Testing Hypotheses of Nonzero Differences in Means: From Westlake to Wellek, and Future Research Prospects

• Activity Number: 335
• Type: Contributed
• Date/Time: Tuesday, August 5, 2014 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Kallappa Koti*+
• Companies: FDA
• Keywords: Comparative bioavailability ; equivalence limits ; PK parameters ; analysis of variance ; TOST procedure ; generalized p-value

Abstract

Increased demand for biosimilar and generic products has raised the importance of equivalence testing to a new higher level. The underlying statistical exercise is to test the null hypotheses of non-zero differences in means. Wellek (2003) has advocated the use of uniformly most powerful invariant (UMPI) tests in the analysis of bioequivalence studies. Rationale behind using UMPI tests is explained. In this article, noncentral t-distribution based intersection-union tests (IUT) are proposed. Power at zero difference is derived. Bioequivalence of biosimilars is discussed. Impact of UMPI tests and intersection-union tests on study designs is discussed. In addition, noncentral t-distribution based intersection-union tests for analyzing QTc interval prolongation data are proposed. Pros and cons are presented. Examples are provided. An alternative to using noncentral t-distribution based tests is briefly described.