AINeutralarXiv – CS AI · 3h ago6/10
🧠
Heterogeneous Causal Discovery of Repeated Undesirable Health Outcomes
Researchers present a novel causal discovery framework that combines multiple structure learning algorithms with heterogeneous effect estimation to identify drivers of undesirable health outcomes across patient subpopulations. Validated through healthcare applications examining emergency department revisits and hospital readmissions, the framework reveals that intervention effectiveness varies significantly by patient characteristics, prioritizing chronic disease management and care coordination as key targets.