|
Abstract:
|
Risk prediction models for survival outcomes are widely applied in medical research to predict future risk for the occurrence of the event. For patients with progressive disease, their biomarker data are measured repeatedly over time. To improve the prediction of disease progression, many dynamic prediction models have been developed which make predictions on real-time basis. As dynamic prediction model updates an individual's risk prediction over time based on new measurements, it is often important to examine how accurate the model-based predictions are at different measurement time and prediction time. In this talk, we propose a two-dimensional area under curve measure for dynamic prediction models and develop associated estimation and inference procedures. The consistent estimation procedures are discussed under two types of biomarker measurement schedules: regular visits and irregular visits. The model parameters are estimated consistently by maximizing a pseudo partial-likelihood function. We apply the proposed method to a renal transplantation study to evaluate and compare the discrimination performance of dynamic prediction models based on longitudinal biomarkers.
|
