Common Practices for Subgroup Identification and Analysis: The Results of a Survey Conducted Within the Pharmaceutical Industry
View Presentation View Presentation
*Cristiana Mayer, JNJ 

Keywords: subgroup identification and analysis; QSPI; survey on common practices

The Subgroup Analysis Working Group was sponsored by the industry group QSPI (Quantitative Sciences in the Pharmaceutical Industry) with the primary goal of creating a unifying and transparent framework for the current state of subgroup analysis/identification statistical methods and their application in drug development. In order to describe the current practices about subgroup identification methodologies used in randomized clinical trials a survey was conducted among statisticians in the pharmaceutical industry in the first quarter of 2014. This talk will present the goals, the design and the results of such survey. This effort also aimed to identify important issues and the most common challenges faced in performing a subgroup analysis in various phases of drug development. Because there are many purposes of subgroup analyses, from evaluating consistency of results across subpopulations, to determining a subgroup of patients with an enhanced treatment effect for personalized medicine, the analysis of subgroups is very common in clinical trials. With the current EMA draft “Guideline on the investigation of subgroups in confirmatory clinical trials” (February 2014) the interest for subgroup identification and analysis has gained momentum. Special attention will be given to the issues in the conduct of subgroup analyses because they will inform and shape future collaborations among statisticians in the field and have an impact on future research on related methodologies. The survey is the first step on the path towards the development of educational tools to facilitate the application and usage of novel subgroup analysis/subgroup identification methods.