|
Activity Number:
|
243
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308484 |
|
Title:
|
Semiparametric Analysis for Recurrent Event Data with Time-Dependent Covariates and Informative Censoring
|
|
Author(s):
|
Chiung-Yu Huang*+ and Jing Qin and Mei-Cheng Wang
|
|
Companies:
|
National Institutes of Health and National Institutes of Health and Johns Hopkins University
|
|
Address:
|
6700A Rockledge Drive, Bethesda, MD, 20817,
|
|
Keywords:
|
Comparable recurrence times ; Frailty ; Informative censoring ; Pairwise pseudolikelihood ; Proportional rate model
|
|
Abstract:
|
We present a semiparametric model of recurrent event data that allows censoring time and recurrent event process to be correlated via frailty. This flexible framework includes both time-dependent and time-independent covariates, while leaving the distributions of frailty and censoring time unspecified. We estimate the effect of time-dependent covariates by constructing a pseudo likelihood based on comparable pairs of event times. For the estimation of the baseline cumulative rate function and the regression coefficients of time-independent covariates, we derive a modified product-limit estimator with bias correction in risk sets, and solve estimating equations formulated based on expected number of observed recurrent events. Numerical studies demonstrate that the proposed methodology performs well for practical sample sizes.
|
