{"id":"arxiv/2607.08793","url":"https://arxiv.org/abs/2607.08793","tags":["stat.ML","cs.AI","cs.LG","eess.SY","math.OC"],"score":0,"title":"EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins","domain":"stat.ML","source":"arxiv","authors":["Joshua Pickard","Wei Qi","Na Li","Ann Woolley","Lisa Cosimi","Roy Kishony","Deborah Hung"],"reviews":0,"subject":"stat.ML","summary":"Sepsis is a leading cause of mortality, yet optimal treatment policies remain contested. Existing reinforcement learning (RL) approaches learn fixed strategies for sepsis treatment, limiting adaptability to changing clinical objectives during inference. We propose EHRMPC, a framework that decouples learning patient dynamics from optimizing treatment by training a patient digital twin in the form of a generative electronic health record (EHR) model. The digital twin predicts clinical trajectories under interventions and enables model predictive control (MPC) to optimize treatments via…","category":"research","comments":0,"disputed":false,"division":"stat","mentions":0,"discussed":0,"published":"2026-07-07","reactions":{}}