Comparison of FINDRISC and the Framingham Model for Predicting CHD Risk: Results from the Kharameh Cohort Study.

Abstract

Background: Coronary heart disease is a leading cause of death worldwide, making early risk identification essential. This study compares the predictive performance of FINDRISC and the Framingham model in the Kharameh Cohort.

Methods: This prospective cohort study included 9,469 participants aged 40-70 years from the Kharameh cohort (a branch of the PERSIAN study) who were free of CHD at baseline and were followed for 6 years to identify incident CHD events. Data on FINDRISC components, as well as laboratory and demographic variables, were collected at enrollment. Logistic regression models were developed using the total FINDRISC score and its individual items, and their predictive performance was compared with the Framingham non‑laboratory risk score using 10‑fold cross‑validation and metrics including AUC‑ROC, sensitivity, and specificity.

Results: During a six‑year follow‑up of 9,469 participants (mean age 51.6 years, 55.6% female), 540 incident CHD events (5.7%) were identified. The total FINDRISC score was significantly higher in individuals who developed CHD (11.35 vs. 9.59, P<0.001). However, the total FINDRISC score alone showed weak discriminatory power (AUC=58.6%, 95% CI: 55.9-61.3%). An item-based FINDRISC model demonstrated improved performance (AUC=71%, 95% CI: 68.9-73%), and further enhancement with gender, systolic blood pressure, and total cholesterol yielded the highest predictive accuracy (AUC=75.2%, 95% CI: 73%-77%). The non‑laboratory Framingham risk score (AUC=73.3%, 95% CI: 71.4-75.1%) showed similar discriminatory performance compared with the enhanced item‑based FINDRISC models.

Conclusion: The enhanced item-based FINDRISC model, incorporating gender and systolic blood pressure showed comparable performance to the Framingham non-laboratory model, providing a simple, accessible screening tool for the Iranian population.