Determining Real-World Data’s Fitness for Use and the Role of Reliability

A new paper by the Duke-Margolis Center for Health Policy provides a systematic framework to characterize the reliability of real-world data (RWD) to support fitness for use in drug development and regulation. RWD is information about patient health status and health care delivery routinely collected from a variety of sources, including wearables and electronic health records (EHRs). The paper advances understanding of how RWD can generate real-world evidence (RWE) that more efficiently evaluates the potential benefits or risks of medical products, is more inclusive of broader patient populations, and more reflective of routine clinical and patient self-care.

“Demonstrating RWD reliability is critical to establishing it as an effective resource for drug development and regulation,” said Mark McClellan, MD, PhD, director of Duke-Margolis. “Our aim is not only to inform the global health care research community on RWD reliability, but also serve as a resource to those designing studies using RWD and to the FDA and its international counterparts as they develop policy.” 

The 21st Century Cures Act requires FDA to establish a program to evaluate its potential use of RWE. In its December 2018 RWE Framework, FDA stated that it plans to issue guidance on assessments of the reliability and relevance of RWD in generating RWE on drug product effectiveness as a program item.

The Center’s paper identifies key concepts that should be evaluated in assessments of reliability for RWD, including completeness, conformance, and plausibility. It also identifies considerations for applying the framework to data from EHRs and person-generated health data.