Data Sheet 1_Development and validation of a risk score heatmap for in-hospital adverse ev

Zizhu Lian developed a 10-variable prediction model for in-hospital adverse events in older adults with acute myocardial infarction. The model was trained on a retrospective cohort of 4,897 patients admitted between January 2019 and December 2024, of whom 434 (8.9%) experienced the composite endpoint. The tool, visualized as a risk score heatmap, achieved an area under the ROC curve of 0.808.

Use Cases

• Predicting in-hospital adverse events based on 10 clinical variables including age, vital signs, and lab results.
• Developing bedside risk stratification tools for older adults with acute myocardial infarction.
• Training machine learning models for clinical outcome prediction using features like NT-proBNP, eGFR, and neutrophil-to-lymphocyte ratio.
• Validating risk scores for composite endpoints involving death, ventricular arrhythmias, cardiogenic shock, or acute heart failure.

Strengths

• Model is based on a cohort of 4,897 patients, providing a substantial sample size.
• The final model incorporates 10 readily available clinical variables, suggesting practical applicability.
• The model demonstrated good discrimination with an AUC of 0.808 (95% CI 0.786–0.830).

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the underlying dataset is unknown, which may limit suitability assessment.
• The data file is a 220.5 KB DOCX document, which is a tiny dataset and may not contain the raw data.

Provenance

Source

Zizhu Lian via figshare.

Collection Method

Retrospective cohort study of consecutive patients aged >60 years with AMI.

Time Range

January 2019 to December 2024.

Freshness

Last updated 2026-05-04 05:33:55; freshness should be verified.