Internal Medicine Resident Loyola Medicine Macneal Hospital Berwyn, Illinois, United States
Objective : Integrating artificial intelligence (AI) into patient medical education enhances awareness and knowledge about the prevention and management of diabetes mellitus. Diabetes mellitus poses a global health challenge, necessitating innovative approaches for effective patient education. Leveraging AI technologies like machine learning algorithms and natural language processing enables personalized and interactive educational modules. For patients with low literacy, traditional written materials may pose challenges in comprehension, potentially leading to misunderstandings or inadequate adherence to medical guidance. AI interventions can bridge this gap by offering alternative communication modes catering to diverse learning preferences.
Methods: Our intervention utilizes readily available patient education materials from reputable websites like CDC's "Living with Diabetes" and Uptodate Patients' Education. We seek to differentiate this traditional approach by introducing a user-friendly AI-driven platform that tailors educational content to individual patient profiles, such as the free versions of Bing AI and ChatGPT 3.5. Common queries in our clinical setting encompass the importance of weight loss, the significance of exercise, the necessity of home blood sugar checks, dietary guidelines, and the meaning of A1c. An endocrinologist from our hospital evaluated responses using the Likert scale for quality in each option (Poor-Fair-Good-Very Good-Excellent), followed by a comparative analysis of the scores.
Results: An exemplary response is distinguished by its restrained utilization of medical terminology, a focused consideration for individuals with constrained health literacy, and a comprehensive expansion in content. Ultimately, Bing AI has evidenced its adeptness as the primary provider of outstanding patient answers, achieving ratings of Good and Very Good in 90% of responses. This is further enhanced by including original websites at the conclusion, facilitating additional patient reading. In comparison, ChatGPT performed well, rating Good in most responses. Notably, websites such as CDC and Uptodate may be limited in information and are not optimized for individuals with low health literacy.
Discussion/Conclusion: An AI-driven platform exhibits the capacity to empower patients with practical insights, promoting a proactive stance toward diabetes prevention and management. The study posits that incorporating AI into patient education shows potential for enhancing healthcare outcomes in chronic diseases, specifically diabetes mellitus, while acknowledging the importance of considering patient health literacy. AI has the capacity to personalize educational content based on individual literacy levels, ensuring that information is presented at an appropriate complexity level. Subsequent research is merited to evaluate its long-term efficacy, scalability, and influence on clinical outcomes.
