Understanding the FDA’s Approach to Real-World Evidence

The ubiquity of computers, mobile devices, wearables and other biosensors is generating a wealth of health-related data with potential value in clinical trial design and post-marketing safety and efficacy studies. A surge in the development of sophisticated data analytics suggests a realistic path forward for the integration of these data to inform clinical and regulatory decision making. The U.S. Food and Drug Administration (FDA) is seeking to provide leadership to realize these goals.

Evaluating Real-World Data and Evidence

As an important new facet of its ongoing commitment to accelerating drug and medical device development and delivering innovations more efficiently to patients, the FDA recently published the anticipated “Framework for FDA’s Real-World Evidence Program.”¹ Pursuant to the goals established by the 21^st Century Cures Act, the agency has outlined its approach to leveraging real-world data (RWD) — data relating to a patient’s health status and/or the delivery of health care routinely collected from a variety of sources — and real-world evidence (RWE) — clinical evidence about the usage and potential benefits or risks of a medical product derived from analysis of RWD — to support both approvals of new indications for approved drugs and post-approval study requirements.

In our increasingly digitized world, valuable RWD exists in a variety of formats, including electronic health records (EHRs), claims and billing data, registries, databases and patient-generated data from wearables and mobile devices. These extant and underutilized sources of RWD can be leveraged to collect data and support a variety of study designs to develop RWE, including observational studies and randomized trials. Additionally, the FDA has articulated the value of RWD in informing and improving the efficiency of clinical trials, via generating hypotheses, identifying drug development tools, assessing trial feasibility and prior probability distributions and assembling geographically distributed research cohorts.

Determining the suitability of RWE to inform regulatory decisions requires assessing both the relevance and reliability of the RWD, as well as the methodology used to generate RWE from the underlying RWD. According to the framework, the most important concerns are: (1) whether the RWD are fit for use, (2) whether the trial or study design provides adequate scientific evidence to address the regulatory question and (3) whether the study conduct meets the agency’s regulatory requirements. The FDA will use this tripartite approach to evaluate forthcoming supplemental applications and to guide their RWE program, which will involve establishing demonstration projects, engaging with stakeholders and developing additional guidance documents to help sponsors determine how best to use RWE.

The FDA’s incipient drive to take advantage of existing sources of RWD is echoed in recent investments and acquisitions of companies involved in curating and assessing EHRs and other data sources, notably Roche’s acquisition of Flatiron Health and Foundation Medicine and the joint venture Syntropy formed by Merck KgaA and Palantir Technologies.³

Applying RWE in the Regulatory Space

To date, the FDA has taken advantage of RWE via the Sentinel Initiative, a national electronic system that leverages a range of RWD sources to proactively monitor the safety of drugs and medical devices that are on the market.² The agency plans to use this experience using RWD as evidence concerning drug safety to provide a framework for its further use in informing effectiveness studies. In particular, the agency is examining the potential of RWE to support labeling changes, including adding or modifying an indication or changing dosing, dose regimens or the route of administration.

One major caveat to the application of RWD sources like EHRs and medical claims data to inform regulatory decision-making is the presence of significant gaps in the kinds of data captured in these sources. Such sources typically record significant events like hospitalization and death, but data related to changes in medical status or potential adverse effects (e.g., quality of life issues, changes in degree of pain, asthma, depression, or anxiety) are not reliably and consistently documented in these data sources, or in some cases are captured but inaccessible or recorded in nonstandard ways. Making the most of these diverse data sources will likely require standardization of documentation practices for patient-reported and other outcomes, as well as the underlying technology to optimize interoperability of data from different sources. This will likely require the establishment of a universal format for RWD with common representation (e.g., terminology, vocabulary, coding schemes); the FDA is actively developing data standards to support this effort. 

A Long Road Ahead

The framework established in the current guidance merely reflects a skeleton to orient the agency’s thinking and guide the development of forthcoming guidances exploring a range of relevant topics in greater detail, including the use of EHRs to measure drug effectiveness, gaps in sources of RWD and potential strategies to address them, the generation of external control arms using RWD, determining how to use RWD to design observational studies and assessing the fitness of such studies to inform regulatory decisions. 

Building upon this framework, the FDA is pursuing a number of pilot projects exploring RWD and RWE to assess the potential to exploit available data and the existing barriers to their pragmatic use. At this stage, perhaps the most essential piece of the initiative is increasing stakeholder engagement, both within and outside of the agency.  

References

1. Framework for FDA’s Real-World Evidence Program. U.S. Food and Drug Administration. Dec. 2018. Web.
2. “FDA’s Sentinel Initiative.” U.S. Food and Drug Administration. 9 Jan. 2019. Web.
3. Terry, Mark. “FDA Doubles Down on Use of Real-World Data for Regulatory Decisions.” BioSpace. 7 Dec. 2018. Web.