Researchers developed a machine learning model to predict coronary heart disease (CHD) risk using common health indicators like blood pressure, age, cholesterol, and fasting glucose. The model was built and tested using data from the Framingham Heart Study, which included 4,240 people for internal validation, and a separate hospital cohort of 200 people for external validation.
The model performed well in both groups. In the internal validation, it achieved an AUC of 0.977 and accuracy of 94.2%. In the external hospital cohort, the AUC was 0.929 with 88.5% accuracy. These results suggest the model can reliably identify people at risk for CHD using information that is routinely collected during doctor visits.
No safety concerns were reported, as this was a data analysis study. The model is not yet ready for widespread clinical use, but it shows promise as a tool to help doctors assess cardiovascular risk without extra tests. More research is needed to confirm its benefits in real-world settings.
For now, this study highlights how existing health data might be used to improve heart disease prediction. If you have concerns about your heart health, talk to your doctor about your individual risk factors.