Evaluating Understandability and Actionability of Artificial Intelligence in Creating a High-Quality Educational Content for Type 2 Diabetes Using Predesigned Prompts

Background: AI has the potential to revolutionize healthcare, with chatbots like ChatGPTemerging as practical tools for creating educational content. Type 2 diabetes has become a significant global health challenge, necessitating effective prevention and management strategies. Patient education is crucial for informed decision-making, but traditional materials often lack accessibility and engagement. Aim: This study aim to evaluates ChatGPT's understandability and actionability in generating high-quality educational content for type 2 diabetes using predesigned prompts. Methods: A qualitative exploratory design was employed to assess the content generated by ChatGPT. The Patient Educational Materials Assessment Tool (PEMAT) focused on five significant health education domains relevant to diabetes. The PEMAT assessment included 17 items for understandability and 7 for actionability, with a score of≥70% required to pass. Results: Prompts with a simple, clear instructional guideline yielded significantly higher scores than those with multiple instructions or single prompt with no addition. This indicates that focused prompt design enhances the quality of AI-generated educational content. Conclusion: This study demonstrates that artificial intelligence, particularly ChatGPT, has the potential to generate high-quality, understandable, and actionable educational content for patients with type 2 diabetes, when guided by simple well-structured, predesigned prompts and supported by scientific source material.