Introduction: This study aimed to evaluate whether ChatGPT provides satisfactory responses to frequently asked questions (FAQs) about joint health and physiotherapy approaches for patients with hemophilia.
Materials and Methods: Fifty questions were reviewed by five expert physiotherapists, who selected the 10 FAQs. Responses were generated using the GPT-4o model of ChatGPT on March 27, 2026. Two non-blinded reviewers independently evaluated the responses using a predefined four-point rating system based on accuracy, completeness, clarity, and clinical relevance. Inter-rater reliability was assessed using the intraclass correlation coefficient (ICC) with a two-way random-effects model for absolute agreement. All responses were analyzed using the Flesch–Kincaid readability index to assess readability.
Results: The median response accuracy score (RAS) was 2, indicating that responses were generally satisfactory and required minimal clarification. Inter-rater reliability between the two reviewers was good (ICC=0.847). The Flesch–Kincaid score was 25, indicating that ChatGPT responses were difficult to read and understand by college graduates. Discussion and Conclusion: ChatGPT demonstrated potential to improve basic knowledge of joint health in hemophilia, with excellent responses to 30% of FAQs. However, responses to physiotherapy-related questions often required additional clarification, particularly regarding exercise prescription parameters. Because physiotherapy interventions are individually tailored and require clinical supervision, ChatGPT may enhance patients’ knowledge but cannot replace supervised physiotherapy practice. Therefore, ChatGPT should be considered a complementary tool rather than a primary clinical decision-making resource. Integrating artificial intelligence with physiotherapy expertise may enhance patient education and hemophilia care.