PURPOSE: This workshop will focus on the key methodological considerations in building external control arms using real-world data and provide recommendations regarding best practices to improve transparency and enable the assessment of analytic validity.

DESCRIPTION: In recent years, high-quality real-world databases have improved our ability to identify clinically relevant cohorts and new methodologic approaches have improved the validity of non-randomized study designs. Rigorously selected real-world cohorts may be able to provide critical evidence in settings when a randomized design is infeasible, insufficient or impractical. The 21st Century Cures Act and recent FDA guidance on the use of real-world data and real-world evidence provide substantial momentum and an increasing number of applications of real-world cohorts being used as comparators for clinical trials to support regulatory and coverage decisions. Although promising, construction of an external control arm using real-world data often requires using rigorous and sophisticated statistical methods to address potential sources of bias. This session will highlight the key methodological considerations as well as the pros and cons of different analytic methods used to build external control arms using real-world data — illustrated both through academic research as well a diverse sample of real-world case studies. In doing so, we’ll provide recommendations to help researchers construct real-world control arms to generate evidence that is robust and transparent so as to best support decision makers’ confidence in the results.