Incorporating Intermediate Binary Responses into Interim Analysis of a Long-Term Binary Endpoint

• Activity Number: 540 - SPEED: Clinical Trial Design, Longitudinal Analysis, and Other Topics in Biopharmaceutical Statistics
• Type: Contributed
• Date/Time: Wednesday, August 1, 2018 : 11:35 AM to 12:20 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Jingjing Chen and Tina Liu and Andrejus Parfionovas* and Cong Han and Xiaopan Yao
• Companies: Takeda Pharmaceuticals and Takeda Pharmaceuticals and Takeda Pharmaceuticals
• Keywords: interim analysis; long-term binary endpoint; intermediate binary response; futility monitoring; sample size re-estimation

Abstract

It is common when an interim analysis is conducted, only a small portion of subjects who have reached the timepoint of primary endpoint, and a large proportion of subjects who have been recruited have not reached the timepoint of the primary endpoint but have undergone some intermediate assessments that are predictive of the primary endpoint due to fast recruitment rate. A common practice is to eliminate such subjects from the interim analysis. For interim analysis that addresses futility monitoring or unblinded sample size re-estimation with lagged outcomes, increased efficiency of interim analysis could be achieved by improving the precision of interim estimates via utilizing most available information by the time of interim analysis. We propose a modeling-based approach to analyze intermediate and long-term data jointly and augment long-term information with intermediate information on the subjects who have not reached the primary timepoint, and compare the proposed method with the existing methods via simulations. It is shown by simulation that the use of intermediate information improves the efficiency where long-term treatment difference is assessed based on interim data.