In Focus: Top Takeaways from FDA Final Guidance on Real-World Evidence
A look at the insights and implications surrounding the agency's recommendations on using RWE from electronic health records and medical claims data.
The FDA’s latest guidance on how trial sponsors should use real-world evidence (RWE) in their studies is an apt reminder that the drug-making business is not for the faint of heart.
In July, the agency finalized its recommendations1 on using RWE from electronic health records (EHRs) and medical claims data; the draft has been circulating for three years. Since 2021, the FDA has released a few other guidances, in draft and final versions, that discuss the use of RWE, be it in the drug development trial process or in postmarketing efforts.
The July guidance is all about provenance: If a piece of data is not traceable to its core—for every patient under consideration for trial inclusion—the agency will likely question its validity as supporting evidence. The document spends time urging sponsors to pay attention to migration—patients leaving one insurer for another. “The validity of findings from a study using these data depends in part on the documentation of the migration of patients into and out of health plans and healthcare systems,” according to the document.
This guidance doesn’t deal with devices, nor does it recommend use of any particular study design, data source, or study methodology. It also notes that a study should not be designed to fit a specific data source, but rather the data sources should be selected to best fit the questions of interest in the trial.2
The same legal reviewer noted that the agency made minimal changes between the draft and final guidances, to the chagrin, most likely, of groups such as Janssen and the American Society of Clinical Oncology (ASCO). In comments made to the FDA in November 2021, Janssen wanted “alternatives” in providing supporting data that didn’t belong to the company, and ASCO asked the agency to define differences between RWE and electronic data capture, which, unlike EHR information, documents disease response, progression, the grading of adverse events, and so on. “These research-specific data often are not captured in EHR data,” ASCO wrote.
When queried about Janssen’s concern, the FDA responded via email: “FDA guidance on regulatory considerations outlines existing regulatory pathways for third parties to submit the relevant datasets directly to FDA with an appropriate right of reference granted to the relevant sponsors.” In a final guidance published last year, the agency said sponsors need to have “agreements in place” with owners of the data should the FDA want to inspect them.3
The July guidance does not mention electronic data capture.
The document also discusses the use of unstructured data. Whatever method the sponsor uses to extract the desired information, especially artificial intelligence, has to be documented, including the source of the data, the assumptions made in the algorithm’s design, whether use of the algorithm was supervised, and the data validation’s methods.
The documents are written broadly for a reason, says Sara Bristol Calvert, director of projects for the Clinical Trials Transformation Initiative. “It is good for establishing a comprehensive set of considerations, and it establishes the commitment of the agency to allow use of RWD (real-world data)/RWE. Those are good things.” And she adds. “There are always questions.”
Some sponsors have tried to cut corners in the use of RWE in their studies. A new study by Burcu et al.4 that looked at the use of RWE in various studies found problems with sample size, missing data, selection bias, and more. “Some studies failed to report essential design elements, including study period and inclusion/exclusion criteria, raising concerns about generalizability, potential selection bias, and confounding bias,” the authors wrote.
As it has for the past few years, the FDA encouraged sponsors to contact the agency with questions before submitting any documentation. One can foresee a lot of requests for just that, so of course the question arises whether the FDA has the staff to do so. In its email, the agency said it “has the capacity to accommodate the scheduling of such meetings, including staff with expertise in fields relevant to [RWD] and [RWE].”
The European Medicines Agency (EMA) also has numerous plans in play for the use of RWE.5 Two years ago, the Europeans put DARWIN (Data Analysis and Real-World Interrogation Network) online. This huge pool of data now has 20 data partners from 13 EU members representing 130 million patients, according to the EMA; agency researchers are using the DARWIN data for its own studies. Other groups are drawing a roadmap for the development of RWE guidance, and DARWIN has compiled catalogs detailing the metadata collected from RWD sources and subsequent studies. “They are intended to help regulators, pharmaceutical companies, and researchers identify and use such data when investigating the use, safety, and effectiveness of medicines,” said an EMA press officer in an email.
Will all this culling, identifying, cleaning, and analyzing be worth it?
“The use of RWE is one way to improve efficiency of clinical trials, but it is only one tool,” says Calvert, speaking about other innovations that may also improve efficiency. But, she adds, if sponsors “can use RWD to answer study questions, it can put less burden on patients and ultimately get new medicines to patients faster.”
Christine Bahls is a Freelance Writer for Medical, Clinical Trials, and Pharma Information
References
1. FDA, Real-World Data: Assessing Electronic Health Records and Medical Claims Data to Support Regulatory Decision-Making for Drug and Biological Products (July 2024). https://www.fda.gov/media/152503/download
2.Ravitz, J.R.; Chen, J.; Gadiock, P.S.; et al. Real-World Data Back in the Spotlight at FDA. McDermott Will & Emery. August 12, 2024. https://www.mwe.com/insights/real-world-data-back-in-the-spotlight-at-fda/
3. FDA, Considerations for the Use of Real-World Data and Real-World Evidence to Support Regulatory Decision-Making for Drug and Biological Products (August 2023). https://www.fda.gov/media/171667/download
4. Alipour-Haris, G.; Liu, X.; Acha, V.; Winterstein, A.G.; Burcu, M. Real-World Evidence to Support Regulatory Submissions: A landscape Review and Assessment of Use Cases. Clin Transl Sci. 2024. 17 (8), e13903. https://www.ncbi.nlm.nih.gov/pmc/articles/PMC11295294/
5. European Medicines Agency; Head of Medicines Agency. Real-World Evidence Provided by EMA: Support for Regulatory Decision-Making. April 10, 2024. https://www.ema.europa.eu/en/documents/other/guide-real-world-evidence-provided-ema-support-regulatory-decision-making_en.pdf
