Comparison of Two Groups for Normally Distributed Outcomes and Test for Equivalence

Abstract

The aim of equivalence trials is to demonstrate that the difference between an experimental treatment and a control with respect to the endpoint of interest lies within a pre-defined interval, the so-called equivalence margin. Prominent examples of such trials are bioequivalence studies which aim to show that rate and extent of drug absorption of a generic medicinal product and the original formulation differ maximally by a pre-defined clinically irrelevant amount. In this chapter, methods for sample size calculation are described when applying the “two one-sided tests procedure” for equivalence assessment. The situation of normally distributed outcomes and the cases that the treatment effect is expressed in terms of the difference or ratio of means, respectively, are considered.