|
Activity Number:
|
286
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Biometrics Section
|
| Abstract - #307059 |
|
Title:
|
Bayesian Decision-Theoretic Adaptive Designs for Clinical Trials
|
|
Author(s):
|
Yi Cheng*+
|
|
Companies:
|
Indiana University South Bend
|
|
Address:
|
1700 Mishawaka Ave., South Bend, IN, 46530,
|
|
Keywords:
|
clinical trial ; group sequential methods ; adaptive design ; Bayesian analysis ; utility functions ; conditional power
|
|
Abstract:
|
In this talk, I propose a Bayesian decision theoretical approach by assessing utilities on group sequentially collected data. The loss function consists of the penalties of making wrong decisions and the cost of enroll patients in the trial. We incorporate estimated conditional power or predicted power into the decision procedure to accommodate desire for good performance in terms of frequentist characteristics regarding the rates of type I and type II errors. The performance of the proposed design is evaluated by numerical studies and by comparison to existing group sequential methods.
|
