Survey Results on Industry Practices and Challenges in Subgroup Analysis in Clinical Trials

This article provides a summary of the results of a recent survey in the pharmaceutical and device industry on the current practice and common issues in the identification and analysis of subgroups. The survey was part of a cross-industry effort sponsored by the Quantitative Sciences in the Pharmaceutical Industry (QSPI) for two purposes: (1) to identify general practices in how statistical subgroup analyses are planned and conducted in clinical trials and (2) collect information on similarities and differences across the prespecified types of subgroup analyses in terms of most common challenges, methodologies, and other characteristics. The survey results help integrate good practices and create a unifying framework for subgroup analyses in industry encompassing those in discovery/early phases directed to biomarkers identification and those at later stages for exploratory, confirmatory, or post hoc evaluations. Results from this survey identify the number one challenge to be the lack of power related to the sample size and number of subgroups tested. The findings further highlight an unmet need for intensified educational efforts focused on sharing statistical methodologies, experiences, and knowledge on the most appropriate approaches for identifying and analyzing subgroups. The scientific debate among industry, regulators and academia should continue to find solutions to unresolved challenges and perceived difficulties of subgroup analyses in clinical trials.