(P-250) Lessons from a Post-authorisation Study of Rivaroxaban Use and Potential Adverse Outcomes in Routine Practice in Four Countries

Background: A post-authorisation study programme of rivaroxaban use in four indications was conducted between 2011–2020.

Objectives: To summarize the lessons learnt from a large post-authorisation study programme that assessed drug use patterns, safety and effectiveness in first-time users of rivaroxaban or standard of care (SOC) in routine practice.

Methods: Data from IQVIA Medical Research Data-UK (IMRD-UK) in the UK, the German Pharmacoepidemiological Research Database (GePaRD), the Dutch PHARMO Database Network and Swedish national health registries were used to calculate unstandardized incidence rates of bleeding events, myocardial infarction, ischaemic stroke and mortality. Nested case-control analyses were also performed for each safety outcome.

Results: The study programme was initiated before the European launch of rivaroxaban in 2011, prior to the use of a common data model allowing one computer script to be executed across sources in a united way; however analytical approaches were harmonised across data sources, complemented by several validation activities. Despite this, some differences remained owing to variations in the data sources, healthcare systems, prescribing environments and minor differences in analytical methods, which may, at least partly, explain the observed variability in the results between the studies; these factors must be considered in similar future studies.
Surveillance studies of newly launched drugs must consider how the evolving prescribing environment may affect physicians’ treatment decisions, due to changes in SOC and the treatment landscape, which can result in a potential time-dependent bias and subgroup channelling of rivaroxaban.

This residual selection bias is likely to affect the results and statistical methods may not fully compensate; the nested case-control approach used here could offer a better analytical approach than a direct comparison between study drug and SOC.

When contrasting the results of real-world studies to those from the pivotal RCTs, one should bear in mind that both the background information and the outcome events are captured in a different manner and that the study populations are distinct. The study format allows assessment of the consistency of the findings between the studies, thereby providing a comprehensive assessment of rivaroxaban use, its safety and effectiveness under routine clinical conditions.

Conclusions: Overall the study programme was successful in its aims, and the design has provided a suitable framework for real-world data generation.