Table 1.

Definitions of key clinical trial terminology.

Terminology	Definition
Estimand	The true target of an estimate for a particular clinical trial objective. It is defined by the subject population, the outcome, the handling of intercurrent events, and the statistical summary measure of effect. Estimates are produced by statistical estimation procedures applied to data and might depend on a variety of assumptions if the estimand of interest is not directly observable, e.g. due to imperfect adherence of subjects to the protocol.
Treatment-policy estimand	The effectiveness of an intervention regardless of events that occur after intervention is administered (e.g. compliance to intervention regime or attrition). The intention-to-treat principle (analyzing all available data from all randomized subjects) is applied when the treatment-policy estimand is desired. If all subjects are followed until the end of the trial the treatment-policy estimand can be estimated without assumptions.
Intercurrent event	An event which occurs after randomization to an intervention which might interfere with the estimation or interpretation of the effect of the intervention (e.g. initiation of a rescue therapy).
Hypothetical estimand	An alternative to the treatment-policy estimand under a particular hypothetical scenario (e.g., the efficacy of an intervention had an intercurrent event, such as initiation of rescue therapy, not occurred). Estimation of hypothetical estimands generally relies on untestable assumptions.