The 21st Century Cures Act, enacted December 2016, mandated that the FDA consider how to use real-world evidence (RWE) to bring innovative treatments to market faster. To meet this statutory requirement, the FDA has provided guidance outlining its policies for using RWE to support the approval of a new drug indication.
What is real-world evidence?
The document starts out by differentiating between real-world data (RWD) and RWE.
- Real-world data: data relating to patient health status and/or the delivery of health care routinely collected from a variety of sources, such as EHR, claims data, patient registries, patient-generated data,
- Real-world evidence: the clinical evidence about the usage and potential benefits or risks of a medical product derived from analysis of RWD.
Existing applications of RWE at the FDA
The FDA notes that real-world evidence is already helpful for informing clinical trials (e.g., by generating hypotheses to tests, biomarker identification, informing a prior in Bayesian statistical models, informing patient characteristics for enrichment or stratification, or identifying geographic areas to recruit people for clinical trials), but this use for RWD and RWE is already well-established.
The FDA has used RWE to inform drug safety (e.g., through the Sentinel system), but use of RWE to support treatment effectiveness is generally limited to oncology or rare disease single arm trials, which require supportive RWE on response rates from chart reviews.
However, FDA is looking to expand the use of RWE to measure treatment effectiveness in some cases:
Specifically, FDA’s RWE Program will evaluate the potential use of RWE to support changes to labeling about drug product effectiveness, including adding or modifying an indication, such as a change in dose, dose regimen, or route of administration; adding a new population; or adding comparative effectiveness or safety information….The RWE Program will involve the establishment of demonstration projects, engagement with stakeholders, the use of internal processes that bring senior leadership input into the evaluation of RWE and promote shared learning and consistency in applying the framework, and the development of guidance documents to assist sponsors interested in using RWE to support their work.
The FDA indicated that its experience using EHR data to inform drug safety (Best Practices for Conducting and Reporting Pharmacoepidemiologic Safety Studies Using Electronic Healthcare Data) will inform it’s use of EHR data to inform drug effectiveness. They plan to issue future guidance building on this document to show how to use EHR data to measure drug effectiveness.
While claims data and EHR data are well-known data sources, they typically do poorly in measuring treatment effects on quality of life. Alternatives the FDA will consider in the future include the use of: mobile technologies, electronic patient reported outcome tools, wearables, and biosensors.
Single arm trials are sometimes needed for rare diseases or to avoid ethical issues. One alternative approach to a randomized controlled trial is to pair a single arm trial with a synthetical control based on RWD. FDA is exploring this option, but challenges exist because: (i) populations may not be comparable, diagnostic criteria or outcomes may not be equivalent, differences in how outcomes are assessed, and variability in follow-up procedures. The only statistical method the mention for addressing any of these issues, however, is propensity score matching; other statistical approaches for determining causal inference–such as regression discontinuity or instrumental variables–were not mentioned in the document.
Observational studies could be used to measure drug effectiveness.
FDA intends to evaluate multiple questions of interest that could affect the ability to draw a reliable causal inference, including, for example, the role of existing evidence (e.g., the natural history of the disease) and how the inclusion of a more diverse population can result in a heterogeneity of treatment effects making it difficult to detect smaller effect sizes.
Clearly, the key issue with observational studies is causal inference, and dealing with unmeasured confounders. Also determining the validity of the data, and identifying any necessary sensitivity analyses and statistical diagnostics will be key. Like a clinical trial, the FDA would likely mandate that a study protocol be posted before the study is executed. A website analogous to ClinicalTrials.gov could be considered, or stakeholders could even be required to post their study protocols to ClinicalTrials.gov. The ISPOR-ISPE guidelines on best practices for measuring real-world effectiveness could serve as key foundational guidelines for FDA to follow.
Next steps/Additional guidance to be issued
What will the FDA do next? They are developing guidance on the following topics:
- Use of EHR to measure drug effectiveness
- Potential gaps in RWD sources and strategies to address them
- Design of clinical trials that include pragmatic design elements (e.g., recruitment/enrollment, intervention approach, outcome assessment)
- Use of RWD to generate external control arms
- Observational study design using RWD, and whether and how these studies could inform regulatory decisionmaking
- Whether existing guidance documents regarding use of electronic source data (e.g., informed consent for electronic data guidance) can be applied for RWD
FDA is also increasing stakeholder engagement both internally–by creating a RWE Subcommittee of CDER’s Medical Policy and Program– and externally through public meetings convened by the Duke Margolis Center and workshops with the National Academy of Sciences, Engineering and Medicine