Abstract
Artificial intelligence (AI), particularly large language models (LLMs), has demonstrated potential to improve patient communication by delivering accurate, readable, and empathetic medical information. In plastic surgery-a specialty where preoperative counseling is essential-patients often seek online information to understand procedures, risks, and recovery. However, limited research exists on ChatGPT-4o's utility in this context.BACKGROUNDArtificial intelligence (AI), particularly large language models (LLMs), has demonstrated potential to improve patient communication by delivering accurate, readable, and empathetic medical information. In plastic surgery-a specialty where preoperative counseling is essential-patients often seek online information to understand procedures, risks, and recovery. However, limited research exists on ChatGPT-4o's utility in this context.This prospective qualitative study evaluated ChatGPT-4o's responses to 25 standardized patient questions across 5 plastic surgery procedures: rhinoplasty, breast augmentation, abdominoplasty, blepharoplasty, and rhytidectomy. Each procedure was queried with 5 common preoperative questions covering indications, alternatives, risks, surgical steps, and recovery. Responses were reviewed independently by a board-certified plastic surgeon and 2 researchers for accuracy, completeness, and appropriateness. Strengths, weaknesses, omissions, and potentially unsafe guidance were identified and summarized.METHODSThis prospective qualitative study evaluated ChatGPT-4o's responses to 25 standardized patient questions across 5 plastic surgery procedures: rhinoplasty, breast augmentation, abdominoplasty, blepharoplasty, and rhytidectomy. Each procedure was queried with 5 common preoperative questions covering indications, alternatives, risks, surgical steps, and recovery. Responses were reviewed independently by a board-certified plastic surgeon and 2 researchers for accuracy, completeness, and appropriateness. Strengths, weaknesses, omissions, and potentially unsafe guidance were identified and summarized.ChatGPT-4o provided generally accurate, well-structured, and patient-friendly answers across all procedures, with no unsafe recommendations. Strengths included clear explanations of surgical rationale, common risks, general procedural steps, and recovery expectations. The model promoted safety and professional consultation. However, notable limitations included a lack of procedural nuance, omission of less common but clinically important risks, failure to tailor guidance to individual variables, and incomplete recovery or postoperative care details.RESULTSChatGPT-4o provided generally accurate, well-structured, and patient-friendly answers across all procedures, with no unsafe recommendations. Strengths included clear explanations of surgical rationale, common risks, general procedural steps, and recovery expectations. The model promoted safety and professional consultation. However, notable limitations included a lack of procedural nuance, omission of less common but clinically important risks, failure to tailor guidance to individual variables, and incomplete recovery or postoperative care details.ChatGPT-4o offers significant promise as a supplementary patient education tool in plastic surgery. Its ability to deliver coherent, empathetic, and accessible responses may help bridge health literacy gaps. However, it should not replace detailed, individualized surgeon-patient discussions. Further refinement and real-world validation are needed to enhance its clinical reliability and integration into patient care.CONCLUSIONSChatGPT-4o offers significant promise as a supplementary patient education tool in plastic surgery. Its ability to deliver coherent, empathetic, and accessible responses may help bridge health literacy gaps. However, it should not replace detailed, individualized surgeon-patient discussions. Further refinement and real-world validation are needed to enhance its clinical reliability and integration into patient care.