GenAI can help simplify medical jargon in discharge notes, study finds
Facts at your Fingertips
By Emily Margosian, Assistant Editor, February 1, 2025
Can artificial intelligence humanize patient discharge summaries? By law, patients have the right to immediate access to discharge notes in their medical records. However, technical language and abbreviations can make notes difficult to understand for the average patient.
A recent study in JAMA Network Open found that the use of large language models (LLMs) showed promise in improving the readability of patient discharge summaries, converting them into patient-friendly language and format (doi:10.1001/jamanetworkopen.2024.0357). The cross-sectional study of 50 discharge summaries found that understandability scores were significantly higher for notes processed by the LLM, although physician review was still required for each note due to concerns regarding safety and accuracy. Two physicians reviewed each patient-friendly discharge summary for accuracy on a six-point scale, with 54% of AI-assisted summaries receiving the best possible rating. Summaries were rated entirely complete in 56% of reviews, and 18% of reviews were noted as safety concerns, mostly involving omissions, but also several inaccurate statements (termed hallucinations).
For a more detailed breakdown of how AI-enhanced discharge summaries stacked up, see the graphic below
