|
Activity Number:
|
239
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #300585 |
|
Title:
|
Optimal Inferences for Proportional Hazards Model with Parametric Covariate Transformations
|
|
Author(s):
|
Chunpeng Fan*+ and Jason P. Fine and Jong-Hyeon Jeong
|
|
Companies:
|
sanofi-aventis and University of Wisconsin-Madison and University of Pittsburgh
|
|
Address:
|
Mail Stop: BRW M-203A, 1041 Rt 202-206, Bridgewater, NJ, 08807-0800,
|
|
Keywords:
|
Cox proportional hazards model ; Parametric covariate transformation ; Optimal test ; Semi-parametric model
|
|
Abstract:
|
Traditional Cox model assumes a log-linear relationship between covariates and the underlying hazard function. However, the linearity may be invalid in real data. We propose a generalized Cox model which uses parametric covariate transformations to recover it. While the proposed generalization may seem simple, the inferential issues are challenging due to the loss of identifiability under no effects of transformed covariates. Optimal tests are derived for certain alternatives. Rigorous parameter inference is established under regularity conditions and non-zero transformed covariate effects. The estimates perform well in simulation studies with realistic sample size and the proposed tests are more powerful than the usual or sup partial likelihood ratio test. A real data is used to illustrate the model building and the better fit of the proposed model, comparing to traditional Cox model.
|
