Department of Biostatistics Seminar/Workshop Series

Utilizing external incidence information for risk prediction with cohort data: Application to WHI Colorectal Cancer Data

Dandan Liu, PhD

Assistant Professor of Biostatistics, Department of Biostatistics, Vanderbilt University School of Medicine

Wednesday, October 31, 1:30-2:30pm, MRBIII Room 1220

Accurate and individualized risk prediction is valuable for successful management of chronic diseases such as cancer and cardiovascular diseases. Large cohort studies provide valuable resources for building risk prediction models, as the risk factors are collected at the baseline and subjects are followed over time until the occurrence of diseases or termination of the study. However, many cohorts are assembled for particular purposes, and hence their baseline risk may differ from the general population. Moreover for rare diseases the baseline risk may not be estimated reliably based on cohort data only. In this paper, we propose to make use of external disease incidence rate for estimating the baseline risk, which increases efficiency. We proposed two sets of estimators and established the asymptotic distributions for both of them. Simulation results show that the proposed estimators are more efficient than the methods that don't utilize the external incidence rates. We applied the method to a large cohort study, the Women's Health Initiative, for estimating colorectal cancer risk.