arXiv proposes patient world model for health system simulation in intensive care
Original title: Learning a patient world model for health system simulation in intensive care
The model is trained on MIMIC-IV and couples time-queryable predictions to ICU calendars for simulating bed capacity and care dispatch.
Designing healthcare strategies requires anticipating clinical and operational consequences. Conventional approaches tailor state transitions to a single policy question. The authors learn a patient world model from longitudinal EHRs using a neural network to serve multiple simulation scenarios.
- Model matches Markov and semi-Markov comparators in aggregate occupancy.
- Learned representation captures 24-h recovery difference missed by compact state models.
- One model renders admission demand, care tasks, and discharge proposals across simulations.
From the original article
1 min read at arxiv.orgDesigning healthcare strategies requires anticipating their clinical and operational consequences before implementation, a task that often relies on health system simulation. Such a simulation necessitates modeling how patient states evolve and generate care demands.
By Yingtai Li · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.