Driving · 2026-10-14
PatientWM: Towards Generalizable Clinical World Models via Unified Patient States
Ziwen Kan, Jie Peng, Bo Yan, Davis Hobley, Zheng Zhang, Yu Tian, Tianlong Chen, Song Wang
AI-assisted summary based on the paper. Consult the linked source for methods and limitations. Benchmark numbers appear after editor verification.
TL;DR
PatientWM is a clinical world-model framework that learns a unified predictive patient state from heterogeneous observations using availability-aware encoders. It conditions next-state prediction on the current state, supplied clinical action, and an ordered bounded history of preceding latent patient-state and action pairs, using typed decoders for task-specific clinical readouts. Experiments on MIMIC-III, MIMIC-IV, and three external datasets report superior clinical prediction, multi-step forecasting, and action-planning evaluation, with transfer/generalization via frozen-state reuse.
Why it matters
The authors address two limits of prior clinical world models: limited-scope patient states tied to specific modalities/targets, and implicit/unspecified history that does not expose ordered state–action transition context. PatientWM’s unified patient state with typed decoders and sequential latent history aims to enable reuse of the learned state for new clinical prediction tasks and improved multi-step/action-conditioned forecasting for sequential clinical decision support.
Method
- Availability-aware observation encoding: fuse only modalities available at each decision time into a unified predictive patient state z_t; typed decoders produce task-specific clinical readouts without defining the state around any single target.
- Sequential latent history: retain a bounded ordered sequence of preceding latent patient-state and clinical-action pairs H_t, and condition each transition on (z_t, a_t, H_t).
- Residual, action-conditioned latent dynamics with recursive rollouts: predict next latent state as z_{t+1} = z_t + Δ_θ(z_t, α_t, H_t), and recursively reuse predicted states while supplying externally recorded actions.
Abstract (from arXiv)
Clinical decisions depend on anticipating how a patient's condition will evolve as treatment is adjusted and new evidence becomes available. Existing medical world models capture patient-state evolution, but often define the evolving state around a particular modality or task, limiting reuse as inputs or prediction targets change. These models also retain history through observation sequences or recurrent or latent summaries rather than exposing preceding patient states and clinical actions as an ordered transition context. This distinction matters because similar current measurements may follow different recent changes, treatment exposures, and observed responses. To tackle these limitations, we propose PatientWM, a clinical world-model framework built around a unified patient state and sequential latent history. Availability-aware encoders integrate observations available at each decision time into the patient state, while typed decoders produce task-specific clinical readouts without defining the state around any single target. PatientWM further conditions each next-state prediction on the current state and supplied action together with a bounded, ordered history of preceding latent patient states and clinical actions, preserving recent treatment and response context. Experiments across MIMIC-III, MIMIC-IV, and three external clinical datasets show superior performance in clinical prediction, multi-step forecasting, and action-planning evaluation. Transfer experiments further demonstrate the generalizability of PatientWM's patient state. Together, these results establish its ability to forecast clinical evolution and support planning.
Related papers
- Forecasting Patient Trajectories from the Electronic Health Records with Action-Conditioned World Models
- Repeated-Measure Leakage, Distribution Shift, and Reliability under Partial Observation in Patient World Models
- Same Predictions, Different Harms: Causal Auditing of Patient World Models
- Patient, Place, Prior (P$^3$): What Counts as Personalization in Medical World Models?
- Understanding Trajectory Heterogeneity in Federated World Model Learning
Publication record
Bibliographic evidence describes where and when a record appeared. It does not validate scientific claims.
First public release: 2026-10-12 · Indexed by Pub-AI: 2026-10-14 UTC
PatientWM: Towards Generalizable Clinical World Models via Unified Patient States
Preprint posted: 2026-10-12 · Original record ↗
- arxiv
- 2610.17059
Explicit source versions (1)
- v1 · 2026-10-12
- Preprint available
Source date: 2026-10-12 · First detected: 2026-10-25 UTC · Historical baseline
Source evidence (1)
- Official evidence ↗ · atom:published (official metadata)
Abstract · arxiv
Clinical decisions depend on anticipating how a patient's condition will evolve as treatment is adjusted and new evidence becomes available. Existing medical world models capture patient-state evolution, but often define the evolving state around a particular modality or task, limiting reuse as inputs or prediction targets change. These models also retain history through observation sequences or recurrent or latent summaries rather than exposing preceding patient states and clinical actions as an ordered transition context. This distinction matters because similar current measurements may follow different recent changes, treatment exposures, and observed responses. To tackle these limitations, we propose PatientWM, a clinical world-model framework built around a unified patient state and sequential latent history. Availability-aware encoders integrate observations available at each decision time into the patient state, while typed decoders produce task-specific clinical readouts without defining the state around any single target. PatientWM further conditions each next-state prediction on the current state and supplied action together with a bounded, ordered history of preceding latent patient states and clinical actions, preserving recent treatment and response context. Experiments across MIMIC-III, MIMIC-IV, and three external clinical datasets show superior performance in clinical prediction, multi-step forecasting, and action-planning evaluation. Transfer experiments further demonstrate the generalizability of PatientWM's patient state. Together, these results establish its ability to forecast clinical evolution and support planning.