A Novel Approach in Subgroup Identification Using a Quantitative Benefit-Risk Index

• Activity Number: 658
• Type: Contributed
• Date/Time: Thursday, August 13, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Ramin Arani* and Ian Hirsch and Jonathan Norton and Christy Chuang-Stein and Weili He and Shihua Wen and Qi Jiang
• Companies: AstraZeneca and AstraZeneca and MedImmune and Pfizer Inc. and Merck and AbbVie and Amgen
• Keywords: Virtual Twin Method ; MCDA Method ; Jackknife method ; Decision Tree

Abstract

In recent years, there has been increasing level of interest and publications in methodologies to carry out structured quantitative Benefit Risk assessments to facilitate the interpretation and decision making process in clinical trials.

Often, the interest is to identify a sub-group of population that optimize benefit-risk index. Currently the approach is to carry out in a univariate approach by focusing either on primary efficacy or safety endpoint. This approach does not consider the totality of information at the same time. In this presentation, we introduce an algorithm which allows calculating patient level contribution to benefit risk index. Subsequently, one's method of choice for sub-group identification, such as tree based methods can be implemented. Example for the proof of concept will be provided utilizing Multi-Criteria Decision Analysis (MCDA) benefit-risk index.