Recommendations for improving causal inference of mediation analysis

Abstract

Mediation analysis is an essential tool for investigating how a treatment causally affects an outcome via intermediate variables. However, violations of the (often implicit) causal assumptions can severely threaten the validity of causal inferences of mediation analysis. Psychologists have recently started to raise such concerns, but the discussions have been limited to mediation analysis with a single mediator. In this article, we examine the causal assumptions when there are multiple possible mediators. We pay particular attention to the practice of exploring mediated effects along various paths linking several mediators. Substantive conclusions using such methods are predicated on stringent assumptions about the underlying causal structure that can be indefensible in practice. Therefore, we recommend that researchers shift focus to mediator-specific indirect effects using a recently proposed framework of interventional (in)direct effects. A vital benefit of this approach is that valid causal inference of mediation analysis with multiple mediators does not necessitate correctly assuming the underlying causal structure among the mediators. Finally, we provide a practical guide with suggestions to improve the research practice of mediation analysis at each study stage. We hope this article will encourage explication, justification, and reflection of the causal assumptions underpinning mediation analysis to improve the validity of causal inferences in psychology research.

Reductions in kinesiophobia and distress after pain neuroscience education and exercise lead to favourable outcomes: a secondary mediation analysis of a randomized controlled trial in primary care

“…This study has several strengths compared with the previous mediation literature in pain rehabilitation. It followed an appropriate reporting guideline 36 and recent recommendations 41,56 to reduce risk of bias specific to mediation analysis with multiple mediators. In contrast to previous studies, 56 the postintervention outcome value was also included as another competing mediator as its omission can result in an overestimation of the indirect effect through other mediators.…”

“…This mediation analysis was conducted and reported following the AGReMA 36 guidelines and recent recommendations for causal inference of mediation analysis with multiple mediators. 41,56…”

“…We justify treatment-mediator, mediator-outcome, and mediator-mediator causal assumptions depicted in the directed acyclic graph with theoretical and empirical-based rationale in Table 1. 41,87…”