Patient education tools have been vital for patient advocacy in their healthcare process. Recent developments have shown the important of patient advocacy and involvement in their healthcare process. For patients of chronic disease, such as Diabetes Mellitus, this involvement can produce significant differences in outcomes. The challenge many clinicians face is how to best provide information to patients. Due to this, patients are often referred to external resources. However, these resources may not always be easily understood. With the introduction of the ChatGPT and AI-based chatbots, patients have access to health information through a unique medium, which may be able to better explain and synthesize information. This study compared the ChatGPT answers to common Diabetes Mellitus questions to the accepted answers from organizations such as the World Health Organization (WHO), Centers for Diease Control & Prevention (CDC), and American Diabetes Association (ADA). The answers were assessed based on accuracy (as compared to the accepted responses), what types of intervention were emphasized, and the Fleish-Kincaid readability score. Based on this,the ChatGPT was able to provide accurate information with a clear emphasis on providing lifestyle based advice. This illustrated the holistic improvement that was present in all of the ChatGPT recommendations. The ChatGPT response readability was deemed more difficult than the official responses in 19/20 questions. The AI responses provided accurate and holistic information across the categories assessed. The personal questions were unique as they allowed for the AI algorithm to construct its own advice. The personalization aspect highlighted in this category is one of the greatest strengths offered by the AI algorithm which cannot be found in the standard sources.
Primary Speaker
Faculty Sponsors