Why the study?
The study was conducted to investigate which clinical features most strongly predict coronary artery disease.
Observational (n=297)
Machine learning analysis of the Cleveland dataset demonstrates that chest pain type, thallium scintigraphy, maximum heart rate, and ST depression are stronger predictors of coronary artery disease than serum cholesterol.
Answers are grounded in this paper and cite their sources.
Hypothesis-generating that clinical variables outperform cholesterol for CAD prediction; prospective validation needed before clinical adoption.
This study applies exploratory data analysis, statistical significance testing (Mann-Whitney U with effect sizes), and machine learning classification to the UCI Cleveland Heart Disease dataset (n=297) to investigate which clinical features most strongly predict coronary artery disease. Key findings:- Chest pain type, thallium scintigraphy result, maximum heart rate, and ST depression are the strongest predictors- Serum cholesterol shows only marginal significance (p=0.047, r=0.134, small effect) confirming its weak predictive value- Random Forest achieved 85% accuracy (95% CI: 76.63–93.33%) with AUC 0.94 and F2 score 0.803- Logistic Regression achieved 83% accuracy (95% CI: 73.33–91.67%) with AUC 0.95 and F2 score 0.797 Full analysis code and notebook available at: https://github.com/lydiaobeng/heart-disease-analysis
No takes yet. Share an insight, caveat, or question.
Lydia Obeng Bema (2026) conducted an observational in coronary artery disease (n=297). Machine learning classification (Random Forest and Logistic Regression) was evaluated on Prediction of coronary artery disease (AUC 0.94, 95% CI 76.63-93.33).
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: