(A-194) Intercurrent Events in Estimands: Where RWE meets RCTs

Background: Intercurrent events (ICEs), defined as events occurring after randomization that affect either the interpretation or the existence of the endpoint, are a key component of the estimand approach established in 2017 by ICH and endorsed by EMA in 2020 and FDA in 2021 for RCTs. ICEs introduce real-world events and require an understanding of real-world data (RWD) and causal modeling, thus providing a need for epidemiological input into the design and conduct of randomized clinical trials (RCT).

Objectives: To evaluate the utilization of ICE in a therapeutic area that is known to have ICE: leukemia and lymphoma. A lack of utilization of ICE strategies in RCT would demonstrate a need for RWE input and support in these trials.

Methods: A comprehensive review of phase III trials occurring after the EMA and FDA guidelines (2021) was conducted in PubMed. Each publication was evaluated for whether an estimand approach, including the definition of ICEs, inclusion of ICE in a CONSORT diagram as well as in the statistical analysis. Descriptive results were tabulated to summarize across all identified publications.

Results: A total of 122 publications were identified. 31 were excluded because the publication did not contain results from a phase III trial. Of the 91 remaining, 35 included ICEs in a CONSORT diagram, but only 1 utilized ICE terminology and corresponding statistical analyses following an estimand approach.

Conclusions: Despite the adoption and endorsement by EMA and the FDA for years, the estimand approach is still rarely utilized in RCTs. Given the expertise in causal methods and nuances of RWD, pharmacoepidemiologists and RWE experts can play a pivotal role in the incorporation of the estimand in phase III trials.