Improving Patient Subgroup Representation with Real-World Data

Real world data (RWD), or patient-level data generated outside of traditional clinical trial settings, has enabled the development and use of real-world evidence (RWE) that is generated and increasingly used in pragmatic trials, externally controlled trials, and observational studies. Subgroups, or subsets of patients with specific clinical or demographic characteristics, may be traditionally excluded from or otherwise unengaged in gold standard randomized clinical trials (RCTs) and could greatly benefit from RWD/E. Studies using RWD/E can enable inclusion of traditionally underrepresented subgroups as there are fewer barriers to research participation than RCTs. Underrepresentation of these subgroups limits the generalizability of research findings, reinforces treatment inequities, perpetuates data bias resulting from overestimations, and further excludes subgroups from benefits of the knowledge and innovation derived from research. Understanding the individual and/or combined influence of clinical and demographic characteristics among patient subgroups holds measurable value, especially in cases where such characteristics can be a proxy for observable variations in treatment effects.

In this paper, we describe opportunities for and challenges to leveraging RWD in order to estimate and measure treatment effects among and across patient subgroups. We also discuss important terminology considerations that accompany measurements of therapeutic effect in subgroups, supplementing clinical trials with RWD for subgroup analysis, leveraging RWD to address shortcomings in trial data, equitable solutions to address data missingness, and leveraging RWD to improve care among patient subgroups.

Read the full paper.