|
Activity Number:
|
239
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 5, 2008 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #301928 |
|
Title:
|
Adjust the Analysis of Group Sequential Trials Using the Bayesian Dynamic Survival Model Combined with Computation Techniques
|
|
Author(s):
|
Jianghua He*+ and Matthew Mayo
|
|
Companies:
|
University of Kansas Medical Center and The University of Kansas Medical Center
|
|
Address:
|
Department of Biostatistics, Kansas City, KS, 66210,
|
|
Keywords:
|
non-proprotional hazards ; Bayesian Dynamic Survival Model ; multiple imputation ; group sequential trials
|
|
Abstract:
|
When the proportional hazards assumption is not satisfied, the Cox model is still often used in medical research. The estimate of treatment effect under such condition depends heavily on the censoring pattern of the data. For survival-analysis-based monitoring of group sequential trials, the large scale of censoring due to staggered entry or loss-to-follow-up may strongly affect the results. We proposed to adjust the effect of censoring using different methods including adjusted weighted average and missing data imputation techniques. The adjusted estimates make full use of the information at the interim analysis stage and have better interpretations. A simulation study is applied to demonstrate the performances of these methods. The adjusted analysis is illustrated with a gastric cancer trial.
|
