EMA Officials Discuss Limits of Real World Evidence
Posted 05 July 2018 | By
A team of European Medicines Agency (EMA) officials are cautious about the utility of data from real world evidence in regulatory decision-making.
In a newly published analysis in BMJ Open
, five officials from EMA's Pharmacovigilance and Epidemiology Department find that "the immediate utilisation of their outputs to support regulatory decision-making is limited, often due to insufficient available information and to discrepancies between outputs and objectives."
Despite these issues, the authors emphasize that real world evidence is needed to supplement data from randomized clinical trials (RCTs).
"There is a need to supplement RCTs with other sources more representative of everyday 'real world' medical practice in order to provide additional insight on the benefit-risk balance," they write.
Regulators have increasingly looked to the use of real world evidence in regulatory decision-making, including to support or supplement premarket decisions.
"Real world databases have the potential to pick up and analyze safety issues and potentially provide information about effect sizes in a real world setting and in sub populations much more quickly, allowing regulators to take action at an earlier stage," EMA and the Heads of Medicines Agencies (HMA) write in the EU Medicines Agencies Network Strategy to 2020
For the analysis, the authors identified a patchwork of real world evidence initiatives funded through programs including the EU Framework Programmes (FP6 and FP7), Horizon 202, the Innovative Medicines Initiative and the European Research Council among others.
Of the 171 initiatives identified, 106 were omitted for being out of scope, with the remaining 65 divided across initiatives intended to develop data sources, methodologies, governance, analytical approaches and infrastructure for using real world evidence.
Each initiative was then assessed for its relevance to regulatory decision-making based on its particular type, and all were assessed for whether the outputs matched the initiatives' objectives.
While the outputs for governance model, methodology and infrastructure initiatives matched the stated objectives in most cases, the outputs for data source and analytical model initiatives were less likely (65% and 58%) to reflect the objectives going in to the initiative.
The authors also found that the vast majority (86%) of data source initiatives did not share data publicly and noted that several therapeutic areas (hematology, gastroenterology and cardiovascular systems products) were not represented, despite being active areas of product development.
Additionally, the authors expressed concern about the sustainability of many of the initiatives, noting that many of the initiatives did not provide details for how they would continue beyond their initial funding.
"There is an apparent lack of an overarching vision for an infrastructure that would provide sustainable access to the data generated by the projects," the authors write.
To improve the usability of the outputs of real world evidence initiatives going forward, the authors say that regulators and other end users of the data should be involved early in the planning of EU-funded initiatives and that a greater emphasis should be placed on interoperability.
EU-funded initiatives for real world evidence: descriptive analysis of their characteristics and relevance for regulatory decision-making