Selecting Critical Boundaries in Group-Sequential Trials with Multiple Endpoints

• Activity Number: 342 - Topics in Adaptive-Seamless and Group Sequential Designs
• Type: Contributed
• Date/Time: Tuesday, July 30, 2019 : 10:30 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Toshimitsu Hamasaki* and Hsien-Ming James Hung, PhD and Chin-Fu Hsiao and Scott R Evans
• Companies: National Cerebral and Cardiovascular Center and Food and Drug Administration and National Health Research Institutes and George Washington University
• Keywords: Familywise error rate; Conjunctive power; Disjunctive power; Multiple endpoints; Group-sequential designs; Critical boundary function

Abstract

We discuss group sequential trials with multiple endpoints. Specifically we focus on the selection of the combination of critical boundaries for testing primary and secondary event-time endpoints, where one is for short-term efficacy assessment and other is for long-term efficacy assessment, or one may require a longer follow-up duration than other. We evaluate the behavior of power and Type I error for combinations of critical boundaries, hazard ratios, the number of analyses, the timing of analyses, and the correlation between the endpoints. We illustrate the calculation of sample sizes and the number of events for combinations of critical boundaries. We discuss considerations and offer guidance on the selection of critical boundaries in group-sequential trials with multiple endpoints.