Assessing the quality and readability of ChatGPT responses to frequently asked questions about trigger finger release

Artificial intelligence, specifically large language models like ChatGPT, is rapidly transforming the healthcare landscape. As ChatGPT becomes more popular for obtaining medical information, there are concerns regarding the accuracy and quality of its content. While prior studies in various medical specialties have yielded mixed results regarding ChatGPT's reliability, little research has focused on its ability to address questions regarding specific orthopedic procedures, such as surgical intervention for stenosing tenosynovitis (trigger finger). This study assessed the accuracy, clarity, and readability of ChatGPT's responses to ten commonly asked patient questions regarding trigger finger release. The questions were obtained from Google's "People also ask" section and queried in ChatGPT 4.0 on September 24, 2024. Responses were evaluated by two authors using a four-point accuracy scale. Additionally, the education level required to understand the responses was assessed using the Flesch-Kincaid scale. ChatGPT's responses achieved an average score of 1.9, falling between "excellent, requiring no clarification" and "satisfactory, requiring minimal clarification." Although the chatbot provided largely accurate information, it produced an incorrect response in one case and displayed occasional factual inaccuracies, particularly regarding treatment recommendations. The average reading level of responses was at a 12th-grade level, which exceeds the recommended 7th-8th-grade level for patient materials. ChatGPT can serve as a useful starting point for patients seeking information about orthopedic procedures like trigger finger release, but healthcare providers should guide patients in validating AI-generated content to enhance medical literacy and ensure accurate understanding.