In a groundbreaking development, artificial intelligence (AI) tools have demonstrated the ability to significantly improve the readability of online patient education materials (PEMs), making them more accessible to a broader audience. A recent study conducted by researchers at NYU Langone Health underscores how these digital tools can play a pivotal role in patient empowerment and improving healthcare outcomes.
Making Sense of Complex Information
Patient education materials on websites of leading health organizations such as the American Heart Association (AHA), American Cancer Society (ACS), and American Stroke Association (ASA) serve as critical resources for individuals seeking to understand their health conditions and treatment options. However, a significant barrier to comprehension is that these materials often exceed the recommended reading level of grade 6, making them difficult for many patients to understand.
The study investigated the ability of AI tools, specifically large language models (LLMs) like ChatGPT, Gemini, and Claude, to simplify these texts without sacrificing their accuracy. These LLMs, trained on extensive internet data, can predict the next word in a sentence with high precision, enabling them to rephrase complex articles into simpler language effectively.
Study Insights and Results
The study analyzed 60 random PEMs from the websites of AHA, ACS, and ASA. Initially, the materials ranked significantly above the ideal readability levels, with average grade-level scores of 10.7, 10, and 9.6, respectively. Following the application of LLMs, these scores improved to more accessible levels: ChatGPT adjusted the text to a grade level of 7.6, Gemini refined it to 6.6, and Claude simplified it to a more comprehensible level of 5.6. Additionally, AI processing resulted in more concise wording, enhancing overall material usability.
Dr. Jonah Feldman, senior author of the study, highlighted the transformative potential of AI in rendering expert-composed medical content more readable, thereby facilitating better patient understanding and health outcomes. The research also showcases how healthcare institutions can effectively employ AI to enhance patient communication.
Practical Implications and Future Directions
The application of AI in health communication extends well beyond rewriting PEMs. It shows promise in generating patient-friendly summaries for various applications, including explaining heart test results and providing responses to electronic queries. Currently, the NYU Langone team is incorporating AI-generated materials into a controlled trial to assess their impact on patient comprehension and satisfaction, particularly concerning hospital discharge instructions.
Dr. Paul Testa, co-author of the study, underlined the ongoing application of these AI tools within NYU Langone, emphasizing the need for real-world evidence from randomized trials to validate the clinical effectiveness of such technology.
Key Takeaways
This study highlights the significant role AI can play in enhancing healthcare communication to be more inclusive and understandable. By bridging the gap between complex medical information and patient comprehension, AI tools have the potential to empower patients, leading to more informed healthcare decisions and improved overall well-being. As AI technology continues to evolve, it is likely to remain a valuable ally in clinical settings worldwide, offering individualized, clear, and supportive healthcare communication.