An Asymptotically Distribution-Free Test to Assess Exchangeability of a Test Treatment and the Standard Treatment

• Activity Number: 468
• Type: Contributed
• Date/Time: Wednesday, August 4, 2010 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Jiuzhou Wang*+ and Steven Chung-Kuei Chang
• Companies: Cephalon Inc. and Cephalon Inc.
• Keywords: Bioequivalence test ; Exchangeability ; Distribution-free ; U-Statistics

Abstract

In the bioequivalence test setting, the null hypothesis is typically that the difference of two treatment means is either smaller than a pre-specified lower equivalence limit or larger than a pre-specified upper equivalence limit. Alternatively, the null hypothesis can be stated in the form of stochastic inequalities to assess the exchangeability of a test treatment and the standard treatment. In this article, we propose an asymptotically distribution-free test to assess the exchangeability of two treatments. We show that the test statistic is asymptotically normally distributed and demonstrate the large sample properties with Monte Carlo simulation. We also compare the proposed test procedure to other methods including the tolerance interval approach and the non-central t-test under various assumptions.