The FDA is clarifying how it plans to evaluate real-world evidence to determine when it could be used to support regulatory decisions for medical devices.
Real-world data collected from sources outside of traditional trials, including retrospective studies, registries, electronic health records and other sources, could provide enough evidence for regulatory decisions, the FDA said in draft guidance issued July 26.
Although real-world evidence (RWE) gathered from real-world data (RWD) elements could constitute valid scientific evidence, the agency said, it is not changing its evidentiary standards in any way.
Rather, the agency explains how it might consider circumstances under which RWD could be used to support regulatory decisions. For example, an IDE may be needed to prospectively collect and use RWD to determine safety and efficacy. The agency acknowledged that data collected in certain settings may lack rigor compared to data collected in clinical trial settings; however, such data may help inform or augment information about devices that could support approval decisions.
The decision to consider real-world evidence comes alongside the agency’s plans to develop a national evaluation system that would leverage RWD to identify safety problems and to better understand risk-benefit profiles ().
The threshold for considering RWD sources will depend on the device and its level of risk. The guidance discusses which types of data could be useful for gaining additional information. For example, disease-specific RWD sources sponsored by patient advocacy groups could be useful for tracking disease progression for poorly understood diseases. And, treatment-specific RWD could be useful in tracking overall outcomes.
RWE may also be useful for the following applications:
RWD could be subject to investigational device exemption regulations if the data collection constitutes a clinical investigation. That determination would be made on a case-by-case basis. If an IDE is required, the FDA would work with the sponsor to determine “the least burdensome approach to facilitate the efficient collection of high-quality data,” the guidance says.
Primary factors for considering reliability of data include how the data were collected; whether the data were collected as complete accurate and adequate for answering the specific question studied; and whether people and processes involved in collecting the data provide adequate assurance that the bias is minimized.
To determine whether data accrual ensures reliability; the FDA will assess:
Comments on the draft guidance are due Oct. 25. Read the guidance here: www.fdanews.com/07-27-16-realworldevidence.pdf. — Tamra Sami