From: PHHP-ANNOUNCE [mailto:[log in to unmask]] On Behalf Of Parks-James,Renee
Sent: Wednesday, April 16, 2014 1:47 PM
To: . PHHP-Announce
Subject: Department of Biostatistics Spring Seminar Series - Friday, April 18, 2014
[cid:[log in to unmask]]
College of Public Health and Health Professions
College of Medicine
Department of Biostatistics
Spring 2014 Seminar Series
Friday, April 18, 2014
10: 00 a.m. - 11:00 a.m.
CTRB, Room 5235
Department of Biostatistics and Bioinformatics
A Generalized Framework for Censored Quantile Regression Based on Counting Process
Censored quantile regression has become a useful alternative to standard analysis of time-to-event data. In this paper, we present a counting process based perspective for a popular censored quantile regression model. The new perspective motivates a generalized framework for censored quantile regression, which accommodates the important survival analysis scenario with recurrent events. The proposed methodology for recurrent events data retains the appealing features of censored quantile regression in interpretation, flexibility and computation. Furthermore, it can flexibly accommodate more complex observation windows of recurrent events as often encountered in observational studies. A general framework for theory, inference and computation is developed to unify the new approach for recurrent events data with some existing censored quantile regression methods. As another useful contribution of this work, we also propose a sample-based covariance estimation procedure, which can be applied to censored quantile regression and the proposed generalization to recurrent events. We demonstrate the utility of our proposals via simulation studies and an application to a dataset from the US Cystic Fibrosis Foundation Patient Registry (CFFPR).