What question did this study set out to answer?

This research aims to evaluate a virtual stroke alert simulator designed for neurology residents and assess its effectiveness in tracking competency.

Abstract Number: Esoc2026a2380 a Novel LLM-Based Approach to Simulating Stroke Alerts for Neurology Residents

View Full Paper

Ask AI

Bookmark

Authors

Randomized trial shows improved stroke alert performance in neurology residents, suggesting enhanced training methods.

This research aims to evaluate a virtual stroke alert simulator designed for neurology residents and assess its effectiveness in tracking competency.
Thirty participants, including neurology residents and attending physicians, completed three LLM-based stroke alert simulations.
Performance was graded by two blinded reviewers based on transcripts of the simulations.
Model fidelity, medical accuracy, and a metric for information control were also evaluated.
In the first simulation, 60% of PGY1s offered TNK after recognizing subtherapeutic warfarin, compared to 75% of PGY2-4s and 83% of stroke attendings.
In the second simulation, fundoscopy was performed by 0% of PGY1s, 30% of PGY2-4s, and 67% of stroke attendings.
The model demonstrated 90% medical accuracy and maintained 100% script adherence during the simulations.

Rayala et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark