Flexible Alpha Spending Functions for Type I Error Control in Clinical Trials with Interim Analysis and Multiple Endpoints

• Activity Number: 60
• Type: Topic Contributed
• Date/Time: Sunday, August 9, 2015 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Abdul Sankoh and Xueying Chen*
• Companies: Synageva BioPharma and Novartis Pharmaceuticals
• Keywords: Alpha spending functions ; Interim analysis ; Multiple endpoints ; Simulations ; Type I error

Abstract

With current advancements in clinical trial design and statistical analysis approaches, it is now not uncommon to design a clinical trial with one or more pre-planned interim analyses and pre-specified primary and key secondary endpoints for which a sponsor desires to make labeling claims. As such designs are becoming more and more complex, necessitating the need to come up with flexible statistical methods for testing different hypotheses under strong control of the nominal significance level. In this presentation, we investigate properties of some flexible alpha spending functions that are commonly used to provide strong control of type I error in clinical trials with interim analysis and pre-specified multiple primary and key secondary endpoints. Simulation studies are conducted to demonstrate the performance of various approaches.