NEJM Perspective Breaks Down Real World Evidence for Devices
Two medical device researchers have outlined in a 2/15 New England Journal of Medicine perspective piece the benefits and pitfalls of FDA relying on real world evidence (RWE) in regulatory decision making. The researchers reflect on a recent guidance on “Use of Real-World Evidence to Support Regulatory Decision-Making for Medical Devices,” noting that such evidence will be “most valuable in the near future for post-marketing, since it is particularly in this setting that our current systems are inadequate for ensuring the safety and evaluating the performance of modern medical devices. The largest gaps in our understanding of devices relate to use outside the context of controlled clinical trials, and it is there that the majority of RWE will be collected. For the past 40 years, the FDA has relied extensively on physicians to report adverse events suspected to be related to a medical device. Unfortunately, clinicians significantly underreport such events: some estimates suggest that only 0.5% of all device-related adverse events are reported to the FDA.”
The researchers caution on RWE’s use to support device-approval decisions. “Perhaps the best initial use of RWE in the premarketing setting is to leverage electronic clinical data resources to support pragmatic clinical trials by embedding the necessary data collection into ongoing registries or [electronic health record] systems,” they write. “Such approaches could significantly reduce the cost and time required to complete a study as compared with traditional randomized clinical trials. Clinical follow-up through routine care or pre-specified electronic patient contact could offer tremendous efficiencies, while also permitting broader patient populations to participate in trials of innovative therapies.”
The researchers add that they see value in using RWE to support expanding populations or conditions for which a device is approved. However, they say there are “important risks in basing initial market-approval decisions for high-risk medical devices on contemporary RWE. Even with the most stringent methods of statistical adjustment, RWE-based analyses are observational in nature, with inherent biases in treatment choice coloring the patterns and associations that might be observed.”