FDA Releases Examples of Real-World Evidence Used in Medical Device Decision Making

FDA has long said that it would consider real-world evidence (RWE) in making regulatory decisions related to medical devices.  Those in industry know that FDA can be very critical of RWE, however, and it is not always clear why RWE was (or was not) acceptable in a particular submission.  Indeed, in 2017, when FDA released its guidance on acceptance of RWE, this author was skeptical as to whether it would actually result in an increase in Agency acceptance of RWE in medical device decision making (see earlier post here).

Earlier this month, FDA released a summary document describing 90 instances in which RWE was used in medical device decision making (see report here).  Interestingly, the examples in the report span between fiscal years 2012 and 2019, making it difficult to identify whether there has been any kind of shift in FDA acceptance of RWE following issuance of the 2017 guidance.

The report is helpful, however, in giving sponsors examples of where FDA has been accepting of clinical data in recent history.  In the press statement announcing release of the report, FDA stated, “When reviewing the use of RWE to support a regulatory decision, the FDA relies on scientifically robust methods and approaches to determine whether the submitted RWE is of sufficient quality to support the regulatory decision.”  (see March 16, 2021 announcement here) While the considerations for acceptance of RWE were made clearer in the 2017 guidance, the guidance included a modest list of six examples of where FDA had accepted RWE to illustrate the criteria in the guidance.  At 183 pages long, this new report can essentially serve as a robust appendix to the guidance giving sponsors a much more comprehensive and wide-ranging list of examples.

The report includes examples from all submission types, including 510(k)s, de novos, HDEs, and PMAs.  It also includes examples of RWE used in both pre- and post-market decision making.  The report is organized into six sections and is separated by device type (therapeutic devices, in vitro diagnostics) and RWE source (Registries, Administrative Claims Data, Medical Records, Other Sources).  The report also discusses each RWE example individually.  While the discussion of the examples does not give the specific reasons why FDA concluded that the RWE was acceptable, the examples should, nonetheless, provide sponsors with analogies upon which it can draw when arguing for its own RWE.

In the report, FDA encourages sponsors to continue to use RWE to support device regulatory submissions throughout the entire product lifecycle.  FDA also advises sponsors who are considering using RWE to consult with FDA via the pre-submission process, as needed, “to understand how to best utilize the RWE to support the marketing claims.”

We commend FDA for making these examples public, although caution sponsors that these examples only represent one-side of the story.  In our experience, there are plenty of examples of cases where FDA has not accepted RWE due to the quality and/or availability of details in the underlying data.  Thus, while this list looks robust and encouraging, sponsors should not consider this as an endorsement by FDA of any and all RWE.