The expanded use of electronic health records (EHRs) across all medical specialties has had a profound impact on observational research over the past decade. As the patient journey is increasingly captured digitally, the volume and diversity of available data has increased exponentially. Historically, these real-world data (RWD) were collected passively through administrative claims databases and were used to explore associations between exposures and outcomes, generating hypotheses to be tested in clinical trials. However, due to the length of time and high cost of drug approvals, various stakeholders including patients, advocacy groups, the medical device and pharmaceutical industry, and regulatory bodies, such as the U.S. Food and Drug Administration (FDA), are looking to these data to do substantially more. Namely, to test a hypothesis and make a causal inference regarding an exposure such as a treatment, and an outcome such as disease progression.
This concept, and the underpinning technology, has been used in the United States for over a decade in specific cases. In 2008 the U.S. FDA launched the Sentinel program which actively monitors FDA-regulated products via EHR, insurance claims data and registries for safety signals.1-3 This marked a significant departure from the traditional passive reporting of adverse events, which improved data accuracy and thereby reduced the length of time to conduct these investigations. In a similar fashion, under the 21st Century Cures Act, the FDA was required to establish a program to evaluate the potential use of RWE to support the approval of new indications for approved drug products as well as to support or satisfy post-approval study requirements.
Stakeholders believe that RWD can potentially support pharmaceutical product label approvals and label expansions, but its adoption is mired in methodological challenges requiring rigorous solutions. Here we discuss two fundamental challenges, data quality and internal validity, and how Cardinal Health researchers are pursuing novel strategies to improve upon methods to achieve high levels of both.