EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Performance evaluation of Meditron3-70B in medical coding: Current limitations and integration perspectives for clinical practice

Galland-Decker, Coralie, Usenbacher, Muaziza, Nunes, Christophe, Bouche, François, Carra, Giorgia, Boillat-Blanco, Noémie und 8 weitere

GMS Medizinische Informatik, Biometrie und Epidemiologie · 17.10.2025 · 21 (Doc15) · Doc15

DOI: 10.3205/mibe000287 ↗

Kurz gefasst

Fachartikel in der Zeitschrift GMS Medizinische Informatik, Biometrie und Epidemiologie, erschienen am 17.10.2025. Verfasst von Galland-Decker und 19 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

The emergence of large language models (LLMs) presents challenges for their integration into clinical practice, particularly for medical coding. This study evaluated the performance of Meditron3-70B, a state-of-the-art open-source medical LLM, in generating SNOMED CT and ICD-10 codes using 200 fictitious emergency department consultation vignettes. Healthcare professionals assessed the accuracy of the outputs. Although Meditron performed well on standard benchmarks such as MedQA, significant shortcomings were observed in the relevance and completeness of the generated diagnostic codes, with only 2% of responses deemed acceptable. While LLMs show promise for supporting clinical decision-making, their current ability to produce accurate and comprehensive medical codes remains limited. Integrating specialized retrieval tools through hybrid approaches could enhance coding accuracy and warrants further investigation in real-world clinical settings.

Schlagwörter

large language modelsemergency room visitsclinical codingSNOMED CTICD-10Large Language ModelsNotaufnahmeklinische KodierungSNOMED CTICD-10

Zitieren

APA

Galland-Decker, Coralie, Usenbacher, Muaziza, Nunes, Christophe, Bouche, François, Carra, Giorgia, Boillat-Blanco, Noémie, Hartley, Mary-Anne, Sallinen, Alexandre, Raisaro, Louis, J., Bastardot, & François (2025). Performance evaluation of Meditron3-70B in medical coding: Current limitations and integration perspectives for clinical practice. GMS Medizinische Informatik, Biometrie und Epidemiologie, 21(Doc15), Doc15. https://doi.org/10.3205/mibe000287

Vancouver

Galland-Decker, Coralie, Usenbacher, Muaziza, Nunes, Christophe, et al. Performance evaluation of Meditron3-70B in medical coding: Current limitations and integration perspectives for clinical practice. GMS Medizinische Informatik, Biometrie und Epidemiologie. 2025;21(Doc15):Doc15. doi:10.3205/mibe000287

BibTeX
@article{GallandDecker2025,
  author = {Galland-Decker and Coralie and Usenbacher and Muaziza and Nunes and Christophe and Bouche and François and Carra and Giorgia and Boillat-Blanco and Noémie and Hartley and Mary-Anne and Sallinen and Alexandre and Raisaro and Louis, Jean and Bastardot and François},
  title = {Performance evaluation of Meditron3-70B in medical coding: Current limitations and integration perspectives for clinical practice},
  journal = {GMS Medizinische Informatik, Biometrie und Epidemiologie},
  year = {2025},
  volume = {21},
  number = {Doc15},
  pages = {Doc15},
  doi = {10.3205/mibe000287}
}
Kostenlos starten →