Estimation of treatment effects in short‐term depression studies. An evaluation based on the ICH E9(R1) estimands framework

Abstract

Estimands aim to incorporate intercurrent events in design, data collection and estimation of treatment effects in clinical trials. Our aim was to understand what estimands may correspond to efficacy analyses commonly employed in clinical trials conducted before publication of ICH E9(R1). We reanalysed six clinical trials evaluating a new anti-depression treatment. We selected the following analysis methods-ANCOVA on complete cases, following last observation carried forward (LOCF) imputation and following multiple imputation; mixed-models for repeated measurements without imputation (MMRM), MMRM following LOCF imputation and following jump-toreference imputation; and pattern-mixture mixed models. We included a principal stratum analysis based on the predicted subset of the study population who would not discontinue due to adverse events or lack of efficacy. We translated each analysis into the implicitly targeted estimand, and formulated corresponding clinical questions. We could map six estimands to analysis methods. The same analysis method could be mapped to more than one estimand. The major difference between estimands was the strategy for intercurrent events, with other attributes mostly the same across mapped estimands. The quantitative differences in MADRS10 population-level summaries between the estimands were 4-8 points. Not all six estimands had a clini-

Consensus Statement for Protocols of Factorial Randomized Trials

“…Understanding the exact treatment effect being estimated is essential to proper interpretation of study results. However, this is not always clear from the study methods alone . A particular issue for factorial trials is that the treatment groups used for comparison are not always the same as those in which there is interest in estimating the treatment effect .…”

Reporting of Factorial Randomized Trials

“…The statistical methods alone are not always sufficient to allow readers to understand the exact treatment effect (estimand) being estimated . In factorial trials, the treatment groups used for comparison are not always the same as those in which there is interest in estimating the treatment effect .…”

“…• Whether the target treatment effect for each main comparison pertains to the effect in the presence or absence of other factors The statistical methods alone are not always sufficient to allow readers to understand the exact treatment effect (estimand) being estimated. [30][31][32] In factorial trials, the treatment groups used for comparison are not always the same as those in which there is interest in estimating the treatment effect. 11,33 For example, many factorial trials use a factorial analysis to compare "all A" vs "all not A" for reasons of efficiency, even though interest really lies in the effect of A alone vs control (the effect of A in the absence of B) or, alternatively, the effect of A plus B vs B alone (the effect of A in the presence of B) if treatment B has been demonstrated to be effective.…”

Tell me what you want, what you really really want: Estimands in observational pharmacoepidemiologic comparative effectiveness and safety studies

“…Another limitation is that we did not explicate how an estimand can inform the choice of study design and analytical approach. We refer to existing papers on this topic, 29,33,41,45–47 particularly to explanations about aligning estimands to estimators in randomized trials 38,48 . An increasingly popular methodological approach is the target trial emulation framework 13,49–51 .…”

“…We refer to existing papers on this topic, 29,33,41,[45][46][47] particularly to explanations about aligning estimands to estimators in randomized trials. 38,48 An increasingly popular methodological approach is the target trial emulation framework. 13,[49][50][51] Target trial emulation provides a structure to estimate a treatment effect of interest using observational data.…”