Researchers at NYU Langone are pioneering the use of GenAI to make discharge summaries more understandable for patients.

NYU Langone Health is pioneering a GenAI project to make discharge summaries easier for patients to understand by translating medical jargon into lay language. One of the first to employ OpenAI’s GPT-4, this research highlights GenAI’s capacity to demystify healthcare communication. Although GenAI operates on probabilities rather than human-like reasoning and may introduce inaccuracies, it is designed to assist healthcare providers, not replace them.

In March 2023, NYU Langone utilized GPT-4 to investigate AI solutions for clinical problems while ensuring patient privacy. Their research involved converting 50 patient discharge notes into language that a 6th-grader could understand. The results showed significant potential for improving patient education. By using GPT-4, they were able to increase the readability score from 13% to 81%, demonstrating the ability of GenAI to enhance health literacy.

Although GPT-4 has shown to be efficient in simplifying medical terminology, it still lacks accuracy and completeness. Therefore, it is necessary to have physician oversight while using this technology. However, this approach can significantly reduce medical paperwork, and ease the burden on healthcare providers without compromising the quality of care. The study highlights the potential of GenAI to save time for healthcare professionals and reduce patient anxiety, with minimal supervision.

NYU Langone is planning to evaluate patient feedback on AI-generated summaries and expand their usage. This initiative demonstrates how GenAI can play a beneficial role in healthcare by refining patient communication and comprehension. Integrating GenAI technologies into patient care processes is expected to significantly enhance the overall healthcare experience.