Dynamic risk prediction for cardiovascular disease: An illustration using the ARIC Study

Abstract

We review methods for using repeated measurements of a time-varying risk predictor in dynamic risk prediction. We compare how the landmarking and joint modeling approaches can incorporate information from repeated measurements of systolic blood pressure in cardiovascular risk prediction and illustrate the methods using data from the Atherosclerosis Risk in Communities Study. We assess predictive accuracy using dynamic measures of discrimination and calibration.