ChatGPT as a Postoperative Tool Following Lower Lid Blepharoplasty

This study aimed to assess the potential role of large language models in the postoperative management of lower lid blepharoplasty patients.

This study employed a cross-sectional design to evaluate the content validity of ChatGPT-4's responses to common postoperative concerns following lower lid blepharoplasty: bleeding, diplopia, dry eyes, pain, lower lid retraction, and bruising/hematoma/orbital hemorrhage. Twelve American Society of Ophthalmic Plastic and Reconstructive Surgery physicians graded the responses for clinical accuracy on a 5-point Likert scale and assessed the appropriateness of management recommendations. Flesch Reading Ease scores and Flesch-Kincaid Grade Level were generated to assess readability. The Patient Education Materials Assessment Tool was utilized to evaluate understandability and actionability of each response.

Average accuracy grades for the 6 responses ranged from 4.33 to 4.67 out of 5, with an overall average of 4.50 ± 0.63. The proportion of graders who deemed ChatGPT-4's patient management recommendations appropriate ranged from 75.0% to 100% across the 6 responses. Flesch Reading Ease scores ranged from 28.9 to 50.2, with Flesch-Kincaid Grade Level scores ranging from grade 9.4 to 13.4 for reading level. The average Patient Education Materials Assessment Tool understandability score was 91.67 ± 0.52%, while all 6 responses received an actionability score of 60%.

ChatGPT-4 showed promise as a supportive tool for postoperative care following lower lid blepharoplasty, with generally accurate and appropriate responses. However, further refinement is needed before it can be considered a reliable clinical resource. Ongoing physician involvement in the development and evaluation of medical AI will be critical to ensuring these technologies are safe, accurate, and capable of meaningfully enhancing patient care.