ABSTRACT
Multiplicity refers to situations in a trial in which multiple statistical tests or analyses create multiple ways to “win” for treatment efficacy or safety. Multiplicity arises in clinical trials such as multiple endpoints, multiple treatment arms, multiple subgroups, multiple time points, multiple analyses, composite endpoints and their components, and combinations. The following notation will be used throughout this chapter. Consider a general multiplicity problem in a confirmatory clinical trial with several objectives, for example, several endpoints or dose–control comparisons. All confirmatory clinical trials involve more than one endpoint. Some are designated as primary and co-primary, some are designated as secondary endpoints, etc. Carefully triaging endpoints to primary and secondary is a very important consideration, but not sufficient in the trial design. The generalization of the Holm–Bonferroni procedure is particularly useful in dealing with complex multiplicity situations. However, it is easier to follow by considering only two endpoints first.
