ChatSLE: consulting ChatGPT-4 for 100 frequently asked lupus questions
Large language models (LLMs) have shown huge potential in answering patient questions with good quality and readability.1 As the shortage of specialists in rheumatology worsens, the question arises as to what extent LLMs could support physicians in providing patients with comprehensive information. Previous studies often do not have a suitable control group to the LLM-generated responses using, for example, replies from doctors on social media forums.2,3 Because these studies compared the content generated by LLMs with texts that were not explicitly created as high-quality patient information, it cannot be determined how well a LLM performs for this purpose.