In regulating and authorizing medicine, what’s the role of RWE in FDA and EMA approvals?
Nowadays, both FDA and EMA use RWE extensively in their regulatory decision-making. A recent analysis showed that in 2020 three quarters of all NDAs and BLAs submitted to FDA included RWE. Of those, RWE informed FDA’s decision-making in most cases as supplemental evidence (to the gold standard of randomized clinical trials), however, in about 10% of submissions RWE provided substantial or primary evidence. While the percentages of MAA submissions to EMA including RWE might not yet be quite as high, the trend is in the same direction. In terms of the indications, RWE is most often used in oncology, but also in infectious disease, neuroscience and endocrinology.
In rare diseases, where conditions are often less well understood and there is no precedence for clinical trials or regulatory submission, natural history studies are crucial, offering information on the severity, symptoms, progression, outcomes, and unmet medical needs of patients. Thus, providing the scientific foundation for the design of well controlled clinical trials, e.g., in terms of endpoints, duration of follow-up, as well as inclusion- and exclusion criteria of the rare and often heterogeneous study population.
Is there a difference in RWD regulatory decision-making between FDA and EMA?
The two main use cases where RWE is most influential in regulatory decision-making include external control groups to provide a comparator for single-arm, nonrandomized pivotal trials, as well as long-term follow-up registries for both safety and effectiveness in the post-approval setting, e.g., also as part of commitments related to conditional marketing authorizations (CMA).
These are quite similar between the FDA and EMA, however, one should not assume that a RWE strategy approved by the FDA will also be accepted by the EMA, and vice-versa; this remains highly situative on a case-by-case basis, and therefore, seeking early dialogues with both agencies is recommended for developers planning to submit RWE as a substantial part of their marketing authorization dossier.
In addition, there are some differences regarding the RWD sources, where in the U.S. RWE is often based on the analysis of claims and electronic health record (EHR) data, and in Europe more frequently on registry data, with the EMA maintaining a particular focus on high-quality registries (including relevant guidance). This is certainly also a reflection of the availability of “regulatory grade” RWD in the different countries, where claims and EHR data are readily commercially available in the U.S. and rather more difficult to obtain in Europe (there are several initiatives ongoing to improve this).
Is RWE becoming more popular in pharmaceutical development or quite the opposite? Why?
As we take the logical next step in advancing RWE for causal inference, in other words: cutting-edge comparative effectiveness research, I am confident that the use of RWE in regulatory decision-making will continue to increase. As Professor Jamie Robins, Epidemiologist at Harvard University, states: “In 2022, we are in the midst of an ongoing causal revolution. Novel approaches to problems previously thought to be intractable are exploding, with much recent highly original work on unmeasured confounding, and discovery of causal structures.”
Furthermore, RWE is equally becoming more important for payer negotiations and in health technology assessment (HTA). Two key trends worth mentioning here are performance- and outcomes-based pricing, which requires the collection of data on how an innovative drug performs in real-world clinical practice, either in individual patients or as a cohort, as well as reimbursement with continued evidence generation, often in the form of long-term registries.
What are the main challenges of implementation of RWE/RWD approach in pharmaceutical development?
Certainly, the data sources, we must continue to work on improving the availability and access to high-quality real-world data that is both reliable and relevant so that it will be fit-for-purpose for the research question at hand as well as decision-making. It is crucial, however, that RWD is generated in a patient-centric way, which includes fully recognizing and respecting patients’ rights to privacy and their ownership of the data. Trust and transparency are vital issues in this regard.
What is the future of RWE in science and research? What can be expected in the next 5 or 10 years?
Predicting the future is always tricky, but one does not have to be a prophet to foresee the significant impact of artificial intelligence (AI) and machine learning (ML), which is already in full swing. These programs will enable the rapid analysis of massive amounts of unstructured healthcare and patient-centered data, including from wearables and other digital health tools, and allow for much more rapid and deep insights in real-time. This can include more advanced solutions such as synthetic patients to complement hard to find and recruit rare disease populations.
We are seeing highly innovative startups and solutions pop up now on a weekly basis, however, the “proof in the pudding” for all these shiny new approaches is whether they are scientifically validated, reliable, relevant, and thus recognized – and ideally qualified – as fit-for-purpose for regulatory and payer decision-making. Watch this space.