Some statistical methods for multiple endpoints in clinical trials

Abstract

This paper summarizes, defines, and discusses multiple endpoints comparison procedures, concepts, and methodologies for applications to clinical trials. We address the more widely used methods of α-level, p-value, and critical value adjustments. We examine global assessment measures such as O'Brien's test and Simes' procedure and contrast them with the α-adjustment procedures of Bonferroni and Holm. We propose a global assessment procedure based on categorization of the individual endpoints to form an overall composite endpoint. Additionally, we discuss a new weighting scheme for Holm's sequentially rejective α-adjustment procedure. Investigation of the correlation between endpoints is examined in relation to adjustment of the α-level. In the context of a clinical trial, the above multiplicity procedures are applied and compared. Finally, some comments concerning ease of use and relevance are summarized for the above methods.