---
dataset_info:
  features:
  - name: pmid
    dtype: int64
  - name: mesh_terms
    dtype: string
  - name: mesh_terms_ui_list
    sequence: string
  - name: title
    dtype: string
  - name: abstract
    dtype: string
  - name: index
    dtype: float64
  - name: in_total_pubmed
    dtype: bool
  - name: text
    dtype: string
  splits:
  - name: train
    num_bytes: 6677736915
    num_examples: 2116914
  download_size: 3420092013
  dataset_size: 6677736915
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
---

# PubMed Cardiovascular Subset Dataset

This dataset is a curated subset of PubMed articles filtered to include those related to Cardiovascular Disease and its associated subcategories ("children"). It is designed to support research and analysis in cardiovascular health, epidemiology, and related biomedical fields by providing structured metadata along with the article text.

## Dataset Overview

- **Source:** Derived from PubMed, the dataset includes articles indexed with relevance to cardiovascular conditions.
- **Scope:** Focused on cardiovascular disease, including its broader categories and subtypes, ensuring that users can perform detailed analyses on relevant biomedical literature.
- **Curation:** Articles were selected using MeSH term filtering, where both the primary term for Cardiovascular Disease and its children terms were used to narrow the selection.

## Dataset Structure

The dataset is structured with the following columns:

- **pmid:**  
  A unique PubMed Identifier for each article. This numeric code allows easy cross-referencing with the full PubMed record.

- **mesh_terms:**  
  A string containing Medical Subject Headings (MeSH) terms with their corresponding codes (e.g., "D000369:Aged, 80 and over"). These terms classify the content and facilitate detailed filtering based on biomedical topics.

- **mesh_terms_ui_list:**  
  A list of MeSH term unique identifiers extracted from the `mesh_terms` field. This format aids in programmatic filtering and analysis by isolating the identifier codes.

- **title:**  
  The title of the research article. It offers a brief overview of the study focus, allowing users to quickly gauge the article's relevance.

- **abstract:**  
  A summary of the article, including background, methods, results, and conclusions. This field provides a concise yet informative snapshot of the research content.

- **index:**  
  Position in pubmed_en. In case of not being contained, it is null.

  
- **in_total_pubmed:**  
  A boolean indicator that shows whether the article is part of the broader PubMed_en dataset.

  
- **text:**  
  A consolidated field combining key elements (often the title and abstract) into one text block.

  
## Example Record

```json
{
  "pmid": 36475546,
  "mesh_terms": "D000369:Aged, 80 and over; D006801:Humans; D005260:Female; D008297:Male; D000072657:ST Elevation Myocardial Infarction; D000088642:Nonagenarians; D012189:Retrospective Studies; D017052:Hospital Mortality; D062645:Percutaneous Coronary Intervention; D006761:Hospitals",
  "mesh_terms_ui_list": [
    "D000369",
    "D006801",
    "D005260",
    "D008297",
    "D000072657",
    "D000088642",
    "D012189",
    "D017052",
    "D062645",
    "D006761"
  ],
  "title": "In-hospital outcomes in nonagenarian patients undergoing primary percutaneous coronary intervention.",
  "abstract": "BACKGROUND\nThe aim of the present analysis was to evaluate the incidence and predictors of in-hospital adverse outcomes in nonagenarian patients undergoing primary percutaneous coronary intervention (pPCI) for ST-segment elevation myocardial infarction (STEMI)...",
  "index": null,
  "in_total_pubmed": false,
  "text": "In-hospital outcomes in nonagenarian patients undergoing primary percutaneous coronary intervention. BACKGROUND\nThe aim of the present analysis was to evaluate the incidence and predictors of in-hospital adverse outcomes in nonagenarian patients undergoing primary percutaneous coronary intervention (pPCI) for ST-segment elevation myocardial infarction (STEMI)..."
}
```