The effect of adding real-world evidence to regulatory submissions on the breadth of population indicated for rare disease medicine treatment by the European Medicines Agency.

Background: Despite calls for the use of additional real-world evidence (RWE) during drug development, rates of inclusion at the regulatory stage remain low. The medicine adoption model suggests that providing additional RWE to regulators would result in a wider indicated population than providing randomised-controlled trial evidence (RCTE) alone. Here, we tested this hypothesis.

Methods: All engagements concerning the 88 orphan drugs approved between 2009 and 2019 on the European Medicines Agency Orphan Register were reviewed between September and December 2019. Engagements were grouped as containing either randomised-controlled trial evidence (RCTE) or RCTE with real-world evidence (RWE). The data on indicatable population (the therapeutic indication requested by an engagement) and indicated population (the therapeutic indication ultimately granted) as well as the median number of criteria limiting the indicated population in each study type (RCTE/RWE) was extracted. A chi-square test assessed the association between the indicated population (as a proportion of the indicatable population) and type of evidence (RCTE with or without RWE) and a Wilcoxon rank sum test assessed the difference between the median number of limiting criteria between RCTE and RWE studies. Prediction modelling extrapolated the results of a power analysis to a level expected to deliver significance and the time this would take.

Results: The review identified 103 engagements, of which three were excluded (one contained only RWE; two contained only systematic literature reviews), leaving 100 engagements for 87 orphan medicines in the final analysis. Only 13% of engagements contained RWE. Although the difference was statistically insignificant, 76.92% of engagements containing RCTE and RWE resulted in a broader indicated population as compared to only 56.32% of those that contained RCTE alone. The median number of limiting criteria from RCTE (37 (28, 43)) and RWE (5 (2, 9)) studies varied significantly (p = 0.005). Modelling suggested that the analysis would achieve sufficient power by 2033-37 at the current RWE adoption rate.

Conclusion: The proportion of the disease population studied in RWE was greater than that in RCTE. The analysis testing the relationship between additional RWE and broader indicated population would achieve adequate power between 2032 and 2037 at the current RWE adoption rate.