Predictive Accuracy of the Framingham Risk Score in the Republic of Korea

BACKGROUND: Multivariate risk prediction models are recommended for assessing cardiovascular disease (CVD) risk and developing effective management plans. The widely used Framingham Risk Score (FRS) requires validation before application, as the original cohorts primarily comprised White, middle-class individuals in the United States. OBJECTIVES: To evaluate the FRS's predictive accuracy over 10 years within the Korean National Health Insurance Service-National Health Screening Cohort. METHODS: This retrospective observational study involved 483,643 eligible participants. Discrimination and calibration were assessed using the c statistic, calibration slope, and intercept, supplemented by graphical calibration plots. RESULTS: Over 10 years, 23,681 men and 16,904 women experienced a CVD event. The FRS showed good discrimination in both men and women. It overestimated CVD risk for men, while showing good calibration in women. DISCUSSION: The observed discrepancy in calibration may be related to differences in risk factor profiles and CVD event rates between the original Framingham cohorts and the Korean validation cohort. In the absence of a Korean-specific model, the FRS remains a practical option for use in primary care. However, to enhance accuracy of risk prediction, the development and validation of a population-specific CVD risk model for Korea are recommended.