Can AI Answer My Questions? Utilizing Artificial Intelligence in the Perioperative Assessment for Abdominoplasty...

Area of Science:

• Medical Informatics
• Artificial Intelligence in Healthcare
• Plastic Surgery

Background:

• Abdominoplasty addresses cosmetic and functional concerns, including divarication of recti, post-weight loss, and post-pregnancy.
• Patient-surgeon communication gaps can impede informed decision-making regarding abdominoplasty.
• Large language models (LLMs) present an opportunity to improve patient information and education.

Purpose of the Study:

• To evaluate the feasibility and efficacy of using LLMs for answering patient queries related to abdominoplasty.
• To compare the performance of four leading LLMs (ChatGPT-3.5, Claude, Gemini, CoPilot) in providing perioperative information.

Main Methods:

• Four LLMs were tested using fifteen unique prompts relevant to abdominoplasty.
• Readability was assessed using Flesch-Kincaid, Flesch Reading Ease, and Coleman-Liau index.
• Quality was evaluated using DISCERN scores and Likert scales by surgical residents and specialists.

Main Results:

• ChatGPT-3.5 required the highest comprehension level, followed by Gemini, Claude, and CoPilot.
• Claude provided the most appropriate and actionable advice.
• CoPilot demonstrated superior patient-friendliness and engagement, incorporating visual aids, though with some limitations.

Conclusions:

• LLMs exhibit distinct differences in readability and reliability for patient education in abdominoplasty.
• Careful selection of LLMs is essential to leverage their advantages and mitigate weaknesses.
• Further research is needed to optimize LLM integration for enhanced patient education in plastic surgery.