Artificial Intelligence in Rhinoplasty Recovery: Linguistic Intelligence and Machine Learning-Driven Insights

Area of Science:

• Medical Informatics
• Artificial Intelligence in Medicine
• Otolaryngology

Background:

• Postoperative care is crucial for rhinoplasty outcomes.
• Patients often seek readily accessible information regarding recovery.
• Large language models (LLMs) like ChatGPT-4 offer potential for patient information dissemination.

Purpose of the Study:

• To evaluate ChatGPT-4's performance as a postoperative information tool for rhinoplasty.
• To assess the accuracy, clarity, relevance, and patient-centeredness of AI-generated responses.
• To compare AI responses with expert otolaryngologist (ENT) evaluations.

Main Methods:

• An observational, cross-sectional simulation study design was employed.
• Ten standardized postoperative rhinoplasty questions were posed to ChatGPT-4.
• Responses were independently rated by blinded ENT specialists using a structured performance instrument and linguistic/statistical analyses.

Main Results:

• ChatGPT-4 achieved high accuracy (90%) and clarity (87%) scores.
• Patient-centered communication scores were lower (77%).
• Machine learning analysis identified clarity, diagnostic depth, and empathy as key factors for high ratings.

Conclusions:

• ChatGPT-4 demonstrates significant potential in providing accurate and clear postoperative rhinoplasty information.
• While AI tools can complement clinician counseling, they do not substitute for individualized medical advice or surgical follow-up.
• Further development is needed to enhance the patient-centered communication capabilities of AI in medical contexts.