|
Activity Number:
|
199
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Monday, July 30, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #308284 |
|
Title:
|
The Impact of Proportional Hazards Assumption on the Late Onset Survival Data
|
|
Author(s):
|
Guoguang Ma*+ and Kathy Harris
|
|
Companies:
|
Amgen Inc. and Amgen Inc.
|
|
Address:
|
1120 Veterans Blvd, South San Francisco, CA, 94080,
|
|
Keywords:
|
survival analysis ; proportional hazards assumption ; Cox model ; log-rank test ; clinical trials
|
|
Abstract:
|
The log-rank and Cox tests are powerful under the assumption of proportional hazards that the relative risk remains constant over time. While these tests can be well powered to detect some differences between treatment groups that do not satisfy the assumption, they can have poor power to detect differences when cumulative incidence curves that are initially equal but later diverge or that initially diverge but later approach one another. This issue is usually raised in the analysis of clinical trials data when an investigational drug has a delayed treatment effect. In this work, we evaluate the impact of the proportional hazards assumption on the late onset survival data through simulation studies where we assume exponential distributions for the survival time and uniform distributions for the dropout time. The log-rank and Cox tests are also compared to the logistic regression model.
|
