Path Dependent Structural Equation Models

TL;DR

This paper introduces Path Dependent Structural Equation Models (PDSEMs), a new framework for causal analysis of systems with state-dependent causal relationships and counterfactual intervention effects in longitudinal data.

Contribution

The paper proposes PDSEMs, extending causal graphical models to handle state-dependent causal structures and counterfactual interventions in dynamic systems.

Findings

Demonstrated causal inference methods within PDSEMs

Simulations validate the model's ability to capture state-dependent causality

Applied the approach to real surgical procedure data

Abstract

Causal analyses of longitudinal data generally assume that the qualitative causal structure relating variables remains invariant over time. In structured systems that transition between qualitatively different states in discrete time steps, such an approach is deficient on two fronts. First, time-varying variables may have state-specific causal relationships that need to be captured. Second, an intervention can result in state transitions downstream of the intervention different from those actually observed in the data. In other words, interventions may counterfactually alter the subsequent temporal evolution of the system. We introduce a generalization of causal graphical models, Path Dependent Structural Equation Models (PDSEMs), that can describe such systems. We show how causal inference may be performed in such models and illustrate its use in simulations and data obtained from a septoplasty surgical procedure.