Dynamic Survival Transformers for Causal Inference with Electronic Health Records

TL;DR

This paper introduces DynST, a deep transformer model that leverages time-varying EHR data to improve causal survival analysis and treatment effect estimation.

Contribution

The paper presents DynST, a novel deep survival transformer that captures complex interactions and time-varying information in EHRs for causal inference.

Findings

DynST accurately estimates causal effects on RMST.

DynST outperforms alternative models in predictive accuracy.

DynST effectively utilizes time-varying data for survival prediction.

Abstract

In medicine, researchers often seek to infer the effects of a given treatment on patients' outcomes. However, the standard methods for causal survival analysis make simplistic assumptions about the data-generating process and cannot capture complex interactions among patient covariates. We introduce the Dynamic Survival Transformer (DynST), a deep survival model that trains on electronic health records (EHRs). Unlike previous transformers used in survival analysis, DynST can make use of time-varying information to predict evolving survival probabilities. We derive a semi-synthetic EHR dataset from MIMIC-III to show that DynST can accurately estimate the causal effect of a treatment intervention on restricted mean survival time (RMST). We demonstrate that DynST achieves better predictive and causal estimation than two alternative models.