Abstract:Mediation analysis aims to decipher the underlying causal mechanisms between an exposure, an outcome, and intermediate variables called mediators. Initially developed for fixed-time mediator and outcome, it has been extended to the framework of longitudinal data by discretizing the assessment times of mediator and outcome. Yet, processes in play in longitudinal studies are usually defined in continuous time and measured at irregular and subject-specific visits. This is the case in dementia research when cerebral and cognitive changes measured at planned visits in cohorts are of interest. We thus propose a methodology to estimate the causal mechanisms between a time-fixed exposure ($X$), a mediator process ($\mathcal{M}_t$) and an outcome process ($\mathcal{Y}_t$) both measured repeatedly over time in the presence of a time-dependent confounding process ($\mathcal{L}_t$). We consider three types of causal estimands, the natural effects, path-specific effects and randomized interventional analogues to natural effects, and provide identifiability assumptions. We employ a dynamic multivariate model based on differential equations for their estimation. The performance of the methods are explored in simulations, and we illustrate the method in two real-world examples motivated by the 3C cerebral aging study to assess: (1) the effect of educational level on functional dependency through depressive symptomatology and cognitive functioning, and (2) the effect of a genetic factor on cognitive functioning potentially mediated by vascular brain lesions and confounded by neurodegeneration.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is how to conduct continuous - time mediation analysis of the causal mechanisms among time - fixed exposure variables (such as educational level or genetic factors), mediation processes (such as depressive symptoms or cognitive functions), and outcome processes (such as functional dependence or cognitive functions) in longitudinal studies. Traditional mediation analysis methods usually assume that the mediator and outcome variables are measured at fixed time points, but many health processes are essentially dynamic and are measured in an irregular manner at different time points. Therefore, this paper proposes a new methodological framework aimed at estimating the causal effects existing in these continuous - time processes, especially in the presence of time - dependent confounding factors. Specifically, the paper proposes the following innovations: 1. **Continuous - time framework**: Unlike traditional methods, this method takes into account the changes of mediator and outcome variables in continuous time, not just at fixed time points. 2. **Time - dependent confounding factors**: This method can handle time - dependent confounding factors, which is a common problem in many practical studies. 3. **Multiple causal effects**: The paper defines three types of causal effects: natural effects, path - specific effects, and stochastic intervention - analog natural effects, and provides identifiability assumptions for these effects. 4. **Dynamic multivariate model**: Use a dynamic multivariate model based on differential equations to estimate these causal effects, which can capture the mutual influence between dynamic processes. Through these innovations, the paper aims to provide a more accurate and flexible method for analyzing complex longitudinal data, especially in brain - aging studies, such as changes in cognitive functions and brain structures. This method helps to better understand biological processes and identify potential treatment or prevention targets.

Continuous-time mediation analysis for repeatedly measured mediators and outcomes

Identification and estimation of mediational effects of longitudinal modified treatment policies

Causal Mediation Analysis for Sparse and Irregular Longitudinal Data

Non-linear Mediation Analysis with High-dimensional Mediators whose Causal Structure is Unknown

Mediation analysis of time‐to‐event endpoints accounting for repeatedly measured mediators subject to time‐varying confounding

Estimation and inference for the mediation effect in a time-varying mediation model

Causal Mediation Analysis in the Context of Clinical Research.

Mediation Analysis with Multiple Exposures and Multiple Mediators

Efficient and flexible causal mediation with time-varying mediators, treatments, and confounders

Latent Class Dynamic Mediation Model with Application to Smoking Cessation Data

Model Selection for Exposure-Mediator Interaction

A Framework for Mediation Analysis with Multiple Exposures, Multivariate Mediators, and Non-Linear Response Models

Longitudinal mediation analysis of time-to-event endpoints in the presence of competing risks

Multivariate Dynamic Mediation Analysis under a Reinforcement Learning Framework

Methods of analysis for survival outcomes with time-updated mediators, with application to longitudinal disease registry data

Clinically Relevant Mediation Analysis using Controlled Indirect Effect

Causal Mediation Analysis With a Binary Outcome and Multiple Continuous or Ordinal Mediators: Simulations and Application to an Alcohol Intervention

Mediation with External Summary Statistic Information (MESSI)

Practical causal mediation analysis: extending nonparametric estimators to accommodate multiple mediators and multiple intermediate confounders

Causal mediation analysis in presence of multiple mediators uncausally related

Estimating Causal Mediation Effects under Correlated Errors