|
Activity Number:
|
199
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #310234 |
|
Title:
|
An Empirical Test for Using a Cox Proportional Hazards Model with Time-Varying Hazard Ratios in the Analysis of Cardiac Events of Hypertensive Patients
|
|
Author(s):
|
D. Das Purkayastha*+ and Lynn Ashcraft
|
|
Companies:
|
Novartis Pharmaceuticals and Novartis Pharmaceuticals
|
|
Address:
|
44 Center Grove Road, Randolph, NJ, 07869,
|
|
Keywords:
|
Time-varying hazard ratio ; Cox proportional hazards model ; Stationarity ; Autocorrelation
|
|
Abstract:
|
A conventional Cox proportional hazards model postulates hazard ratios to be constant over time. If we consider time dependent covariates in the model with an assumption of time-varying hazard ratios as functions of weighted lag effects of change-point hazards, then the proportionality is disturbed in the model. In such models, in order to justify the clinical data and the model used for analysis, it is imperative to test for the stationarity of such a series of hazard ratios. This paper explains their stationarity conditions for their applications in the analyses of hazard models with an empirical exposition with 24 weeks of data of cardiac events from 200 hypertensive patients. Along with the empirical illustration, the paper also provides an analytical check of the applicability of a Cox model and stationarity conditions for hazard ratios with clinical data.
|
