Predictive Causal Inference via Spatio-Temporal Modeling and Penalized Empirical Likelihood

TL;DR

This paper presents a novel integrated framework combining spatio-temporal modeling and penalized empirical likelihood to improve predictive causal inference in complex biomedical data, addressing limitations of traditional methods.

Contribution

It introduces a combined HMM and MTGCN approach that jointly models spatial and temporal data for more accurate causal effect estimation in biomedical applications.

Findings

Enhanced bias correction and predictive accuracy demonstrated in simulations

Effective modeling of latent disease dynamics in clinical domains

Framework adaptable to complex spatiotemporal biomedical data

Abstract

This study introduces an integrated framework for predictive causal inference designed to overcome limitations inherent in conventional single model approaches. Specifically, we combine a Hidden Markov Model (HMM) for spatial health state estimation with a Multi Task and Multi Graph Convolutional Network (MTGCN) for capturing temporal outcome trajectories. The framework asymmetrically treats temporal and spatial information regarding them as endogenous variables in the outcome regression, and exogenous variables in the propensity score model, thereby expanding the standard doubly robust treatment effect estimation to jointly enhance bias correction and predictive accuracy. To demonstrate its utility, we focus on clinical domains such as cancer, dementia, and Parkinson disease, where treatment effects are challenging to observe directly. Simulation studies are conducted to emulate latent disease dynamics and evaluate the model performance under varying conditions. Overall, the proposed framework advances predictive causal inference by structurally adapting to spatiotemporal complexities common in biomedical data.