|
Activity Number:
|
471
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Wednesday, August 1, 2007 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #308882 |
|
Title:
|
Bootstrap Test for Assessing the Efficacy of Drug Combinations
|
|
Author(s):
|
Liwen Xi*+ and Guoyong Jiang and Jianjun Li
|
|
Companies:
|
Johnson & Johnson PRD and Cephalon, Inc. and Merck & Co., Inc.
|
|
Address:
|
Welsh and McKean Roads, Spring House, PA, 19477-0776,
|
|
Keywords:
|
bootstrap-min test ; min-test ; composite hypothesis
|
|
Abstract:
|
In clinical trials, a new therapy is often comprised of a combination of two or more drugs. It is necessary to demonstrate that the combination therapy is more efficacious than each of its drug constituents alone. A classical statistical method for assessing the superiority of a combination therapy is the min test, whose critical value is determined conservatively. We propose the bootstrap-min test for assessing the efficacy of a combination therapy which entails using the bootstrap technique, resulting in an increase in power. In addition, this new test does not have any distribution requirements for the outcome variable of interest; therefore, it offers more flexibility compared with the classical min test. Power and type-I error rate comparisons are carried out using simulations between the new bootstrap and currently available tests.
|
