Can ChatGPT-4o Educate the Public About Vasectomy?: A Google Trends–Based Expert Panel Assessment (Preprint)

BACKGROUND ChatGPT-4o, a state-of-the-art large language model (LLM), has gained widespread public attention for its capacity to provide health information. However, its effectiveness in addressing sensitive medical topics such as vasectomy has not been systematically evaluated. OBJECTIVE This study aimed to assess the accuracy, comprehensiveness, and public suitability of ChatGPT-4o’s responses to frequently asked public questions about vasectomy, as identified through Google Trends. METHODS A total of seven experts (four urologists, two public health professionals, and one fertility nurse) evaluated ChatGPT-4o's answers to 15 high-frequency vasectomy-related questions. Evaluations were based on six criteria—medical accuracy, completeness, clarity, tone, public usefulness, and recommendability—using a 5-point Likert scale. Descriptive statistics, Kruskal–Wallis tests, and inter-rater reliability (ICC) were used for analysis. RESULTS Mean ratings across all criteria ranged from 3.75 to 4.04, indicating generally favorable assessments. Clarity and tone received the highest average scores, while medical accuracy showed moderate variability. The Kruskal–Wallis test revealed no statistically significant differences between expert groups (p > .05). However, the ICC value (–0.01) suggested poor inter-rater agreement, reflecting variation in expert perspectives. CONCLUSIONS ChatGPT-4o responses were generally evaluated as clear, appropriate in tone, and useful for the general public. Nonetheless, differences in expert evaluation and the lack of layperson input highlight the need for cautious integration of AI-generated health content. Expert oversight and user-centered design are recommended to enhance the reliability and accessibility of such tools in public health education. CLINICALTRIAL Not applicable. This study is a retrospective, non-interventional, multidisciplinary expert panel evaluation and did not require trial registration.