Abstract: SA-PO002
The Role of Artificial Intelligence (AI) Chatbots in Educating Patients with CKD on Intermittent Fasting
Session Information
- Augmented Intelligence, Large Language Models, and Digital Health
October 26, 2024 | Location: Exhibit Hall, Convention Center
Abstract Time: 10:00 AM - 12:00 PM
Category: Augmented Intelligence, Digital Health, and Data Science
- 300 Augmented Intelligence, Digital Health, and Data Science
Authors
- Shah, Joel N., Texas Tech University Health Sciences Center El Paso, El Paso, Texas, United States
- Ahmad, Adeel, Texas Tech University Health Sciences Center El Paso, El Paso, Texas, United States
- Thongprayoon, Charat, Mayo Foundation for Medical Education and Research, Rochester, Minnesota, United States
- Cheungpasitporn, Wisit, Mayo Foundation for Medical Education and Research, Rochester, Minnesota, United States
- Acharya, Prakrati C., Texas Tech University Health Sciences Center El Paso, El Paso, Texas, United States
Background
Intermittent fasting (IF) is a growing nutritional intervention trend that may provide health and weight loss benefits in patients. However, research is limited regarding the effects of IF in CKD patients. AI Chatbots hold enormous potential in patient education regarding various lifestyle modifications. This study aims to provide an analysis of the effectiveness of AI Chatbots on patient education regarding the effects of IF in CKD patients.
Methods
A Nephrologist generated 25 questions related to IF and CKD. Responses were generated from three AI-based Chatbots, Chat GPT 3.5, Chat GPT 4.0, and Claude 3 Sonnet. Responses for each question were generated twice across all three Chatbots with >24 hours between each. Responses were evaluated and compared to the current literature regarding IF and CKD for accuracy.
Results
The 3 AI Chatbots each generated accurate responses though the vocabulary, use of paraphrasing, and formatting differed. Each chatbot’s individual responses were roughly similar though some key differences were noted. Chat GPT 4.0 noted “[Patients] should be fully informed about...the lack of extensive research specifically in CKD populations.” in response number 2 for a question, an answer absent from the 1st response. Claude 3 Sonnet was noted to have large variations between its responses, including one question in which electrolyte imbalances were categorized in the 1st response but replaced by a discussion of proteinuria in the 2nd. Chat GPT 4.0, Chat GPT 3.5, and Claude 3 Sonnet referenced limited literature or the need for more extensive research in 32%, 42%, and 40% of the responses generated respectively.
Conclusion
AI Chatbots provide a novel way to educate patients with CKD on lifestyle modifications including IF. Chat GPT 3.5, Chat GPT 4.0, and Claude 3 Sonnet provided mostly accurate responses with evidence of active learning between the 2 responses for each. Chat GPT 3.5, despite being the free version of Chat GPT, provided a greater percentage of responses correctly addressing the limited literature or the need for more research regarding IF and CKD. AI Chatbots require additional technological modifications and advancements to improve patient education regarding lifestyle interventions. Additionally, further research is necessary to establish the effects of IF in CKD patients.