|
Activity Number:
|
284
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, July 31, 2007 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #308497 |
|
Title:
|
Early Termination of Two-Stage, Phase II, One-Sample Cancer Clinical Trials
|
|
Author(s):
|
William E. Barlow*+
|
|
Companies:
|
Cancer Research and Biostatistics
|
|
Address:
|
1730 Minor Ave, Seattle, WA, 98101,
|
|
Keywords:
|
Phase II ; clinical trial ; early termination ; negative binomial
|
|
Abstract:
|
Two-stage Phase II single sample designs (Simon, 1989) can result in early termination if response is less than expected. If the trial proceeds to Stage 2, the MLE of the response rate is biased , but a UMVUE estimator can be constructed (Jung & Kim, 2004). We propose a fixed sample size for the first stage, but a variable sample size for the second stage. The trial terminates at the maximum allowed sample size or when the number of responses results in rejection of the null hypothesis in favor of the alternative. We use a modification of the Jung & Kim estimator that allows for a truncated negative binomial model in the second stage resulting in an unbiased estimator. Exact confidence intervals may be computed. This design has an expected sample size smaller or equal to that of the fixed design, but the same power and test size.
|
- The address information is for the authors that have a + after their name.
- Authors who are presenting talks have a * after their name.
Back to the full JSM 2007 program |
