Boosting the precision of mediation analyses of randomised experiments through covariate adjustment

Abstract

Analyses of randomised experiments frequently include attempts to decompose the intention-to-treat effect into a direct and indirect effect, mediated by given intermediaries, with the aim to shed light onto the treatment mechanism. Methods from causal mediation analysis have facilitated this by allowing for arbitrary models for the outcome and the mediator. They thereby generalise the traditional approach to direct and indirect effects, which is essentially limited to linear models. The default maximum likelihood methods make use of a model for the conditional distribution of the mediator, given treatment and baseline covariates, but are prone to bias when that model is misspecified. In randomised experiments, specification of such model can be easily avoided, but at the expense of a sometimes major efficiency loss when those baseline covariates are predictive of the mediator. In this article, we develop a compromise approach: it makes use of a model for the mediator to optimally extract information from the baseline covariate data but is insulated from the impact of misspecification of that model; it achieves this by exploiting the known randomisation probabilities. Simulation studies and the analysis of a randomised study show major efficiency gains and confirm our theoretical findings that the default methods from causal mediation analysis are sometimes, although not always, reasonably robust to model misspecification. Copyright © 2017 John Wiley & Sons, Ltd.

Beyond total treatment effects in randomised controlled trials: Baseline measurement of intermediate outcomes needed to reduce confounding in mediation investigations

“…A change score approach was again found to be suffering from bias when baseline measures predicted subsequent change. Finally in line with the study by Vandenberghe et al, 26 we found that for continuous mediators and outcomes results were robust against misspecification of the mediator model.…”

A Pre-Implementation Enhancement Strategy to Increase the Yield of Training and Consultation for School-Based Behavioral Preventive Practices: a Triple-Blind Randomized Controlled Trial

“…To examine the hypothesized change mechanisms of SC-PIES (RQs 3 and 4), we sequentially explored the mediation effect of ITI on the relationship between study conditions and intervention fidelity, followed by a secondary model where teachers' fidelity was hypothesized to mediate the relationships between their ITI and class-wide AET. To control for baseline status, the mediation models were fitted with change scores that were calculated by subtracting pretest values from the posttest (Vandenberghe et al, 2017). The nonparametric bootstrapping mediation analysis was used rather than the causal steps approach (Baron & Kenny, 1986) because of its superior capacities to handle small samples, estimate robust standard errors, and detect mediation effects without significant total effects (Hayes, 2017).…”

Counterfactual mediation analysis in the multistate model framework for surrogate and clinical time‐to‐event outcomes in randomized controlled trials

“…Hsu et al outlined a conceptual framework for a counterfactual mediation model for time‐varying exposures and time‐varying surrogates 36 . Vandenberghe et al described a counterfactual‐based model for surrogate evaluation with a binary mediator and a continuous or binary true endpoint 37 . They proposed several estimators using propensity score‐based models and generalized linear models.…”