|
Activity Number:
|
334
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #309328 |
|
Title:
|
Midcourse Selection of the Primary Endpoint in Clinical Trials
|
|
Author(s):
|
Zhilong Yuan*+ and Yang Song and Xiaolong Luo and George Chi
|
|
Companies:
|
Johnson & Johnson PRD and Johnson & Johnson PRD and Johnson & Johnson PRD and Johnson & Johnson PRD
|
|
Address:
|
920 Route 202 South, Raritan, NJ, 08860,
|
|
Keywords:
|
Endpoint Selection ; Adaptive Design ; Clinical Trial ; Type I Error Rate
|
|
Abstract:
|
Mid-course selection of endpoints is often encountered in clinical trials. Several group sequential and adaptive methods have been developed to address this issue but few has explicitly incorporated the selection rules, hence the loss of power. We formulate endpoint selection in the closed testing framework, and characterize the corresponding type I errors and powers. We then propose flexible alpha allocation and apply it to the global test to optimize the power, which is subsequently illustrated in a cancer clinical trial. A generalization to conditional error functions is also discussed which allows the incorporation of other flexible design features such as sample size re-estimation.
|
