Enhancing Patient Education with AI-Driven Simplification of Medical Materials

Recent advancements in artificial intelligence (AI) have demonstrated significant potential to improve the accessibility of online patient education materials (PEMs). A study led by researchers at NYU Langone Health examined how large language models (LLMs), such as ChatGPT, Gemini, and Claude, can optimize the readability of medical information without sacrificing accuracy. These tools work by predicting and replacing complex language with simpler, clearer phrases, making health information easier for a broad audience to understand.

The study focused on PEMs from well-known health organizations like the American Heart Association, American Cancer Society, and American Stroke Association. It revealed that many of these materials are written at reading levels above the recommended grade 6, often reaching levels around grades 10 to 11, which can pose challenges for many patients in comprehension and decision-making. By prompting LLMs to simplify the text, researchers successfully reduced the reading level to approximately grade 6, thereby enhancing patient understanding. Additionally, the AI rewrite performance resulted in more concise content, potentially making it easier for patients to grasp key information quickly.

According to senior author Dr. Jonah Feldman, the use of AI models holds promise for transforming how healthcare organizations communicate with patients. He emphasizes that even well-intentioned educational materials can benefit from AI-driven refinement, ensuring they are truly patient-friendly. This approach not only empowers patients but also supports better health outcomes.

The study's findings suggest that AI tools can be integrated into healthcare practices to improve the clarity of discharge instructions, consent forms, and other critical documentation. Currently, NYU Langone is implementing these AI techniques in clinical trials to generate simplified hospital discharge summaries, aiming to enhance patient comprehension, satisfaction, and postdischarge care. Experts highlight that validating these AI applications through rigorous clinical trials is essential to confirm their effectiveness and safety.

This research underscores the broader potential of AI to revolutionize healthcare communication—not just within the United States but globally—by making complex medical information accessible and understandable for all patients.