|
Activity Number:
|
220
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2004 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #302143 |
|
Title:
|
Optimal Three-stage Study Design for Single-arm Phase II Clinical Trials
|
|
Author(s):
|
Muhammad Jalaluddin*+ and Enayet Talukder
|
|
Companies:
|
Pfizer Inc. and Pfizer Inc.
|
|
Address:
|
50 Pequot Ave., New London, CT, 06320,
|
|
Keywords:
|
clinical trials ; multistage clinical trials ; exact binomial distribution ; sample size
|
|
Abstract:
|
Optimal two-stage study designs are commonly used in single-arm phase II clinical trials. One of the objectives of these designs is to minimize exposure of a new therapy when there is no beneficial treatment effect. It is achieved by choosing sample sizes and decision rules such that the expected sample size under null hypothesis is minimum. This expected sample size could still be large. This raises ethical questions because the efficacy level of a new therapy is usually unknown at the early stage of development and can potentially be less efficacious than standard therapy. We have addressed these issues by introducing an optimal three-stage study design. This design requires pre-specification of minimum and desired efficacy levels. We initiate subsequent stages of a study only if data from completed stage(s) suggest better efficacy of the new therapy compared to pre-specified minimum and desired efficacy levels. Therefore, this design allows early termination of a study when the new therapy performs poorly. We have demonstrated the application of this design and have compared it with the existing designs under various set-ups.
|
