Multiplicity Issues in Confirmatory Clinical Trials for Drug Development

• Activity Number: 315
• Type: Invited
• Date/Time: Tuesday, July 31, 2012 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Robert O'Neill*+ and Mohammad Huque*+ and Hsien-Ming James Hung*+ and Gary Koch*+ and Ralph D'Agostino*+ and Jason Hsu*+ and Alex Dmitrienko*+ and Frank Bretz*+ and Walter Offen*+ and Willi Maurer*+
• Companies: FDA and FDA and FDA and The University of North Carolina at Chapel Hill and Boston University and The Ohio State University and Quintiles, Inc. and Novartis and Eli Lilly and Company and Novartis
• Keywords: confirmatory ; decision making ; strong control ; type I error probability

Abstract

In drug development, confirmatory clinical trials are usually planned based on the information obtained from exploratory learning trials. Definitive evidences for supporting drug licensure are expected to be generated from confirmatory trials, also known as adequate and well-controlled trials. In this spirit, exploratory trials, if carefully conducted, may provide supportive information. Adequate and well-controlled trials are therefore critical to provide scientifically rigorous evidential standard and for regulatory decision-making. It is important to realize that failure to deal with multiplicity issues can exaggerate the nominal statistical significance of a finding resulted from multiple ways to win for the purpose of establishing the claim(s) for drug approval. Multiplicity considerations are less critical for earlier-stage exploratory clinical trials, where it is common to conduct a variety of assessments to generate several more hypotheses that guide the design of subsequent confirmatory trials. The US Food and Drug Administration draft guidance on multiple endpoints in clinical trials is anticipated to release for public comments in the near future. The guidance addresses