A new study by NYU Langone Health reveals that AI tools significantly enhance the readability of patient education materials from major organizations, promoting patient empowerment and better health outcomes.
Artificial Intelligences making significant strides in health care, with a new study from NYU Langone Health demonstrating that AI can substantially improve the readability of online patient education materials (PEMs).
The research focused on PEMs from the websites of the American Heart Association (AHA), American Cancer Society (ACS) and American Stroke Association (ASA). These materials are crucial for helping patients make informed decisions about their health but often are written at a reading level too high for many patients to understand.
The study, recently published in the Journal of Medical Internet Research, evaluated the capabilities of three large language models (LLMs) — ChatGPT, Gemini and Claude — in optimizing the readability of these materials.
These generative AI tools simplify complex texts by predicting the next word in a sentence based on extensive data. This process enables the AI to rewrite complicated articles in a more accessible language.
Study Highlights
The researchers randomly selected 60 PEMs from the AHA, ACS and ASA websites. The original readability scores were significantly higher than the recommended sixth-grade level, averaging grades 10.7, 10 and 9.6, respectively.
After being processed by the LLMs, the readability of these materials improved markedly:
- ChatGPT reduced the average reading level to 7.6.
- Gemini lowered it to 6.6.
- Claude improved it to an average of 5.6.
Word counts also decreased, making the materials more concise and easier to understand.
“Our study shows that widely used large language models have the potential to transform patient education materials into more readable content, which is essential for patient empowerment and better health outcomes,” senior author Jonah Feldman, a medical director of transformation and informatics at NYU Langone, said in a news release.
Feldman, who also serves as an assistant professor at NYU Grossman Long Island School of Medicine, emphasized that even expertly crafted education materials could benefit from AI-driven enhancements.
The research showcases how health care organizations can leverage AI to make clinical communication more patient-friendly.
Previous studies have shown AI’s potential in generating patient-focused explanations of medical results, drafting responses to electronic health queries and creating understandable summaries of complex reports.
“The breadth of possible AI offerings shows how technology can be leveraged to transform the patient experience across health care systems, and not just in the United States,” added co-author Paul Testa, a chief health informatics officer at NYU Langone and clinical professor at NYU Grossman School of Medicine..
NYU Langone is already incorporating these AI tools in a randomized controlled trial to test their efficacy in crafting patient-friendly hospital discharge summaries. The goal is to see if clearer instructions improve patient comprehension and satisfaction post-discharge.
“Generating real-world evidence through randomized trials is crucial for validating the effectiveness of AI tools in clinical settings,” added co-author Jonah Zaretsky, an associate chief of medicine at NYU Langone Hospital – Brooklyn and clinical assistant professor at NYU Grossman School of Medicine. “This approach ensures that the AI-generated documentation is not only accurate but also genuinely beneficial for patients and their families.”
This study was self-funded by NYU Langone, demonstrating their commitment to advancing health care through innovative technologies. Besides Feldman, Testa and Zaretsky, the research team included lead author John Will and co-authors Mahin Gupta and Aliesha Dowlath.
Source: NYU Langone Health