|
Activity Number:
|
27
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, August 2, 2009 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #303847 |
|
Title:
|
Comparison of the Branching Group Sequential Design with Frequent Adaptation Methods in an Equivalence Study
|
|
Author(s):
|
Yevgen Tymofyeyev*+ and James A. Bolognese and Keaven Anderson
|
|
Companies:
|
Merck & Co., Inc. and Cytel, Inc. and Merck & Co., Inc.
|
|
Address:
|
RY 34-A304, Rahway, NJ, 07065,
|
|
Keywords:
|
adaptive design ; group sequential design ; dose-ranging study
|
|
Abstract:
|
We consider a dose-ranging study where the objective is to find a dose level that provides a response such that the corresponding 95% confidence interval is within given clinically meaningful limits. The proposed design is based on group sequential methodology and constituted by a series of decision rules organized as a tree to define early stopping and switching dose levels as the trial progresses. Frequent (or continuous) adaptation type of designs such as "up-and-down" methods are also appropriate for dose-ranging studies, so we compare these approaches in terms of type I and II errors, sample size, probability of correct dose identification, estimation bias and operational complexity. The branching group sequential design is much simpler to implement due to less frequent adaptation and has competitive performance characteristics compared to designs using continuous adaptation.
|
