Artificial Intelligence in Patient Education: ChatGPT for Penile Prosthesis Surgery Counseling, a Likert Scale Analysis

Le résumé fourni par la source

INTRODUCTION: Erectile dysfunction (ED) significantly impacts quality of life. For patients unresponsive to conservative treatments, penile prosthesis implantation represents an effective solution with high satisfaction rates. However, the complexity and sensitive nature of the procedure often generate multiple patient concerns. Artificial intelligence (AI) chatbots such as ChatGPT are increasingly used for medical information, although concerns remain regarding accuracy and misinformation. This study aimed to evaluate the clarity, accuracy, and clinical appropriateness of ChatGPT's responses to frequently asked questions (FAQs) on penile prosthesis surgery. METHODS: Fifteen representative FAQs were identified from patient forums and educational platforms. Questions were submitted to ChatGPT (GPT-4, November 2024) in a single session. A panel of seven expert urologists evaluated responses using a four-point Likert scale across four domains: clarity, medical accuracy, relevance, and risk of misinformation. RESULTS: All responses achieved mean scores above 3, indicating overall clinical acceptability. The highest ratings were observed for relevance (3.40 ± 0.54) and clarity (3.37 ± 0.57). Slightly lower scores were reported for medical accuracy (3.19 ± 0.67) and risk of misinformation (3.18 ± 0.59). A strong positive correlation was found between clarity and relevance (ρ = 0.80), while medical accuracy correlated with lower misinformation risk (ρ = 0.77). CONCLUSIONS: ChatGPT provides generally clear, relevant, and clinically acceptable responses for patient education in penile prosthesis surgery, although minor clarifications are often required. It may serve as a supportive tool to complement traditional counseling but cannot replace physician-patient communication, which remains essential for individualized care and informed decision-making.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.