A clinical environment simulator for dynamic AI evaluation - PubMed
7 hours ago
- #AI in healthcare
- #Clinical simulation
- #LLM evaluation
- Proposes the Clinical Environment Simulator (CES) for dynamic evaluation of clinical LLMs in digital hospital settings.
- CES features a 'hospital engine' for real-time resource tracking and a 'patient engine' for disease progression simulation.
- Evaluates LLMs on temporal reasoning, resource-aware decision-making, and operational resilience through realistic EHR interfaces.
- Shifts focus from static datasets to dynamic, integrated healthcare system evaluations.
- Authors disclose various competing interests, including affiliations with AI and healthcare companies.