Combining Survival Trials Using Aggregate Data Based on Misspecified Models: Method and Application

• Activity Number: 658
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Yabing Mai* and Tinghui Yu and Sherry Liu and Xiaofei Hu
• Companies: Merck and FDA/CDRH and FDA/CDRH and
• Keywords: Combining survival trials ; misspecified model ; harmonic average ; time-to-event ; aggregate data

Abstract

It is often necessary to assess the treatment effect of a therapeutic product based on outcome data observed from a mixed population of high heterogeneity, as such data may be collected from multiple independent clinical trials. For clinical trials with time-to-event data, the non-linear structure of the (log) hazard ratio estimates based on a Cox proportional hazard model leads to challenges in the statistical analysis of combining such clinical trials. We have proposed efficient estimates together with a version of Wald test for the combined hazard ratio using only aggregate data. In this talk, we introduce our method from the perspective of real practice and demonstrate its application in various areas of clinical trial development, from trial design to data analysis, with real data example and simulation.