Optimizing Anesthesia Care: Leveraging Large Language Models for Effective Preoperative Planning

Area of Science:

• Anesthesiology
• Artificial Intelligence
• Clinical Informatics

Background:

• The rapid advancement of artificial intelligence (AI) necessitates research into its clinical applications.
• A gap exists in understanding the utility of large language models (LLMs) within clinical practice, specifically in anesthesia.
• Evaluating AI tools for healthcare applications is crucial for safe and effective integration.

Purpose of the Study:

• To assess the effectiveness of ChatGPT, a large language model, in generating anesthesia care plans.
• To identify the strengths and limitations of using AI for drafting critical patient care documents.

Main Methods:

• First-year nurse anesthesiology residents inputted key anesthesia care plan variables into ChatGPT.
• The AI-generated anesthesia care plans were independently reviewed by three expert faculty members.
• Analysis focused on the accuracy and completeness of ChatGPT's output regarding essential anesthetic variables.

Main Results:

• ChatGPT successfully generated appropriate anesthetic plans based on provided medical history and home medications.
• The AI model demonstrated limitations in adequately addressing critical variables such as ventilator settings, fluid management, and medication dosages.
• The quality of ChatGPT's output was contingent upon the clarity and specificity of the input provided by the user.

Conclusions:

• Large language models like ChatGPT show potential as assistive tools in creating anesthesia care plans.
• Effective utilization of AI in clinical settings requires clear, detailed input from knowledgeable healthcare professionals.
• Further research is needed to refine AI capabilities for comprehensive anesthesia care plan generation.