Improving the Power to Detect Indirect Effects in Mediation Analysis

TL;DR

This paper introduces two new intersection-union test methods that improve the power to detect indirect effects in causal mediation analysis while maintaining proper error control, demonstrated through simulations and a proteomic study.

Contribution

The paper proposes novel intersection-union test methods that enhance power in mediation analysis, addressing limitations of existing tests.

Findings

Proposed methods outperform existing tests in simulation studies.

New tests maintain proper type I error control.

Application to proteomic data illustrates practical utility.

Abstract

Causal mediation analysis seeks to determine whether an independent variable affects a response variable directly or whether it does so indirectly, by way of a mediator. The existing statistical tests to determine the existence of an indirect effect are overly conservative or have inflated type I error. In this article, we propose two methods based on the principle of intersection-union tests that offer improvements in power while controlling the type I error. We demonstrate the advantages of the proposed methods through extensive simulation. Finally, we provide an application to a large proteomic study.