Dataset for prompt strategies and anaesthesia
This repository contains supplementary materials supporting a study evaluating large language model responses to 17 standardised clinical scenarios across major domains of perioperative anaesthesia. The repository includes: Clinical Scenarios: The complete set of 17 standardised clinical scenarios covering airway management, regional anaesthesia, obstetric anaesthesia, cardiovascular and respiratory anaesthesia, critical care, paediatric anaesthesia, and acute pain management. Reference Answer Checklists: Guideline-based checklists outlining essential clinical management steps and safety-critical considerations for each scenario. Evaluation Scores: Item-level assessment scores assigned to the large language model responses, providing the underlying data used for statistical analysis. R Scripts: The R code used for statistical analyses, enabling verification and reproduction of the reported results. These materials are made publicly available to promote methodological transparency, reproducibility, and independent verification of the study findings. The repository provides access to the clinical scenarios, reference standards, evaluation data, and statistical analysis code used in the study.
Authors
- Gül Çakmak (ORCID: https://orcid.org/0000-0001-6900-0293)
- Ceren Önal (ORCID: https://orcid.org/0000-0003-2196-2500)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-15
- DOI
- https://doi.org/10.5281/zenodo.24245800
- Primary Topic
- Artificial Intelligence in Healthcare and Education
- Type
- article
- Field-Weighted Citation Impact
- 0.00