Comparison of Framingham Risk Score and Native Model of Cardiovascular Risk Prediction Among Residents of Mashhad Metropolis, Iran, 2015

Cardiovascular diseases are the leading cause of adult mortality in many developing countries. This study aims to compare the estimation of the ten-year relative risk of cardiovascular events using the Framingham criteria with a native model.

This population-based cross-sectional study was conducted in 2014, focusing on the adult population (≥16 years) of Mashhad. Stratified random cluster sampling was employed to gather participants' information based on Framingham's criteria. Data mining, utilizing the decision tree algorithm design, was evaluated using Rapidminer v5.3 software and the cross-validation method.

Out of 2978 individuals, 1930 (64.9%) were women and 1041 (35.1%) were men, with a mean age of 43.5±14.7. Applying the Framingham criteria, the ten-year risk levels of cardiovascular disease were estimated as follows: 77.8% at a low-risk level, 13.4% at a medium-risk level, and 8.8% at a high-risk level. Regarding data mining, model number (1) achieved an accuracy of 79.56%, indicating that the predicted risk levels using the Framingham algorithm matched the observed values at 95.24% for the low-risk level, 90.8% for the medium-risk level, and 33.13% for the high-risk level. As for model number (2), an accuracy of 82.78% was obtained, with the matching values being 98.20% for the low-risk level, 0.42% for the medium-risk level, and 53.01% for the high-risk level.

The Framingham criteria demonstrate limited effectiveness in predicting medium and high-risk levels in the Mashhad population. According to the local model, smoking and high blood pressure in adulthood are the most significant factors in predicting the risk of cardiovascular diseases in young individuals.