Abstract

Mediation analyses of randomized trials and observational studies can generate evidence about the mechanisms by which interventions and exposures may influence health outcomes. Publications of mediation analyses are increasing, but the quality of their reporting is suboptimal.