Alert button
Picture for Prayag Chatha

Prayag Chatha

Alert button

Dynamic Survival Transformers for Causal Inference with Electronic Health Records

Add code
Bookmark button
Alert button
Oct 25, 2022
Prayag Chatha, Yixin Wang, Zhenke Wu, Jeffrey Regier

Figure 1 for Dynamic Survival Transformers for Causal Inference with Electronic Health Records
Figure 2 for Dynamic Survival Transformers for Causal Inference with Electronic Health Records
Figure 3 for Dynamic Survival Transformers for Causal Inference with Electronic Health Records
Figure 4 for Dynamic Survival Transformers for Causal Inference with Electronic Health Records

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.

* Accepted to the NeurIPS 2022 Workshop on Learning from Time Series for Health 
Viaarxiv icon
Github IconAccess Paper or Ask Questions
Twitter IconFacebook IconLinkedin IconWhatsapp IconMessenger Icon