Online Program Home
  My Program

All Times EDT

Abstract Details

Activity Number: 440 - Let’s Make Everyone and Everything Count! Benefit-Risk Assessment Challenges, Lessons and Impacts in the Age of Big Data from Clinical Trials to Real-World Evidence
Type: Topic Contributed
Date/Time: Thursday, August 6, 2020 : 10:00 AM to 11:50 AM
Sponsor: Biopharmaceutical Section
Abstract #309733
Title: Do Birds of a Feather Flock Together? a Question of Credible Subgroup Identification and Inference on Benefitting Populations in Clinical Trials
Author(s): Duy Ngo* and Patrick Schnell and Richard Baumgartner and Shahrul Mt-Isa and Jie Chen and Dai Feng
Companies: Western Michigan University and The Ohio State University College of Public Health and Merck and Merck and Merck & Co., Inc. and AbbVie
Keywords: Bayesian ; subgroup analysis; credible subgroup; efficacy endpoints; benefit-risk ; multiplicity
Abstract:

Identification of populations benefitting from treatment is emerging as an important desideratum of regulatory and health technology assessment agencies. This is also reflected in increased research of statistical methods with different emphasis on exploratory and confirmatory aspects, whereas inherent simultaneous inference (multiplicity) issues when flagging the benefitting patients represent a major challenge. Recently developed Bayesian credible subgroup method avoids an overall test for treatment-covariate interaction. It leverages joint posterior distribution of the parameters of the statistical models (or regression surfaces) to obtain inferences about the benefitting subgroup. In our contribution we will focus on the Bayesian credible subgroup approach and show how it provides for effective multiplicity control. We will demonstrate its utility with commonly used efficacy endpoints, including continuous, binomial and survival outcomes, particularly from the practitioner’s point of view. We will also discuss further extensions to multiple endpoints to stratify populations with suitable benefit-risk profiles.


Authors who are presenting talks have a * after their name.

Back to the full JSM 2020 program