A NEW study suggests that ChatGPT4 can provide responses to patient questions about systemic lupus erythematosus (SLE) that are comparable to, and in some areas superior to, those of human specialists. The AI chatbot outperformed rheumatologists in patient satisfaction and empathy, raising the possibility of AI-powered support for medical consultations.
The study analysed 95 commonly asked lupus-related questions in both English and Chinese, comparing answers generated by ChatGPT4 and five practicing rheumatologists. A panel of specialists and patients with lupus scored responses across six domains, including scientific accuracy, logic, completeness, patient satisfaction, and empathy.
ChatGPT4 responses were rated higher than those from rheumatologists in satisfaction and empathy by specialists. Patients found ChatGPT4’s responses to be on par with rheumatologists in all domains. The chatbot performed particularly well in Chinese, demonstrating greater logical consistency, comprehensibility, and completeness. However, ChatGPT4 struggled slightly with comprehensibility in English.
The study highlights AI’s potential role in supporting patient education and engagement in rheumatology, particularly for chronic conditions like lupus. While AI-driven chatbots aren’t a replacement for doctors, they could help bridge gaps in accessibility and provide supplemental information between consultations.
Aleksandra Zurowska, EMJ
Reference
Xu D et al. ChatGPT4’s proficiency in addressing patients’ questions on systemic lupus erythematosus: a blinded comparative study with specialists, Rheumatology. 2024;63(9):2450-6.
