Leveraging Large Language Models (LLM) for the Plastic Surgery Resident Training: Do They Have a Role?

Area of Science:

• Medical Education
• Artificial Intelligence in Surgery
• Plastic Surgery Training

Background:

• Large language models (LLMs) possess capabilities for text recognition, summarization, translation, prediction, and generation.
• ChatGPT4, a transformer-based LLM, utilizes deep learning for text prediction.
• The potential of LLMs as teaching assistants (TAs) in plastic surgery education requires exploration.

Purpose of the Study:

• To investigate the feasibility and effectiveness of utilizing LLMs as TAs in plastic surgery residency programs.
• To identify specific roles and tasks where LLMs can contribute to surgical education.

Main Methods:

• Literature review to identify existing TA roles.
• Development of a list of suitable tasks for LLM implementation.
• Creation and application of specific prompts for ChatGPT to generate educational content.
• Evaluation of generated content by study authors for accuracy and usefulness, with interobserver agreement analysis.

Main Results:

• Eight distinct TA roles for LLMs in plastic surgery education were identified.
• Generated content was assessed for usefulness and accuracy by eight study authors.
• 100% interobserver agreement was achieved for content accuracy, usefulness, and clarity in interactive case studies, preoperative consultation simulations, and ethical considerations generation.

Conclusions:

• LLMs, specifically ChatGPT, can accurately and readably generate educational content for plastic surgery residents.
• LLMs can function as valuable TAs, providing automated content synthesis for summaries, Q&A, and case scenarios.
• Integration of LLMs into residency programs offers a dynamic, interactive, and individualized learning experience, fostering critical thinking through feedback and reflection.