Around the Globe

FDA’s Real World Evidence Program Moving Forward
Jacqueline Corrigan-Curay
David Martin
Office of Medical Policy, CDER, FDA

n December 2018, FDA published a Framework for its real world evidence (RWE) program. As articulated in the Framework, three central considerations will drive FDA’s evaluation of RWE: whether the real world data (RWD) are fit for use; whether the trial or study design used to generate RWE can provide adequate scientific evidence to answer or help answer the regulatory question; and whether the study conduct meets FDA regulatory requirements. To inform this evaluation, FDA is engaging external and internal stakeholders, funding demonstration projects, and developing guidance for industry.

What is the RWE Program?

FDA launched its RWE program to evaluate the potential use of RWE to help support the approval of new indications for drugs already approved under section 505(c) of the Federal Food, Drug, and Cosmetic Act, or to help support or satisfy post-approval study requirements, consistent with expectations outlined in the 21st Century Cures Act. FDA defines RWE to be clinical evidence about the usage and potential benefits or risks of a medical product derived from analysis of real world data. RWD are data relating to patient health status and/or the delivery of healthcare routinely collected from a variety of sources, including electronic health records (EHRs), claims data, and data collected outside of the healthcare system that informs on health, e.g., mobile technology.

Stakeholder Engagement

Beyond considering comments on the RWE Framework, FDA has been actively engaging external stakeholders in various forums. In the past year, for example, FDA has held two public meetings: the first on leveraging RWD in randomized clinical trials, and the second an annual review of considerations for developing RWE for regulatory use including initiatives to improve and assess data quality, use of mobile technologies and sensors, and evaluation of observational methods to draw causal inference about product effectiveness. In addition, FDA leadership has participated in more than 40 engagements, including presenting at the DIA Global Annual Meeting and serving on the planning committee for the past two annual DIA RWE meetings. CDER has also produced two RWE Small Business and Industry Assistance webinars, one on the RWE Framework and the second on FDA’s open-source MyStudies mobile application for use in clinical trials. These interactions are valuable to ongoing educational initiatives in this dynamic area, and to identify key outstanding questions regarding use of RWE.

FDA is also engaging regularly with regulators, including the EMA, to share knowledge and provide updates on current activities with respect to RWE policy development.

Engagement of internal stakeholders is another critical part of FDA’s RWE program. Recognizing the importance of this initiative and the need to facilitate consistent practices, leadership from CDER and CBER created a steering committee of Center leadership and subject matter experts to oversee policy development and to provide resources and guidance to review divisions evaluating novel applications for use of RWE. This committee not only supports reviewers in their assessment of proposals but also oversees policy development and demonstration projects and has benefited from various presentations by groups exploring new methodologies to collect and curate RWD that address data quality and relevance challenges.

Demonstration Projects

An important feature of FDA’s RWE program is supporting demonstration projects that can address knowledge gaps and support guidance development. FDA is currently supporting demonstration projects focused on understanding data quality, improving RWE tools such as data standards and mobile technologies, and evaluating RWE approaches to study design and data analysis. For example, regarding data quality, FDA funded the HARMONY-OUTCOMES Ancillary Study to examine whether and how EHRs at selected sites in a Glaxo SmithKline cardiovascular outcome trial for albiglutide could have been used to identify eligible participants, populate baseline characteristics of participants into an electronic case report form, and identify clinical endpoints. Results of this study should be available in the next year, and FDA is expanding demonstration projects in this area.

In addition, FDA is supporting the University of California in San Francisco’s Carol Franc Buck Breast Care Center on their development of a system to both improve the capture of structured data in the EHR and to enable such data to be seamlessly used as part of an FDA-regulated clinical trial — eliminating the need for duplicate entry, and potentially saving time, money, and improving care.

Within the FDA Catalyst Program, FDA developed the MyStudies mobile application, a tool that can collect data directly from participants outside of the healthcare system and at their convenience. To examine the role of patient-generated data using mobile technologies for real world trials and registries, FDA MyStudies has been re-branded and configured to support the Childhood Arthritis and Rheumatology Research Alliance Limit Juvenile Idiopathic Arthritis Trial and the Crohn’s and Colitis Foundation Inflammatory Bowel Disease (IDB) Plexus Platform. The app will collect the uveitis endpoint as well as questions regarding medication administration in Limit JIA, and collect patient reported outcomes within IBD Plexus.

Understanding methods for generating RWE will be critical for its use in regulatory decisions. Use of RWD to generate evidence on product effectiveness does not mean abandoning tools and methods of traditional trials to obtain valid results and minimize bias. Random treatment assignment and other methods to control bias will likely remain critical to generation of reliable evidence of effectiveness. Integration of RWD into clinical trials, which may lead to greater integration of research in clinical practice settings, presents new research opportunities but new challenges as well.

To increase our understanding of such trials, FDA is supporting demonstration projects under the FDA-Catalyst program. IMPACT AFib, the first FDA-Catalyst randomized trial, is evaluating whether an educational intervention can improve appropriate use of oral anticoagulants and lead to improved clinical outcomes. FDA has also partnered with PCORI to support the RELIANCE trial, a real world clinical trial enrolling patients with chronic obstructive pulmonary disease in which participating sites will capture the endpoints (all-cause hospitalization and all-cause mortality) through a call center, portal, and EHRs. For eligible trial participants, FDA-Catalyst will also link Medicare claims data to augment the assessment of trial endpoints.

FDA understands that great interest exists in exploring whether non-interventional designs (designs where treatment is assigned during clinical practice, and analytic techniques are then used to assemble “comparable” populations and control for potential confounders that may drive clinical treatment selection) can support a finding of product effectiveness. Examples exist wherein non-interventional studies produce the same causal inference as randomized trials, whereas conclusions in other examples have differed. For FDA, the key issue is understanding when concordance can be expected or when bias from unmeasured confounding could lead to false conclusions. A series of demonstration projects are being supported—most notably the Duplicate project by researchers at Harvard Medical School and the Brigham and Women’s Hospital—who, in close consultation with FDA staff, are attempting to replicate the results of 30 completed randomized controlled trials (RCTs) and also to conduct seven additional non-interventional studies in advance of the results from corresponding RCTs. FDA has also recently expanded this work to an additional sample of prospective replications through its Yale-Mayo CERSI. Preliminary results from these efforts are expected by summer of 2020. In addition, FDA is supporting efforts to inventory existing sensitivity analysis methods that might be useful for assessing uncontrolled confounding in non-interventional (observational) RWE studies and use clinical trial data to evaluate these methods. FDA intends to engage stakeholders in discussions of the results of these projects.

Moving Forward: RWE Guidance for Industry

All these RWE program activities will inform FDA guidance development regarding RWE. As outlined in the RWE Framework, FDA proposes guidance to address assessment of data quality, use of RWD in randomized trials, the potential role of non-interventional studies, and regulatory considerations. FDA’s approach will also consider flexibilities needed to allow for the consideration of variable RWD sources such as from EHR, claims data, and data captured using digital tools. FDA, like our stakeholders, is excited about the prospect of utilizing RWD to enhance the efficiency of evidence generation by providing guidance on how to utilize such data to generate robust RWE that meets FDA’s regulatory standards. We will continue to work with our stakeholders, researchers, and patients to realize this vision.

References available upon request.