Aim: The study aimed to evaluate the performance of ChatGPT in the context of inflammatory bowel disease (IBD) and to determine its reliability and usefulness for both patients and healthcare professionals.
Methods: The study involved 20 specific questions related to the two main types of IBD: Crohn’s disease (CD) and ulcerative colitis (UC). These questions were divided into two sets: one for healthcare professionals and one for patients. The answers were evaluated using a seven-point Likert-type scale for reliability and usefulness.
Results: The highest scores for both reliability and usefulness were obtained from questions sourced from professionals (5.00±1.21 for reliability and 5.15±1.08 for usefulness). CD-related questions scored 4.70±1.26 (reliability) and 4.75±1.06 (usefulness), while UC-related questions scored 4.40±1.21 (reliability) and 4.55±1.31 (usefulness). Notably, the reliability scores for answers to professional questions were significantly higher than those for patient questions.
Conclusion: While ChatGPT has shown potential in terms of reliability and usefulness for IBD-related queries, there are still areas of improvement. Enhancing ChatGPT with more detailed and current information could make it a valuable resource for both patients and medical professionals.
Discussion: The study found that ChatGPT exhibits partial reliability and usefulness in the field of IBD, with some limitations. It could be a more reliable source for healthcare professionals compared to patients. The study also highlighted the need for developers to update and enhance ChatGPT, ensuring the information provided comes from reliable sources.