Conditional Partial Likelihood for Event Data with Sparsely Measured Longitudinal Information
主 题: Conditional Partial Likelihood for Event Data with Sparsely Measured Longitudinal Information
报告人: Ye Shen, Assistant Professor（Department of Epidemiology and Biostatistics, University of Georgia）
时 间: 2014-06-13 14:00-15:00
地 点: 理科1号楼1479（统计中心活动）
Recurrent event data are quite common in biomedical and epidemiological studies. A large portion of these data also contain sparsely observed information on surrogate markers. It has been shown that popular methods using Cox model with longitudinal outcomes as time dependent covariates may lead to biased results, especially when longitudinal outcomes are measured with error. Thus, how to appropriately incorporate longitudinal information into recurrent event models becomes an important research question. Our research is motivated by a psychiatry study conducted at The Yale Stress Center (YSC). In the study, we collect both longitudinal and recurrent event data. Based on a joint modeling framework which models the correlation between processes using latent random effect terms, we propose a two-stage conditional partial likelihood (CPL) method to model the rate function of recurrent event process conditional on sparsely observed longitudinal information.