Current and future state of evaluation of large language models for medical summarization tasks
Large Language Models have expanded the potential for clinical Natural Language Generation (NLG), presenting new opportunities to manage the vast amounts of medical text. However, their use in such high-stakes environments necessitate robust evaluation workflows. In this review, we investigated the current landscape of evaluation metrics for NLG in healthcare and proposed a future direction to address the resource constraints of expert human evaluation while balancing alignment with human judgments.
Authors
- Karen Wong (ORCID: https://orcid.org/0000-0002-9109-0173)
- Brian W. Patterson (ORCID: https://orcid.org/0000-0002-4584-3808)
- Nicholas Pellegrino (ORCID: https://orcid.org/0000-0002-2102-2431)
- Frank Liao (ORCID: https://orcid.org/0000-0002-7577-7530)
- Majid Afshar (ORCID: https://orcid.org/0000-0002-6368-4652)
- Emma Croxford (ORCID: https://orcid.org/0009-0001-9117-7009)
- Elliot First
- Yanjun Gao (ORCID: https://orcid.org/0000-0002-9341-7360)
- Cherodeep Goswami (ORCID: https://orcid.org/0009-0008-9805-4694)
- Graham Wills
Institutions
- University of Wisconsin–Madison (US)
- University of Wisconsin Health (US)
- University of Colorado Anschutz Medical Campus (US)
- Epic Systems (United States) (US)
Publication Details
- Journal
- npj Health Systems
- Published
- 2025-02-03
- DOI
- https://doi.org/10.1038/s44401-024-00011-2
- Citations
- 39
- Primary Topic
- Topic Modeling
- Type
- article
- Field-Weighted Citation Impact
- 96.76
Funders
- U.S. National Library of Medicine