Testing Multiple Endpoints in Group Sequential Designs

• Activity Number: 486
• Type: Contributed
• Date/Time: Wednesday, August 7, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Guohui Liu*+ and Yi Liu
• Companies: Millennium: The Takeda Oncology Company and Millennium : The Takeda Oncology Company
• Keywords: clinical trial, family wise error rate, multiple endpoints, group sequential design

Abstract

Clinical trials with interim stopping for the primary endpoint usually uses alpha spending functions in a group sequential setting. This guarantees the strong control of family wise error rate (FWER). However, when multiple secondary endpoints are included in the design and tested along with the primary endpoint at interim or final analysis, the FWER can be violated if secondary endpoints are tested at full alpha level even if they are only tested when primary endpoint is significant (see Huang et al. 2007 and Glimm et al. 2009 for example). As a remedy, Huang et al. suggested using the same critical values as the primary endpoint for the secondary endpoints at each time point to be conservative. In this presentation, we show that if the secondary endpoints use the same alpha spending function as the primary endpoint, FWER can still be maintained while power is improved.