EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins
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 inference-time planning over simulations.
- ▪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 inference-time planning over simulations.
Opening excerpt (first ~120 words) tap to expand
Statistics > Machine Learning arXiv:2607.08793 (stat) [Submitted on 7 Jul 2026] Title:EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins Authors:Joshua Pickard, Wei Qi, Na Li, Ann Woolley, Lisa Cosimi, Roy Kishony, Deborah Hung View a PDF of the paper titled EHR-MPC: Inference-Time Control for Sepsis Treatment with Generative Patient Digital Twins, by Joshua Pickard and 6 other authors View PDF HTML (experimental) Abstract: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.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
