Large language model discharge summary preparation using real‐world electronic medical record data shows promise
The efficacy of large language models (LLMs) in discharge summary preparation using real clinical documentation remains novel. Our study aimed to test the efficacy of two LLMs to generate DC summaries which were scored using a validated discharge summary scoring metric. The models performed nearly identically, with the llama3:instruct model having a mean score of 19.1/31 (SD: 2.42) compared to 19.2/31 (SD: 3.48) when produced by llama3:70b. Using LLMs to aid in the generation of discharge summaries may help to reduce the overall clinical administrative workload.
Authors
- Jarrod A. Marks (ORCID: https://orcid.org/0000-0003-4279-3806)
- Stephen Bacchi (ORCID: https://orcid.org/0000-0001-5130-8628)
- Toby Gilbert (ORCID: https://orcid.org/0000-0002-3431-9672)
- Samuel Gluck (ORCID: https://orcid.org/0000-0001-7883-517X)
- Oliver Kleinig (ORCID: https://orcid.org/0000-0003-3320-4424)
- Lewis Hains (ORCID: https://orcid.org/0009-0003-0577-9105)
- Ashwin Murugappa
Institutions
- Harvard University (US)
- Lyell McEwin Hospital (AU)
- Harvard University Press (US)
- University of Massachusetts Boston (US)
- Massachusetts General Hospital (US)
- Dana-Farber/Harvard Cancer Center (US)
- The University of Adelaide (AU)
Publication Details
- Journal
- Internal Medicine Journal
- Published
- 2025-05-28
- DOI
- https://doi.org/10.1111/imj.70073
- Citations
- 2
- Primary Topic
- Hospital Admissions and Outcomes
- Type
- article
- Field-Weighted Citation Impact
- 4.51
Funders
- University of Adelaide