Activity Number:
|
92
|
Type:
|
Topic Contributed
|
Date/Time:
|
Monday, August 12, 2002 : 10:30 AM to 12:20 PM
|
Sponsor:
|
Biopharmaceutical Section*
|
Abstract - #301134 |
Title:
|
A More Powerful Average Bioequivalence Analysis for the 2x2 Crossover
|
Author(s):
|
Devan Mehrotra*+ and Catalina Stefanescu
|
Affiliation(s):
|
Merck Research Laboratories and Cornell University
|
Address:
|
785 Jolly Rd, Bldg C, Blue Bell, Pennsylvania, 19422, USA
|
Keywords:
|
ANCOVA ; Bootstrap ; Biased estimator
|
Abstract:
|
Power/sample size calculations for 2x2 crossover studies to establish average bioequivalence are usually done under the belief that the true (geometric) means of the test and control treatments are identical. If we really believe that the true means are identical, then standard analyses are arguably inefficient, because they fail to capitalize on our prior belief. On the other hand, even small departures from the identical means assumption could make the actual power to show bioequivalence notably smaller than the planned power. We propose an innovative analytic approach that resolves these problematic issues. Our approach uses simple tools from the areas of covariate adjustment and bootstrapping. Simulation results and a numerical example are used to illustrate the improved power and utility of the proposed new methodology.
|