OBJECTIVES: Data external to clinical trials (i.e., ECAs) are increasingly being utilized to estimate comparative treatment effect in the absence of an appropriate within-trial comparative arm. We examined the frequency with which recent applications submitted to EMA include data from ECAs, the associated therapeutic areas, methods used to derive ECAs, and feedback received regarding the ECAs.
METHODS: The EMA website was searched for European Public Assessment Reports (EPARs) published between 1 January 2022 and 23 May 2023. The following were excluded: medical devices, vaccines, diagnostic technologies, hybrid medicines, and biosimilars. The clinical efficacy section of each EPAR was reviewed; further, each identified EPAR was searched using the terms “external control,” “observational,” and “real world.” Data were extracted from EPARs in which ECAs were mentioned.
RESULTS: Of 54 eligible EPARs, 11 (20%) included an ECA. Among the 11 EPARs, 9 (82%) were for rare diseases. Eight (73%) EPARs included ECAs as comparators for single-arm trials; the remainder were compared to randomized clinical trials. ECAs were derived from retrospective medical record reviews (n=5), prospective observational studies (n=4), historical trials (n=2), clinical databases (n=2), and literature review (n=1): 3 EPARs included 2 ECA sources. Matching techniques were used in 4 EPARs (36%), naïve comparisons (direct or adjusted indirect) were used in 7 (64%). The most common feedback on the use of ECAs was concerns about presence of selection bias and unobserved confounding. Evidence from ECAs was considered acceptable in 4 EPARs, supportive in 6 EPARs, and non-supportive in 1 EPAR.
CONCLUSIONS: The EMA considers evidence submitted from ECAs as acceptable when bias in the ECA is considered not to impact the overall conclusions, and population characteristics between the clinical trial and ECA are comparable.
