Sequential designs for equivalence studies.

Sequential designs are increasingly being used in major clinical trials concerning life-threatening diseases. So far most applications have concerned trials designed to establish whether an experimental treatment is superior to a control. However, many trials are conducted with the objective of showing that an experimental treatment is equivalent to a control. This paper concerns the application of sequential designs to equivalence trials. Criteria for claiming equivalence are reviewed and compared, and methods first developed in the context of bioequivalence are described. Appropriate sequential procedures are identified. A simulated example, based on a clinical comparison of bronchodilators, is used to illustrate both the double triangular test and a comparable procedure constructed from alpha-spending functions.