Exploratory Subgroup Analysis: Subgroup Identification Approaches in Clinical Trials

• Activity Number: 302
• Type: Topic Contributed
• Date/Time: Tuesday, August 6, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Ilya Lipkovich*+ and Alexei Dmitrienko
• Companies: and Quintiles
• Keywords: personalized medicine ; subgroup analysis ; biomarker identification

Abstract

In these talk we contrast traditional methodology of subgroup analysis developed in the clinical statistics community over the past 15 years with emerging approaches to subgroup identification/search under a general framework of personalized medicine that were stimulated by machine learning and data mining methods. We illustrate some challenges of applying data mining methodology to subgroup identification with a novel recursive partitioning procedure SIDEScreen and discuss its key elements including (1) generation of multiple promising subgroups based on different splitting criteria, (2) choice of optimal values of complexity parameters via cross-validation, (3) evaluation of variable importance and using it for building a two-stage procedure with improved operating characteristics, and (4) addressing Type I error rate inflation using a resampling-based method.