Premium
Individual bioequivalence testing under 2×3 designs
Author(s) -
Chow SheinChung,
Shao Jun,
Wang Hansheng
Publication year - 2002
Publication title -
statistics in medicine
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.996
H-Index - 183
eISSN - 1097-0258
pISSN - 0277-6715
DOI - 10.1002/sim.1056
Subject(s) - bioequivalence , food and drug administration , computer science , sample size determination , quality (philosophy) , econometrics , product (mathematics) , statistics , test (biology) , reliability engineering , risk analysis (engineering) , medicine , mathematics , engineering , pharmacology , philosophy , geometry , epistemology , bioavailability , paleontology , biology
In recent years, as more generic drug products become available, it is a concern not only whether generic drug products that have been approved based on the regulation of average bioequivalence will have the same quality, safety and efficacy as that of the brand‐name drug product, but also whether the approved generic drug products can be used interchangeably. In its recent draft guidance, the U.S. Food and Drug Administration (FDA) recommends that individual bioequivalence (IBE) be assessed using the method proposed by Hyslop, Hsuan, and Holder to address drug switchability. The FDA suggests that a 2×4 cross‐over design be considered for assessment of IBE, while a 2×3 cross‐over design may be used as an alternative design to reduce the length and cost of the study. Little or no information regarding the statistical procedures under 2×3 cross‐over designs is discussed in the guidance. In this paper, a detailed statistical procedure for assessment of IBE under 2×3 cross‐over designs is derived. The main purpose of this paper, however, is to derive an IBE test under an alternative 2×3 design and show that the resulting IBE test is better than that under a 2×3 cross‐over design and is comparable to or even better than that under a 2×4 cross‐over design. Our conclusions are supported by theoretical considerations and empirical results. Furthermore, a method of determining the sample sizes required for IBE tests to reach a given level of power is proposed. Copyright © 2002 John Wiley & Sons, Ltd.