PHP327 - CATEGORISATION OF OBSERVATIONAL STUDIES REQUESTED BY THE EMA AND SUBMITTED TO THE EUROPEAN UNION (EU) ELECTRONIC REGISTER OF POST-AUTHORIZATION STUDIES (PAS)

Abstract

The EU PAS Register is a publicly available register of non-interventional post-authorisation studies. Study types include: active surveillance, observational study, clinical trial and other. Here we aim to categorise all studies submitted as observational and that were requested by the EMA, in terms of their design and types of variables collected. We searched the ENCEPP website on 08-June-18 using following search terms: observational study, study requested by a regulator (EMA), finalised, adults. 30% of observational studies were requested by the EMA (123/385). 83% were labelled as risk management plan (RMP) category 3 (required), 12% as category 1 (imposed as condition of marketing authorisation), 1% as category 2 and 18% did not require a RMP or were “not applicable”. The number of studies requested by the EMA increased from 3 (2006) to 27 (2016). In terms of scope, 61% listed risk assessment, 59% drug utilization, 9% effectiveness evaluation, and 8% disease epidemiology (most studies listed multiple options). Most (72%) were cohort studies, followed by surveys (11%), case-control studies (7%), cross-sectional studies (5%), and case series (3%). 66% used retrospective data, 30% prospective data (4% were hybrid). 72% involved longitudinal data (28% cross-sectional). 58% used databases (mostly primary care electronic patient registries), 29% collected data prospectively, 6% retrospectively (chart review), and 11% were surveys. Most studies collected safety (60%), followed by drug usage (58%), effectiveness (20%), and healthcare resource use or patient reported outcomes (3%). We observed a sharp increase in observational studies requested by the EMA, demonstrating the increasing use of real-world data to inform decision-making. Studies are typically part of an RMP, predominantly category 3. Approximately 60% of studies focus on risk assessment and cover safety endpoints. The majority are retrospective cohort studies using existing databases and approximately 30% are prospective observational cohort studies.