What is your estimand? Navigating marginal and conditional treatment effects

At ISPOR Europe 2025 in Glasgow, I also joined Shannon Cope (PRECISIONheor), David M. Phillippo (Bristol University) and Jeroen P. Jansen (University of California San Francisco) in an exciting issue panel about marginal versus conditional estimands and implications for health economic modeling.

My presentation was largely based on the assessment of clinical efficacy/effectiveness, and my co-presenters delved more deeply into implications for economic evaluation using decision analytical modeling and cost-effectiveness analysis. I defined and clarified the interpretation of different marginal and conditional average treatment effects. I later outlined potential estimand mismatches under different hypothetical outcome-generating mechanisms, with and without effect modification. While these mismatches have primarily been recognized for non-collapsible summary measures, they are also problematic for some collapsible measures in the presence of effect modification. Finally, I discussed implications for transportability and evidence synthesis. I have co-authored a paper with David covering these topics.

The day before the talk, I had the pleasure of joining Antonia Morga (Astellas Pharma) in presenting our poster “Can Estimands Support Aligned Evidence Generation for EU HTA? Reflections on their Role in JCA, PICO Alignment and Beyond”. Our work, developed on behalf of the EFPIA/EFSPI Estimands Implementation Working Group on HTA and RWE, was selected among the Top 5% of posters and featured in a Poster Tour Presentation. We are honored that our abstract was recognized as a Top 5% Finalist for the ISPOR Europe 2025 Research Presentation Awards, a great acknowledgment of our collective effort and collaboration to advance the discussion on how estimands can enhance evidence alignment for HTA.