Estimands in Health Technology Assessments - Methodological Considerations and Recommendations

The ICH E9 (R1) Addendum on Estimands and Sensitivity Analysis in Clinical Trials aligns the target of estimation (objective) with the method of estimation (estimator) and interpretation of results from clinical studies. Confirmatory clinical trials represent a critical source of evidence for health technology assessment (HTA), regardless of HTA archetype. While HTAs have traditionally focused on the internal validity of clinical trials, the estimands framework offers a unique opportunity to address external validity considerations for HTA purposes. We illustrate how the HTA policy aims are translated into research questions through the specification of a PICO (population, intervention, comparator, outcome) framework. We examine the current adoption of the Addendum by HTA Agencies, with specific focus on EUnetHTA 21 methodological guidelines for the EU HTA. We introduce the estimands framework and demonstrate how the different attributes of an estimand (population, treatment condition, variable of interest, summary measure and strategies for intercurrent event) can be used to strengthen external validity through case studies. The adoption of estimands in HTA guidelines is still in its early stages, and there is no consensus on their definition and application. However, some agencies are exploring the use of estimands in their assessments. Using some case studies, we show that strategies departing from the treatment policy estimand may help in minimizing the risk of external validity bias, e.g., when the trial population does not conform to the target population, or the trial’s rescue medication differs from the standard of care in routine clinical practice. The estimands framework plays an important role in healthcare decision-making by precisely describing the question of interest. Early consideration of estimands can bring benefits such as improved external validity and better-informed choices. Further research is needed to establish best practices and guidelines for the use of estimands in HTA.

In Value in Health, 26(12)