Evaluation of cardiovascular risk assessment models with respect to the clinical interpretation of atherosclerosis in a different type II diabetes cohort.

Abstract

Epidemiological studies on cardiovascular risk developed many assessment models which are widely available for the public use. As many arterial occlusive diseases are developed from atherosclerosis in their early stage, it is meaningful to evaluate such models with respect to the clinical interpretation of atherosclerosis so as to promote the preventive care of vascular diseases. Our study aims to make use of the data collection form from the Hong Kong Chinese type II diabetes to evaluate and compare the performance of the risk assessment models of ARIC, FHS, UKPDS using ROC curve.We found that ARIC's Stroke model gives the best performance whose AUC is 0.646 in model for Black. UKPDS's Stroke has the lowest AUC, 0.497. It was found that ARIC model for the Black Americans has superior performance with respect to the cohort in Hong Kong based on the ROC analysis.