A groundbreaking study from NYU Langone Health has shown that artificial intelligence (AI) can significantly improve the readability of patient education materials (PEMs). These improvements, driven by Artificial Intelligence, make them more accessible to a broader audience. This improvement is essential, as PEMs often exceed the recommended grade level for readability, hindering many patients’ ability to comprehend important healthcare information.
The research was published in the Journal of Medical Internet Research. It evaluated the readability of materials from three major health organizations. These organizations are the American Heart Association (AHA), the American Cancer Society (ACS), and the American Stroke Association (ASA). Powered by Artificial Intelligence, these materials are critical for helping patients make informed decisions about their health care. Yet, earlier studies showed that they often surpass the recommended grade 6 reading level. This makes them challenging for many individuals to understand.
AI MODELS ENHANCE READABILITY WITHOUT COMPROMISING ACCURACY
Researchers conducted the study to assess the ability of three large language models (LLMs). These models are ChatGPT, Gemini, and Claude. These Artificial Intelligence models simplified the texts of 60 randomly selected PEMs from the AHA, ACS, and ASA websites. The LLMs were prompted to reduce the complexity of the language without sacrificing the accuracy of the content.
The findings showed that the original readability scores of the materials were much higher than the recommended grade 6 reading level, with average scores of 10.7, 10, and 9.6 for AHA, ACS, and ASA materials, respectively. After using the AI tools to optimize the content, the results were significantly improved:
- ChatGPT reduced the reading level to 7.6 (grade level).
- Gemini brought the reading level to 6.6.
- Claude simplified it to a 5.6 grade level.
Additionally, the word count of the materials was significantly reduced, making the content more concise and digestible for patients.
IMPACT ON PATIENT EMPOWERMENT AND HEALTH OUTCOMES
Dr. Jonah Feldman, senior author of the study and medical director of transformation and informatics at NYU Langone, emphasized that the results demonstrate the potential of AI tools. These artificial intelligence tools can transform patient education materials into more readable content. This transformation is crucial for patient empowerment and improved health outcomes.
“These AI-driven improvements can enhance even expert-composed materials. These materials are already patient-directed. This demonstrates that technology can significantly influence clinical communication,” said Feldman, who is also an assistant professor at NYU Grossman Long Island School of Medicine.
REAL-WORLD APPLICATIONS OF AI IN HEALTHCARE
The study also highlights how healthcare organizations can use AI to make clinical communication more patient-friendly. Previous research demonstrated AI’s capabilities to generate patient-focused explanations for heart test results. It can draft responses to electronic health queries. Additionally, it can produce summaries of complex medical reports with Artificial Intelligence.
As Paul Testa, MD, JD, MPH, chief health informatics officer at NYU Langone, pointed out, these AI tools are already being put into practice. The team is conducting a randomized controlled trial using AI-generated patient-friendly summaries for hospital discharge instructions. The goal is to improve patient comprehension and satisfaction, ensuring better post-discharge care and smoother transitions.
“The potential applications for Artificial Intelligence in improving patient education are vast. We are committed to generating real-world evidence to validate these tools’ effectiveness,” said Testa, who is also a clinical professor at NYU Grossman School of Medicine.
THE FUTURE OF AI IN PATIENT COMMUNICATION
This study offers a powerful example of how Artificial Intelligence technology can be leveraged to enhance patient communication in healthcare. As Jonah Zaretsky, MD, associate chief of medicine at NYU Langone Hospital — Brooklyn, noted, generating real-world evidence through randomized trials is essential for confirming that AI-driven tools are not only accurate but also truly beneficial for patients and their families.
The integration of Artificial Intelligence into healthcare systems has the potential to revolutionize how information is communicated to patients, ensuring that critical medical knowledge is accessible, understandable, and actionable for everyone, regardless of their background or literacy level.
