Exploring the Use of ChatGPT in Delivering Evidence-Based Information to Patients with Thyroid Eye Disease
Hossein Ghahvehchian1 *, Mohammad Taher Rajabi2 , Seyed Mohsen Rafizadeh2
- Eye Research Center, The Five Senses Health Institute, Rassoul Akram Hospital, Iran University of Medical Sciences, Tehran, Iran
- Farabi Comprehensive Center of Excellence in Ophthalmology, Department of Oculofacial Plastic and Orbital Surgery, Tehran University of Medical Sciences, Tehran, Iran
Abstract: Thyroid eye disease (TED) is a prevalent autoimmune inflammatory condition that can significantly impact patients' quality of life. However, there is often a diagnostic lag in TED patients, possibly due to inadequate awareness. Patient counseling and education play a critical role in TED management. The emergence of artificial intelligence (AI), particularly ChatGPT, a large language model developed by OpenAI, has the potential to enhance patient education and care in various medical fields, including ophthalmology.
Methods: In this study, ChatGPT-3.5 was utilized to simulate a doctor-patient consultation regarding thyroid eye disease. Eight commonly asked questions in ophthalmology clinics were posed to ChatGPT, based on the Consensus Statement by the American Thyroid Association and the European Thyroid Association, as well as the 2021 European Group on Graves' orbitopathy (EUGOGO) clinical practice guidelines. The chat sessions were conducted on May 9, 2023, and the responses provided by ChatGPT were evaluated for accuracy, informativeness, and accessibility.
Results: ChatGPT offered patients with TED timely and reliable medical advice on common questions. The responses were detailed yet easily understandable, avoiding excessive medical terminology. The answers provided by ChatGPT were consistent with established guidelines, demonstrating its potential as a tool to enhance patient comprehension regarding thyroid eye disease. However, limitations were observed when ChatGPT was prompted to provide scientific references, as three out of five references provided were fabricated and not found in the relevant medical literature.
Conclusion: ChatGPT shows promise in delivering medical information to patients seeking knowledge on thyroid eye disease. However, caution must be exercised due to the limitations of AI-generated information. ChatGPT's training data is limited to information available until September 2021, potentially leading to outdated knowledge in rapidly evolving fields like TED. Additionally, the quality and representativeness of the data may impact the accuracy and completeness of the AI-generated information. Therefore, patients are advised to consult with qualified healthcare providers, including ophthalmologists, endocrinologists, and oculoplastic surgeons, to ensure personalized and up-to-date advice for the management of thyroid eye disease.
