Biometric Colloquium
LOUISE JESPERSEN
Institute of Medical Statistics, Center for Medical Data Science, Medical University of Vienna,
Vienna, Austria
Causal inference and the hypothetical estimand in randomised controlled trials
November 26th, 2025 at 09:15 pm
Seminarraum Center for Medical Data Science (previously CeMSIIS),
Spitalgasse 23, Room 88.03.512
Medical University of Vienna, 1090 Wien
Host: Felix Herkner
Abstract:
The presence of post-baseline intercurrent events in randomised controlled trials complicates the estimation of treatment effects, especially when interest lies in the direct or biological effect of a treatment rather than the effect of prescribing it. In line with the ICH E9(R1) guideline, which emphasises the need for clear strategies to handle intercurrent events, we explore methods from causal inference to address these challenges within the hypothetical estimand framework.
When estimating treatment efficacy under a hypothetical strategy, the common practice is to use a Mixed Model for Repeated Measures (MMRM) restricted to data collected before the intercurrent event. In this work, we investigate whether it is possible to make use of the "contaminated" data instead of discarding it, by applying a semiparametric causal approach Longitudinal Targeted Maximum Likelihood Estimation (LTMLE). Through simulations and empirical analysis, we assess how these methodologies perform under varying amounts of intercurrent events.
