/j.biocon.2021.108984
108984
Article
Review: COVID-19 highlights the importance of camera traps for wildlife conservation research and management
Blount J. David a1
Chynoweth Mark W. b1
Green Austin M. a⁎1
Şekercioğlu Çağan H. ac
a School of Biological Sciences, University of Utah, 257 South 1400 East, Salt Lake City, UT 84112-0840, USA
b Department of Wildland Resources, Utah State University, Uintah Basin, 320 North Aggie Blvd., Vernal, UT 84078, USA
c College of Sciences, Koç University, Rumelifeneri, İstanbul, Sarıyer, Turkey
⁎ Corresponding author.
1 These authors contributed equally to this work.
2 2 2021
4 2021
2 2 2021
256 108984108984
21 9 2020
11 1 2021
16 1 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
COVID-19 has altered many aspects of everyday life. For the scientific community, the pandemic has called upon investigators to continue work in novel ways, curtailing field and lab research. However, this unprecedented situation also offers an opportunity for researchers to optimize and further develop available field methods. Camera traps are one example of a tool used in science to answer questions about wildlife ecology, conservation, and management. Camera traps have long battery lives, lasting more than a year in certain cases, and photo storage capacity, with some models capable of wirelessly transmitting images from the field. This allows researchers to deploy cameras without having to check them for up to a year or more, making them an ideal field research tool during restrictions on in-person research activities such as COVID-19 lockdowns. As technological advances allow cameras to collect increasingly greater numbers of photos and videos, the analysis techniques for large amounts of data are evolving. Here, we describe the most common research questions suitable for camera trap studies and their importance for biodiversity conservation. As COVID-19 continues to affect how people interact with the natural environment, we discuss novel questions for which camera traps can provide insights on. We conclude by summarizing the results of a systematic review of camera trap studies, providing data on target taxa, geographic distribution, publication rate, and publication venues to help researchers planning to use camera traps in response to the current changes in human activity.
Keywords
COVID-19
Camera traps
Conservation biology
Data analysis
Biodiversity monitoring
Wildlife ecology
Remote sensing
Tropical biology
==== Body
pmc1 Introduction
The spread of COVID-19 across the globe has led to lockdowns of cities, towns, villages, and protected areas, resulting in drastic alterations to human activity. Most notably, these restrictions have led to extensive decreases in human mobility, a period some are calling the “anthropause” (Rutz et al., 2020). However, this change in human activity across the globe has not been consistent, and there exists wide variation in how certain locales have responded to the pandemic (Bennett et al., 2020; Rutz et al., 2020; Zellmer et al., 2020). Specifically, the extent to which local governments enforce restrictions varies by country or municipality, resulting in a gradient of human activity and mobility changes (Kleinschroth and Kowarik, 2020; Rutz et al., 2020). This combination of overall decreases in human mobility and natural variation in human traffic across different geographies, specifically in cities and natural areas, offers a unique opportunity for scientists to study and understand human-wildlife interactions on an unparalleled scale (Bates et al., 2020; Corlett et al., 2020; Saraswat and Saraswat, 2020).
However, scientific research is also susceptible to the effects of lockdowns. Researchers have been forced to sideline, cancel, or postpone their projects as a result of the COVID-19 restrictions (Pennisi, 2020). For many, long-term monitoring projects have been put on hold, field research grants have been postponed, and travel has been restricted. With the number of cases of COVID-19 continuing to grow worldwide and extending these restrictions indefinitely, the need for projects that allow monitoring without extensive time in the field, human involvement, and regular maintenance has never been more urgent. Camera traps provide a useful method for conservation ecologists to both continue important research and to investigate the novel wildlife conservation and ecology questions posed by this current pandemic.
Camera traps, or trail cameras, are motion- and heat-activated remote sensing devices that come in a wide variety of settings and are equipped with myriad triggering mechanisms, photo/video capabilities, sensor levels, flash types, housing, and other specifications (Rovero et al., 2013). The majority of modern cameras use a passive infrared sensor (PIR) that can detect the differences in heat and motion (Rovero et al., 2013). Most camera traps are relatively low-cost tools for research and management, with negligible impacts on target species or the environment (O'Connell et al., 2011; Burton et al., 2015; Steenweg et al., 2017; Caravaggi et al., 2017; Wearn and Glover-Kapfer, 2019).
Over the past couple of decades, the use of camera traps in conservation biology, ecology, and biodiversity assessments has grown significantly (McCallum, 2013; Burton et al., 2015; Steenweg et al., 2017; Kays et al., 2020). However, without appropriate research design, advanced planning, and power analyses, conservation biologists often collect high volumes of data that they are unable to use to either inform the management of vulnerable species and systems or to answer the conservation questions that initiated their research (Hebblewhite and Haydon, 2010). Since camera traps are relatively easy to setup and maintain in the field, researchers may inherit a false sense of security with their use as scientific research tools and neglect important aspects of the study design process. For efficient use of time and resources, researchers must distinguish the reasons for camera trapping research programs before deployment and choose appropriate study designs and analyses for conservation monitoring programs (Jones et al., 2013; Wearn and Glover-Kapfer, 2019; Kays et al., 2020).
In this review, we summarize the most common camera trap objectives, specifically: 1) documenting species presence/richness, 2) evaluating relative abundance, 3) estimating density, 4) estimating occupancy, and 5) quantifying activity patterns. We discuss the appropriate use of camera traps for each of these objectives and illustrate how such studies are designed and conducted, as well as how the resultant data are analyzed. We then provide, within each objective, key wildlife conservation and ecology questions associated with the current “anthropause” (Rutz et al., 2020) that camera traps are well-suited to tackling. We conclude by providing results from a systematic review of camera trap studies, providing information on publication rate, popular publication venues, target taxa, and geographic distribution of studies worldwide. We also provided a summary on current camera trap initiatives both actively gathering data and looking for additional collaborators. Our goal is to inform conservation biologists of the advantages and disadvantages of camera traps; assist in the appropriate choice of method and study design; discuss how camera traps can be critical for studying wildlife behavior, activity, and dispersal in the time of COVID-19; and review the contribution of camera trapping studies to conservation biology.
1.1 Biological objectives of camera trap studies
Camera traps have been used to study many species and objectives in animal ecology (O'Connell et al., 2011; Burton et al., 2015). Here we focus on the most common study objectives: presence, relative abundance, density, occupancy, and activity. Early in the use of camera traps for research purposes, few studies went beyond baseline assessments of population size and structure (Linkie et al., 2010). However, conservationists are increasingly using camera traps to test hypotheses and address a range of questions including human impacts on wildlife (Main and Richardson, 2002; Magle et al., 2012; Gallo et al., 2017; Parsons et al., 2018; Fidino et al.,2020; Parsons et al., 2019), biodiversity monitoring over space and time (Waldon et al., 2011; Gilbert et al., 2020), reproductive ecology (Farhadinia et al., 2009), interspecific interactions (Rota et al., 2016), animal behavior (Caravaggi et al., 2017; Rowcliffe, 2017; Caravaggi et al., 2020), and nest predation (Bayne and Hobson, 1997; Beck and Terborgh, 2002; Vilardell et al., 2012).
1.1.1 Documenting species presence
Documenting species presence or absence is crucial to the discovery (Rovero et al., 2008), rediscovery (Yamada et al., 2010), and confirmation (Lhota et al., 2012) of range expansions of both native (Chynoweth et al., 2015) and invasive (Naderi et al., 2020) species. Studies of species presence are pertinent to monitoring elusive and endangered species, and photos of these species are also invaluable for education and public outreach. Effects of human activity on species and ecosystem dynamics in remote and rural areas (Muhly et al., 2011; Gallo et al., 2017; Parsons et al., 2018) and conservation threats, such as the impact of poachers on wildlife populations, can also be monitored (Jenks et al., 2012).
Photographic evidence often renders species presence indisputable. However, photos of animals can be misinterpreted or indecipherable, leading to spurious claims of new species (Meijaard et al., 2006). These claims, along with apparent range expansions, rediscoveries, and retractions, may be a result of the lack of baseline information and insufficient density of camera traps (Dobson and Nowak, 2010). Researchers must acknowledge that non-detection is not the same as absence, as individual species' detection probabilities, which are almost always <1 (i.e., not all species in the area will be perfectly detected), may result in a species that is actually present within a study area going undetected during sampling (Tilson et al., 2004; MacKenzie et al., 2017).
Though several established methods effectively document species presence, comparison studies suggest that camera traps have higher probabilities than hair tunnels (O'Connell et al., 2006; Paull et al., 2012), cubby boxes (O'Connell et al., 2006), patrol observations (Burton, 2012), and line-transect surveys (Trolle et al., 2008) for detecting smaller, solitary, and nocturnal species (Wearn and Glover-Kapfer, 2019). Studies incorporating track plates used to document the presence of animals by recording footprints have shown that species richness and recording rates correlate with camera trapping results (Espartosa et al., 2011). In some cases, track plates were more effective (Hackett et al., 2007) and detected more individuals (Rosas-Rosas and Bender, 2012). Yet, with technological advances, remote cameras require less maintenance and may be more cost effective than track plates for studies >1 year (Ford et al., 2009). An alternative method for detecting presence is genotyping by scat collection, which has produced consistent (Galaverni et al., 2011) and sometimes better (Harrison et al., 2002) detection probabilities than camera traps. Finally, under proper weather conditions, snow tracking surveys have the highest probability of detection for species active in the winter (Gompper et al., 2006).
Study design for documenting species presence does not necessarily need to be systematic and can be targeted at specific sites or use species-specific baits to maximize detection probability. Furthermore, there is no minimum for the number of cameras a researcher should deploy for this type of analysis, but having more cameras increases detection probability and allows one to investigate across multiple different habitats, human influence levels, and other variables. Increasing the number of camera traps also decreases the amount of time the study will need to stay active in order to sample the species of interest. With species presence data, all that is needed is the photographs from the camera traps and the GPS locations of the cameras themselves. This is the simplest form of camera trap data used for research purposes.
1.1.2 Relative Abundance Index (RAI)
Camera trap data can be used to generate a Relative Abundance Index (RAI), which is typically calculated by summing detections (usually in the form of the number of “independent” photographs, where independence is denoted as a set amount of time that needs to pass before a photograph is deemed a new detection) for each species, dividing by the total number of active camera days, and multiplying this fraction by 100 (O'Brien, 2011). This approach is attractive to conservationists because of its simplicity. However, it has been criticized for being an inappropriate and unreliable method (Sollmann et al., 2013a). This index is known to produce biased estimates based on heterogeneous detection probabilities (Jennelle et al., 2002; Sollmann et al., 2013a), and as a result, its use needs to be justified as the only reasonable alternative to other methods (O'Brien, 2011).
The application of the RAI relies on the assumption that the index is directly related to true species abundance (O'Brien et al., 2003). The majority of RAI studies aim to estimate abundance at a single point in time at a specific site (e.g., protected area), but this index has also been used as an abundance proxy to study a variety of ecological processes including habitat use (Bowkett et al., 2008), human impacts on wildlife (Kinnaird and O'Brien, 2012), temporal population dynamics (Jenks et al., 2011), and activity patterns (Ramesh et al., 2012).
As discussed above, the main issue with RAI is that of detectability. Detectability varies among and within species and is considered a major source of bias (Larrucea et al., 2007; MacKenzie et al., 2017). Variations in detection probability due to species differences in behavior, life history characteristics, natural rarity, home range size, and temporal activity patterns have all been shown to bias RAI estimates (Sollmann et al., 2013a). With independently derived abundance estimates in a double sampling design, RAIs can be calibrated for a particular system or study area (O'Brien et al., 2003), but this also requires continuous re-calibration and results do not translate outside that area.
Study design for RAI surveys should aim to limit the effect of variation in detection probabilities to account for the main deficiency of this approach (Sollmann et al., 2013a). Once the study area is determined, cameras should be placed at distances smaller than the home range diameter of the target species to prevent false negatives. The number of cameras necessary depends on study area extent and target species, but should cover the area uniformly and at a great enough density to maximize detection probability. Furthermore, feature-focused designs (study designs focusing on trails, roads, streams, and other habitat features to increase the rate of detection) should proceed with extreme caution because of the inability to account for the differences in detectability across these features. Instead, studies using RAIs should consider setting up cameras randomly across the study area, where individual camera sites are most representative of the surrounding habitat. To calculate RAI or trap rate, only species presence data and trapping effort (the number of active camera days) are needed. If covariates of habitat structure, human influence, and other site variables are to be included in analysis, the GPS coordinates of each camera will also be needed. Methods described in the previous section can be used in lieu of camera traps to calculate RAIs (Jhala et al., 2011), but these methods may require more extensive fieldwork and physical trapping effort than camera traps.
1.1.3 Density estimates and individual recognition
Density estimates are a common objective of camera trapping studies and may be the most sought-after population parameter (O'Connell et al., 2011; Burton et al., 2015). Density estimates allow for easy comparisons between sites and years or extrapolation to larger areas (Bellan et al., 2013). If individuals can be identified in a population, capture-recapture methods can produce reliable estimates for a study area. Three main types of capture-recapture population models are used to estimate abundance: (i) closed—no birth, death, immigration or emigration (O'Brien, 2011), (ii) open—losses and recruitments are allowed (Gutiérrez-González et al., 2012), and (iii) spatially explicit—including spatial characteristics such as home range and individual mobility (Gardner et al., 2010; Royle et al., 2014; Royle et al., 2018; Green et al., 2020).
The sampling area for density estimate studies is typically set up in a grid-like system, with the outermost trap locations representing the study area boundary. To estimate the effective sampling area, the simplest approach is to draw a concave polygon by connecting the outermost trap locations in a geographic information system. However, this fails to include ingress from outside animals and outward movement from animals inside the polygon. A more appropriate approach is to estimate a buffer around this polygon. Though no consensus exists on calculating this area, a buffer of mean maximum distance moved (MMDM) of the target species is common. MMDM can be estimated from camera trap data, spatially explicit capture-recapture (SECR) models, or estimates based on auxiliary telemetry data. The MMDM method may inflate density estimates (Soisalo and Cavalcanti, 2006), which has led to the arbitrary but frequently used ½MMDM approach. Neither has a theoretical basis (Obbard et al., 2010). Auxiliary telemetry data, typically available from other studies on target species, is most effective at estimating MMDM (Dillon and Kelly, 2008; Núñez-Pérez, 2011).
Early in camera trapping science, two landmark papers estimated density of tigers by identifying individuals with unique pelage characteristics (Karanth, 1995; Karanth and Nichols, 1998). This approach has been extended to a variety of species to identify individuals based on spots (Jackson et al., 2006), stripes (Singh et al., 2010), muzzle markings (Mazzolli, 2010) and other forms of unique pelage (Caruso et al., 2012). Additionally, capture-recapture methods are possible if animals are captured and tagged with artificial markings such as ear tags or GPS collars (Jordan et al., 2011; Weckel and Rockwell, 2013). However, individual identification can be subject to researcher bias (Oliveira-Santos et al., 2010), and efforts have been made to incorporate a more rigorous Bayesian approach to individual identification (Stafford and Lloyd, 2011). Bilateral photo identification records from single trap stations can introduce inconsistencies due to bilateral asymmetry in coat patterns, but modeling approaches to combine left- and right-sided photos are being developed to address this (McClintock et al., 2013). Currently, a common and simple solution is to modify study design to include two cameras at each station (Negrões et al., 2012). Furthermore, study design issues related to sampling area, camera spacing, and detection probability may introduce significant biases (Dillon and Kelly, 2007; Foster and Harmsen, 2012), and recent literature on study design should be consulted before project implementation (Royle et al., 2018; Efford and Boulanger, 2019; Green et al., 2020).
Several reviews have focused on analysis techniques (Sharma et al., 2010; Obbard et al., 2010; Foster and Harmsen, 2012; Royle et al., 2018; Green et al., 2020), improving current capture-recapture analysis (Royle et al., 2009), and developing new techniques including Bayesian inferences for arbitrary sample sizes (Gardner et al., 2010) and maximum likelihood approaches (O'Brien and Kinnaird, 2011; Efford et al., 2019). In particular, SECR models use a hierarchical approach to model both detection probability and home range location and have produced more accurate density measurements in most studies (Kalle et al., 2011; Blanc et al., 2013; Royle et al., 2014; Royle et al., 2018; Green et al., 2020). Currently, these advances in density estimators have been used for relatively few species that can be individually identified by coat patterns (Green et al., 2020). Techniques to estimate density without individual identification have been proposed (Carbone et al., 2001; Rowcliffe et al., 2008; Manzo et al., 2012; Chandler and Royle, 2013), but have not been without criticisms (Foster and Harmsen, 2012). Finally, spatial mark-resight models require only partially marked populations, extending the number of species whose density can be estimated through camera trapping to species only partially individually identifiable and have garnered much attention in recent years (Sollmann et al., 2013b; Jimenez et al., 2017; Whittington et al., 2018).
1.1.4 Occupancy analysis
Reliable density estimates require rigorous study design and large quantities of resources. An alternative approach is occupancy modeling, an established method to model the probability of a site being occupied by a species (MacKenzie et al., 2002; O'Connell and Bailey, 2011; MacKenzie et al., 2017). Occupancy uses presence/absence (or, more appropriately, detection/non-detection) data from independent replicate surveys under the assumption that the population is closed during the survey period. Results provide estimates on the proportion of area occupied by a species. Conversely, if temporal or spatial closure is violated, the parameter estimated becomes the proportion of study sites used by a species. In addition, surveys can be conducted over time and space to elucidate how habitat covariates impact species occurrence. A major advantage of occupancy modeling is that it explicitly estimates and models detection probability (Jones, 2011; MacKenzie et al., 2017). Generally, there is a positive relationship between occupancy and abundance, and occupancy has been used as a proxy for abundance in studies of niche partitioning (Di Bitetti et al., 2010), impact of human disturbance (Mohamed et al., 2013), and predator-prey dynamics (Silva-Rodríguez and Sieving, 2012). However, when spatial and/or temporal closure is violated, occupancy should not be used as a proxy for abundance, as species with varying home ranges and densities will produce biases in multi-species estimates. Typically, occupancy requires smaller sample sizes and is therefore typically less expensive and time-consuming than density estimation (MacKenzie et al., 2017). A rich literature exists on modeling species occupancy, and a wide variety of presence/absence (detection/non-detection) data (Vojta, 2005; MacKenzie et al., 2006) have been used in a number of camera trap studies (Erb et al., 2012; Gopalaswamy et al., 2012a; Schuette et al., 2013; Burton et al., 2015; MacKenzie et al., 2017).
Study design for occupancy models requires a camera array that provides a representative sample of the study area, or a sample design where habitat heterogeneity is incorporated into analysis covariates. Occupancy models allow for stations to be shifted between units, given that they are present at each location long enough to collect sufficient data (O'Connell and Bailey, 2011; MacKenzie et al., 2017). Cameras should be spaced at a distance greater than the minimum of the diameter of the target species' home range, unless the goal of the study is to estimate habitat use instead of the proportion of area occupied. Information on environmental conditions for each site also need to be collected if researchers choose to include habitat covariates in their occupancy model. Finally, prior work has focused on camera trap design for occupancy studies, and we direct readers to these for more detailed information on study design criteria such as the number of sites, spatial replicates, and others design parameters (MacKenzie and Royle, 2005; Guillera-Arroita et al., 2010; Whittington et al., 2018; Kays et al., 2020).
1.1.5 Activity analysis
Camera trap data can be used to elucidate diel and seasonal activity patterns and understand interspecific competition and niche partitioning (Linkie and Ridout, 2011; Rota et al., 2016; Parsons et al., 2016; Frey et al., 2017). Camera traps allow researchers to record multiple species over long periods with minimal disturbance (Ramesh and Kalle, 2013). Much work has been done with sympatric species, such as felids and canids (Foster et al., 2013; Athreya et al., 2013) and on observations of predator-prey dynamics (Weckel et al., 2006; Ford and Clevenger, 2010; Linkie and Ridout, 2011). Especially important for conservation, human impact on animal activity, including human-wildlife coexistence, has also been investigated (Carter et al., 2012; Wang et al., 2015; Gaynor et al., 2018). However, co-occurrence does not necessarily equal coexistence (Harihar et al., 2013), and camera trapping data may fail to capture important factors that determine species activity and distribution. Specifically, camera traps can tell researchers where an individual animal is or has been, but it cannot tell them how that individual got there.
Though camera traps enable researchers to gain new insights into the activity patterns of wild animals, pairing cameras with other approaches can produce more reliable data. Most telemetry collars are now equipped with an activity sensor that uses triaxial accelerometers that record movement of an animal's neck at very high temporal resolutions. Combined with movement data from GPS location and camera trap data, detailed animal activity can be observed. More recently, National Geographic Crittercams© and BBC's animal cameras have been deployed on large mammal species to document previously unknown activity and behavior (Şekercioğlu, 2013; PBS, 2018).
Study design for activity surveys focuses on documenting temporal and seasonal presence data and therefore should strive to maximize detection probabilities for target species. Camera placement on game trails and other areas frequented by animals may increase captures, but may also produce bias in activity estimates if species concentrate on certain features during specific times of the day. To avoid biases associated with detection, study design must aim to have equal detection probabilities between species or account for these presumed differences with both species-specific and site-specific covariates. Recent reviews on activity and behavior analysis should be consulted for further information on study design recommendations and potential sources of bias when designing camera trap studies for activity analysis (Caravaggi et al., 2017; Frey et al., 2017).
2 Methods
2.1 Literature review
We conducted a systematic search of peer-reviewed literature published between 1975 and 2020 using search terms related to camera traps and animal ecology (Table 1 ). Every combination of these terms was used in a search of the ISI Web of Knowledge Complete Collection database search engine. Each article was reviewed to confirm that it discussed camera traps. The database included: publication year, article title, journal name, target taxa, study country, paper type, and primary objective. Finally, to assess the recent extent to which camera traps have been used for urban ecology research, we conducted a keyword search of abstracts from papers published after 1 January 2018 referencing either urban or suburban study sites.Table 1 Camera trap and animal ecology keywords used in literature search of the ISI Web of Knowledge.
Table 1Camera trap terms Animal ecology terms
“Camera Trap*” Wildlife
“Game Camera*” Birds
“Trail Camera*” Mammals
“Remote Photography” Reptiles
Amphibians
3 Results
Our review of camera trapping studies from 1975 to 2020 (Supplementary Table 1) reveals that publications have increased at a rapid rate, with over 500 articles published in 2020 alone (Fig. 1). In the ISI Web of Knowledge Complete Collection, there were no camera trapping papers published prior to 1993. Our literature search resulted in 3326 papers published across 433 journals, with 258 journals having more than two articles and 75 journals having >10 articles (see top 10 journals in Table 2). Research was conducted in 124 countries (Fig. 2). Target taxa included mammals, birds, herpetofauna, and multiple other taxa (Table 3). The majority of articles covered studies of mammals (85.9%), most of which belonged to the order Carnivora (45.9%), the majority of which were felids (52.4%). A considerable number of studies focused on multiple species, with 26.8% of mammal studies including species from multiple taxa (22.2% of total dataset). This is an underestimate, however, given the presence of multi-species studies focusing entirely within the same order (e.g., 17.7% of carnivore studies included multiple families, Supplementary Table 1). The primary objective of most studies was investigating different methodology (11.9%). Papers also had primary objectives of estimating density (11.7%), species presence (9.5%), relative abundance (9.1%), occupancy (8.8%), activity (7.6%), and species richness (5.1%). Most recent studies were conducted in protected or non-urban areas, as a keyword search of all abstracts for papers published since 2018 (n = 1375) shows that only 1.1% (n = 15 papers) targeted urban or suburban environments.
4 Discussion
4.1 Importance of camera trapping for conservation research and community engagement
In the past decade, much work has been done to improve the scientific rigor of camera trapping studies. Camera trapping science is evolving rapidly, and scientists and practitioners emphasize that carefully executed study designs can yield informative parameters, as described in the sections above. However, the potential of simple, inexpensive camera deployments to revolutionize conservation projects with budgetary restrictions should also be recognized. It has been suggested that there are two categories of camera trap studies: (1) science, understanding how an ecosystem works, and (2) management, moving an ecosystem from less to more desirable states (Nichols et al., 2011). We assert that conservation outreach, community science, and environmental education constitute a third category (Adler et al., 2020). While other experts have suggested that photos are the means to an end goal of informing the larger process of science and management (Nichols et al., 2011), we also affirm the value of photographic records of elusive species. For example, the authors' existing conservation project in eastern Turkey initially deployed four camera traps at a study site in 2006. The documentation of an unexpectedly high relative abundance of large carnivores and the scarcity of their prey species has led to national and international support for a large-scale monitoring project for mammals and catalyzed the government to designate Turkey's first wildlife corridor. A conservation success in a country experiencing a major biodiversity crisis (Şekercioğlu et al., 2011), the project has since evolved into a more rigorous study with a network of 40 camera traps being systematically deployed over a multi-year period.Fig. 1 Camera trapping studies by year published from 1990 to 2020 based on a systematic search of key terms in ISI Web of Knowledge.
Fig. 1
Fig. 2 The global distribution of camera trapping studies published from 1990 to 2020 based on a systematic search of key terms in ISI Web of Knowledge.
Fig. 2
Table 2 Number of camera trapping articles published in the top ten journals from 1990 to 2020 based on a systematic search of key terms in ISI Web of Knowledge.
Table 2Journal Number of articles
PLoS One 157
Oryx 136
Biological Conservation 115
Wildlife Research 86
Journal of Mammalogy 84
European Journal of Wildlife Research 78
Mammalia 73
Wildlife Society Bulletin 68
Mammalian Biology 67
Journal of Wildlife Management 60
Table 3 Proportion of target taxa in camera trapping studies published from 1990 to 2020 based on a systematic search of key terms in ISI Web of Knowledge.
Table 3Taxa Percent of total articlesa
Mammal 82.9
Bird 5.9
Multiple taxa 5.7
Herpetofauna 1.7
Insect 0.2
Within mammal order diversity Percent of mammal category
Carnivore 45.9
Multiple Orders 26.8
Ungulate 13.2
Rodent 5.5
Primate 4.3
Marsupial 1.8
Other 2.3
Within carnivore family diversity Percent of carnivore category
Felidae 52.4
Multiple Families 17.7
Canidae 11.8
Ursidae 7.1
Mustelidae 6.8
Hyaenidae 1.3
Other 2.9
a Percent of total articles does not add up to 100% (96.4%) because review articles were present in the database.
Camera trap photos and videos are also effective public outreach tools that raise awareness about important study sites, vulnerable species, and conservation priorities of local and global organizations or governmental agencies. A single photo published via social and traditional media can deliver important conservation messages to millions of people. The authors share camera trap photos and project updates on Facebook and Instagram, where a single photo can be viewed by over 25,000 individuals in a five-day period. Public outreach opportunities extend to citizen science approaches (Adler et al., 2020) in which members of the public deploy cameras or identify species in camera trap photos. Several large-scale camera trapping efforts, such as the Tropical Ecology Assessment and Monitoring Network (TEAM; www.teamnetwork.org; also see Fegraus et al., 2011) Smithsonian Wild (see http://siwild.si.edu), eMammal (https://emammal.si.edu), Wildlife Insights (https://www.wildlifeinsights.org/home), EUROMAMMALS (https://euromammals.org), and the Urban Wildlife Information Network (https://urbanwildlifeinfo.org) have already made progress through citizen science and multi-city collaboration efforts.
4.1.1 Guiding questions during COVID-19
The effects of COVID-19 lockdowns will most likely have a marked impact on where animals go and what habitats they access. Simple comparison studies on species presence before, during, and after a marked change in human activity can help scientists understand the effects of human influence on wildlife distribution. Camera traps can help identify which species' ranges extend, contract, or stay the same in the face of changing human influence and global decreases in human travel. Anecdotal evidence of wildlife reclaiming cities during lockdown may be a sign of species range extensions (Sahagun, 2020). However, as Zellmer et al. (2020) point out, this may also be a result of people using time previously spent commuting or gathering socially to observe the wildlife that have always been present in their cities (Garrard et al., 2008). Furthermore, some locales are actually seeing great increases in recreational traffic on public lands and urban greenspaces, as people seek out leisure activities that adhere to the CDC's current social distancing guidelines (Samuelsson et al., 2020; Slater et al., 2020; Razani et al., 2020; Rice et al., 2020). Camera traps have already been used to observe urban species often overlooked by the public (Magle et al., 2019), and future camera trap studies could contribute to the growing literature of new species accounts in urban areas (Feinberg et al., 2014; Hartop et al., 2015). Therefore, empirical investigations of species either reclaiming urban habitat during COVID-19 shutdowns or retreating from habitat in areas experiencing increases in recreational traffic are needed before anecdotal observations can be substantiated. Furthermore, these changes in species distribution, colonization, and retreat can be modeled across geographies, identifying climatic, socioeconomic, and environmental factors that may be influencing these changes. These baseline investigations will provide scientists with a better mechanistic understanding of how human traffic effects the wildlife use of habitat within a wildland-urban interface (Martinuzzi et al., 2015).
One challenge for researchers will be to document and quantify the variety of impacts COVID-19 restrictions have had on human activity and behavior in order to understand how wildlife communities respond. The results from our review demonstrate the geographic disparity that exists regarding where camera trap studies occur (Fig. 2). This is unsurprising, as many research fields follow similar patterns. For researchers in countries with high numbers of camera trap studies (e.g., USA), opportunities already exist to collaborate or conduct meta-analyses (see below), with datasets across a variety of pandemic-caused lockdown scenarios. This will make it easier to detect the effect of this anthropause on wildlife. Our review also highlights regions of the world where camera trap publications are lacking or lagging, and opportunities may exist to encourage camera trap use or peer-reviewed publication of results in these areas.
Documenting changes in species presence across a landscape can only answer questions on how species distributions may change in response to changes in human mobility and traffic. However, it cannot discern the extent to which species use a particular habitat. Using camera traps to calculate species-specific, community-specific, and study-site specific RAIs as a relative measure of habitat preference, scientists can go beyond investigations of species presence, range expansions, and range contractions and examine the scope of this change. Are species, communities, and populations colonizing new areas in relatively large numbers, or are these colonization events restricted to vagrant individuals or small groups? RAI can also be used for measuring the differences in species temporal activity and species-species interactions. Specifically, investigating questions concerning how individual species change their activity and behavior in response to the COVID-19 pandemic can help scientists understand the extent that species have adapted to living in urban and semi-urban areas, as well as identify novel changes in behavior in response to an unprecedented change in human influence. As mentioned above, other methods exist to calculate RAI, but none are as well-suited during current field research restrictions as are camera traps. However, it is important to note that using RAI to investigate changes in habitat use, relative abundance, and behavior should be done with caution, and investigators should pay close attention to the concerns noted above.
Density is a state variable of conservation concern, and is often considered the pinnacle of biodiversity assessment (Williams et al., 2002; Royle et al., 2014). With density estimates, scientists and wildlife managers have the most accurate measure of population assessment available. During COVID-19, quantifying potential changes in density is one of the best ways to assess how wildlife react to human influence. Long-term studies continuing their work through and after the pandemic will be in a position to measure how specific species react to changes in human activity in response to the pandemic, including increases in hiking and recreational traffic, vehicle traffic, CO2 emissions, air pollution, anthropogenic food resources, and other impacts specific to the effects of lockdowns. Camera traps are valuable tools in measuring density of both fully marked and partially marked species and do not require direct capture, substantially lowering the amount of work required in the field. For this reason, it is important for researchers interested in collecting density estimates to consider study design and study species carefully, as only a subset of the wildlife community will be available for sampling, and density estimation requires large amounts of data from potentially rare and elusive species (Royle et al., 2018; Green et al., 2020).
Occupancy modeling is one of the fastest growing analysis methodologies in camera trapping. Like much of the methodology described above, occupancy modeling is particularly powerful when researchers are interested in distributions, species-species interactions, and habitat use dynamics and preferences. A strength of occupancy modeling is that the analysis directly accounts for detection error and allows for both detection probability and occupancy to vary in response to site-level and species-specific covariates, as opposed to simple presence/absence surveys and RAIs. Occupancy modeling is less-intensive and does not require as many resources as density surveys; however, it also lacks the capacity to directly monitor population size through time. Since the analysis does not require individual recognition, it is ideal for surveys looking at multiple species (Rich et al., 2016). This feature makes occupancy modeling an exciting avenue for investigating how species-species interactions may change in response to the effects of COVID-19 lockdowns, as well as how entire communities may be changing as a result of the global change in human mobility.
One of the most pressing ecological questions during the COVID-19 pandemic is the effect the sudden change in human impact is having on wildlife behavior. Camera traps are perfectly suited to capture these changes, especially in regards to temporal activity and spatiotemporal behavior (Frey et al., 2017). Temporal activity changes have already been noted during periods of increase in human activity (Gaynor et al., 2018), with animals shifting their activity to become more nocturnal. Investigating if the reverse trend holds during decreases in human traffic is of paramount importance during this period of global decrease in human mobility. Furthermore, utilizing the variation in shutdown intensity across municipalities will illuminate whether these changes are species-specific, vary across climates and socioeconomic status, and are linear or threshold-based. Finally, camera traps can be used to investigate how species' spatiotemporal activity changes in response to biotic and abiotic factors (Parsons et al., 2016; Naidoo and Burton, 2020). From this information, it is possible to elucidate whether species are attracted to or avoid one another (Parsons et al., 2016), providing valuable insight on predator-prey dynamics and interspecies interactions. This information could be used to investigate how species are using space and time in relation to human influence (Ditmer et al., 2020; Fidino et al.,2020; Naidoo and Burton, 2020) and how this relationship has changed in the wake of COVID-19.
As explained above, camera traps have been used extensively to understand the large-scale spatial distribution of species. However, camera traps can also be used for studying the fine-scale behavior of species (Caravaggi et al., 2017), including social interactions (Leuchtenberger et al., 2014) and anti-predator behavior (Carthey and Banks, 2016). Most modern camera traps come with a video setting, making it possible to capture short video clips each time the camera is triggered. This technology allows researchers to establish experimental and quasi-experimental systems to elucidate how wildlife react to different environmental stimuli. To date, a lot of work has been done to study wildlife behavior in natural settings (Caravaggi et al., 2017), but less work has been done on elucidating the fine-scale effects of human influence on wildlife behavior (but see Gallo et al., 2019; Breck et al., 2019). We see this as one of the most fascinating and hereto underutilized opportunities for camera trap research in the future, especially during this era of rapid global change. However, detector-caused bias must be considered when using camera traps for fine-scale behavioral research (Caravaggi et al., 2020), and practitioners should consider the possibility that the cameras themselves may elicit a response (i.e., it is necessary to establish an experimental control to test the behaviroal effect cameras may have on wildlife). Furthermore, using camera traps to gather videos can drain batteries and fill up memory cards faster than using them for photos, which may result in the need for more intensive fieldwork. This may not be possible in many areas experiencing restrictions in fieldwork activity.
Finally, although COVID-19 has had a marked effect on conservation, research, tourism, and ecotourism activities (Bakar and Rosbi, 2020; Buckley, 2020), COVID-19 movement bans do not seem to have stopped illegal activity, and in some cases, there is evidence that such activity has increased (McNamara et al., 2020; Buckley, 2020). Banning researchers, ecotourists, wildlife managers, rangers, and government personnel from going out into the field results in fewer people to notice and report poachers, illegal loggers, collectors, and developers. This further highlights the importance of camera traps as conservation tools. GPS/GSM camera traps can email photos to law enforcement officials upon taking them, facilitating the identification of poachers and triggering an immediate enforcement response. For example, in 2019, Panthera updated and released their newest version of the PoacherCam (Panthera, New York City, USA), which features a small, easily-hidden, motion-activated trail camera with the capacity to take images in the field, process images for photos of humans using built-in Artificial Intelligence technology, and wirelessly send human photos to authorities. The technology allows officials to setup cameras outside or near the peripheries of protected areas and use them to catch poachers before they are able to make a kill. Other technology and camera trap companies have followed suit (see https://www.resolve.ngo/trailguard.htm), which will make these highly-specialized cameras more affordable for deployment anywhere in the world.
A final set of guiding questions focuses on the target taxa of most camera trapping studies. Our review highlights the overwhelming focus of camera trap studies on mammals (83.2%), specifically carnivores (46.0% of mammal category). Within the carnivore category, over half of the published papers focused on felids. This indicates that more data may exist for species that are particularly sensitive to human activity and environmental change (Ripple et al., 2014). These species could serve as indicator species for wildlife response to our current anthropause. Finally, our results highlight the lack of target species in other taxa, suggesting we may be able to apply camera trapping as a method to a broader range of taxa during periods of field research shutdowns.
4.2 Future of camera trapping in conservation biology and ecology: during and beyond COVID-19
In the past two decades, camera trapping has emerged as an important method for conservation biology and ecology research, and the rapid increase in studies using these tools is likely to continue. Research questions on presence/absence and basic ecology of animals are valuable to conservation efforts. However, further development of study designs, analyses, and standardization of reporting camera trap results is needed (Meek et al., 2014; Steenweg et al., 2017). Currently, few studies go beyond baseline assessments (Linkie et al., 2010), but as the price of equipment decreases, broad scale landscape ecology studies can incorporate camera traps to address novel questions in conservation biology (Erb et al., 2012; Rich et al., 2016, Rich et al., 2017).
Our literature review highlights the benefits of camera traps as a low cost, low maintenance, and largely non-invasive monitoring tool for conservation biology research and applied conservation projects. Many studies in our review documented understudied species in remote areas and significant camera trapping findings contributed to the conservation of species and ecosystems. Using the growing body of literature, conservationists can ensure they are defining questions a priori and making inferences using appropriate analyses and statistical techniques.
As more sophisticated studies are designed, camera traps will help shape large-scale conservation agendas, especially across protected areas (Kinnaird and O'Brien, 2012; Li et al., 2012). Camera trapping has been increasingly discussed as a method for conservation hotspot analyses (Kouakou et al., 2011), monitoring biodiversity (Waldon et al., 2011; Steenweg et al., 2017), comparing human dominated landscapes to natural areas (Cassano et al., 2012; Gallo et al., 2017; Parsons et al., 2018; Parsons et al., 2019), and assessing how animals respond to fluctuations in human activity (Harihar et al., 2009; Mohamed et al., 2013; Gallo et al., 2017). Camera traps have already helped biologists document unexpected wildlife presence in human-dominated landscapes (Athreya et al., 2013). If global camera trapping efforts can be standardized (Ahumada et al., 2011) and coordinated, camera traps could contribute to a comprehensive global mammal conservation strategy (Rondinini et al., 2011). However, our review highlights important geographic gaps in where camera trap studies are occurring. Furthermore, if data management issues are addressed, meta-analyses of current data could be pursued for regional analysis of abundance and diversity (Ordeñana et al., 2010).
In the unprecedented era of COVID-19, conservation biology and ecology research has been hampered by rigid restrictions on travel and field research. The resulting landscape is forcing researchers the world over to develop creative new ways of investigating our natural world (Maas et al., 2020). In this arena, camera traps offer a path forward. Camera traps are cost-effective, non-invasive survey tools that require little maintenance in the field and can be left monitoring for long periods of time. Furthermore, due to ongoing CDC social distancing guidelines and local government restrictions on mobility, researchers are experiencing at least periodic bans on fieldwork. For example, the authors lost an entire field season of live-trapping and bird banding in Utah due to a university-wide field research ban, but our active camera trapping projects continued.
With improving battery life and higher capacity memory cards, camera traps deployed in the field do not require extensive maintenance and can be left alone to monitor for long periods of time. Certain models are capable of being deployed in the field for up to a year with the ability to wirelessly send photographs without manual checks. Because camera traps can continue to collect critical wildlife data despite travel and research restrictions, they are currently more crucial to scientific research than ever before. Researchers can plan for potential restrictions by changing the settings on cameras to collect fewer images per trigger or program cameras to be active during certain times in order to prolong battery life and data storage capacity. If researchers carefully determine study objectives, the data collection (i.e., camera settings) can be selected to maximize how long cameras can remain in the field without service. As previously mentioned, one potential limitation would be studies that require video recording, typically for study objectives focused on animal behavior.
This, however, poses a very important question: how does one start a camera trapping project during the pandemic, and are there active projects that are looking for additional collaborators? First, the research team must clearly articulate the question they hope to address and decide if camera trapping is a valid tool for their research (i.e., can camera traps help answer the particular question of interest?). If so, the team must then decide on a proper study design and camera model. This will vary based on target species, research goals, study area, and other parameters, but spending more time on design and implementation beforehand will pay big dividends in the long-run (Rovero et al., 2013; Kays et al., 2020). Specifically, running simulation and power analysis using readily available software may aid researchers in avoiding a sub-optimal design (Gerrodette, 1987; Steenweg et al., 2016; Efford and Boulanger, 2019), especially if the goal is to monitor trends through space or time (Green et al., 2020). We outline only the basics of camera trap study design in the above sections, and we strongly encourage new researchers to consult earlier work on this topic (MacKenzie and Royle, 2005; Guillera-Arroita et al., 2010; Frey et al., 2017; Green et al., 2020; Kays et al., 2020). Furthermore, there are multiple research groups that have continued work during the COVID-19 pandemic, with existing datasets going back years before COVID-19. Here, we highlight a few of the larger collaborations and research groups. Each of these groups has continued their work during the pandemic, and they plan to continue doing so long after the pandemic has passed. We hope that by highlighting these organizations, we encourage conservation researchers throughout the globe to harness the power of camera traps as a research tool, take advantage of the already existing infrastructure to investigate conservation questions at a large scale, and leverage the power of collaboration and rigorous study design to advance camera trapping to the forefront of ecological research.
The Urban Wildlife Information Network (UWIN; Magle et al., 2019), led by the Lincoln Park Zoo's Urban Wildlife Institute (https://urbanwildlifeinfo.org), is a partnering organization of researchers that use camera traps and other non-invasive monitoring techniques to study the ecology and behavior of urban wildlife. With partnering institutions across all of North America, UWIN is a good collaboration to study the effects of COVID-19 lockdown on urban-adapted species. Furthermore, UWIN is actively looking for research groups to join their team, especially groups outside of the United States and Canada.
eMammal (https://emammal.si.edu) is an online data management tool built specifically for camera trap research throughout the globe. The project offers resources for practitioners in the form of data entry, upload, and review tools; project page development; training for camera trappers and volunteers; study design and camera selection recommendations; and data formatting, analysis, and visualization pipelines. Furthermore, eMammal is the host of Snapshot USA (https://emammal.si.edu/snapshot-usa), a nationwide collaboration dedicated to examining trends in mammal communities across a gradient of human influence and climatic conditions (Cove et al., 2021). Snapshot USA was active for the 2020 field season, even under COVID-19 restrictions.
EUROMAMMALS (https://euromammals.org) is the umbrella project for multiple continent-wide, species-specific studies looking at understanding the movement ecology of European mammals across different habitats and anthropogenic influence levels. Although not initially a camera trapping initiative, the project has grown to incorporate camera trap technology into their protocol.
Finally, Wildlife Insights (https://www.wildlifeinsights.org) is another collaborative effort between conservation organizations, professional scientists, and Google (Ahumada et al., 2020). Although currently in the beta stages of development, Wildlife Insights looks to address four major barriers to most camera trap practitioners, including data entry, data sharing, data analysis, and hardware issues, by leveraging an easy-to-use online interface and project builder with the Artificial Intelligence capabilities of big technology companies.
4.2.1 How current camera trapping efforts can inform future work
Our review highlights multiple aspects of camera trap research that may be utilized by future investigators, especially during lockdown. First, much of the current camera trapping research has originated in the Americas, with fewer studies published in Asia and Austraila and even fewer published in Africa and Europe. This discrepancy offers a major opportunity for researchers in these less-studied areas, especially during the pandemic. Many of these understudied areas are home to both rapid economic and urban development (e.g., southeast Asia and southern Africa) and older, established urban centers (e.g., Europe), providing an opportunity to advance urban wildlife research with comparative, multi-city analyses. Assessing the differences in urban wildlife response across age and structure of development and the effects these factors have on wildlife distribution and behavior in response to lockdown are fascinating avenues for future research. Furthermore, the rapid growth of organized networks like the ones mentioned above provide both the framework and resources to conduct such studies at both larger and more localized scales.
Second, as mentioned previously, the breakdown of taxa studied hitherto highlights an important advantage of camera trap research, especially during a pandemic that has resulted in lockdowns across the globe. Most camera trap research is currently focused on rare, elusive, and carnivorous mammals (Table 3), species likely to be affected by changes in human activities. Carnivorous mammals are often considered both exceptionally important to natural ecosystems and human-intolerant (Ripple et al., 2014). These species can be difficult to study through other means, given their elusive behavior, natural rarity, and wariness of human beings. In fact, a recent review of density estimation research with camera traps found that for many of the mammalian carnivores studied, estimates of density represented the first and only ever reported for that species (Green et al., 2020). Furthermore, camera traps are increasingly used to study mammalian carnivores in urban areas (Gallo et al., 2017; Gallo et al., 2019; Breck et al., 2019), with research networks and collaborations around the globe dedicated to studying the effects of human influence on these species (e.g., UWIN and Snapshot USA). This means researchers can leverage both these large networks and camera traps to understand the effect lockdown has had on mammalian carnivore distribution and behavior, especially along urban, suburban, and exurban gradients, something that would be exceedingly difficult using other methodologies.
This review highlights three recurring takeaways for researchers using camera traps in the future. First, detection probability and survey effort are frequently ignored, but these elements are fundamental to the inferences that can be made from camera trapping data. Scientists, managers and conservationists should be careful when comparing or applying results from camera trapping studies that do not address these issues. Second, camera trapping may not be the most suitable method to address all given conservation biology questions, and although they are effective, they are not a panacea. Many reliable methods can document the presence of species and lead to accurate estimations of population parameters. Some of the most successful studies in our review used camera traps in conjunction with other techniques to generate estimates of target species density (Gopalaswamy et al., 2012b) and a more holistic picture of population dynamics (Palomares et al., 2012). Finally, camera traps, especially in the time of COVID-19, are invaluable, cost-effective, and relatively low-labor research and remote monitoring tools for conservation biology and ecology. They can be used to help answer questions on a broad range of topics. If used effectively, camera traps may be used to continue research on novel questions during lockdown that would otherwise be impossible.
Given the current benefits and future prospects of camera traps to address objectives in animal ecology and conservation biology, the continued surge in peer-reviewed publications over the last few years is encouraging. This surge is only expected to increase during the COVID-19 pandemic. Our literature review used only the Web of Science™ database and did not include gray literature from conservation organizations or government agencies. Inclusion would have increased our sample size of papers, but we believe that we would have reached similar conclusions. In light of the current popularity of camera traps, biologists must carefully define their questions and objectives before field data collection. With careful planning and study design, COVID-19 is likely to put camera traps at the forefront of ecology and conservation biology to help us further understand the impact of human activity on wildlife and generate solutions to promote human-wildlife coexistence.
The following is the supplementary data related to this article.Supplementary Table 1
Full dataset of articles included in the systematic review of camera trap literature within the ISI Web of Knowledge Complete Collection.
Supplementary Table 1
Credit authorship contribution statement
Austin M. Green: Conceptualization, Methodology, Validation, Investigation, Data Curation, Writing – Original Draft, Supervision, Project Administration; J. David Blount: Methodology, Validation, Formal Analysis, Investigation, Data Curation, Writing – Original Draft, Visualization; Mark W. Chynoweth: Conceptualization, Methodology, Validation, Investigation, Data Curation, Writing – Original Draft, Visualization; Cagan H. Sekercioglu: Conceptualization, Methodology, Validation, Resources, Writing – Review & Editing, Supervision, Project Administration.
Declaration of competing interest
The authors have no competing interests to declare.
Acknowledgements
This work was greatly improved by the comments and suggestions of three anonymous reviewers and the handling editor. AG would like to thank the Global Change and Sustainability Center, the Sustainable Campus Initiative Fund, the National Geogaphic Society, and the 10.13039/100007747 University of Utah Graduate Research Fellowship Program. Part of this work was also supported by the 10.13039/100000001 National Science Foundation Graduate Research Fellowship under Grant No. 2010094953 for MWC. DB and ÇHŞ thank Hamit Batubay Özkan and Barbara J. Watkins for their generous support, as well as the Conservation Ecology Graduate Fellowship and the Environmental Studies Graduate Fellowship. This research did not receive any other specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
==== Refs
References
Adler F.R. Green A.M. Şekercioğlu Ç.H. Citizen science in ecology: a place for humans in nature Ann. N. Y. Acad. Sci. 1469 2020 52 64 32216157
Ahumada J.A. Silva C.E.F. Gajapersad K. Hallam C. Hurtado J. Martin E. McWilliam A. Mugerwa B. O'Brien T. Rovero F. Sheil D. Spironello W.R. Winarni N. Andelman S.J. Community structure and diversity of tropical forest mammals: data from a global camera trap network Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 366 2011 2703 2711 21844049
Ahumada J.A. Fegraus E. Birch T. Flores N. Kays R. O'Brien T.G. Palmer J. Schuttler S. Zhao J.Y. Jetz W. Kinnarid M. Kulkarni S. Lyet A. Thau D. Duong M. Oliver R. Dancer A. Wildlife insights: a platform to maximize the potential of camera trap and other passive sensor wildlife data for the planet Environ. Conserv. 47 2020 1 6
Athreya V. Odden M. Linnell J.D.C. Krishnaswamy J. Karanth U. Big cats in our backyards: persistence of large carnivores in a human dominated landscape in India PLoS One 8 2013 e57872
Bakar N.A. Rosbi S. Effect of coronavirus disease (COVID-19) to tourism industry Int. J. Adv. Eng. Res. Sci. 7 2020
Bates A.E. Primack R.B. Moraga P. Duarte C.M. COVID-19 pandemic and associated lockdown as a “global human confinement experiement” to investigate biodiversity conservation Biol. Conserv. 248 2020 108665 32549587
Bayne E.M. Hobson K.A. Comparing the effects of landscape fragmentation by forestry and agriculture on predation of artificial nests Conserv. Biol. 11 1997 1418 1429
Beck H. Terborgh J. Groves versus isolates: how spatial aggregation of Astrocaryum murumuru palms affects seed removal J. Trop. Ecol. 2002 18
Bellan S.E. Gimenez O. Choquet R. Getz W.M. A hierarchical distance sampling approach to estimating mortality rates from opportunistic carcass surveillance data Methods Ecol. Evol. 4 2013 361 369
Bennett N.J. Finkbeiner E.M. Ban N.C. Belhabib D. Jupiter S.D. Kittinger J.N. Mangubhai S. Scholtens J. Gill D. Christie P. The COVID-19 pandemic, small-scale fisheries and coastal fishing communities Coast. Manag. 48 2020 336 347
Blanc L. Marboutin E. Gatti S. Gimenez O. Abundance of rare and elusive species: empirical investigation of closed versus spatially explicit capture-recapture models with lynx as a case study J. Wildl. Manag. 77 2013 372 378
Bowkett A.E. Rovero F. Marshall A.R. The use of camera-trap data to model habitat use by antelope species in the Udzungwa Mountain forests, Tanzania Afr. J. Ecol. 46 2008 479 487
Breck S.W. Poessel S.A. Mahoney P. Young J.K. The intrepid coyote: a comparison of bold and exploratory behavior in coyotes from urban and rural environments Sci. Rep. 9 2019 2104 30765777
Buckley R. Conservation implications of COVID-19: effects via tourism and extractive industries Biol. Conserv. 247 2020 108640 32501298
Burton A.C. Critical evaluation of a long-term, locally-based wildlife monitoring program in West Africa Biodivers. Conserv. 21 2012 3079 3094
Burton A.C. Neilson E. Moreira D. Ladle A. Steenweg R. Fisher J.T. Bayne E. Boutin S. REVIEW: wildlife camera trapping: a review and recommendations for linking surveys to ecological processes J. Appl. Ecol. 52 2015 675 685
Caravaggi A. Banks P.B. Burton A.C. Finlay C.M.V. Haswell P.M. Hayward M.W. Rowcliffe M.J. Wood M.D. A review of camera trapping for conservation behaviour research Remote Sens. Ecol. Conserv. 3 2017 109 122
Caravaggi A. Burton A.C. Clark D.A. Fisher J.T. Grass A. Green S. Hobaiter C. Hofmeester T.R. Kalan A.K. Rabaiotti D. Rivet D. A review of factors to consider when using camera traps to study animal behavior to inform wildlfie ecology and conservation Conserv. Sci. Pract. 2 2020 e239
Carbone C. Christie S. Conforti K. Coulson T. Franklin N. Ginsberg J.R. Griffiths M. Holden J. Kawanishi K. Kinnaird M. Laidlaw R. Lynam A. Macdonald D.W. Martyr D. McDougal C. Nath L. O'Brien T. Seidensticker J. Smith D.J.L. Sunquist M. Tilson R. Shahruddin W.N. The use of photographic rates to estimate densities of tigers and other cryptic mammals Anim. Conserv. 4 2001 75 79
Carter N.H. Shrestha B.K. Karki J.B. Pradhan N.M.B. Liu J. Coexistence between wildlife and humans at fine spatial scales Proc. Natl. Acad. Sci. U. S. A. 109 2012 15360 15365 22949642
Carthey A.J.R. Banks P.B. Naivete is not forever: responses of a vulnerable native rodent to its long term alien predators Oikos. 125 2016 918 926
Caruso N. Manfredi C. Vidal E.M.L. Casanaveo E.B. Lucherinio M. First density estimation of two sympatric small cats, Leopardus colocolo and Leopardus geoffroyi, in a shrubland area of Central Argentina Ann. Zool. Fenn. 49 2012 181 191
Cassano C.R. Barlow J. Pardini R. Large mammals in an agroforestry mosaic in the Brazilian Atlantic forest Biotropica. 44 2012 818 825
Chandler R.B. Royle J.A. Spatially explicit models for inference about density in unmarked or partially marked populations Ann. Appl. Stat. 7 2013 936 954
Chynoweth M.W. Çoban E. Şekercioğlu C.H. Conservation of a new breeding population of Caucasian lynx (Lynx lynx dinniki) in eastern Turkey Turk. J. Zool. 39 2015 541 543
Corlett R.T. Primack R.B. Devictor V. Maas B. Goswami V.R. Bates A.E. Koh L.P. Regan T.J. Loyola R. Pakeman R.J. Impacts of the coronavirus pandemic on biodiversity conservation Biol. Conserv. 246 2020 108571 32292203
Cove M.V. Kays R. Bontrager H. Bresnen T. Frerichs R.K. McShea W.J. SNAPSHOT USA 2019: the first coordinated national camera trap survey of the United States Ecology 2021 In press
Di Bitetti M.S. De Angelo C.D. Di Blanco Y.E. Paviolo A. Niche partitioning and species coexistence in a Neotropical felid assemblage Acta Oecol. 36 2010 403 412
Dillon A. Kelly M.J. Ocelot Leopardus pardalis in Belize: the impact of trap spacing and distance moved on density estimates Oryx. 41 2007 469 477
Dillon A. Kelly M.J. Ocelot home range, overlap and density: comparing radio telemetry with camera trapping J. Zool. 275 2008 391 398
Ditmer M.A. Stoner D.C. Francis C.D. Barber J.R. Forester J.D. Choate D.M. Ironside K.E. Lonsghore K.M. Hersey K.R. Larsen R.T. McMillan B.R. Olsen D.D. Andreasen A.M. Beckmann J.P. Holton P.B. Messmer T.A. Carter N.H. Artificial nightlight alters the predator-prey dynamics of an apex carnivore Ecography. 43 2020 1 13
Dobson A. Nowak K. Does this photo make my range look big? Anim. Conserv. 13 2010 347 349
Efford M.G. Boulanger J. Fast evaluation of study designs for spatially explicit capture-recapture Methods Ecol. Evol. 10 2019 1529 1535
Erb P.L. McShea W.J. Guralnick R.P. Anthropogenic influences on macro-level mammal occupancy in the Appalachian Trail corridor PLoS One 7 2012 e42574
Espartosa K.D. Pinotti B.T. Pardini R. Performance of camera trapping and track counts for surveying large mammals in rainforest remnants Biodivers. Conserv. 20 2011 2815 2829
Farhadinia M.S. Mahdavi A. Hosseini-Zavarei F. Reproductive ecology of the Persian leopard, Panthera pardus saxicolor, in Sarigol National Park, northeastern Iran Zool. Middle East. 48 2009 13 16
Fegraus E.H. Lin K. Ahumada J.A. Baru C. Chandra S. Youn C. Data acquisition and management software for camera trap data: a case study from the TEAM network Ecol. Inform. 6 2011 345 353
Feinberg J.A. Newman C.E. Watkins-Colwell G.J. Schlesinger M.D. Zarate B. Curry B.R. Shaffer H.B. Burger J. Cryptic diversity in metropolis: confirmation of a new leopard frog species (Anura: Ranidae) from New York City and surrounding Atlantic coast regions PLoS One 9 2014 e108213
Fidino M. Gallo T. Lehrer E.W. Murray M.H. Kay C. Sander H.A. MacDougall B. Salsbury C.M. Ryan T.J. Angstmann J.L. Belaire J.A. Dugelby B. Schell C. Stankowich T. Amaya M. Drake D. Hursh S.H. Ahlers A.A. Williamson J. Hartley L.M. Zellmer A.J. Simon K. Magle S.B. Landscape-scale differences among cities alter common species’ responses to urbanization Ecol. Appl. 2020 e2253
Ford A.T. Clevenger A.P. Validity of the prey-trap hypothesis for carnivore-ungulate interactions at wildlife-crossing structures Conserv. Biol. 24 2010 1679 1685 20825447
Ford A.T. Clevenger A.P. Bennett A. Comparison of methods of monitoring wildlife crossing-structures on highways J. Wildl. Manag. 73 2009 1213 1222
Foster R.J. Harmsen B.J. A critique of density estimation from camera-trap data J. Wildl. Manag. 76 2012 224 236
Foster V.C. Sarmento P. Sollmann R. Tôrres N. Jácomo A.T.A. Negrões N. Fonseca C. Silveira L. Jaguar and puma activity patterns and predator-prey interactions in four Brazilian biomes Biotropica. 45 2013 373 379
Frey S. Fisher J.T. Burton A.C. Volpe J.P. Investigating animal activity patterns and temporal partitioning using camera-trap data: challenges and opportunities Remote Sens. Ecol. Conserv. 3 2017 123 132
Galaverni M. Palumbo D. Fabbri E. Caniglia R. Greco C. Randi E. Monitoring wolves (Canis lupus) by non-invasive genetics and camera trapping: a small-scale pilot study Eur. J. Wildl. Res. 58 2011 47 58
Gallo T. Fidino M. Lehrer E.W. Magle S.B. Mammal diversity and metacommunity dynamics in urban green spaces: implications for urban wildlife conservation Ecol. Appl. 27 2017 2330 2341 28833978
Gallo T. Fidino M. Lehrer E.W. Magle S. Urbanization alters predator-avoidance behaviours J. Appl. Ecol. 88 2019 793 803
Gardner B. Reppucci J. Lucherini M. Royle J.A. Spatially explicit inference for open populations: estimating demographic parameters from camera-trap studies Ecology. 91 2010 3376 3383 21141198
Garrard G.E. Bekessy S.A. McCarthy M.A. Wintle B.A. When have we looked hard enough? A novel method for setting minimum survey effort protocols for flora surveys Austral. Ecol. 33 2008 986 998
Gaynor K.M. Hojnowski C.E. Carter N.H. Brashares J.S. The influence of human disturbance on wildlife nocturnality Science. 360 2018 1232 1235 29903973
Gerrodette T. A power analysis for detection trends Ecology. 68 1987 1364 1372
Gilbert N.A. Clare J.D.J. Stenglein J.L. Zuckerberg B. Abundance estimation of unmarked animals based on camera-trap data Conserv. Biol. 2020 10.1111/cobi.13517
Gompper M.E. Kays R.W. Ray J.C. LaPoint S.D. Bogan D.A. Cryan J.R. A comparison of noninvasive techniques to survey carnivore communities in Northeastern North America Wildl. Soc. Bull. 34 2006 1142 1151
Gopalaswamy A.M. Karanth K.U. Kumar N.S. Macdonald D.W. Estimating tropical forest ungulate densities from sign surveys using abundance models of occupancy Anim. Conserv. 15 2012 669 679
Gopalaswamy A.M. Royle J.A. Delampady M. Nichols J.D. Karanth K.U. Macdonald D.W. Density estimation in tiger populations: combining information for strong inference Ecology. 93 2012 1741 1751 22919919
Green A.M. Chynoweth M.W. Şekercioğlu Ç.H. Spatially explicit capture-recapture through camera trapping: a review of benchmark analyses for wildlife density estimation Front. Ecol. Evol. 8 2020 563477
Guillera-Arroita G. Ridout M.S. Morgan J.T. Design of occupancy studies with imperfect detection Methods Ecol. Evol. 1 2010 131 139
Gutiérrez-González C.E. Gómez-Ramírez M.A. López-González C.A. Estimation of the density of the near threatened jaguar Panthera onca in Sonora, Mexico, using camera trapping and an open population model Oryx. 46 2012 431 437
Hackett H.M. Lesmeister D.B. Desanty-Combes J. Montague W.G. Millspaugh J.J. Gompper M.E. Detection rates of eastern spotted skunks (Spilogale putorius) in Missouri and Arkansas using live-capture and non-invasive techniques Am. Midl. Nat. 158 2007 123 131
Harihar A. Pandav B. Goyal S.P. Responses of tiger (Panthera tigris) and their prey to removal of anthropogenic influences in Rajaji National Park, India Eur. J. Wildl. Res. 55 2009 97 105
Harihar A. Chanchani P. Sharma R.K. Vattakaven J. Gubbi S. Pandav B. Noon B. Conflating “co-occurrence” with “coexistence” Proc. Natl. Acad. Sci. U. S. Am. 110 2013 E109
Harrison R.L. Barr D.J. Dragoo J. A comparison of population survey techniques for swift foxes (Vulpes velox) in New Mexico Am. Midl. Nat. 148 2002 320
Hartop E.A. Brown B.V. Disney R.H.L. Opportunity in our ignorance: urban biodiversity study reveals 20 new species and one new nearctic record for Megaselia (Diptera: Phoridae) in Los Angeles (California, USA) Zootaxa. 3941 2015 451 484 25947525
Hebblewhite M. Haydon D.T. Distinguishing technology from biology: a critical review of the use of GPS telemetry data in ecology Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 365 2010 2303 2312 20566506
Jackson R.M. Roe J.D. Wangchuk R. Hunter D.O. Estimating snow leopard population abundance using photography and capture–recapture techniques Wildl. Soc. Bull. 34 2006 772 781
Jenks K.E. Chanteap P. Damrongchainarong K. Cutter P. Cutter P. Redford T. Lynam A.J. Howard J. Leimgruber P. Using relative abundance indices from camera-trapping to test wildlife conservation hypotheses – an example from Khao Yai National Park, Thailand Trop. Conserv. Sci. 4 2011 113 131
Jenks K.E. Howard J. Leimgruber P. Do ranger stations deter poaching activity in national parks in Thailand? Biotropica. 44 2012 826 833
Jennelle C.S. Runge M.C. MacKenzie D.I. The use of photographic rates to estimate densities of tigers and other cryptic mammals: a comment on misleading conclusions Anim. Conserv. 5 2002 119 120
Jhala Y. Qureshi Q. Gopal R. Can the abundance of tigers be assessed from their signs? J. Appl. Ecol. 48 2011 14 24
Jimenez J. Nunez-Arjona J.C. Rueda C. Gonzalez L.M. Garcia-Dominguez F. Munoz-Igualada J. Lopez-Bao J.V. Estimating carnivore community structures Science. 7 2017 41036
Jones J.P.G. Monitoring species abundance and distribution at the landscape scale J. Appl. Ecol. 48 2011 9 13
Jones J.P.G. Asner G.P. And S.H.M.B. Karanth K.U. The “why”, “what” and “how” of monitoring for conservation Macdonald D.W. Willis K.J. Key Topics in Conservation Ecology First 2013 John Wiley and Sons, Ltd. 329 343
Jordan M.J. Barrett R.H. Purcell K.L. Camera trapping estimates of density and survival of fishers Martes pennanti Wildl. Biol. 17 2011 266 276
Kalle R. Ramesh T. Qureshi Q. Sankar K. Density of tiger and leopard in a tropical deciduous forest of Mudumalai Tiger Reserve, southern India, as estimated using photographic capture–recapture sampling Acta Theriol. 56 2011 335 342
Karanth K.U. Estimating tiger Panthera tigris populations from camera-trap data using capture—recapture models Biol. Conserv. 71 1995 333 338
Karanth K.U. Nichols J.D. Estimation of tiger densities in India using photgraphic captures and recaptures Ecology. 79 1998 2852 2862
Kays R. Arbogast B.S. Baker-Whatton M. Beirne C. Boone H.M. Bowler M. Burneo S.F. Cove M.V. Ding P. Espinosa S. Goncalves A.L.S. Hansen C.P. Jansen P.A. Kolowski J.M. Knowles T.W. Lima M.G.M. Millspaugh J. McShea W.J. Pacific K. Parsons A.W. Pease B.S. Rovero F. Santos F. Schuttler S.G. Sheil D. Si X. Snider M. Spironello W.R. An empirical evaluation of camera trap study design: how many, how long and when? Methods Ecol. Evol. 11 2020 700 713
Kinnaird M.F. O'Brien T.G. Effects of private-land use, livestock management, and human tolerance on diversity, distribution, and abundance of large african mammals Conserv. Biol. 26 2012 1026 1039 23082891
Kleinschroth F. Kowarik I. COVID-19 crisis demonstrates the urgent need for urban greenspaces Front. Ecol. Environ. 18 2020 318 319 32834788
Kouakou C.Y. Boesch C. Kuehl H.S. Identifying hotspots of chimpanzee group activity from transect surveys in Taï National Park, Côte d'Ivoire J. Trop. Ecol. 27 2011 621 630
Larrucea E.S. Brussard P.F. Jaeger M.M. Barrett R.H. Cameras, coyotes, and the assumption of equal detectability J. Wildl. Manag. 71 2007 1682 1689
Leuchtenberger C. Zucco C.A. Ribas C. Magnusson W. Mourao G. Activity patterns of giant otters recorded by telemetry and camera traps Ethol. Ecol. Evol. 26 2014 19 28
Lhota S. Loken B. Spehar S. Fell E. Pospěch A. Kasyanto N. Discovery of Miller's grizzled langur (Presbytis hosei canicrus) in Wehea forest confirms the continued existence and extends known geographical range of an endangered primate Am. J. Primatol. 74 2012 193 198 24006537
Li S. McShea W.J. Wang D. Lu Z. Gu X. Gauging the impact of management expertise on the distribution of large mammals across protected areas Divers. Distrib. 18 2012 1166 1176
Linkie M. Ridout M.S. Assessing tiger-prey interactions in Sumatran rainforests J. Zool. 284 2011 224 229
Linkie M. Guillera-Arroita G. Smith J. Rayan D.M. Monitoring tigers with confidence Integrat. Zool. 5 2010 342 350
Maas B. Grogan K.E. Chirango Y. Harris N. Lievano-Latorre L.F. McGuire K.L. Moore A.C. Ocampo-Ariza C. Palta M.M. Perfecto I. Primack R.B. Rowell K. Sales L. Santos-Silva R. Aparecida S. Sterling E.J. Vieira R.R.S. Wyborn C. Toomey A. Academic leaders must support inclusive scientific communities during COVID-19 Nat. Ecol. Evol. 4 2020 997 998 32493950
MacKenzie D.I. Royle J.A. Designing occupancy studies: general advice and allocating survey effort J. Appl. Ecol. 42 2005 1105 1114
MacKenzie D.I. Nichols J.D. Lachman G.B. Droege S. Royle J.A. Langtimm C.A. Estimating site occupancy rates when detection probabilities are less than one Ecology. 83 2002 2248 2255
MacKenzie D.I. Nichols J.D. Royle J.A. Pollock K.H. Bailey L.L. Hines J.E. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence 2017 Elsevier London, United Kingdom
Magle S.B. Hunt V.M. Vernon M. Crooks K.R. Urban wildlife research: past, present, and future Biol. Conserv. 155 2012 23 32
Magle S.B. Fidino M. Lehrer E.W. Gallo T. Mulligan M.P. Rios M.J. Ahlers A.A. Angstmann J. Belaire A. Dugelby B. Gramza A. Hartley L. MacDougall B. Ryan T. Salsbury C. Sander H. Schell C. Simon K. St Onge S. Drake D. Advancing urban wildlife research through multi-city collaboration Front. Ecol. Environ. 17 2019 232 239
Main M.B. Richardson L.W. Habitat management response of wildlife to prescribed fire in southwest Florida pine flatwoods Wildl. Soc. Bull. 30 2002 213 221
Manzo E. Bartolommei P. Rowcliffe J.M. Cozzolino R. Estimation of population density of European pine marten in central Italy using camera trapping Acta Theriol. 57 2012 165 172
Martinuzzi S. Stewart S.I. Helmers D.P. Mockrin M.H. Hammer R.B. Radelhoff V.C. The 2010 Wildland-urban Interface of the Conterminous United States 2015 U.S. Department of Agriculture Forest Service Washington, D.C., USA
Mazzolli M. Mosaics of exotic forest plantations and native forests as habitat of pumas Environ. Manag. 46 2010 237 253
McCallum J. Changing use of camera traps in mammalian field research: habitats, taxa and study types Mammal Rev. 43 2013 196 206
McClintock B.T. Conn P. Alonso R. Crooks K.R. Integrated modeling of bilateral photo-identification data in mark-recapture analyses Ecology. 94 2013 1464 1471 (130315134921006) 23951706
McNamara J. Robinson E.J.Z. Abernathy K. Iponga D.M. Sackey H.N.K. Wright J.H. Milner-Gulland E.J. COVID-19, systemic crisis, and possible implications for the wild meat trade in sub-Saharran Africa Environ. Resour. Econ. 76 2020 1045 1066
Meek P.D. Ballard G. Claridge A. Kays R. Moseby K. O'Brien T. O'Connell A. Sanderson J. Swann D.E. Tobler M. Townsend S. Recommended guiding principles for reporting on camera trapping research Biodivers. Conserv. 23 2014 2321 2343
Meijaard E. Chener A.C.K.I.T. Smeenk C. “New Bornean carnivore” is most likely a little known flying squirrel Mammal Rev. 36 2006 318 324
Mohamed A. Sollmann R. Bernard H. Ambu L.N. Lagan P. Mannan S. Hofer H. Wilting A. Density and habitat use of the leopard cat (Prionailurus bengalensis) in three commercial forest reserves in Sabah, Malaysian Borneo J. Mammal. 94 2013 82 89
Muhly T.B. Semeniuk C. Massolo A. Hickman L. Musiani M. Human activity helps prey win the predator-prey space race PLoS One 6 2011 e17050
Naderi M. Coban E. Kusak J. Kemahli Aytekin M.C. Chynoweth M. Agirkaya I.K. Guven N. Coban A. Sekercioglu C.H. The first record of raccoon dog (Nyctereutes procyonoides) in Turkey Turk. J. Zool. 44 2020 209 213
Naidoo R. Burton A.C. Relative effects of recreational activities on a temperate terrestrial wildlife assemblage Conserv. Sci. Pract. 2 2020 e271
Negrões N. Sollmann R. Fonseca C. Jácomo A.T.A. Revilla E. Silveira L. One or two cameras per station? Monitoring jaguars and other mammals in the Amazon Ecol. Res. 27 2012 639 648
Nichols J.D. Karanth K.U. O'Connell A.F. Science, conservation, and camera traps O'Connell A.F. Nichols J.D. Karanth K.U. Camera Traps in Animal Ecology 2011 Springer Japan, Tokyo 45 56
Núñez-Pérez R. Estimating jaguar population density using camera-traps: a comparison with radio-telemetry estimates J. Zool. 285 2011
Obbard M.E. Howe E.J. Kyle C.J. Empirical comparison of density estimators for large carnivores J. Appl. Ecol. 47 2010 76 84
O'Brien T.G. Abundance, density, and relative abundance O'Connell A.F. Nichols J.D. Karanth K.U. Camera Traps in Animal Ecology 2011 Springer Tokyo, Japan 71 96
O'Brien T.G. Kinnaird M.F. Density estimation of sympatric carnivores using spatially explicit capture–recapture methods and standard trapping grid Ecol. Appl. 21 2011 2908 2916
O'Brien T.G. Kinnaird M.F. Wibisono H.T. Crouching tigers, hidden prey: Sumatran tiger and prey populations in a tropical forest landscape Anim. Conserv. 6 2003 131 139
O'Connell A.F. Bailey L.L. Inference for occupancy and occupancy dynamics O'Connell A.F. Nichols J.D. Karanth K.U. Camera Traps in Animal Ecology 2011 Springer Tokyo, Japan 191 205
O'Connell A.F. Talancy N.W. Bailey L.L. Sauer J.R. Cook R. Gilbert A.T. Estimating site occupancy and detection probability parameters for meso- and large mammals in a coastal ecosystem J. Wildl. Manag. 70 2006 1625 1633
O'Connell A.F. Nichols J.D. Karanth U.K. Camera Traps in Animal Ecology 2011 Springer Japan Tokyo
Oliveira-Santos L.G.R. Zucco C.A. Antunes P.C. Crawshaw P.G. Is it possible to individually identify mammals with no natural markings using camera-traps? A controlled case-study with lowland tapirs Mammal. Biol. Z. Säugetierkunde. 75 2010 375 378
Ordeñana M.A. Crooks K.R. Boydston E.E. Fisher R.N. Lyren L.M. Siudyla S. Haas C.D. Harris S. Hathaway S.A. Turschak G.M. Miles A.K. Van Vuren D.H. Effects of urbanization on carnivore species distribution and richness J. Mammal. 91 2010 1322 1331
Palomares F. Godoy J.A. López-Bao J.V. Rodríguez A. Roques S. Casas-Marce M. Revilla E. Delibes M. Possible extinction vortex for a population of Iberian lynx on the verge of extirpation Conserv. Biol. 26 2012 689 697 22731698
Parsons A.W. Bland C. Forrester T. Baker-Whatton M.C. Schuttler S.G. McShea W.J. Costello R. Kays R. The ecological impact of humans and dogs on wildlife in protected areas in eastern North America Biol. Conserv. 203 2016 75 88
Parsons A.W. Forrester T. Baker-Whatton M.C. McShea W.J. Rota C.T. Schuttler S.G. Millspaugh J.J. Kays R. Mammal communities are larger and more diverse in moderately developed areas eLife. 7 2018 e38012
Parsons A.W. Rota C.T. Forrester T. Baker-Whatton M.C. McShea W.J. Schuttler S.G. Millspaugh J.J. Kays R. Urbanization focuses carnivore activity in reamining natural habitats, increasing species interaction J. Appl. Ecol. 2019 10.1111/1365-2664.13385
Paull D.J. Claridge A.W. Cunningham R.B. Effective detection methods for medium-sized ground dwelling mammals: a comparison between infrared digital cameras and hair tunnels Wildl. Res. 39 2012 546 553
Pennisi E. Pandemic robs field scientists of ‘once-in-a-lifetime’ moments Science 15 April 2020
Public Broadcasting Service (PBS) Observing a brown bear's daily routine with Cagan Sekercioglu URL: https://www.pbs.org/wnet/nature/animals-cameras-observing-brown-bears-daily-routine-cagan-sekercioglu/16063/ 2018 (Arlington, Virginia)
Ramesh T. Kalle R. Activity pattern of sloth bear Melursus ursinus (Mammalia: Ursidae) in Mudumalai Tiger Reserve, Western Ghats, India J. Theatened Taxa 5 2013 3989 3992
Ramesh T. Kalle R. Sankar K. Qureshi Q. Spatio-temporal partitioning among large carnivores in relation to major prey species in Western Ghats J. Zool. 287 2012 269 275
Razani N. Radhakrishna R. Chan C. Public lands are essential to public health during a pandemic Pediatrics. 146 2020 e20201271
Rice W.L. Mateer T.J. Reigner N. Newman P. Lawhon B. Taff B.D. Changes in recreational behaviors of outdoor enthusiasts during the COVID-19 pandemic: analysis across urban and rural communities J. Urban Ecol. 2020 10.31235/osf.io/prnz9
Rich L.N. Davis C.L. Farris Z.J. Miller D.A.W. Tucker J.M. Hamel S. Farhadinia M.S. Steenweg R. Di Bitetti M.S. Thapa K. Kane M.D. Sunarto S. Robinson N.P. Paviolo A. Cruz P. Martins Q. Gholikhani N. Taktehrani A. Whittington J. Widodo F.A. Yoccoz N.G. Wultsch C. Harmsen B.J. Kelly M.J. Assessing global patterns in mammalian carnivore occupancy and richness by integrating local camera trap surveys Glob. Ecol. Biogeogr. 26 2017 918 929
Rich L.N. Miller D.A.W. Robinson H.S. McNutt J.W. Kelly M.J. Using camera trapping and hierarchical occupancy modelling to evaluate the spatial ecology of an African mammal community J. Appl. Ecol. 53 2016 1225 1235
Ripple W.J. Estes J.A. Beschta R.L. Wilmers C.C. Ritchie E.G. Hebblewhite M. Status and ecological effects of the world's largest carnivores Science 343 2014 1241484 24408439
Rondinini C. Rodrigues A.S.L. Boitani L. The key elements of a comprehensive global mammal conservation strategy Philos. Trans. R. Soc. Lond. Ser. B Biol. Sci. 366 2011 2591 2597 21844038
Rosas-Rosas O.C. Bender L.C. Population status of jaguars (Panthera onca) and pumas (Puma concolor) in Northeastern Sonora, Mexico Acta Zool. Mexic. 28 2012 86 101
Rota C.T. Ferreira M.A.R. Kays R.W. Forrester T.D. Kalies E.L. McShea W.J. Parsons A.W. Millspaugh J.J. A multispecies occupancy model for two or more interacting species Methods Ecol. Evol. 7 2016 1164 1173
Rovero F. Rathbun G.B. Perkin A. Jones T. Ribble D.O. Leonard C. Mwakisoma R.R. Doggart N. A new species of giant sengi or elephant-shrew (genus Rhynchocyon) highlights the exceptional biodiversity of the Udzungwa Mountains of Tanzania J. Zool. 274 2008 126 133
Rovero R. Zimmerman F. Berzi D. Meek P. “Which camera trap type and how many do I need?” A review of camera features and study designs for a range of wildlife research applications Hystrix Ital. J. Mammal. 24 2013 148 156
Rowcliffe J.M. Key frontiers in camera trapping research Remote Sens. Ecol. Conserv. 3 2017 107 108
Rowcliffe J.M. Field J. Turvey S.T. Carbone C. Estimating animal density using camera traps without the need for individual recognition J. Appl. Ecol. 45 2008 1228 1236
Royle J.A. Nichols J.D. Karanth K.U. Gopalaswamy A.M. A hierarchical model for estimating density in camera-trap studies J. Appl. Ecol. 46 2009 118 127
Royle J.A. Chandler R.B. Sollmann R. Gardner B. Spatial Capture-Recapture 2014 Academic Press Waltham
Royle J.A. Fuller A.K. Sutherland C. Unifying population and landscape ecology with spatial capture-recapture Ecography. 41 2018 444 456
Rutz C. Loretto M. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R.B. Ropert-Coudert Y. Tucker M.A. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evol. 4 2020 1156 1159 32572222
Sahagun L. Coyotes, falcons, deer and other wildlife are reclaiming L.A. territory as humans stay at home Los Angeles Times https://www.latimes.com/environment/story/2020-04-21/wildlife-thrives-amid-conronavirus-lockdown 2020
Samuelsson K. Barthel S. Colding J. Macassa G. Giusti M. Urban Nature as a Source of Resilience During Social Distancing Amidst the Coronavirus Pandemic 2020 (OSF Preprints. doi.org/10.31219.osf.io/3wx5a)
Saraswat R. Saraswat D.A. Research opportunities in pandemic lockdown Science. 368 2020 594 595 32354838
Schuette P. Wagner A.P. Wagner M.E. Creel S. Occupancy patterns and niche partitioning within a diverse carnivore community exposed to anthropogenic pressures Biol. Conserv. 158 2013 301 312
Şekercioğlu Ç.H. Anderson S. Akçay E. Bilgin R. Can Ö.E. Semiz C. Tavşanoğlu Ç. Yokeş M.B. Soyumert A. İpekdal K. Sağlam İ.K. Yücel M. Nüzhet Dalfes H. Turkey's globally important biodiversity in crisis Biol. Conserv. 144 2011 2752 2769
Sekercioglu C. Happy 125th National Geographic: Brown Bears Film Their Lives with Turkey’s First CritterCams 2013 http://newswatch.nationalgeographic.com/2013/01/13/happy-125th-national-geographic-brown-bears-film-their-lives-with-turkeys-first-crittercams/
Sharma R.K. Jhala Y. Qureshi Q. Vattakaven J. Gopal R. Nayak K. Evaluating capture-recapture population and density estimation of tigers in a population with known parameters Anim. Conserv. 13 2010 94 103
Silva-Rodríguez E.A. Sieving K.E. Domestic dogs shape the landscape-scale distribution of a threatened forest ungulate Biol. Conserv. 150 2012 103 110
Singh P. Gopalaswamy A.M. Karanth K.U. Factors influencing densities of striped hyenas (Hyaena hyaena) in arid regions of India J. Mammal. 91 2010 1152 1159
Slater S.J. Christiana R.W. Gustat J. Recommendations for keeping parks and green space accessible for mental and physical health during COVID-19 and other pandemics Prevent. Chronic Dis. 17 2020 E59
Soisalo M.K. Cavalcanti S.M.C. Estimating the density of a jaguar population in the Brazilian Pantanal using camera-traps and capture–recapture sampling in combination with GPS radio-telemetry Biol. Conserv. 129 2006 487 496
Sollmann R. Mohamed A. Samejima H. Wilting A. Risky business or simple solution - relative abundance indices from camera-trapping Biol. Conserv. 159 2013 405 412
Sollmann R. Gardner B. Chandler R.B. Shindle D.B. Onorato D.P. Royle J.A. O'Connell A.F. Using multiple data sources provides density estimates for endangered Florida panther J. Appl. Ecol. 50 2013 961 968
Stafford R. Lloyd J.R. Evaluating a Bayesian approach to improve accuracy of individual photographic identification methods using ecological distribution data Comput. Ecol. Softw. 1 2011 49 54
Steenweg R. Whittington J. Hebblewhite M. Forshner A. Johnston B. Petersen D. Shepherd B. Lukacs P.M. Camera-based occupancy monitoring at large scales: power to detect trends in grizzly bears across the Canadian Rockies Biol. Conserv. 201 2016 192 200
Steenweg R. Hebblewhite M. Kays R. Ahumada J. Fisher J.T. Burton C. Townsend S.E. Carbone C. Rowcliffe J.M. Whittington J. Brodie J. Royle J.A. Switalski A. Clevenger A.P. Heim N. Rich L.N. Scaling up camera traps: monitoring the planet's biodiversity with networks of remote sensors Front. Ecol. Environ. 15 2017 26 34
Tilson R. Defu H. Muntifering J. Nyhus P.J. Dramatic decline of wild South China tigers Panthera tigris amoyensis: field survey of priority tiger reserves Oryx. 38 2004 40 47
Trolle M. Noss A.J. Cordeiro J.L.P. Oliveira L.F.B. Brazilian tapir density in the pantanal: a comparison of systematic camera-trapping and line-transect surveys Biotropica. 40 2008 211 217
Vilardell A. Capalleras X. Budó J. Pons P. Predator identification and effects of habitat management and fencing on depredation rates of simulated nests of an endangered population of Hermann's tortoises Eur. J. Wildl. Res. 58 2012 707 713
Vojta C.D. Old dog new tricks-innovations with presence-absence information J. Wildl. Manag. 69 2005 845 848
Waldon J. Miller B.W. Miller C.M. A model biodiversity monitoring protocol for REDD projects Trop. Conserv. Sci. 4 2011 254 260
Wang Y. Allen M.L. Wilmers C.C. Mesopredator spatial and temporal responses to larege predators and human development in the Santa Cruz Mountains of California Biol. Conserv. 190 2015 23 33
Wearn O.R. Glover-Kapfer P. Snap happy: camera traps are an effect sampling tool when compared with alternative methods R. Soc. Open Sci. 6 2019 181748 31032031
Weckel M. Rockwell R.F. Can controlled bow hunts reduce overabundant white-tailed deer populations in suburban ecosystems? Ecol. Model. 250 2013 143 154
Weckel M. Giuliano W. Silver S. Jaguar (Panthera onca) feeding ecology: distribution of predator and prey through time and space J. Zool. 270 2006
Whittington J. Hebblewhite M. Chandler R.B. Generalized spatial mark-resight models with an application to grizzly bears J. Appl. Ecol. 55 2018 157 168
Williams B.K. Nichols J.D. Conroy M.J. Analysis and Management of Animal Populations 2002 Academic Press San Diego, California, USA
Yamada F. Kawauchi N. Nakata K. Abe S. Kotaka N. Takashima A. Murata C. Kuroiwa A. Rediscovery after thirty years since the last capture of the critically endangered Okinawa spiny rat Tokudaia muenninki in the northern part of Okinawa Island Mammal Stud. 35 2010 243 255
Zellmer A.J. Wood E.M. Surasinghe T. Putman B.J. Pauly G.B. Magle S.B. Lewis J.S. Kay C.A.M. Fidino M. What can we learn from wildlife sightings during the COVID-19 global shutdown? Ecosphere. 11 2020 e03215
| 0 | PMC9746925 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 Apr 2; 256:108984 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.108984 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00045-8
10.1016/j.biocon.2021.108993
108993
Article
New social trails made during the pandemic increase fragmentation of an urban protected area
Primack Richard B. a
Terry Carina b⁎
a Department of Biology, Boston University, 5 Cummington Mall, Boston, MA 02215, United States of America
b Department of Earth and Environment, Boston University, 685 Commonwealth Ave, Boston, MA 02215, United States of America
⁎ Corresponding author.
16 2 2021
3 2021
16 2 2021
255 108993108993
16 10 2020
17 1 2021
22 1 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Urban protected areas are an important resource to people and wildlife, providing many ecosystem services. During the initial phase of the COVID-19 pandemic lockdown during March–June 2020, there was a major increase in the number of hikers and bicyclists in urban protected areas, including the Webster Woods in Newton, Massachusetts (USA), an 82.5-ha protected area. The Webster Woods is one of the largest protected areas near the center of Boston and is widely used in conservation textbooks as an example of the effects of habitat fragmentation on the amount of undisturbed habitat. Prior to the pandemic, the Webster Woods had been extensively fragmented by paved roads, dirt roads, and trails, with little interior habitat remaining. During the first four months of the pandemic, hikers and bicyclists made 4.9 km of new social (or informal) trails, an increase of 36%. This recent fragmentation represents a dramatic increase in the level of human impact on the area, reducing the amount of interior habitat from 3.2 to 2.1 ha. Levels of human activity returned to pre-pandemic levels in autumn 2020 and city officials have started closing access to some of the new trails, allowing vegetation to regrow. It is possible that similar increases in social trails and associated fragmentation have occurred in other protected areas (especially those in urban areas) around the world due to the pandemic, and these disturbances should be evaluated for their effects on plant and animal populations.
Keywords
Social trails
Habitat fragmentation
COVID-19
Protected areas
Urban park
==== Body
pmc1 Introduction
Urban protected areas provide essential ecosystem services, creating habitat for local wildlife as well as opportunities for people to observe and enjoy nature (Secretariat of the CBD 2008; Millennium Ecosystem Assessment 2005; Mexia et al. 2018). The importance of urban protected areas was illustrated dramatically during the COVID-19 pandemic when city residents crowded into nearby protected areas to relax and exercise (Rushing 2020; Adler and Hendricks 2020). In some parts of the world, managers had to close urban protected areas to prevent over-crowding, virus transmission, and damage to the parks (Friedman et al. 2020).
With greater visitation to parks during the pandemic, hikers and walkers widened paths as people tried to avoid close contact (Primack, personal observation; Miller-Rushing, personal observation). Hikers and bicyclists also began to make new paths, perhaps as a way to avoid crowded existing trails and to explore and experience new areas. It is unknown to what extent these new social trails (i.e., informal trails created casually or deliberately by foot or bike traffic) increased the degree of fragmentation of urban protected areas, if these trails are temporary or permanent, or if government agencies need to take action to close these social trails.
Social trails are considered problematic in many protected areas because they can damage sensitive vegetation, increase erosion, alter microclimates and increase fragmentation (Wimpey and Marion 2011; Ballantyne and Pickering 2015; Barros and Pickering 2017; Havlick et al. 2016; Lucas 2020). Habitat fragmentation created by social trails can have unintended negative consequences for wildlife and other aspects of biodiversity (Corlett et al. 2020; Rutz et al. 2020; Bates et al. 2020) by reducing habitat, altering environmental conditions, and inhibiting the movement of species (Saunders et al. 1991; Haddad et al. 2015; Fahrig 2017; Laurance et al., 2007, Laurance et al., 2011). The edges of social trails can provide entry points for invasive species. People and dogs walking on social trails can also disturb wildlife, particularly during times of reproduction (Miller et al. 1998; Reed and Merenlender 2011; Bötsch et al. 2018).
The Webster Woods, in Newton, Massachusetts (USA) represents a case study of the creation of new social trails and fragmentation during the COVID-19 pandemic. Changes in the trail system, including social trails, have been well mapped over the past 50 years, so new social trails can be readily identified and compared to past changes in trails. The vegetation of the Woods is predominantly deciduous oak-maple forest with abundant deciduous ericaceous shrubs and scattered wildflowers. It is likely that increased fragmentation of the Woods will result in some species increasing and others decreasing in abundance as the amount of edge habitat increases and the amount of interior habitat decreases. A past field study from these Woods demonstrated that bumblebees foraging on flowering shrubs only rarely crossed a road, even though they were capable of doing so (Bhattacharya et al. 2003). The Webster Woods is also an interesting case because it provides the basis for the idealized model park used in several university textbooks to illustrate the effects of habitat fragmentation on undisturbed habitats (for example, Primack 1993; Sher and Primack 2019).
In this study, we use the Webster Woods as a case study to address two questions: (1) How much has the trail network increased during the COVID-19 pandemic? (2) How much has habitat fragmentation increased during the COVID-19 pandemic? The results of this analysis can be used to suggest ways to manage this area to reduce the effects of fragmentation and can stimulate researchers to investigate the impacts of COVID-19 on other protected areas around the world, particularly those in urban areas.
2 Methods
2.1 Study site
The study site comprises the undeveloped woodland known as the Webster Woods (also sometimes known as the Hammond Woods), in Newton, Massachusetts, in metropolitan Boston. The Woods constitute an area of 82.5 ha, and parcels of it are owned by the City of Newton and the Commonwealth of Massachusetts. The area is contained within a five-sided block of busy urban roads. Within the area bounded by these roads are residential neighborhoods, with the Webster Woods in the undeveloped interior. The Webster Woods likely remained undeveloped due to the difficult-to-develop landscape of rock outcrops and ledges, surface boulders, swamps, and thin soil.
The Woods are transected east-to-west by a commuter railroad and north-to-south by a four-lane highway (Fig. 1 ). Also present in the Woods, bordering the highway, is a former synagogue, built in 1957, and its irregularly shaped parking lots. The railroad and highway divide the Woods into four unequal parts. Within each of these parts, the area is further divided by: (1) dirt roads, essentially wide paths, all established prior to 1972, (2) hiking trails established prior to 1972, and (3) social trails created between 1972 and 2019. Between March and June 2020, state health regulations and recommendations limited travel and indoor school and business activities, including the uses of commercial gyms, in Massachusetts. During this time, the number of people walking and using trail bikes in the woods increased substantially, in the process creating more social trails in the Woods. Some of the new trails appear to have been deliberately created by trail bikers to enter new parks of the Woods and to experience steep terrain. These new trails are predominantly about 1 m wide, but can vary from 0.3 to 2 m wide, and many of them have exposed soil and often significant erosion. After June 2020, the number of visitors to the Woods substantially declined, though numbers were still somewhat above pre-pandemic levels, and no new trails were created.Fig. 1 Maps of the Webster Woods, Newton, MA, with grey areas indicating the regions within 50 m of an edge. A. Showing the boundaries, railroad, highway, and synagogue (the whole property shown in black). B. With the addition of dirt roads (green) and trails (blue) as of 1972. C. with the addition of new trails as of 2019 (red). D. with the addition of new trails as of 2020 (orange). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 1
2.2 Mapping the trails
We mapped the dirt roads and trails (including social trails) using a detailed topographic map, official trail maps, and personal field notes. We mapped the trails that existed in three different years: 1972, 2019, and 2020. In addition, in 1971 and 1972 the flora of the Webster Woods was inventoried, creating a detailed account of the Woods, including the dirt roads and trails (Primack 1972).
2.3 Analyzing fragmentation
We analyzed fragmentation of the Webster Woods using the following five snapshots:1. Before the establishment of the railroad, highway, and synagogue (and its associated parking lots)
2. Establishment of the railroad, highway, and synagogue (and its associated parking lots)
3. Railroad, highway, synagogue, dirt roads, and trails in 1972
4. Railroad, highway, synagogue, dirt roads, and trails in 2019
5. Railroad, highway, synagogue, dirt roads, and trails in July 2020
Whenever we refer to the synagogue in our analysis, we are referring to the entire property, which includes the building and parking lots.
For each snapshot, we calculated the following:• The length of the boundary around the Woods, both prior to fragmentation and after the establishment of the railroad, highways, and synagogue
• The length of the dirt roads and trails
• The area of edge habitat, using a 50-m buffer extending from the boundary of the Woods and on either side of dirt roads and trails
While different edge effects can vary considerably in the distance they penetrate into adjacent undisturbed habitat, research on habitat fragmentation suggests that microclimate and ecological effects are most pronounced within 50 m of forest edges (Laurance et al. 2011).
We initially used Google Maps to delineate trails, park outline, highway, railroad, and synagogue boundary. We then exported the maps as KML files and analyzed them using R version 3.5.2 (R Core Team 2018). We analyzed the data using the packages sf and units (Pebesma 2018; Pebesma et al. 2016), and created maps using ggplot2 (Wickham 2016).
3 Results
3.1 How much has the trail network increased during the COVID-19 pandemic?
Due to its irregular shape, the boundary of the Webster Woods—without the highway, railroad, and synagogue—is 5585 m (Table 1 ). The combined outlines of the parkway, railroad, and synagogue have a total length of 3089 m, and divide the Webster Woods into four parts. The dirt roads and trails in the Woods in 1972 had a length of 8409 m. Between 1972 and 2019, an additional 5348 m of social trails had been added by hikers and bicyclists, bringing the total length of dirt roads and trails to 13,757 m. During the COVID-19 pandemic in 2020, 4948 m of new social trails were created, an increase of 36%. During the four months of the pandemic, almost as many new social trails were created as had been created during the previous 48 years.Table 1 The length of boundaries, the length of dirt roads and trails, and the area of interior habitat that existed at five different snapshots in time. The area of interior habitat is calculated with a 50 m buffer.
Table 1Snapshot time period Individual length
(m) Total length (m) Interior habitat (ha) Interior habitat (%)
Boundaries:
Prior to fragmentation 5585 5585 61.0 71.3
After establishment of railroad, parkway, synagogue 3089 8674 37.4 43.7
Dirt roads and trails:
1972 8409 8409 8.2 9.5
2019 5348 13,757 3.2 3.7
2020 4948 18,705 2.1 2.4
3.2 How much has fragmentation increased during the COVID-19 pandemic?
Considering only the boundaries of the Webster Woods and using a 50-m buffer, 61 ha (71.3%) of the Webster Woods is interior habitat (Table 1). When we add the railroad, highway, and synagogue, the Woods has 37.4 ha (43.7%) interior habitat. When we add the dirt roads and trails that existed in 1972, the amount of interior habitat declines to 8.2 ha (9.5%). Interior habitat declines to 3.2 ha (3.7%) when we add dirt roads and trails that existed in 2019, and to 2.1 ha (2.4%) when we add new social trails created in 2020.
4 Discussion
Beginning in March 2020, the number of hikers and bicyclists in the Webster Woods increased substantially, leading to the creation of roughly 5 km of new social trails. This represents an increase of 36% over the 14-km trail network that existed in 2019. The length of new social trails created during the first four months of the pandemic was comparable to the length of new social trails created in the 47 years from 1972 to 2019. Because the Woods were already so fragmented as of 2019, these new trails created during the pandemic had relatively little impact on the amount of interior habitat. This would likely also be the case in other small protected areas that might have experienced similar amounts of social trail creation. However, larger protected areas (e.g., urban parks like Middlesex Fells in Boston or Stanley Park in Vancouver) could have experienced greater declines in interior habitat area. While we focus on trail length and fragmentation in our analysis, these new trails would have also had the effect of decreasing habitat connectivity and increasing habitat isolation.
In this study, for convenience we treated all boundaries, trails, dirt roads, the highway, and the railroad as equivalent in their contributions to fragmentation and impacts on interior habitat. This is certainly an over-simplification. The highway is the noisiest place. The asphalt surface and cars dramatically alter the microclimate near the road. The railroad is very loud when trains pass and is bordered by a chain link fence that prevents the movement of large ground-dwelling animals. Trails and dirt roads alter microclimate conditions the least, but they were used extensively during the pandemic by joggers, mountain bikers, and dog walkers—with many dogs off leash—which can alter the behaviors of wildlife and hikers (Reed and Merenlender, 2008, Reed and Merenlender, 2011; Davis et al. 2010; Reilly et al. 2017; Bötsch et al. 2018). This increased level of fragmentation might also alter the sense of tranquility that many people seek when they visit an urban park.
Many new social trails appeared to result from teenagers and young adults using mountain bikes to access isolated areas of the park and steeper terrain. As the pandemic restrictions began to ease in July 2020 and when signs prohibiting biking in the woods were posted, the level of mountain biking in the woods appeared to diminish (personal observation). By the end of 2020, the number of walkers and bicyclists in the woods appeared to be far below the March–June levels, and no additional social trails were created. The second wave of COVID-19 restrictions in December 2020 did not appear to result in an increase in visitors to the Woods, possibly due to the colder weather (daytime temperatures regularly around 0 degrees C) dampening some people's willingness to walk and bike outside.
With the arrival of autumn and winter 2020, a layer of fallen leaves and snow covered the woodland floor. It is possible that many of the new trails created in 2020 will disappear and will no longer be used. However, some of these social trails continue to be used, even through autumn and winter, and will likely persist, especially where they extend the trail system into previously inaccessible areas and provide convenient short cuts between older trails.
The Newton Conservation Commission has implemented management options to close certain new social trails that are perceived to be particularly damaging because they have increased erosion on steep slopes. Actions being taken include blocking entry points with fallen tree limbs and rock walls and posting signs saying trails are closed. The Commission is also considering allowing certain social trails to remain, where they improve visitor flow and access. The recovery of the disused and closed trails will likely take at least several years in places via the re-sprouting of low ericaceous shrubs where they have been damaged. Active plant restoration is probably not needed at this site.
Our findings suggest that during the early stages of the COVID-19 pandemic in Massachusetts, USA, this urban protected area experienced dramatic increases in social trail creation and fragmentation because of unusually heavy use for recreation and exercise. We present just one case study, but anecdotal reports from other protected areas suggest this might be a widespread phenomenon in other urban parks (O'Neill, personal observation) and national parks (Miller-Rushing, personal observation). It is possible that other parks might have experienced reduced visitor usage due to more restricted pandemic lockdowns and parks being closed to public access, which may have reduced social trail creation. We suggest that researchers and managers investigate expansion of social trails in other protected areas and take the opportunity to assess impacts to microclimate, human recreational use, plant populations, and wildlife behavior. We also recommend that researchers and managers assess methods for closing social trails and restoring native habitats. Parks elsewhere might also have experienced other changes in human activity, such as increased recreational fishing, increased dog-walking, decreased trail maintenance by volunteers, and decreased removal of invasive plant species, which should be examined for their ecological and conservation impacts.
CRediT authorship contribution statement
Richard Primack: all project aspects; Carina Terry: formal analysis, visualization, validation, writing – review and editing.
Declaration of competing interest
The authors of this paper certify that they have NO conflicts of interests associated with the creation and publication of this manuscript.
Acknowledgments
We would like to thank Jennifer Steel and Doug Greenfield of the City of Newton for advice and maps. Tara Miller, Abraham Miller-Rushing, and Lucy Zipf provided helpful comments on the manuscript.
==== Refs
References
Secretariat of the Convention on Biological Diversity, 2008. Protected Areas in Today's World: Their Values and Benefits for the Welfare of the Planet. Montreal, Technical Series, no. 36, i-vii + 96.
Wimpey J. Marion J.L. A spatial exploration of informal trail networks within Great Falls Park, VA J. Environ. Manag. 92 2011 1012 1022 10.1016/j.jenvman.2010.11.015
Adler E. Hendricks M. Safe? KC Park Is Packed Even as COVID-19 Crisis Pushes Social Distancing Kansas City Star www.kansascity.com/news/coronavirus/article241529276.html 2020
Ballantyne M. Pickering C.M. Differences in the impacts of formal and informal recreational trails on urban forest loss and tree structure J. Environ. Manag. 159 2015 94 105 10.1016/j.jenvman.2015.05.007
Barros A. Pickering C.M. How networks of informal trails cause landscape level damage to vegetation Environ. Manag. 60 2017 57 68 10.1007/s00267-017-0865-9
Bates, A.E., Primack, R.B., Moraga, P., Duarte, C.M., 2020. COVID-19 pandemic and associated lockdown as a “global human confinement experiment” to investigate biodiversity conservation. Biological conservation, 108665.
Bhattacharya M. Primack R.B. Gerwein J. Are roads and railroads barriers to bumblebee movement in a temperate suburban conservation area? Biol. Conserv. 109 2003 37 45
Bötsch Y. Tablado Z. Sheri D. Kéry M. Graf R.F. Jenni L. Effect of recreational trails on forest birds: human presence matters Front. Ecol. Evol. 6 2018 175 10.3389/fevo.2018.00175
R Core Team, 2018. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. www.R-project.org/.
Corlett R.T. Primack R.B. Devictor V. Maas B. Goswami V.R. Bates A.E. Koh L.P. Regan T.J. Loyola R. Pakeman R.J. Impacts of the coronavirus pandemic on biodiversity conservation Biol. Conserv. 246 2020 108571 32292203
Davis C.A. Leslie D.M. Jr. Walter W.D. Graber A. Mountain biking trail use affects reproductive success of nesting golden-cheeked warblers Wilson Journal of Ornithology 122 3 2010 465 474
Fahrig L. Ecological responses to habitat fragmentation per se Annu. Rev. Ecol. Evol. Syst. 48 2017 1 23 10.1146/annurev-ecolsys-110316-022612
Freidman W. Allen J.G. Lipsitch M. Keep parks open The benefits of fresh air outweigh the risks of infection. Washington Post. April 13, 2020 2020
Haddad N.M. Brudvig L.A. Clobert J. Davies K.F. Gonzalez A. Holt R.D. Lovejoy T.E. Sexton J.O. Austin M.P. Collins C.D. Cook W.M. Damschen E.I. Ewers R.M. Foster B.L. Jenkins C.N. King A.J. Laurance W.F. Levey D.J. Margules C.R. Melbourne B.A. Nicholls A.O. Orrock J.L. Song D.-X. Townshend J.R. Habitat fragmentation and its lasting impact on Earth’s ecosystems Sci. Adv. 1 2015 e1500052 10.1126/sciadv.1500052
Havlick D.G. Billmeyer E. Huber T. Vogt B. Rodman K. Informal trail creation: hiking, trail running, and mountain bicycling in shortgrass prairie J. Sustain. Tour. 2016 10.1080/09669582.2015.1101127
Laurance W.F. Nascimento H.E. Laurance S.G. Andrade A. Ewers R.M. Harms K.E. Luizão R.C. Ribeiro J.E. Habitat fragmentation, variable edge effects, and the landscape-divergence hypothesis PLoS One 2 10 2007 e1017 10.1371/journal.pone.0001017
Laurance W.F. Camargo J.L. Luizão R.C. Laurance S.G. Pimm S.L. Bruna E.M. Stouffer P.C. Williamson G.B. Benítez-Malvido J. Vasconcelos H.L. Van Houtan K.S. Zartman C.E. Boyle S.A. Didham R.K. Andrade A. Lovejoy T.E. The fate of Amazonian forest fragments: a 32-year investigation Biol. Conserv. 144 2011 56 67 10.1016/j.biocon.2010.09.021
Lucas E.K. A Review of Trail-related Fragmentation, Unauthorized Trails, and Other Aspects of Recreation Ecology in Protected Areas 2020 California Fish and Wildlife Recreation Special Issue 95 125
Mexia T. Vieira J. Príncipe A. Anjos A. Silva P. Lopes N. Freitas C. Santos-Reis M. Correia O. Branquinho C. Pinho P. Ecosystem services: urban parks under a magnifying glass Environ. Res. 160 2018 469 478 10.1016/j.envres.2017.10.023 29078140
Millennium Ecosystem Assessment Ecosystems and Human Well-being: Synthesis 2005 Island Press Washington, DC
Miller S.G. Knight R.L. Miller C.K. Influence of recreational trails on breeding bird communities Ecol. Appl. 8 1998 162 169 10.2307/2641318
Pebesma, E., 2018. Simple features for R: standardized support for spatial vector data. The R Journal 10 (1), 439-446, doi:10.32614/RJ-2018-009.
Pebesma, E., Mailund, T., Hiebert, J., 2016. Measurement units in R. The R Journal 8(2), 486–494. doi:10.32614/RJ-2016-061.
Primack, R.B., 1972. The changing flora of the Hammond Woods. Harvard University, unpublished undergraduate thesis.
Primack R.B. Essentials of Conservation Biology 1993 Sinauer Associates Sunderland, MA
Reed S.E. Merenlender A.M. Quiet, nonconsumptive recreation reduces protected area effectiveness Conserv. Lett. 1 2008 146 154 10.1111/j.1755-263X.2008.00019.x
Reed S.E. Merenlender A.M. Effects of management of domestic dogs and recreation on carnivores in protected areas in northern California Conserv. Lett. 25 2011 504 513 10.1111/j.1523-1739.2010.01641.x
Reilly M.L. Tobler M.W. Sonderegger D.L. Beier P. Spatial and temporal response of wildlife to recreational activities in the San Francisco Bay ecoregion Biol. Conserv. 207 2017 117 126
Rushing E. Illegal Parking, Public Defecation, Overflowing Trash Cans: Crowds Cause Mess at Parks During Coronavirus Pandemic The Philadelphia Inquirer www.inquirer.com/news/coronavirus-covid-19-state-national-parks-closed-crowded-social-distancing-valley-forge-20200403.html?__vfz=medium%3Dsharebar 2020, April 3
Rutz C. Loretto M.-C. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nature Ecology & Evolution 2020 1 4 31900451
Saunders D.A. Hobbs R.J. Margules C.R. Biological consequences of ecosystem fragmentation: a review Conserv. Biol. 5 1991 18 32 www.jstor.org/stable/2386335?origin=JSTOR-pdf
Sher A.A. Primack R.B. An Introduction to Conservation Biology Second edition 2019 Sinauer Associates Sunderland, MA
Wickham, H., 2016. ggplot2: elegant graphics for data analysis. Springer-Verlag, New York, New York.
| 0 | PMC9746928 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 Mar 16; 255:108993 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.108993 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00048-3
10.1016/j.biocon.2021.108996
108996
Short Communication
Listening to cities during the COVID-19 lockdown: How do human activities and urbanization impact soundscapes in Colombia?
Ulloa Juan Sebastian ⁎
Hernández-Palma Angélica
Acevedo-Charry Orlando
Gómez-Valencia Bibiana
Cruz-Rodríguez Cristian
Herrera-Varón Yenifer
Roa Margarita
Rodríguez-Buriticá Susana
Ochoa-Quintero Jose Manuel
Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Avenida Paseo Bolívar 16-20, Bogotá, Colombia
⁎ Corresponding author at: Instituto de Investigación de Recursos Biológicos Alexander von Humboldt, Avenida Paseo Bolívar 16-20, Bogotá, Colombia.
10 2 2021
3 2021
10 2 2021
255 108996108996
15 9 2020
8 1 2021
16 1 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Noise is one of the fastest growing and most ubiquitous type of environmental pollution, with prevalence in cities. The COVID-19 confinement in 2020 in Colombia led to a reduction in human activities and their associated noise. We used this unique opportunity to measure the impacts of noise on urban soundscapes, and explore the effects of urbanization intensity independently of human activity. We launched a community science initiative inviting participants to collect audio recordings from their windows using smartphones. Recordings were taken during severe mobility restrictions (April), and during a period of lightened restrictions (May–June). From the data collected, we measured changes in sound pressure levels (SPL), acoustic structure (soundscape spectro-temporal characteristics), and human perception between the two periods. A 12% increase in human activities had a detectable acoustic footprint, with a significant increase of SPL (2.15 dB, 128% increase), a shift towards dominance of low-frequency broadband signals, and a perceived dominance of human-made over wildlife sounds. Measured changes in SPL and acoustic structure were directly proportional to urbanization; however, perception of these changes was not. This gap may be associated with a masking effect generated by noise or a disconnect of humans from nature in large cities. The mobility restrictions created a chance to better understand the impacts of urbanization and human activities on the soundscape, while raising public awareness regarding noise pollution effects on people and wildlife. Information analyzed here might serve in urban planning in developing countries where urban expansion is occurring in a rapid, unplanned fashion.
Keywords
Urban acoustics
Noise pollution
Biological sounds
Community science
Human perception
Noise masking
==== Body
pmc1 Introduction
The soundscape, which refers to human and natural sounds in a landscape (Pijanowski et al., 2011), is composed of rich acoustic textures with information about the surrounding environment, and it is crucial to define our sense of place (Stocker, 2013). Human activities are transforming the soundscapes, producing an acoustic overload that is ubiquitous and louder than most natural sounds (Schafer, 1993). Noise pollution is an emerging environmental issue that has been shown to have adverse effects on human health (WHO, 2011), as well as on wildlife behavior and communication (Barber et al., 2010; Shannon et al., 2016; Brumm and Slabbekoorn, 2005); these impacts on animal communities can ultimately alter the ecological services they provide (Francis et al., 2012).
Noise, anthropic sounds that can be physically harmful or distracting to humans and wildlife (Francis et al., 2009), can alter the soundscape structure and inhibit the perception of sounds by people and wildlife, a phenomenon known as masking (Barber et al., 2010). In human health assessments and urban planning, noise has been considered using primarily sound pressure levels (SPL) (Warren et al., 2006). Although noise is a noticeable element of urban acoustics, it is not the only component characterizing city soundscapes. Therefore, a better management of noise pollution depends on a more integral understanding of the impacts of human activities, complementing sound pressure measurements with other facets of the soundscape.
During 2020, human confinement due to the COVID-19 pandemic created dramatic changes in city life (Rutz et al., 2020). It also created an opportunity to measure the impact of human activities on urban soundscapes under a before-after scenario never experienced before (Bates et al., 2020). With most people staying at home, noise dropped significantly (Zambrano-Monserrate et al., 2020), providing a cleaner background to document urban soundscapes independent of human activity. It also created the possibility of exploring the effects of urbanization intensity, a common challenge in urbanization studies (Joo et al., 2011; Kuehne et al., 2013). Finally, the COVID-19 pandemic also represented an opportunity to involve the community in data collection, raise awareness about the environmental impacts of noise pollution (Sonne and Alstrup, 2019), and to evaluate urban dwellers' sensitivity to soundscape changes.
Considering these opportunities, we aimed to characterize the impact of human activities on urban soundscapes in Colombia by testing two main hypotheses: 1) the acoustic impact of human activities is proportional to urbanization intensity, with highly urbanized cities showing the biggest input of anthropophony; 2) the masking effect of anthropic noise negatively affects human perception to changes in the soundscape. Through a community science initiative, we conducted a standardized acoustic sampling throughout Colombia during the most severe mobility restrictions due to COVID-19, and during the following period of lightened restrictions. Using the information provided by the participants, we evaluated three different perspectives: 1) changes in SPL, 2) changes in acoustic structure, and 3) changes in human perception.
2 Materials and methods
2.1 Sampling protocol and data curation
Data collection was carried out from April 02 to June 17, 2020, in two distinct periods. The first period (April 02–27) represented the Full lockdown (FL), when government policies announced a mandatory closure of all non-essential workplaces, and limited outdoor recreational activities and social gatherings. Compared with a baseline taken between January–February of the same year, this period saw mobility reduced by an average −72.08 percentual points (Google, 2020; Table A1). The second collection period (May 01–June 17) represents a Partial lockdown (PL), with a partial mobility re-activation of around 12.48 percentual points from the FL period (Table A1).
For at least two days per week, during sunrise (0500–0700 h) and sunset (1700–1900 h), participants collected 90-second audio recordings from their windows motivated by a community science campaign led by Instituto de Investigación de Recursos Biológicos Alexander von Humboldt (Colombia) called "How does your city sound? Soundscapes from your window" ("¿Cómo suena mi ciudad? Paisajes sonoros desde tu ventana"). Recordings were made using the free application for smartphones Voice Record Pro® (WAV, 24 kHz sampling rate, 16-bit depth, mono channel). Uploads were accompanied by online forms asking participants about the presence of 12 soundscape components (wildlife: insects, amphibians, birds, mammals; anthropic: motorized transportation, construction, loudspeakers, human voices, domestic animals; abiotic: rain, wind, thunder), as well as the dominant perceived component in each recording.
A total of 202 participants from all over the country submitted 4556 recordings (Table 1 , Fig. A1). We then selected participants that had at least six suitable recordings (sampling rate > 22 kHz, audio length > 60 s) per period, which reduced the dataset to 62 participants with 1909 recordings, from three major cities and a pool of other 19 smaller cities. Finally, we trimmed the beginning and end of each recording to 60 s and re-sampled all files to 22.050 kHz to have homogeneous file formats among participants. Recordings and accompanying forms were deposited at the Instituto de Investigación de Recursos Biológicos Alexander von Humboldt data repository (https://doi.org/10.15472/enzm9u).Table 1 Summary of sampling sites with number of samples and participants, spatial, and demographic variables. The reported area is the official urban perimeter without considering suburbs. Total samples indicate the total number of recordings per city, number of participants in parenthesis. Selected samples refer to the final number of samples used in the analysis, number of participants in parenthesis. The average number of trees is the count of urban trees within a buffer of 200 m centered at each participant's location. The vegetation was quantified using the Normalized difference vegetation index (NDVI); low values indicate less vegetation and higher vegetation more vigorous. Data sources detailed in Table A2.
Table 1Urban intensity City Urban perimeter area (km2) Population (millions) Elevation (m) Total samples (participants) Selected samples (participants) Avg. number of trees (sd) Avg. NDVI value (sd)
High Bogotá 1587 7.2 2600 1190 (56) 711 (21) 334.3 (426.1) 2681 (858.6)
Intermediate Cali 619 1.8 1018 336 (22) 181 (7) 180.7 (244.6) 4052 (1271.7)
Intermediate Medellín 380 2.4 1495 968 (37) 191 (10) 285.2 (692.1) 4218 (1346.0)
Low Other
(19 small cities) 46–2393 <0.6 15–2758 2062 (87) 826 (24) NA NA
2.2 Data analysis
Change in SPL was estimated from changes in fitted values of root-mean-square amplitude (RMS) from each recording using a linear mixed model (LMM) with period (FL, PL) and city as fixed factors, and participant ID and time of day (am/pm) as random effects. AIC criterion corrected for small sample size (AICc) was used as a model selection procedure. We also fitted independent models for each city, keeping the same random structure. Root-mean-square amplitude was then transformed to sound pressure in decibels (Eq. (A1) and associated text for details).
Changes in acoustic structure were estimated through displacement differences between periods on a descriptive bidimensional space. Following Ulloa et al. (2018) and using all suitable recordings, we computed a spectrogram (512-sample window, no overlap between windows), and a set of 64 features depicting spectro-temporal patterns of the spectrogram that were derived by using 16 bidimensional wavelets (Morlet family, 8 scales, and 2 orientations) at four frequency bands (from 0 to 11 kHz in steps of 2.75 kHz). We used the t-distributed Stochastic Neighbor Embedding (t-SNE) (van der Maaten and Hinton, 2008) to project the data in a bidimensional space. In this space (t-SNE), samples with predominant anthropic noise located to the left, bird sounds to the right, insect sounds towards the bottom, and samples with few distant sounds towards the top (Fig. A2). A permutational multivariate analysis of variance (PERMANOVA) with Euclidean distance and permutations constrained to sample location, was used to test for displacement differences on t-SNE between periods and cities.
To further test for the effects of urban greenspaces on acoustic displacement, we selected data from the three major cities and modeled displacement (distance and angle from the FL centroid to the PL centroid, at each participant's location) against city and a set of standardized environmental covariates (number of trees, NDVI; Table 1). Although we considered other covariates (Table A2), only the selected ones had good spatial resolution to test for inter-city variation. Different variance structures (varIdent, and varExp) were tested to account for heteroscedasticity; the best model was selected with AICc.
Using the online forms, we computed a Soundscape Perception Index (SPI) as the combination of scores from each soundscape component (wildlife or anthropic, abiotic omitted) present in the recordings, plus the score of the dominant perceived component. Each component was given a score of 0.2 and SPI was computed as follows:SPI=Σwildlife components+1−Σanthropic components2
The index varies from 0 to 1, with 1 indicating full dominance of anthropophony and 0 full dominance of wildlife sounds. We estimated SPI changes using the same modeling approach as for SPL.
Statistical analyses were performed in program R (R Core Team, 2020). Signal processing and audio characterization was done in Python 3 (Van Rossum and Drake, 2009).
3 Results
Between FL and PL periods, the full model, which included period and city, was the best model explaining overall changes in SPL (Table A3). The model indicated an overall significant increase in RMS amplitude, equivalent to 2.15 dB (128% increment, Fig. 1a). Similar directions of change were found using city-specific models (Table A4); although magnitudes were different. Change order decreases in sequence from Bogotá, Cali, other cities and Medellín, with Cali and Medellín being more variable (Fig. 1a). These differences were also evident on the bidimensional acoustic space where Bogotá falls in areas dominated by low frequency traffic noise, Medellín, and Cali show a balanced mix of sounds, and other smaller cities had mid to high-frequency wildlife sounds (Fig. 2 ).Fig. 1 Parameter estimates and 95% confidence intervals for the best fit model predicting changes during Partial Lockdown with respect to Full Lockdown in: a) root-mean-square (RMS), b) Soundscape Perception Index (SPI), c) displacement distance on t-distributed Stochastic Neighbor Embedding (t-SNE) space, and d) displacement angle on t-SNE space.
Fig. 1
Fig. 2 Embedding of acoustic data in a common descriptive feature space evidences the change of acoustic structure in Colombian cities between two sampling periods: full lockdown (FL) and partial lockdown (PL). A t-distributed Stochastic Neighbor Embedding (t-SNE) was used to visualize a 2D projection from the full 64-dimensional acoustic feature space, where x and y axes are dimensions 1 and 2 of the t-SNE. Each sampling site has a dedicated figure with FL samples in orange, PL samples in purple, and centroids of sampling periods are denoted by a label and an ellipse of point dispersion. In this space, samples with predominant anthropic noise located to the left, bird sounds to the right, insect choruses towards the bottom, and samples with few distant sounds towards the top. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)
Fig. 2
Displacement on t-SNE between periods was significantly different among cities (PERMANOVA_F = 5.58, R2 = 0.03, p = 0.03) and showed a marginal effect of counts of trees (F = 2.95, R2 = 0.02, p = 0.08) but no from NDVI (F = 0.37, R2 = 0.002, p = 0.54). During PL all cities moved towards areas with more anthropophony with magnitude of displacement decreasing (without statistical significance) in the sequence from Bogotá, Cali, and Medellín (Fig. 1c, Table A5), with Bogotá showing the sharpest turn (Fig. 1d, Table A5). Our analysis also showed that differences could be associated with trees around sampling points (t = −2.19, p = 0.03) with displacement magnitude decreasing as the tree density increases (β= −0.004, SE = 0.002).
Interestingly, the estimated changes between periods were perceived differently by participants in different cities. In congruence with sound pressure, SPI became more anthropic during PL (Fig. 1b). Nevertheless, city ranking on perceived change differed from sound pressure with decreasing SPI changes from Cali, Medellín, Bogotá, and other cities; with Cali and Medellín being the most variable (Tables A6, A7). The sound components most frequently reported were birds (FL = 83%, PL = 82%) and motorized transportation (FL = 72%, PL = 81%). The latter was also the component with the strongest change between periods showing a 12.5% increase (Fig. A3).
4 Discussion
We characterized the impact of human activities on urban soundscapes in Colombia using a community science initiative during the COVID-19 lockdown. As confinement restrictions were eased (PL), we found a significant increase of SPL, a shift towards dominance of low-frequency broadband signals (0–2.75 kHz), and a perceived dominance of human-made sounds over wildlife sounds. Following our expectations, increasing human activities had an effect on the acoustic environment, which was proportional to urbanization intensity, with the most urbanized city (Bogotá) having the strongest change. However, perception of these changes was not in line with the measured changes, supporting our second hypothesis.
Our study provides the baseline impacts of urbanization and human activities on soundscapes. A 12% increase in human activities had a detectable mark on soundscapes, with cities significantly shifting towards higher sound pressure levels and dominance of anthropic sounds. The most urbanized city (Bogotá) had the sharpest change in SPL and acoustic structure, yet one of the weakest perceived changes, with a shift in soundscape perception comparable to the least urbanized sites in our sample. Previous studies have found a positive relationship between human activity and sound levels in urban areas (Mennitt and Fristrup, 2016). In particular, traffic noise is known to be closely related to urban density (Salomons and Pont, 2012). The higher levels of anthropic noise in the most urbanized city generates a dense background that masks and impairs the perception of sounds, resulting in the so-called lo-fi soundscape (Schafer, 1993). In this lo-fi system overloaded with acoustic signals, little can emerge with clarity, perspective is reduced, and changes are harder to perceive. Hence, the difference in magnitude between measured and perceived change could be explained by this masking effect of noise.
Alternatively, humans could be less in tune with their surroundings in highly urbanized cities, which translates into a lower sensitivity to changes in their soundscape. It has been argued that urbanization promotes the separation of humans from nature (Turner et al., 2004), leading to people living in larger cities to be less perceptive to changes compared to people from smaller cities (Miller, 2005). Either by the masking effect or the decreased sensitivity, elevated noise levels can affect the perception of wildlife in large cities, disrupting human-nature interactions and undermining positive attitudes of people towards nature (Soga and Gaston, 2016).
Acoustic structure results showed an inversely proportional relationship between urbanization intensity and wildlife sounds. A recent meta-analysis found that the density of species in cities is best explained by anthropic features (land cover, city age) rather than by environmental factors such as geography, climate, and topography (Aronson et al., 2014). Both of our moderately urbanized cities (Medellín and Cali) showed a similar response in acoustic structure between periods despite the former being ~30% more populated than the later. One possible explanation for this is the buffering effect of greenspaces in cities (Fang and Ling, 2003); both of these moderately urbanized cities have similar vegetation indices that are comparably higher than the most urbanized city in our sample, suggesting that greener cities could promote a higher diversity of natural sounds.
Although the use of smartphones allowed us to reach a large audience and cover a larger sampling area, it is important to consider that: 1) different brands have variability in the recording quality (microphone and pre-amplifiers), and 2) these sensors are less sensitive to low frequencies (<300 Hz). As a solution for the first aspect, we used paired samples and compared the change measured by each recording device individually. As for the second aspect, since anthropic noise is characterized by energy at low frequencies, our results should be regarded as conservative; therefore, changes in SPL and acoustic structure are likely higher than measured.
Colombia is a developing country with an immense biodiversity. Most of the cities in the country are rapidly growing, prioritizing infrastructure over greenspaces. These urban greenspaces serve as wildlife refugia, noise barriers, and provide opportunities for human-nature interactions. Adequate soundscape management to mitigate anthropic noise in such green spaces could increase the use of these habitats by wildlife, which can in turn facilitate human-nature interactions, ultimately fostering the conditions for human well-being and wildlife in a positive feedback loop (Levenhagen et al., 2020). The dramatic decrease in human activities and their associated noise due to the COVID-19 confinement was a unique opportunity to measure not only the impacts of human activities in the acoustic environment, but also to highlight the importance of greenspaces to buffer noise pollution. Moreover, this was also an occasion for community volunteers to open their ears and experience the surrounding wildlife in cities from their windows, showing that even indoor activities can be designed to strengthen their connection with nature (Collins et al., 2020). While largely underexplored, community science initiatives that incorporate active listening have the potential to raise public awareness regarding urban wildlife and noise pollution effects (Sonne and Alstrup, 2019; Kuehne et al., 2013), engaging volunteers in the establishment and maintenance of more suitable human and wildlife habitats.
CRediT authorship contribution statement
Juan Sebastian Ulloa: Conceptualization, Data acquisition – curation, Formal analysis, Investigation, Methodology, Supervision, Validation. Angélica Hernández-Palma: Data curation, Formal analysis, Investigation, Methodology, Writing - review & editing. Orlando Acevedo-Charry: Investigation, Writing - review & editing. Bibiana Gómez-Valencia: Formal analysis, Investigation, Methodology, Writing - review & editing. Cristian Cruz-Rodríguez: Data acquisition – curation, Visualization, Writing - review & editing. Yenifer Herrera-Varón: Data curation, Visualization, Writing - review & editing. Margarita Roa: Data curation, Methodology, Visualization. Susana Rodríguez-Buriticá: Conceptualization, Formal analysis, Validation, Writing - review & editing. Jose Manuel Ochoa-Quintero: Conceptualization, Methodology, Writing - review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Appendix A Supplementary data
Supplementary material
Image 1
Acknowledgements
Data collection was realized by the active participation of hundreds of people who selflessly contributed records to the citizen science initiative ¿Cómo suena mi ciudad? Paisajes sonoros desde tu ventana, which had the support of the Instituto de Investigación de Recursos Biológicos Alexander von Humboldt and the Red Ecoacústica Colombiana. We thank Carolina Gómez Navarro, Paola Olaya Arenas, and Charles Hathcock for language corrections and comments on the manuscript.
Funding sources
This work was supported by the Colombian Ministry of Environment and Sustainable Development (resolución 0041, 2020). JSU was funded by Colciencias Postdoctoral Grant (811-2018).
Appendix A Supplementary data to this article can be found online at https://doi.org/10.1016/j.biocon.2021.108996.
==== Refs
References
Aronson M.F.J. La Sorte F.A. Nilon C.H. Katti M. Goddard M.A. Lepczyk C.A. Warren P.S. Williams N.S.G. Cilliers S. Clarkson B. Dobbs C. Dolan R. Hedblom M. Klotz S. Kooijmans J.L. Kühn I. MacGregor-Fors I. McDonnell M. Mörtberg U. Pyšek P. Siebert S. Sushinsky J. Werner P. Winter M. A global analysis of the impacts of urbanization on bird and plant diversity reveals key anthropogenic drivers Proc. R. Soc. B 281 2014 20133330 10.1098/rspb.2013.3330
Barber J.R. Crooks K.R. Fristrup K.M. The costs of chronic noise exposure for terrestrial organisms Trends Ecol. Evol. 25 2010 180 189 19762112
Bates A.E. Primack R.B. Moraga P. Duarte C.M. COVID-19 pandemic and associated lockdown as a “Global Human Confinement Experiment” to investigate biodiversity conservation Biol. Conserv. 248 2020 108665 10.1016/j.biocon.2020.108665 32549587
Brumm H. Slabbekoorn H. Acoustic communication in noise Advances in the Study of Behavior 2005 Academic Press 151 209 10.1016/S0065-3454(05)35004-2
Collins M. Dorph R. Foreman J. Pande A. Strang C. Young A. A Field at Risk: The Impact of COVID-19 on Environmental and Outdoor Science Education (Policy Brief) 2020 Lawrence Hall of Science, University of California Berkeley, California
Fang C.-F. Ling D.-L. Investigation of the noise reduction provided by tree belts Landsc. Urban Plan. 63 2003 187 195 10.1016/S0169-2046(02)00190-1
Francis C.D. Ortega C.P. Cruz A. Noise pollution changes avian communities and species interactions Curr. Biol. 19 2009 1415 1419 10.1016/j.cub.2009.06.052 19631542
Francis C.D. Kleist N.J. Ortega C.P. Cruz A. Noise pollution alters ecological services: enhanced pollination and disrupted seed dispersal Proc. R. Soc. B 279 2012 2727 2735 10.1098/rspb.2012.0230
Google LLC Google COVID-19 community mobility reports [WWW document]. URL https://www.google.com/covid19/mobility/ 2020
Joo W. Gage S.H. Kasten E.P. Analysis and interpretation of variability in soundscapes along an urban–rural gradient Landsc. Urban Plan. 103 2011 259 276 10.1016/j.landurbplan.2011.08.001
Kuehne L.M. Padgham B.L. Olden J.D. The soundscapes of lakes across an urbanization gradient PLoS One 8 2013 e55661 10.1371/journal.pone.0055661
Levenhagen M.J. Miller Z.D. Petrelli A.R. Ferguson L.A. Shr Y. (Jimmy) Gomes D.G.E. Taff B.D. White C. Fristrup K. Monz C. McClure C.J.W. Newman P. Francis C.D. Barber J.R. Ecosystem services enhanced through soundscape management link people and wildlife People and Nature 2020 pan3.10156 10.1002/pan3.10156
van der Maaten L. Hinton G. Visualizing data using t-SNE J. Mach. Learn. Res. 9 2008 2579 2605
Mennitt D.J. Fristrup K.M. Influence factors and spatiotemporal patterns of environmental sound levels in the contiguous United States Noise Control Eng. J. 64 2016 342 353
Miller J.R. Biodiversity conservation and the extinction of experience Trends Ecol. Evol. 20 2005 430 434 16701413
Pijanowski B.C. Farina A. Gage S.H. Dumyahn S.L. Krause B.L. What is soundscape ecology? An introduction and overview of an emerging new science Landsc. Ecol. 26 2011 1213 1232 10.1007/s10980-011-9600-8
R Core Team R: A Language and Environment for Statistical Computing 2020 R Foundation for Statistical Computing Vienna, Austria
Rutz C. Loretto M.-C. Bates A.E. Davidson S.C. Duarte C.M. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R.B. Ropert-Coudert Y. Tucker M.A. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evol. 4 2020 1156 1159 10.1038/s41559-020-1237-z 32572222
Salomons E.M. Pont M.B. Urban traffic noise and the relation to urban density, form, and traffic elasticity Landsc. Urban Plan. 108 2012 2 16
Schafer R. The soundscape: our sonic environment and the tuning of the world Destiny Books Distributed to the Book Trade in the United States by American International Distribution Corp 1993 Vt. United States Rochester
Shannon G. McKenna M.F. Angeloni L.M. Crooks K.R. Fristrup K.M. Brown E. Warner K.A. Nelson M.D. White C. Briggs J. A synthesis of two decades of research documenting the effects of noise on wildlife Biol. Rev. 91 2016 982 1005 26118691
Soga M. Gaston K.J. Extinction of experience: the loss of human–nature interactions Front. Ecol. Environ. 4 2016 94 101
Sonne C. Alstrup A.K.O. Using citizen science to speed up plastic collection and mapping of urban noise: lessons learned from Denmark Mar. Pollut. Bull. 149 2019 110591 10.1016/j.marpolbul.2019.110591 31550571
Stocker M. Hear Where We Are: Sound, Ecology, and Sense of Place 2013 Springer-Verlag New York 10.1007/978-1-4614-7285-8
Turner W.R. Nakamura T. Dinetti M. Global urbanization and the separation of humans from nature Bioscience 54 2004 585 590
Ulloa J.S. Aubin T. Llusia D. Bouveyron C. Sueur J. Estimating animal acoustic diversity in tropical environments using unsupervised multiresolution analysis Ecol. Indic. 90 2018 346 355 10.1016/j.ecolind.2018.03.026
Van Rossum G. Drake F.L. Python 3 Reference Manual 2009 CreateSpace Scotts Valley, CA
Warren P.S. Katti M. Ermann M. Brazel A. Urban bioacoustics: it’s not just noise Anim. Behav. 71 2006 491 502 10.1016/j.anbehav.2005.07.014
WHO Burden of Disease From Environmental Noise: Quantification of Healthy Life Years Lost in Europe 2011 World Health Organization. Regional Office for Europe
Zambrano-Monserrate M.A. Ruano M.A. Sanchez-Alcalde L. Indirect effects of COVID-19 on the environment Sci. Total Environ. 728 2020 138813 10.1016/j.scitotenv.2020.138813 32334159
| 0 | PMC9746930 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 Mar 10; 255:108996 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.108996 | oa_other |
==== Front
Biol Conserv
Biol Conserv
Biological Conservation
0006-3207
0006-3207
Elsevier Ltd.
S0006-3207(21)00024-0
10.1016/j.biocon.2021.108972
108972
Policy Analysis
How does the beach ecosystem change without tourists during COVID-19 lockdown?
Soto E.H. an⁎
Botero C.M. bn
Milanés C.B. cn
Rodríguez-Santiago A. dn
Palacios-Moreno M. en
Díaz-Ferguson E. fn
Velázquez Y.R. gn
Abbehusen A. hn
Guerra-Castro E. ijn
Simoes N. jkn
Muciño-Reyes M. jln
Filho J.R. Souza mn
a Centro de Observación Marino para Estudios de Riesgos del Ambiente Costero (COSTAR), Facultad de Ciencias del Mar y de Recursos Naturales, Universidad de Valparaíso, Viña del Mar, Chile
b Escuela de Derecho, Universidad Sergio Arboleda, Santa Marta, Colombia
c Universidad de La Costa, Departamento Civil y Ambiental, Barranquilla, Colombia
d Taller Ecológico de Puerto Rico, Boquerón, Puerto Rico
e Universidad Del Pacífico, Guayaquil, Ecuador
f Estación Científica Coiba (Coiba AIP), Ciudad del Saber, Clayton, Panamá
g Centro de Estudios Multidisciplinarios de Zonas Costeras (CEMZOC), Universidad de Oriente, Santiago de Cuba, Cuba
h Universidade Católica do Salvador, Centro de Ecologia e Conservação animal, ECOA, Salvador, Bahia, Brazil
i Escuela Nacional de Estudios Superiores Unidad Mérida, Universidad Nacional Autónoma de México, Mérida, Yucatán, México
j Laboratorio Nacional de Resiliencia Costera, Laboratorios Nacionales, CONACYT, Mexico
k Unidad Multidisciplinaria de Docencia e Investigación Sisal (UMDI-SISAL), Facultad de Ciencias, Universidad Nacional Autónoma de México, Sisal, Yucatán, México
l Posgrado en Ciencias Biológicas, Universidad Nacional Autónoma de México, Ciudad de México, México
m Instituto Federal de Educação, Ciência e Tecnologia Baiano - IFBAIANO, Bahia, Brazil
n Proplayas Network
⁎ Corresponding author at: Facultad de Ciencias del Mar y de Recursos Naturales, Universidad de Valparaíso, Avenida Borgoño 16344, P.O. Box: Casilla 5080 Reñaca, Viña del Mar, Chile.
21 1 2021
3 2021
21 1 2021
255 108972108972
19 9 2020
7 1 2021
11 1 2021
© 2021 Elsevier Ltd. All rights reserved.
2021
Elsevier Ltd
Since January 2020 Elsevier has created a COVID-19 resource centre with free information in English and Mandarin on the novel coronavirus COVID-19. The COVID-19 resource centre is hosted on Elsevier Connect, the company's public news and information website. Elsevier hereby grants permission to make all its COVID-19-related research that is available on the COVID-19 resource centre - including this research content - immediately available in PubMed Central and other publicly funded repositories, such as the WHO COVID database with rights for unrestricted research re-use and analyses in any form or by any means with acknowledgement of the original source. These permissions are granted for free by Elsevier for as long as the COVID-19 resource centre remains active.
Urban tourist beach ecosystems provide the essential service of recreation. These ecosystems also support critical ecological functions where biodiversity conservation is not usually a priority. The sudden lockdown due to the COVID-19 pandemic created a unique opportunity to evaluate the effects of human absence in these urban-coastal ecosystems. This study examined bioindicators from 29 urban tourist beaches in seven Latin-American countries and assesses their response to lockdown about some relevant anthropogenic stressors such as pollution, noise, human activities, and user density. The presence of animals and plants, as well as the intensity of stressors, were assessed through a standardized protocol during lockdown conditions. Additionally, the environmental conditions of the beaches before and during lockdown were qualitatively compared using multivariate non-parametric statistics. We found notable positive changes in biological components and a clear decrease in human stressors on almost all the beaches. Dune vegetation increased on most sites. Similarly, high burrow densities of ghost crabs were observed on beaches, except those where cleaning activity persisted. Because of the lockdown, there was an exceptionally low frequency of beach users, which in turn reduced litter, noise and unnatural odors. The observed patterns suggest that tourist beaches can be restored to natural settings relatively quickly. We propose several indicators to measure changes in beaches once lockdown is relaxed. Adequate conservation strategies will render the recreational service of tourist beaches more environmental-friendly.
Graphical abstract
Unlabelled Image
Keywords
Tourist beaches
Bioindicators
Stressors
Coronavirus
Coastal biodiversity
Wildlife conservation
==== Body
pmc1 Introduction
Sandy beaches are one of the most common ecosystems of the coastal zone. They provide a wide range of ecosystem services, harbor unique biological diversity and deliver important socio-ecological value (Schlacher et al., 2014a; Olds et al., 2018). However, they are affected by natural threats and anthropic pressures, which are the main changing factors influencing these ecosystems (Defeo et al., 2009; Reyes-Martínez et al., 2015). Climate change translates to increasing impacts of sea-level rise, waves, storms, erosion and landward recession of the shoreline (Jones et al., 2007; Harley et al., 2006). A warming climate also brings rising temperatures, changes in rainfall magnitude, and occurrence of extreme weather events like hurricanes (Becken, 2016). At the same time, pressures mostly related to rapid human population growth and urban expansion exacerbate the negative impacts of human activities such as pollution, construction, exploitation, and recreation on tourist beaches (Brown and McLachlan, 2002; Schlacher et al., 2016).
Tourist beaches are specifically affected by humans, and the extent and type of impacts may differ between urban and more natural tourist beaches. Urban beaches suffer from loss of habitat, beach erosion, sediment plumes, disruption of sand transport, overall pollution, resources overexploitation, maritime accidents, overcrowding, lack of services, loss and change of species, disruption of ecosystem structure and function, and reduced ecosystem resilience (Brown and McLachlan, 2002; Defeo et al., 2009). More natural beaches, located remote from the urban watershed may still suffer from anthropogenic stresses. These include loss of biodiversity, disruption of ecosystem function, saltwater intrusion from freshwater extraction, invasive species, loss of habitat, reduced plant coverage loss, algae accumulation, beach degradation, water pollution from mining and other activities, and eutrophication (Canteiro et al., 2018; Dodds and Holmes, 2019b).
The great majority of tourist beaches record the presence of improperly disposed solid wastes from recreational activities. Commercial tourism activities generate a continuous stream of noise pollution and effluents (Cristiano et al., 2020; Souza and Silva, 2015). In addition, coastal processes deliver a constant supply of floating solid waste, some originating thousands of kilometers away from the receiving beaches. However, on more natural beaches the main human impacts are occupation and construction in the back-beach area (access roads, second homes, small businesses, etc.). This results in the suppression of native vegetation, artificial structures to stabilize sandy terraces and dune cords, all leading to a decrease in important ecosystem services (Souza Filho et al., 2019).
Although scarce, the few studies comparing impacts on protected non-tourist and urbanized-tourist beaches illustrate the negative influence of intensive use on biological communities, resulting in a reduction in the density and diversity of species (Martins, 2007; Gheskiere et al., 2005). Veloso et al. (2006) compared urbanized and natural beaches, finding species richness similar for both. However, the density of some species was lower in urbanized beaches and some species were more vulnerable to trampling. Tourist impacts can be so damaging that Niefer (2002) suggests continuous monitoring of visitors, the establishment of information centers, educational trails, and environmental education projects. Recently, Guerra-Castro et al. (2020) used an open beach biodiversity assessment methodology to establish solid baselines to compare in cases of extreme impacts applied in Mexican tourist and non-tourist beaches.
Beaches are social-ecological systems where interactions among species and humans shape community structure and biodiversity. Human activities often generate negative effects on coastal ecosystems (Halpern et al., 2008; Andrés et al., 2017). Even short periods of reduced human activities on beaches increases species abundance and diversity (Davis, 2019). The reduction of beach use due to COVID-19 lockdowns resulted in cleaner beaches and less environmental noise (Zambrano-Monserrate et al., 2020) with longer-term ecological implications as yet unknown.
Proximity to urban centers or other attractions enables many tourists to visit a nearby beautiful beach (Amyot and Grant, 2014; Dodds and Holmes, 2019a). However, visitors often do know that various forms of beach wildlife suffer from leisure and recreational activities that disturb critical habitat (Schlacher and Thomson, 2012). Tourist beaches have not been recognized as priority areas for conservation, perhaps due to the absence of an extensive vegetation cover or the perception of limited biodiversity (Blankensteyn, 2006; Vilar de Araujo et al., 2008). One example is the Playas Villamil National Recreational Area in Ecuador (2472 ha and 14 km of beaches), which attracts many tourists throughout the year (Zambrano-Monserrate et al., 2018). Urban tourist beaches are highly frequented, and the visitor is unaware of the wildlife importance, therefore the beach becomes a biologically desolate ecosystem and valued only for the landscape and recreational aspects and the Playas Villamil is no exception.
Some recreational activities could seem harmless; however, others can generate deleterious impacts on the beach environment, wildlife, and resources (Marion et al., 2016). The first response to human disturbances by wildlife is usually behavioral, indicating that changes in ethology are good indicators of anthropogenic impacts (Schlacher et al., 2013). Effects can become severe, such as global species extinction or disruption of communities and ecosystems (Green and Giese, 2004: Halpern et al., 2019). Some common impacts are disturbances, habitat loss and modification, decrease in population, and displacement from critical resources like food or water (Marion, 2019). Wildlife seeking to avoid human presence must change their habits or adapt to the new circumstances (Gaynor et al., 2018). These impacts are evident in urban beaches because of the high tourism pressure and urban development (Ariza et al., 2007; Marion et al., 2016).
The sharing of living spaces among humans and wildlife proves problematic since anthropic stressors often drive migrations, redistribution, and life cycle shifts in nature (Gaynor et al., 2018). This complex coexistence results in the need to establish different coastal ecosystem management schemes, with goals often associated with biological conservation to avoid adverse effects caused by human interaction (Jarratt and Davies, 2019) and enabling in some cases the development of scientific research. However, the management of highly visited urban tourist beaches usually promotes services, infrastructure, and leisure activities for tourists (Botero et al., 2018). These decisions do not account for the protection of wildlife and set priorities of economic benefits to satisfy human needs (Mendoza-González et al., 2018). Therefore, the lack of management strategies will affect wildlife and its conservation, generating impacts such as habitat alteration, biodiversity loss, unintended interactions, information gaps, distribution changes and invasive species (Defeo et al., 2009).
Urban tourist beaches generally arise in places of natural beauty associated with human settlements (Mendoza-González et al., 2018). These environments commonly present a wide variety of environmental stressors such as noise, pollution, human activities, and odors that, according to their magnitude and occurrence, impact nature in different ways (Araújo et al., 2018). Generally, an urban beach with high tourist densities and stressors negatively alter the natural environment, affecting biodiversity, ecological processes, and ecosystem services (Marshall et al., 2014). The effects of anthropogenic pressures are known (Brown and McLachlan, 2002) and some have been studied like Brazilian beaches where litter and solid wastes abundances changed significantly, according to tourist's presence or urbanization degree (Suciu et al., 2017: Araújo et al., 2018). Even specific methodologies have been developed to assess tourism impacts on natural protected areas (Canteiro et al., 2018). However, the response of these ecosystems to the lack of stressors or its reduction is currently unknown.
Tourist beaches with visitors all year often show a decrease in their anthropogenic pressures during low season (Reyes-Martínez et al., 2015); nevertheless, the total absence of environmental stressors should be highly unlikely. Yet knowing the effects of eliminating anthropogenic impacts would help establish a baseline condition closer to natural conditions. Moreover, studies assessing tourist beaches under pandemic and lockdown conditions are novel (Zielinski and Botero, 2020). Therefore, the current COVID-19 lockdown allows a unique and unprecedented opportunity to study this response as a “Global Human Confinement Experiment” (Bates et al., 2020; Rutz et al., 2020). This new period of dramatic and unusual slowdown in human activity, that many call the “anthropause,” could provide important insights into human-wildlife interactions (Rutz et al., 2020). The current world scenario also allows unconventional assessment and valuation of stressors and bioindicators, providing new knowledge for management and wildlife conservation. Similar studies prove useful to evaluate the ecological condition, variations and effects in sandy beach systems, being a relevant tool for coastal managers and biodiversity conservationists (McLachlan et al., 2013; Schlacher et al., 2014).
Considering the scenarios described above, conditions on tourist beaches will probably change under lockdown restrictions. Therefore, we set the hypothesis that urban tourist beaches should exhibit better environmental conditions, driven by a decrease in their anthropogenic stressors, and the improvement of selected bioindicators. To test this hypothesis, the goals of this study were: a) Determine the environmental conditions of tourist beaches during lockdown and pandemic time, b) describe the presence or absence of selected stressors and bioindicators under lockdown conditions, c) compare stressors and bioindicators results with previous records and d) assess changes on wildlife and stressors of tourist beaches due to COVID-19 lockdown.
2 Materials and methods
2.1 Study area
The study area consisted of twenty-nine tourist beaches at seven countries from South America, Central America, and the Caribbean, covering the West Tropical Atlantic and the East Tropical Pacific coasts. All beaches were monitored by members of Proplayas Network. Fieldwork was made between July 2nd and August 1st, 2020, when all beaches were closed due to lockdown restrictions. Information about study sites is detailed in Fig. 1 and Supplementary material A.
2.2 Beach background and COVID-19 context
Most of the beaches studied belong to sandy beach ecosystems. Most beaches are microtidal (<2 m of tidal ranges) and mesotidal (between 2 and 4 m of tidal ranges), except for macrotidal beaches (>4 m tidal range) monitored in Panama and Ecuador. All beaches show a low slope and are mainly composed of biogenic sand. Tourist areas may encompass kilometers long to hundreds of meters wide. All beaches present a high tourist influx, mainly during weekends, high season, holidays, Easter and summer under typical conditions. Recreational facilities, tourist services, and infrastructure, in some cases as resorts and apartment buildings, are present for all beaches, hence they are considered as highly modified ones. Anthropogenic pressures, such as vehicles, fishing, and maritime traffic, are common. Some beaches have biological, historical, or cultural importance such as turtle nesting and spawning areas for various animals, or they are in cities with heritage categories. The beaches studied are highly impacted by human presence and generally show absent or reduced dune vegetation (e.g., beachgrass) and regional fauna (e.g., ghost crabs) (Defeo, 2009). Main features of each studied beach are described in Supplementary 1. The world spreading of new coronavirus has obligated all-level government authorities to establish lockdown and restrictions measures, including banning access to tourist beaches.
Therefore, beaches were monitored after authorization by local authorities. Each team followed sanitary and safety protocols.Fig. 1 Location of beaches in the study area.
Fig. 1(Source: Images designed using Google Earth.)
2.3 Protocols and fieldwork
We used the beach typology proposed by Williams and Micallef (2009), which emphasizes the managerial aspects. Therefore, urban, village, and resort beaches were prioritized over rural and remote beaches, because the former usually have the highest tourist pressure (Williams and Micallef, 2009). Most of the studied beaches are of the urban type (21 of the 29). The others included six village beaches, one resort beach and one rural beach (Supplementary A). As an exception, Caracas Beach (Puerto Rico) and El Estero Santa Catalina Beach (Panamá), despite being tourist beaches, are also protected areas for wildlife conservation. Previous knowledge of each beach was also considered to compare the results with normal beach conditions. Surveys took place between 10:00 h and 15:00 h, at the time when the beach is traditionally more crowded. Likewise, we avoided fieldwork during the weekends to reduce possibility of encountering a higher number of illegal visitors at beaches defying lockdown and restriction measures.
2.3.1 Bioindicators
Categories for bioindicators were chosen considering broad taxonomic attributes of the fauna and flora. For fauna, the presence of the following animals was recorded: crabs, lizards, turtles, iguanas, opportunistic birds, sea birds, and domestic animals. For flora, the presence of seaweed, seagrasses, beachgrass, shrubbery, vines, mangroves, and other trees was recorded. As all the beaches studied met the requisite of being highly crowded in typical tourism seasons, user density was expected to affect typical flora and fauna and could increase the presence and extent of opportunistic species associated with human activities. Data were registered in situ in a digital format designed for the mobile application termed Kobo Collect (Supplementary 2) and linked to a Web cloud. We performed monitoring of biological indicators in three defined zones (parallel to the waterline) within the touristic use zone of each beach. These zones were (1) Active: area of sand strip closest to the waterline, dedicated to the circulation of bathers. (2) Rest: an area dedicated to users' rest and sunbathing. (3) Service: an area for shops and services. Records of bioindicators were completed at 100 m transects in each of the defined zones. To generate evidence about the environmental beaches condition during the lockdown, short videos were taken at both ends and the midpoint of each transect. The taxonomic resolution of bioindicators was supported with photographs of each observed organism.
2.3.2 Stressors
An anthropogenic stressor is an action that generates direct or indirect pressure on the beach. To assess the potential effects of this human pressure we chose seven stressors based on their ease and relevance of measurement: 1. Noise; 2. odor; 3. litter; 4. user density; 5. activities; 6. infrastructure; and 7. anthropogenic threats. The first three stressors are part of the Environmental Beach Quality Index proposed by Botero et al. (2015), and the fourth is considered a parameter that affects all the other metrics of this index (called ‘meta-parameter’, such as the concept of ‘meta-data’). The last three stressors named -Activities; Infrastructure and Anthropogenic threats-, were considered as tourist stressors with a relevant incidence in COVID-19 lockdown. The importance of making this analysis was to check how many anthropogenic activities were carried out on beaches during the lockdown, identifying specific potential pressures of tourist activity over these ecosystems. Each stressor had its survey format on the virtual platform Kobo Toolbox (https://www.kobotoolbox.org/), which was linked to the Kobo Collect application, where each researcher registered the data gathered on the field. The surveys were performed in the same three zones defined for the bioindicators (Supplementary 2).
2.4 Data processing, visualization, and statistical analyses
The data collected in the fieldworks was downloaded from the Kobo Toolbox Cloud and organized in a tidy structure (details about this structure are expanded by Wickham, 2014). We used bar graphs to visualize the frequency of occurrence of each indicator by beach zone. We pooled all of the data from beaches since the research focused on before and during effects of lockdowns rather than differences due to spatial variability. All data processing was done with the tidyverse packages (Wickham et al., 2019) of the statistical software R (R Core Team, 2019).
To evaluate potential changes in the environmental conditions of the beaches generated by the lockdown, a qualitative assessment instrument was designed based on four anthropogenic stressors (noise, odors, litters, and activities) and one general stressor for the biological component of the beach. Each indicator was rated on a semi-quantitative ordinal scale from 1 to 5, where 5 is the worst condition for that indicator in the beach. The characteristics of each value for each indicator are described in an assessment tool available as Supplementary material (Supplementary 3). This instrument allows controlling the subjectivity of assessment among researchers. The pre-COVID assessments were assigned based on literature, data, and observations made in previous studies. The assessments during lockdown were applied after visiting the beaches. We recognized the superiority of using quantitative data obtained with standardized procedures, such as the developed by the PROPLAYAS network (Botero et al., 2015) as well as the MBON Pole to Pole network (2019), instead of qualitative assessments. However, carrying out such quantitative methodologies proved infeasible during the lockdown. Restrictions on the number of researches and the time they could stay on site, as well as the impossibility to process chemical and biological samples in our institutional facilities suggests the efficacy of the semi-quantitative ordinal approach we adopted.
Valuations were arranged in a matrix with beaches as rows (before and during lockdown) and indicators as columns. Then, non-parametric multivariate statistics based on similarities were used to compare the beaches before and during the lockdown. We explored several similarity/distance indices before definitive analyzes (i.e., Gower, Bray-Curtis, χ2, Binomial deviance, Euclidean distances, Chord distance), all resulting in patterns of similarities with very high correlation (Spearman rank correlation coefficients all >0.9). For clarity of interpretation, we decided to use the Gower similarity index. This is a quantitative symmetric index that allows the use of qualitative descriptors, like those employed in this study (Legendre and Legendre, 2012). Pairwise similarities were estimated for each pair of rows, then an Analysis of Similarities (ANOSIM) based on 9999 permutations (Clarke, 1993) was performed to test the null hypothesis of differences in valuation of stressors and bioindicators before and during COVID-19 lockdown. A non-metric multidimensional scaling (nMDS) was plotted to represent the patterns of similarity between moments. The contribution of each indicator to the dissimilarity between moments was identified with the routine Similarity Percentage (SIMPER) (Clarke, 1993). These statistical analyses were done with PRIMER v7 (Clarke et al., 2014).
3 Results
3.1 Biological indicators recorded during the lockdown
We found the most bioindicators in the service zone of urban beaches. Thirteen of the fourteen groups of monitored bioindicators were present and the only one absent was seagrass on this beach zone. Crabs were found to be a widespread component of the fauna, particularly in the active zone. Most of the crabs belong to the species Ocypode quadrata and were present in 13 of 29 beaches (Fig. 2 ). Seabirds and non-seabirds were also recorded frequently in this zone (16/29 beaches), particularly Fregata magnificens, Leucophaeus atricilla, Phalacrocorax sp., and Larus delawarensis. Fauna was rare (found in 5 out of 20 beaches) in the rest zone, however, when some animals were recorded these were mostly domestic animals, seabirds and opportunistic birds. Turtle nests were registered on only three beaches, all in the rest zone. In this zone, beach grass was the most frequent bioindicator, present in 11 of 29 beaches. The occurrence of opportunistic species was higher in the service zone, especially pigeons (Columba livia), common grackles (Quiscalus mexicanus) and domestic species such as dogs (Fig. 2).Fig. 2 Bioindicators observed in 3 zones of 29 recreational beaches in Latin America during the COVID-19 lockdown. Bars represent the number of beaches presenting each category. Details with the full list of species are provided in Supplementary 4.
Bioindicators observed in 3 zones of 29 recreational beaches in Latin America during the COVID-19 lockdown. Bars represent the number of beaches presenting each category. Details with the full list of species are provided in Supplementary 4.
Fig. 2
Beach grass, shrubbery, and vines appeared frequently in the service zone, growing in areas of regular transit of people; these bioindicators were registered on 13 beaches. Trees were also prevalent, especially coconut palms, palm trees, and beach grape trees. Besides opportunistic birds, lizards and iguanas were the most frequent fauna bioindicators in the service area (Fig. 2). A higher beachgrass coverage was observed on many beaches and it was also noted that dunes were forming or increasing in size. This new habitat availability could lead to a possible improved condition for increased activity of ghost crabs and other species. Details of bioindicators are provided in Supplementary 4.
3.2 Changes on anthropogenic stressors during the lockdown
3.2.1 Noise
Unnatural noises were almost imperceptible on most beaches (Supplementary 5.1). The signs of its existence were concentrated on urban beaches, caused by motorized vehicles (such as motorcycles, cars, cleaning vehicles) and construction. Noises were sometimes detected from alarm sirens and music from restaurants in three urban beaches, as well as music from festivals, was perceived in one urban beach. Motor vehicles and watercraft noises were only recorded on three village-type beaches.
3.2.2 Odor
Unnatural odors were rarely perceived on studied beaches (Supplementary 5.2.A). With some exceptions, various types of unnatural odor associated with human activity were recognized in village and urban beaches (e.g., smoke, garbage, fuel). The odor sources reported were mainly waste containers and restaurants, closely followed by wrack in urban beaches. Nevertheless, none of the odor sources was reported in more than four beaches (13.33%) (Supplementary 5.2.B). Unnatural odor categories were not perceived in rural and resort beaches.
3.2.3 Litter
Litter was either absent or in exceptionally low abundance for most of the beaches studied (Supplementary 5.3.A). Potentially harmful litter such as broken glasses, glass bottles, syringes and knives appeared in low abundance in 10 beaches. Cigarette butts and polystyrene were similarly registered with low abundances in a few beaches. Conversely, gross and small vegetable items such as tree-trunks, branches, leaves and, stranded vegetation from shallow waters (e.g. seagrass and algae) dominated the large particulate matter for most of the beaches. This type of litter would not be causally related to human activities (Supplementary 5.3.A). The primary source of litter was associated with recreation in the active zone, while the secondary source corresponded to commercial facilities (Supplementary 5.3.B). Both results could indicate that some activities considered as litter sources were not interrupted during lockdown (e.g. transport, trading). However, it should be noted that some sources, such as commercial facilities, are only not related to the tourism activity on the beach.
3.2.4 User density
User density was low in most studied beaches. During the monitoring, less than five beaches (>15%) had visitors in the three beach zones (Fig. 3A). Fieldwork observations showed the overall number of users was <10, mainly on urban and village beaches, and in the active and rest beach zones (Fig. 3A). At least ten beaches (33%) had public visitors (local inhabitants), and seven had the presence of authorities during monitoring. A relevant finding was the spatial distribution of users on the beach, which showed clearly how the active zone was empty in many cases, while the service zone had the highest occupation.Fig. 3 Users density and activities in 29 recreational beaches in Latin America during the COVID-19 lockdown. (The data are presented considering the type of beach and the beach zone. In A, user density is represented as the frequency of beaches without people, few people (<10), and many people (11–20) for each type of user (V = visitors, P = peddling, A = authorities). In B, the frequency of beaches for each type of recreational activity. In C, the frequency of beaches for each type of commercial activity. In D, the frequency of beaches with maintenance activities. Empty panels in Resort and Rural beaches indicate that this zone could not be evaluated or was absent.)
Fig. 3
3.2.5 Activities
We observed recreational activities on 13 urban beaches in three different zones (Fig. 3A and B). The frequency of these activities changed according to the beach zone. For example, people running, swimming, or socializing were mainly registered in the active beach zone, while runners, children playing with sand, and people practicing sports were mostly registered in the rest beach zone. Recreational activities regarding people relaxing and socializing were mainly registered in the service zone (Fig. 3B). Commercial activities were practically not observed on the four types of beaches analyzed (Fig. 3C), while maintenance activities such as cleaning and vigilance were recorded in only eight urban beaches (Fig. 3D).
3.2.6 Infrastructure
This stressor included two types of infrastructure: first, those related directly with the COVID-19 measures, and second, buildings, promenades and structures constructed before the pandemic. The most common infrastructure presents on the beaches analyzed was the seafront. This type of infrastructure is referred to as a public walk located along the coast which is quite common in most of the Caribbean and Latin American beaches. Kiosks and low-rise facilities are usually frequent on the seafront at service zones on the beaches in normal conditions. However, during COVID-19 lockdown, only two urban beaches had facilities on the active beach zone, while ten beaches had the presence of low and mid-rise kiosks in the services zone and the seafront which were associated with commercial activities. During COVID-19 lockdown new infrastructures recently built with local materials (guano, wooden pitchforks) were present in the four types of beaches analyzed, with greater predominance in the village and urban beaches. Details are provided in Supplementary 5.4.
3.2.7 Anthropogenic threats
Activities that affect the marine and terrestrial zones of the beach environment were analyzed as anthropogenic threats. Only six beaches showed signs of such. The threats were invasive species, vehicles on the beach, and sand extraction. The latter was evident on four urban beaches. The potential for invasive species and wastewater discharges appeared at seven beaches: four urban and three villages. Details are provided as Supplementary 5.5.
3.3 Potential changes in beach condition due to COVID-19 lockdown
The similarity patterns recorded in the qualitative assessment were significantly different between before/during lockdown conditions (ANOSIM test, R = 0.875, p < 0.0001). The magnitude of the ANOSIM R statistic was relatively high (the maximum possible value is 1), indicating a solid differentiation between these conditions. The dissimilarities between beaches considering the condition were plotted in Fig. 4A. Besides, the multivariate dispersion was greater for the condition “before” than “during” the lockdown, indicating the variability of the characteristics of use and conservation of the beaches in the region. The dispersion is lower during the lockdown because most of the indicators were consistently scored with the minimum possible values (Fig. 4B). That is, the various beaches converged to a better state of health during the lockdown.Fig. 4 Analyses of the qualitative assessments of bioindicators and anthropogenic stressors in 29 beaches in Latin America, before and during the COVID-19 lockdown. In A, a non-Metric Multidimensional Scaling based on Gower similarities between beaches considering the five indicators of biological conditions and anthropogenic stressors. Labels indicate the beach's name and in parentheses the letter of the corresponding country. Several beaches overlap, so the labels were arranged or grouped to facilitate graphic representation. In the upper left corner, the ANOSIM test result is indicated. In B, the frequency of beaches for each type indicator, before and during the lockdown. The indicators get worse as they approach 5. The definitions of each value for each indicator are detailed in Supplementary material 3.
Analyses of the qualitative assessments of bioindicators and anthropogenic stressors in 29 beaches in Latin America, before and during the COVID-19 lockdown. In A, a non-Metric Multidimensional Scaling based on Gower similarities between beaches considering the five indicators of biological conditions and anthropogenic stressors. Labels indicate the beach's name and in parentheses the letter of the corresponding country. Several beaches overlap, so the labels were arranged or grouped to facilitate graphic representation. In the upper left corner, the ANOSIM test result is indicated. In B, the frequency of beaches for each type indicator, before and during the lockdown. The indicators get worse as they approach 5. The definitions of each value for each indicator are detailed in Supplementary material 3.
Fig. 4
The SIMPER analysis shows that the indicators with greater change before/during the lockdown were the Noise and Activities, both contributing to 51% of differences (Table 1 ). Both indicators, on average, decrease from 3.4 to 3.2 and from 1.5 to 1.3, respectively. As expected, most beaches presented little if any human activity during the lockdown. Surveillance and maintenance accounted for most of the activity.Table 1 Contribution of each indicator to the differences in the perception of changes in Latin-American beaches before and during the COVID-19 lockdown, using the Similarity Percentage routine (SIMPER) to the Gower dissimilarities and the one-way ANOSIM test.
Table 1Indicator Average before Average during Sq.Dist/SD Contrib% Cum.%
Noise 3.4 1.3 1.33 27.5 27.5
Activities 3.2 1.5 1.02 23.3 50.8
Bioindicators 3.4 2.1 0.95 23.0 73.8
Odors 2.6 1.0 1.56 15.0 88.8
Litter 2.7 1.5 1.05 11.2 100.0
Before COVID-19, most beaches scored high for user density and activities such as the presence of off-road vehicles. Similarly, unnatural noise was more intense and frequent for all beaches (grades 3 and 4) before lockdown, mostly music emanating loudspeakers and off-road vehicles. Litter and odors accounted for the lowest contribution to differences (Table 1). During the lockdown, most of the beaches had no or little litter contrasting with the extensive refuse typical of pre-lockdown conditions. However, this indicator was very variable because some beaches were consistently under cleaning programs, even during the lockdown, such as beaches in Cancún, México. Similarly, unnatural odors such as smoke, garbage, fuel, body lotions, food, smoke from grills or campfires, among others were quite frequent (grade 3) before COVID-19, but practically absent (grade 1) during the lockdown.
The biological component contributes to 23% of differences in scores between before and during lockdown, decreasing on average from 3.4 to 2.1. During lockdown most of the beaches scored a qualitative biological assessment of 1 or 2, indicating a presence of dune vegetation, coconut palm, wildlife (reptiles, mammals, birds), coastal crustaceans (ghost crab, blue crab) and turtle nests. Conversely, before the COVID-19 lockdown, many beaches had low presence or absence of natural biological traits, but the presence of opportunistic fauna (grades 3, 4, and 5).
4 Discussion
4.1 Signs of bioindicator recovery due to the COVID-19 lockdown
The potential recovery of some bioindicators was observed in most beaches. The consistent presence of crabs and, specifically ghost crabs, in the active zone was evident. It is highlighted not only due to their reported occurrence but also because researchers noticed crabs being more active than usual and even more confident in human presence. Ghost crabs are known to be important bioindicators (Blankensteyn, 2006) and species of the genus Ocypode are semiterrestrial invertebrates inhabiting areas from the waterline to the dunes (Lucrezi and Schlacher, 2014). The recovery of this species supports the notion of resilient capacity as suggested by Stelling-Wood et al. (2016) for Australian urban beaches. Likewise, the appearance of beach grass, as well as the extension of vines and shrubs in areas where they were not seen before could also be considered an indicator of the dune vegetation resilience (Rickard et al., 1994). In essence, pre-lockdown human disturbance maintains the beach community in an earlier sere of succession.
The records opportunistic birds like grackles were frequent in our beaches during lockdown. (Gilby et al., 2021) found great increases of Torresian crows (Corvus orru) on Australian beaches during lockdown. Hence, the recovery of some bioindicators demonstrates the resilience capacity of tourist beaches. Despite variability in the data, indicators of ecosystem health clearly improved during the lockdown. We note that positive changes in the presence, abundance, diversity, and activity of the main flora and fauna occurred during lockdown. We can say this with confidence as we are well acquainted with the beaches studied in our respective countries.
Managers and conservationists should give special attention the presence of protected and sensitive species in high tourist beaches as highlighted by Boudouresque et al. (2017) and Steven and Castley (2013). A remarkable finding in our study was the presence of turtles and iguanas during lock-down, including threatened species. Records in these species during lockdown occurred in well-established reserves and protected areas, as in Puerto Rico and Panama. This suggests insufficiency in the level of protection afforded by these reserves during normal times. This finding should cause mangers to reconsider the extent of human activity allowed in protected areas. Anthropogenic activities may threaten the population stability of sensitive species (Brown and McLachlan, 2002; Defeo et al., 2009; Canteiro et al., 2018). Hockings et al. (2020) explored the factors that reserves, and protected areas could use to build a more sustainable future for people and nature after COVID-19 pandemic.
4.2 Decline of human stressors due to the COVID-19 lockdown
Consistent with our hypothesis, we documented an overall reduction in anthropogenic stressors during the lockdown. This suggests that the restrictions and confinement measures proved effective in, limiting and preventing access and human activities, and hence generating a positive effect on environmental conditions. Positive effects derived from the COVID-19 lockdown are documented for different environments (Manenti et al., 2020; Derryberry et al., 2020). Considering all the measured stressors, we believe that physical disturbances generated by human activities had the most remarkable effects on biological beach traits and functioning. The improvement of biological indicators during the COVID-19 lockdown supports this contention. It is widely documented that the constant disturbances generated by pedestrian traffic and off-road vehicles on the beach frequently alter the natural substrate stability and prevents the establishment and growth of dune vegetation, as well as of benthic macrofauna (Bom and Colling, 2020; Rickard et al., 1994). The lack of tourists also generated a great change in the landscape of several beaches: Acapulco (Mexico), Salinas (Ecuador) and Barcelona (Spain) now look cleaner and with crystal waters (Zambrano-Monserrate et al., 2020). Even the physical disturbance caused by just a few tourists might produce a loss of biodiversity on beaches, especially in sensitive microscopic animals (meiofauna) that live buried in the sand (Martínez et al., 2020). This effect may cascade up to higher trophic levels such as macrofaunal that depend on the sensitive meiofauna (Afghan et al., 2020), thus generating a negative effect on the ecosystem functioning (Bracken et al., 2008).
The COVID-19 lockdown drove a reduction of noise, odor, and litter on the studied beaches. Different types of anthropogenic stressors produce different impacts on biota (Birk et al., 2020), and noise is an emerging pollutant that affects marine organisms (Peng et al., 2015). Noise during lockdown diminished due to lower human presence. This included reductions in recreational and commercial activities, maritime traffic, fishing, and construction. Unnatural sounds alter avian communities, reducing nesting species richness, and contributing to the success of urban-adapted birds (Francis et al., 2009). It is also a problem for the population and the environment. Noise pollution is associated with various diseases and altered ecosystems (Zambrano-Monserrate and Ruano, 2019).
The absence of open restaurants, food sellers, and cooking tourists on the beaches also reduced the presence and intensity of unnatural odors and organic wastes. This may explain diminished opportunistic fauna on beaches during lockdown. The food orders may have served as attractants prior to lockdown. However, we think some odors would not be associated with lockdown effect.
Another aspect to highlight was the low density of litter on all beaches during the lockdown. It may indicate that the few residues found are remnants of recreational activities carried out before the pandemic, as well as residues found in the active zone are subject to intense mobility due to tides and coastal currents brought from other locations. On the other hand, on many beaches, constant cleaning was perceived, especially on the shores of the Riviera Maya, Mexico. Although this cleaning activity is common in the region throughout the year, the absence of tourists due to lockdown contributed to keeping it cleaner. However, cleaning is usually supported using rakes, and in some places with sand cleaning machines. This disrupts infauna habitat, preventing establishment of ghost crabs and various other fauna (Ocaña et al., 2020). Studies carried out by Lucrezi et al. (2009) and Schlacher et al. (2016) using ghost crabs as indicators confirm human disturbances contribute to the loss of habitat for these species. Although the harmful effects of litter on marine life are well documented (Kühn et al., 2015), beach raking is not a solution that protects biological diversity of the beaches despite this activity removes beach litter.
DPPE (disposable personal protective equipment) and other single-use plastic items have been promoted during the pandemic as a mandatory policy in public spaces, including beaches. These items are emerging pollutants of beaches, such as the discarded masks found in Praia da Ribeira, Brazil. Masks, contaminated gloves, used or expired medications plus other items could be mixed with ordinary wastes, magnifying marine litter issues and the consequences to marine biodiversity and population health (Canning-Clode et al., 2020). The safe management and disposal of common and biohazard waste on beaches will be complex during a pandemic. This situation is especially worrying because the industry has taken the opportunity to repeal disposable bag bans, even though single-use plastic can still harbor viruses and bacteria (Bir, 2020). The reopening of the beaches could make this type of litter more frequent. We encourage monitoring this situation since many beaches already allow access to tourists.
4.3 Perspectives and projections for biological conservation and management
Results from this study suggest that tourist beaches offer important biodiversity that requires conservation initiatives (McLachlan et al., 2013). Some of the studied beaches also support essential ecological processes such as turtle nesting and spawning areas, bird reproduction, nesting and feeding, crab habitat, and increased habitat for plants that in some cases were not recorded before the pandemic. Caracas beach (Puerto Rico), located inside the Vieques U.S. Fish and Wildlife National Reserve, recorded turtle nesting zones for endangered species such as the Green Turtle (Chelonia mydas). Itapuã beach in Brazil showed spawning areas for sea turtles, while El Estero Santa Catalina beach in Panama as part of the buffer zone at Coiba National Park supported a high diversity of wildlife.
Recreation is the primary service provided by tourist beaches to society (Bessa et al., 2014; Zhang et al., 2015) and conservation aims are secondary. However, tourist sandy beaches are not only crucial for recreation; they also need conservation (Jaramillo, 2012; McLachlan et al., 2013). These ecosystems also provide food, economic services, biodiversity, maintenance and regulation of natural processes, aesthetic and cultural values and many other benefits and services (Schlacher et al., 2014a). Therefore, it is necessary to propose mechanisms for setting conservation targets (Harris et al., 2014) for those beaches that meet selected ecological criteria. The recovery capacity observed in several beaches related to wildlife repopulation could be a criterion. Opportunistic species were common on urban and village beaches, so any conservation action should consider the interactions with native and non-native species.
On tourist beaches, initiatives with a focus on wildlife conservation may be new and attractive. However, they may be complex and difficult to develop, since tourist beach management traditionally focuses on protecting infrastructure and sediment structural maintenance (Schlacher et al., 2006). The conservation of habitats, species, and ecological functions is often a minor aspect (Peterson and Bishop, 2005). However, our results would allow thinking that this new scenario called ‘anthropause’ opens a window for biological conservation purposes that promote unique global experiments in nature (Bates et al., 2020; Rutz et al., 2020). Even more so for urban tourist beaches where the possibility to conduct studies without anthropogenic stressors are impossible in practice.
New initiatives and perspectives should define first the conservation targets focusing on species, communities, functions, processes, services, and ecosystems. Nature-based tourism proves to be a sustainable activity under certain circumstances (Winter et al., 2020), providing education and good practices for visitors and local communities as well as economic and political support for wildlife conservation (Wilson and Tisdell, 2003). New concepts like Wildlife Conservation Tourism (W.C.T.) may also be implemented as an ecological conservation strategy that prioritizes endangered species through meaningful interactions with tourists (Boyes, 2016). The current pandemic offers to all actors a unique opportunity to design and consolidate the transition towards a greener and more sustainable tourism (Ioannides and Gyimóthy, 2020), rethinking the way to promote beach tourism based on recreation but also being mindful of the protection of coastal ecosystems biodiversity.
This study evidences an increase in frequency and magnitude of living organisms in Latin American beaches during the lockdown and reveals a reduction in anthropogenic stressors. Our results support the findings of Martínez et al. (2020), who consider it essential to restrict access to beaches in tourist areas to preserve biodiversity. Therefore, better practices of sustainable tourism (i.e., maximum load capacity and better waste disposal) must emerge and be implemented to minimize the impact of human activities. Executive actions should be adjusted to the different regional contexts considering active public participation of the local communities (Milanés et al., 2020). Regarding beach management, the current legal framework (GORC, 2000) and coastal policies of each Latin American country should be considered, and some of these regulations must be adapted (Milanés et al., 2019). The natural characteristics of the beaches and their level of anthropogenic impact should also be considered to carry out effective planning of the coastal areas (Moraes and Milanés, 2020; Batista-Milanés, 2018). The general lockdown limited the possibility of performing exhaustive and more quantitative analyses (Manenti et al., 2020), yet such monitoring is important to verify the trends revealed in this study.
5 Conclusions
The new environmental setting derived from the COVID-19 lockdown generated conspicuous changes in the biological community of beaches previously impacted by human activity that should be assessed with caution. Yet, our results correspond to only a “snapshot” of an unprecedented condition that should be followed over time. Long-term and more robust studies focused on vulnerable species, functioning, and ecosystem services would allow knowing if tourist beaches can become appropriate places for effective biological conservation. Despite the extensive history of human activity on these beaches prior to lockdown, these ecosystems displayed the ability to recover, with increases in biodiversity and system functionality as response to lower environmental alteration by stressors. This supported our initial hypothesis. This suggests an impressive resilience of these environments not previously evaluated to support biodiversity and possibly its conservation.
The positive implications of how most of the bioindicators changed during lockdown are likely to be temporary, and it is currently not clear how conservation will fare in the aftermath of the pandemic. We hope that information gathered in this study may contribute for conservation strategies of sensitive ecosystems like sandy beaches of high tourist use. Adequate monitoring of bioindicators is necessary for a more effective coastal management system, which seeks to consider the conservation of biodiversity, ecosystem balance, and the maintenance of essential services for humankind such as leisure and recreation provided by tourist beaches.
The following are the supplementary data related to this article.Supplementary 1
Description of beaches.
Supplementary 1
Supplementary 2
Kobo collect forms.
Supplementary 2
Supplementary 3
Beach evaluation criteria
Supplementary 3
Supplementary 4
Species list
Supplementary 4
Supplementary 5
Anthropogenic stressors.
Supplementary 5
Supplementary A
Table 1. Information summary of beaches analyzed.
Supplementary A
CRediT authorship contribution statement
E.H. Soto: Conceptualization, Methodology, Data curation, Writing - Original draft, Writing - Review & editing, Project administration;
C.M. Botero: Conceptualization, Methodology, Validation, Investigation, Data curation, Writing - Original draft, Writing - Review & editing;
C.B. Milanés: Conceptualization, Methodology, Data curation, Writing - Original draft, Writing - Review & editing, Visualization;
Rodríguez-Santiago: Investigation, Data curation, Writing - Original draft, Writing - Review & editing, Visualization.
M. Palacios-Moreno: Investigation, Data curation, Writing - Original draft, Visualization.
E. Díaz-Ferguson: Investigation, Writing, Review & editing.
Y.R. Velázquez: Investigation.
Abbehusen: Investigation, Writing - Original draft.
E. Guerra-Castro: Formal analysis, Investigation, Data Curation, Writing - Original draft, Writing - Review & editing, Visualization.
N. Simoes: Investigation, Writing - Review & editing.
M. Muciño-Reyes: Investigation.
J.R. Souza Filho: Investigation, Writing - Review & editing.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
Authors want to express their gratitude to the PROPLAYAS Network because all of them are members. Thanks to this network we were able to carry out this research, working as a harmonic research team. Authors also thank Richard Primack, Amanda Bates, Carlos Duarte and Vincent Devictor, editors of Special Issue for considering this research. Many thanks to all local authorities in each country for allowing our field works under confinement restrictions. Our warmest acknowledgements José Julio Casas, Paco López, Omar Anchaluisa, Alondra Gabriel, Erick Bermúdez and Camila Palacios for field works support and assistance. Finally we thanks four anonymous reviewers and Dr. Benjamin Cuker for improving the English language, editing and style of the manuscript. E.Guerra-Castro was supported by funding from DGAPA Universidad Nacional Autónoma de México, México, Grant ID PAPIIT-IA206320.
1 The international group named by the acronym “Proplayas” is an Ibero-American network formed by members of the academia, civil society, scientists, activists, officials, and businesspeople. All of them working on beach issues (www.proplayas.org).
==== Refs
References
Afghan A. Cerrano C. Luzi G. Calcinai B. Puce S. Pulido Mantas T. Roveta C. Di Camillo C.G. Main anthropogenic impacts on benthic macrofauna of sandy beaches: a review J. Mar. Sci. Eng. 8 2020 405
Amyot J. Grant J. Environmental Function Analysis: a decision support tool for integrated sandy beach planning Ocean Coast. Manag. 102 2014 317 327 10.1016/j.ocecoaman.2014.10.009
Andrés M. Barragan J.M. Garcia-Sambria J. Relationships between coastal urbanization and ecosystems in Spain CITIES 68 2017 8 17
Araújo M. Silva-Cavalcanti J. Costa M. Anthropogenic litter on beaches with different levels of development and use: a snapshot of a coast in Pernambuco (Brazil) Front. Mar. Sci. 5 2018 233 10.3389/fmars.2018.00233
Ariza E. Sardá R. Jiménez J.A. Mora J. Ávila C. Beyond performance assessment measurements for beach management: application to Spanish Mediterranean beaches Coast. Manag. 36 1 2007 47 66 10.1080/08920750701682023
Bates, A.E., Primack, R., Moraga, P., Duarte, C., 2020. COVID-19 pandemic and associated lockdown as a “Global Human Confinement Experiment” to investigate biodiversity conservation. Biol. Conserv. 248, 108,665. doi.org/10.1016/j.biocon.2020.108665.
Batista-Milanés C. Coastal risk Finkl C. Makowski C. Encyclopedia of Coastal Science 2nd Encyclopedia of Earth Sciences Series 1 2018 Springer Nature Cham, Switzerland 524 534
Becken S. Climate Change Impacts on Coastal Tourism. CoastAdapt Impact Sheet 2016 National Climate Change Adaptation Research Facility Gold Coast 6
Bessa F. Gonçalves S.C. Franco J. André J. Cunha P. Marques J. Temporal changes in macrofauna as response indicator to potential human pressures on sandy beaches Ecol. Indic. 41 2014 49 57
Bir B. https://www.aa.com.tr/en/health/single-use-items-not-safest-option-amidcovid-19/1787067 2020
Birk S. Chapman D. Carvalho L. Impacts of multiple stressors on freshwater biota across spatial scales and ecosystems Nat. Ecol. Evol. 4 2020 1060 1068 10.1038/s41559-020-1216-4 32541802
Blankensteyn A. O uso do caranguejo maria-farinha Ocypode quadrata (Fabricius, 1787) (Crustacea, Ocypodidae) como indicador de impactos antropogênicos em praias arenosas da Ilha de Santa Catarina, Santa Catarina, Brasil Rev. Bras. de Zool. 23 3 2006 870 876 10.1590/S0101-81752006000300034
Bom F. Colling L. Impact of vehicles on benthic macrofauna on a subtropical sand beach Mar. Ecol. 41 2020 e12595 10.1111/maec.12595
Botero C. Pereira C. Tosic M. Manjarrez G. Design of an index for monitoring the environmental quality of tourist beaches from a holistic approach Ocean Coast. Manag. 108 2015 65 73 10.1016/j.ocecoaman.2014.07.017
Botero C.M. Cabrera J.A. Zielinski S. Tourist beaches Finkl C. Makowski C. Encyclopedia of Coastal Science Encyclopedia of Earth Sciences Series 2018 Springer Cham 10.1007/978-3-319-48,657-4_401–1
Boudouresque C.F. Ponel P. Astruch P. Barcelo A. Blanfuné A. Geoffroy D. Thibaut T. The high heritage value of the Mediterranean sandy beaches, with a particular focus on the Posidonia oceanica “banquettes”: a review Sci. Rep. Port-Cros Natl. Park 31 2017 23 70
Boyes K. Applying Wildlife Conservation Tourism to Marine Endangered Species: Identifying Indicators for Triple Bottom Line Sustainability 2016 University of Washington (Master of Marine Affairs Thesis, 78 pages)
Bracken M. Friberg S. Gonzalez-Dorantes C. Williams S. Functional consequences of realistic biodiversity changes in a marine ecosystem Proc. Natl. Acad. Sci. U. S. A. 105 2008 924 928 18195375
Brown A. McLachlan A. Sandy shore ecosystems and the threats facing them: some predictions for the year 2025 Env. Conserv. 29 2002 62 77 10.1017/S037689290200005X
Canning-Clode J. Sepúlveda P. Almeida S. Monteiro J. Will COVID-19 containment and treatment measures drive shifts in marine litter pollution? Front. Mar. Sci. 7 2020 691 10.3389/fmars.2020.00691
Canteiro M. Córdova-Tapia F. Brazeiro A. Tourism impact assessment: a tool to evaluate the environmental impacts of touristic activities in Natural Protected Areas Tour. Manag. Pers. 28 2018 220 227 10.1016/j.tmp.2018.09.007
Clarke K. Non-parametric multivariate analyses of changes in community structure Austral. Jour. Ecol. 18 1993 117 143 10.1111/j.1442-9993.1993.tb00438.x
Clarke K. Gorley R. Somerfield P. Warwick R. Change in Marine Communities: An Approach to Statistical Analysis and Interpretation Plymouth 2014 Ltd PRIMER-E
Cristiano S.C. Rockett G.C. Portz L.C. Souza Filho J.R. Beach landscape management as a sustainable tourism resource in Fernando de Noronha Island (Brazil) Mar. Poll. Bull. 150 2020 1 13 10.1016/j.marpolbul.2019.110621
Davis R.A. Human impact on coasts Finkl C.W. Makowski C. Encyclopedia of Coastal Science Encyclopedia of Earth Sciences Series 2019 Springer Cham 10.1007/978-3-319-93,806-6_175
Defeo O. McLachlan A. Schoeman D.S. Schlacher T.A. Dugan J. Jones A. Lastra M. Scapini F. Threats to sandy beach ecosystems: a review. Est. Coast Shelf Sci. 81 2009 1 12 10.1016/j.ecss.2008.09.022
Derryberry E.P. Phillips J.N. Derryberry G.E. Blum M.J. Luther D. Singing in a silent spring: birds respond to a half-century soundscape reversion during the COVID-19 shutdown Science 340 2020 574 579
Dodds R. Holmes M. Beach tourists; what factors satisfy them and drive them to return Ocean Coast. Manag. 168 2019 158 166 10.1016/j.ocecoaman.2018.10.034
Dodds R. Holmes M.R. Preferences at city and rural beaches: are the tourists different? Jour. of Coast Res. 36 2 2019 393 402 10.2112/JCOASTRES-D-19-00048.1
Francis, C., Ortega, C., Cruz, A., 2009. Noise pollution changes avian communities and species interactions. Curr. Biol. 19 (16), 1415–1419. DOI: doi:10.1016/j.cub.2009.06.052.
Gaynor K. Hojnowski C. Carter N. Brashares J. The influence of human disturbance on wildlife nocturnality Science 360 2018 1232 1235 10.1126/science.aar7121 29903973
Gheskiere, T., Vincx, M., Weslawski, J.M., Scapini, F., Degraer, S., 2005. Meiofauna as descriptor of tourism-induced changes at sandy beaches. Mar Environ Res. 60 (2), 245–65. DOI: 10.1016/j.marenvres.2004.10.006. Epub 2004 Dec 7. PMID: 15757751.
Gilby B.L. Henderson C.J. Olds A.D. Ballantyne J.A. Bingham E.L. Elliott B.B. Jones T.R. Kimber O. Mosman J.D. Schlacher T.A. Negative ecological consequences of animal redistribution on beaches during COVID-19 lockdown Biol Conserv. 253 2021 108926
GORC Decreto-Ley 212. Gestión de la Zona Costera. Official Gazette of the Republic of Cuba. Citma 2000 Cuba La Habana 18
Green R. Giese M. Negative effects of wildlife tourism on wildlife. Chapter 5, part 2004 Wildlife Tourism. Impacts, Management and Planning. Editor, Karen Higginbottom In 2
Guerra-Castro E. Hidalgo G. Castillo R. Muciño-Reyes M. Noreña-Barroso E. Quiroz-Deaquino J. Mascaro M. Simoes N. Sandy beach macrofauna of Yucatán State (Mexico) and oil industry development in the Gulf of Mexico: first approach for detecting environmental impacts Front. Mar. Sci. 7 2020 589656 10.3389/fmars.2020.589656
Halpern B.S. Walbridge S. Selkoe K.A. Kappel C.V. Micheli F. D’Agrosa C. Bruno J.F. Casey K.S. Ebert C. Fox H.E. Fujita R. Heinemann D. Lenihan H.S. Madin E.M.P. Perry M.T. Selig E.R. Spalding M. Steneck R. Watson R. A global map of human impact on marine ecosystems Science 319 2008 948 10.1126/science.1149345 18276889
Halpern, B.S., Frazier, M., Afflerbach, J., Lowndes, J.S., Micheli, F., O'Hara, C., Scarborough, C., Selkoe, K.A., 2019. Recent pace of change in human impact on the world's ocean. Sci. Rep. 9, 11,609. doi.org/10.1038/s41598-019-47,201-9.
Harley C. Hughes A. Hultgren K. Miner B. Sorte C. Thornber C. Rodriguez L. Tomanek L. Williams S. The impacts of climate change in coastal marine systems Ecol. Lett. 9 2006 228 241 10.1111/j.1461-0248.2005.00871.x 16958887
Harris L. Nel R. Holness S. Sink K. Schoeman D. Setting conservation targets for sandy beach ecosystems Est. Coast. Shelf. Sci. 150 2014 45 57
Hockings, M., Dudley, N., Elliott, W., Ferreira, M., Mackinnon, K., Pasha, M., Phillips, A., Stolton, S., Woodley, S., Appleton, M., Chassot, O., Fitzsimons, J., Galliers, C., Golden Kroner, R., Goodrich, J., Hopkins, J, Jackson, W., Jonas, H., Long, B., Yang, A. 2020. COVID-19 and protected and conserved areas. Parks. 26, 7–24. 10.2305/IUCN.CH.2020.PARKS-26-1MH.en.
Ioannides D. Gyimóthy S. The COVID-19 crisis as an opportunity for escaping the unsustainable global tourism path Tour. Geogr. 22 3 2020 624 632 10.1080/14616688.2020.1763445
Jaramillo E. Ecological implications of extreme events on exposed sandy beaches: insights from the 2010 Chilean earthquake. In: VI th International Sandy Beach Symposium. Public Presentation, SouthAfrica June 26th 2012
Jarratt D. Davies N. Planning for climate change impacts: coastal tourism destination resilience policies Tour. Plann. and Develop 17 4 2019 423 440 10.1080/21568316.2019.1667861
Jones A. Gladstone W. Hacking N. Australian sandy-beach ecosystems and climate change: ecology and management Zoologist 34 2 2007 190 202 10.7882/AZ.2007.018
Kühn S. Bravo E. van Franeker J. Deleterious effects of litter on marine life Bergmann M. Gutow L. Klages M. Marine Anthropogenic Litter 2015 Springer International Publishing Cham 75 116
Legendre P. Legendre L. Numerical Ecology 2012 Elsevier Amsterdam
Lucrezi S. Schlacher T.A. The ecology of ghost crabs Oce. and Mar. Biol: An Ann. Rev. 52 2014 201 256
Lucrezi S. Schlacher T.A. Walker S. Monitoring human impacts on sandy shore ecosystems: a test of ghost crabs (Ocypode spp.) as biological indicators on an urban beach Environ. Monit. Assess. 152 2009 413 424 10.1007/s10661-008-0326-2 18563608
Manenti, R., Mori, E., Di Canio, V., Mercurio, S., Picone, M., Caffi, M., Brambilla, M., Ficetola, G., Rubolini, D., 2020. The good, the bad and the ugly of COVID-19 lockdown effects on wildlife conservation: insights from the first European locked down country. Biol. Conserv. 249, 108,728. doi.org/10.1016/j.biocon.2020.108728.
Marion J. Impacts to Wildlife: Managing Visitors and Resources to Protect Wildlife. Contributing Paper. Prepared for the Interagency Visitor Use Management Council, March 2019 Edition One 2019
Marion J. Leung Y. Eagleston H. Burroughs K. A review and synthesis of recreation ecology research findings on visitor impacts to wilderness and protected natural areas Jour. of Forest. 114 3 2016 352 362 10.5849/jof.15-498
Marshall F. Banks K. Cook G. Ecosystem indicators for Southeast Florida beaches Ecol. Indic. 44 2014 81 91 10.1016/j.ecolind.2013.12.021
Martínez A. Eckert E.M. Artois T. Careddu G. Casu M. Curini-Galletti M. Gazale V. Gobert S. Ivanenko V. Jondelius U. Marzano M. Pesole G. Zanello A. Todaro M.A. Fontaneto D. Human access impacts biodiversity of microscopic animals in sandy beaches Commun. Biol. 3 2020 175 10.1038/s42003-020-0912-6 32313088
Martins G.A.L. A macrofauna bentônica das praias arenosas expostas do Parque Nacional de Superagüi – PR: Subsídios ao Plano de Manejo Dissertação (Mestrado em Ecologia e Conservação) 2007 Setor de Ciências Biológicas da Universidade Federal do Paraná. Curitiba Programa de Pós Graduação em Ecologia e Conservação
MBON Pole to Pole Sampling protocol for assessment of marine diversity on sandy beaches Marine Biodiversity Observation Network Pole to Pole of the Americas 2019 10.25607/OBP-665 (14 pp.)
McLachlan A. Defeo O. Jaramillo E. Short A.D. Sandy beach conservation and recreation: guidelines for optimising management strategies for multi-purpose use Oc. Coast. Manag. 71 2013 256 268 10.1016/j.ocecoaman.2012.10.005
Mendoza-González G. Martínez M. Guevara R. Pérez-Maqueo O. Garza-Lagler M. Howard A. Towards a sustainable sun, sea, and sand tourism: the value of ocean view and proximity to the coast Sustainability 10 2018 1012 10.3390/su10041012
Milanés C. Pereira C. Botero C. Improving a decree law about coastal zone management in a small island developing state: the case of Cuba Marine Policy 101 2019 93 107 10.1016/j.marpol.2018.12.030 (March)
Milanés C. Planas J. Pelot R. Núñez J. A new methodology incorporating public participation within Cuba's ICZM program. Oce. Coast Manag. 186 105 2020 101 10.1016/j.ocecoaman.2020.105101
Moraes, F., Milanés, C., 2020. Os limites espaciais da zona costeira para fins de gestão a partir de uma perspectiva integrada. Cap. 1 pp. [22–50]. In: Souto, R.D. (org.) Gestão Ambiental e Sustentabilidade em Áreas Costeiras e Marinhas: Conceitos e Práticas. Vol. 1. Rio de Janeiro: Instituto Virtual para o Desenvolvimento Sustentável - IVIDES.org, 2020. [260 p].
Niefer I.A. Análise do perfil dos visitantes das Ilhas do Superagüi e do Mel: Marketing como instrumento para um turismo sustentável Tese (Doutorado em Engenharia Florestal) 2002 Universidade Federal do Paraná. Curitiba Setor de Ciências Agrárias
Ocaña F. de Jesús A. Hernández H. Co-occurring factors affecting ghost crab density at four sandy beaches in the Mexican Caribbean Reg. Stu. in Mar. Sci. 36 101 2020 310 10.1016/j.rsma.2020.101310
Olds A.D. Vargas-Fonseca E. Connolly R.M. Gilby B.L. Huijbers C.M. Hyndes G.A. Layman C.A. Whitfield A.K. Schlacher T.A. The ecology of fish in the surf zones of ocean beaches: a global review Fish and Fisher. 19 2018 78 89
Peng, C., Zhao, X., Liu, G., 2015. Noise in the sea and its impacts on marine organisms. Int. J. Environ. Res. Public Health, 12, 12,304–12,323; doi:10.3390/ijerph121012304.
Peterson C. Bishop M. Assessing the environmental impacts of beach nourishment. BioScience 55 887e896 2005
R Core Team R: A Language and Environment for Statistical Computing 2019 R Foundation for Statistical Computing Vienna, Austria URL https://www.R-project.org/
Reyes-Martínez M.J. Lercari D. Ruíz-Delgado M.C. Sánchez-Moyano J.E. Jiménez-Rodríguez A. Pérez-Hurtado A. García-García F.J. Human pressure on sandy beaches: implications for trophic functioning Estuar. Coasts 38 2015 1782 1796 10.1007/s12237-014-9910-6
Rickard, C., McLachlan, A., Kerley, G.,1994. The effects of vehicular and pedestrian traffic on dune vegetation in South Africa. Oc. Coast. Manag. 23 (3), 225–247. DOI: doi:10.1016/0964-5691(94)90021-3.
Rutz C. Loretto M. Bates A. Davidson S. Duarte C. Jetz W. Johnson M. Kato A. Kays R. Mueller T. Primack R. Ropert-Coudert Y. Tucker M. Wikelski M. Cagnacci F. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife Nat. Ecol. Evo. 4 2020 1156 1159 10.1038/s41559-020-1237-z
Schlacher T. Thomson L. Beach recreation impacts benthic invertebrates on ocean-exposed sandy shores Biol. Conserv. 147 1 2012 123 132 10.1016/j.biocon.2011.12.022
Schlacher T. Schoeman D. Lastra M. Jones A. Dugan J. Scapini F. McLachlan A. Neglected ecosystems bear the brunt of change Ethol. Ecol. Evol. 18 2006 349e351
Schlacher T. Nielsen T. Weston M. Human recreation alters behaviour profiles of non-breeding birds on open-coast sandy shores Estuar. Coast. Shelf Sci. 118 2013 31 42 10.1016/j.ecss.2012.12.016
Schlacher T. Schoeman D. Jones A. Dugan J. Hubbard D. Defeo O. Peterson C. Weston M. Maslo B. Olds A. Scapini F. Nel R. Harris L. Lucrezi S. Lastra M. Huijbers C. Connolly R. Metrics to assess ecological condition, change, and impacts in sandy beach ecosystems Jour. of Env. Manag. 144 2014 322 335 10.1016/j.jenvman.2014.05.036
Schlacher T.A. Jones A.R. Dugan J.E. Weston M.A. Harris L.L. Schoeman D.S. Hubbard D. Scapini F. Nel R. Lastra M. McLachlan A. Peterson C.H. Open-coast sandy beaches and coastal dunes. Chapter 5. In: Lockwood, J.L., Maslo, B. (Eds.), Coastal Conservation. Cambridge University Press, Cambridge, pp 37e94 2014
Schlacher T.A. Lucrezi S. Connolly R.M. Peterson C.H. Gilby B.L. Maslo B. Olds A.D. Walker S.J. Leon J.X. Huijbers C.M. Weston M.A. Turra A. Hyndes G.A. Holt R.A. Schoeman D.S. Human threats to sandy beaches: a meta-analysis of ghost crabs illustrates global anthropogenic impacts Estuar. Coast. and Shelf Sci. 169 2016 56 73
Souza Filho J.R. Silva I.R. Nunes F.N. Avaliação qualitativa dos serviços ecossistêmicos oferecidos pelas praias da APA Lagoa Encantada/Rio Almada, Bahia, Brasil Caminhos de Geografia Uberlândia 20 72 2019 15 32 10.14393/RCG207241182
Souza, J.L, Silva, I.R., 2015. Avaliação da qualidade ambiental das praias da ilha de Itaparica, Baía de Todos os Santos, Bahia. Soc. & Nat. 27 (3), 469–484. DOI: 10.1590/1982-451,320,150,308.
Stelling-Wood T.P. Clark G.F. Poore A.G.B. Responses of ghost crabs to habitat modification of urban sandy beaches Mar. Env. Res. 116 2016 32 40 10.1016/j.marenvres.2016.02.009 26970686
Steven R. Castley J.G. Tourism as a threat to critically endangered and endangered birds: global patterns and trends in conservation hotspots Biodivers. Conserv. 22 2013 1063 1082 10.1007/s10531-013-0470-z
Suciu M. Tavares D. Costa L. Silva M. Zalmon I. Evaluation of environmental quality of sandy beaches in southeastern Brazil. Mar. Poll Bull. 199 2017 133 142 10.1016/j.marpolbul.2017.04.045
Veloso V.G. Silva E.S. Caetano C.H.S. Cardoso R.S. Comparison between the macroinfauna of urbanized and protected beaches in Rio de Janeiro State, Brazil Biol. Conserv. 127 2006 510 515
Vilar de Araujo C.C. Melo Rosa D. Fernandes J.M. Densidade e distribuição espacial do caranguejo Ocypode quadrata (Fabricius, 1787) (Crustacea, Ocypodidae) em três praias arenosas do Espírito Santo, Brasil Biotemas 21 2008 73 80 10.5007/2175-7925.2008v21n4p73
Wickham H. Tidy data J. Stat. Softw. 59 10 2014 1 23 10.18637/jss.v059.i10 26917999
Wickham H. Averick M. Bryan J. Chang W. McGowan L. François R. Welcome to the Tidyverse Journal of Open Source Software 4 43 2019 1686 10.21105/joss.01686
Williams A. Micallef A. Beach Management: Principles and Practice 2009 Earthscan Publishers London
Wilson C. Tisdell C. Conservation and economic benefits of wildlife-based marine tourism: sea turtles and whales as case studies Human Dimensions of Wildlife: An International Journal 8 1 2003 49 58 10.1080/10871200390180145
Winter P.L. Selin S. Cerveny L. Bricker K. Outdoor recreation, nature-based tourism and sustainability. Sustainability 12 1 2020 81
Zambrano-Monserrate, M., Ruano, M., 2019. Does environmental noise affect housing rental prices in developing countries? Evidence from Ecuador. Land Use Policy 87,104,059.
Zambrano-Monserrate M. Silva-Zambrano C. Ruano M. The economic value of natural protected areas in Ecuador: a case of Villamil Beach National Recreation Area Ocean Coast. Manag. 157 2018 193 202 10.1016/j.ocecoaman.2018.02.020
Zambrano-Monserrate M. Ruano M. Sánchez-Alcalde L. Indirect effects of COVID-19 on the environment Sci.Total Env. 728 138 2020 813 10.1016/j.scitotenv.2020.138813
Zhang F. Wang X.A. Nunes P. Ma C. The recreational value of gold coast beaches, Australia: an application of the travel cost method Eco. Serv. 11 2015 106 114 10.1016/j.ecoser.2014.09.001
Zielinski S. Botero C.M. Beach tourism in times of COVID-19 pandemic: critical issues, knowledge gaps and research opportunities Int. J. Environ. Res. Public Health 17 2020 7288 10.3390/ijerph17197288 33036176
| 0 | PMC9746931 | NO-CC CODE | 2022-12-15 23:21:58 | no | Biol Conserv. 2021 Mar 21; 255:108972 | utf-8 | Biol Conserv | 2,021 | 10.1016/j.biocon.2021.108972 | oa_other |
Subsets and Splits