EUVIMEDEUVIMED entdecken
FallberichtOpen Access · CC BY-NC 4.0

Retrieval-Augmented Generation for ICD-10 Coding in German Clinical Texts – A Technical Case Report

Mario Krumscheid, Johannes Blömer, Matthias Becker

Studies in Health Technology and Informatics · 2025

DOI: 10.3233/shti251397 ↗

Kurz gefasst

Fallbericht in der Zeitschrift Studies in Health Technology and Informatics, 2025. Verfasst von Mario Krumscheid und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Computer Science, Kempten University of Applied Sciences, Kempten, Germany. Hauptthemen laut MEDLINE: International Classification of Diseases, Natural Language Processing, Clinical Coding, Information Storage and Retrieval. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction: Manual ICD-10 coding of German clinical texts is time-consuming and error-prone. This project aims to develop a semi-automated pipeline for efficient coding of unstructured medical documentation. State of the art: Existing approaches often rely on fine-tuned language models that require large datasets and perform poorly on rare codes, particularly in low-resource languages such as German. Concept: The proposed system integrates Named Entity Recognition, semantic and lexical retrieval, abbreviation resolution, and context-aware normalization within a Retrieval-Augmented Generation (RAG) framework using a compact generative model. Implementation: The pipeline utilizes Sentence-BERT embeddings, FAISS indexing, and the Mistral-Small-Instruct model. ICD codes are assigned through a combination of semantic similarity and generative refinement among the top retrieval candidates. Lessons learned: Major sources of error were found in semantic retrieval and diagnosis normalization. Future improvements should focus on domain-specific German embeddings, more robust abbreviation handling, and enhanced context-aware prompting to increase accuracy and usability in clinical environments.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

International Classification of DiseasesNatural Language ProcessingClinical CodingInformation Storage and RetrievalElectronic Health RecordsGermanySemanticsHumansVocabulary, Controlled

Stichwörter der Autor:innen

Artificial IntelligenceClinical CodingInternational Classification of DiseasesMedical Informatics ApplicationsNatural Language Processing

Autor:innen und Institutionen

  • Department of Computer Science, Kempten University of Applied Sciences, Kempten, Germany.
  • Bavarian Center for Digital Health and Social Care, Kempten, Germany.
  • Charité - Universitätsmedizin Berlin, Freie Universität Berlin and Humboldt-Universität zu Berlin, Berlin, Germany.

Zitieren

APA

Krumscheid, M., Blömer, J., & Becker, M. (2025). Retrieval-Augmented Generation for ICD-10 Coding in German Clinical Texts – A Technical Case Report. Studies in Health Technology and Informatics. https://doi.org/10.3233/shti251397

Vancouver

Krumscheid M, Blömer J, Becker M. Retrieval-Augmented Generation for ICD-10 Coding in German Clinical Texts – A Technical Case Report. Studies in Health Technology and Informatics. 2025. doi:10.3233/shti251397

BibTeX
@article{Krumscheid2025,
  author = {Krumscheid, Mario and Blömer, Johannes and Becker, Matthias},
  title = {Retrieval-Augmented Generation for ICD-10 Coding in German Clinical Texts – A Technical Case Report},
  journal = {Studies in Health Technology and Informatics},
  year = {2025},
  doi = {10.3233/shti251397}
}
Kostenlos starten →