CommentaryAn introduction to mediation analyses of randomised controlled trials

Mediation analyses of randomised controlled trials can be used to investigate the mechanisms by which health interventions cause outcomes. In this article we provide a brief introduction to mediation analysis in the context of randomised controlled trials. We introduce common target effects, causal assumptions, estimation approaches, and illustrate these concepts using a published mediation analysis of the Systolic Blood Pressure Intervention Trial (SPRINT). Well-conducted mediation analyses of randomised trials can provide meaningful insights to guide clinical and policy decisions.

Source: Journal of Clinical Epidemiology - February 15, 2021 Category: Epidemiology Authors: Aidan G. Cashin, Hopin Lee Source Type: research