How Reliable Are The ChatGPT Responses To Frequently Asked Questions About Talus Osteochondritis Dissecans?

KARAPINAR S. E., Dincer R.

Süleyman Demirel Üniversitesi Tıp Fakültesi Dergisi, cilt.32, sa.3, ss.198-205, 2025 (TRDizin)

Özet

Objective: Talus OCD (Osteochondritis Dissecans) is a painful ankle pathology that affects the subchondral bone and cartilage. This study aimed to determine whether artificial intelligence could provide accurate and sufficient responses to frequently asked questions about this pathology. Material and Method: ChatGPT (Chat Generative Pre-Trained Transformer) version 4.0 was asked 10 questions selected from the frequently asked questions on the websites of various orthopaedic institutions. The responses were analyzed with the Mika and Discern accuracy scores. Readability was based on the Flesch-Kincaid Grade Level, Gunning-Fog Index, Coleman-Liau Index, Automatic Readability Index, Simple Measure of Gobbledygook Index, Gobbledygook Index, Fry Readability Graph, and the Raygor Readability Estimate. The Reading grade levels were stated as USA grade levels. Results: According to the Mika scoring system, the mean score of the responses was 2.1. The general responses were satisfactory and required minimal explanation. The mean Discern score was 48, seen to be an average response given. In the index measurements for readability, the mean points were 15.14, which was equivalent to the reading level of a 3rd year university student. Conclusion: Artificial intelligence is a promising technological development. The continuously developing world of medicine does not yet have the most up-to-date data, and the readability level was seen to be very high. It was seen that the responses were made without showing evidence, which was lacking in the literature.