On embedding-based automatic mapping of clinical classification system: handling linguistic variations and granular inconsistencies
Journal of the American Medical Informatics Association · 2026
DOI: 10.1093/jamia/ocag004 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of the American Medical Informatics Association, 2026. Verfasst von Santosh Purja Pun und 3 weiteren Autor:innen. Beteiligte Einrichtung: School of Computer, Data and Mathematical Sciences, Western Sydney University, Sydney, NSW 2196, Australia. Hauptthemen laut MEDLINE: International Classification of Diseases, Natural Language Processing. Die Arbeit nennt 37 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: Western Sydney University and Australian Consortium for Classification Development Postgraduate Research Scholarship, Western Sydney University, Australian Consortium for Classification Development Postgraduate Research Scholarship.
Abstract (Originalsprache)
Objectives
Mapping clinical classification systems, such as the International Classification of Diseases (ICD), is essential yet challenging. While the manual mapping method remains labor-intensive and lacks scalability, existing embedding-based automatic mapping methods, particularly those leveraging transformer-based pretrained encoders, encounter 2 persistent challenges: (1) linguistic variation and (2) varying granular details in clinical conditions. Materials and Methods We introduce an automatic mapping method that combines the representational power of pretrained encoders with the reasoning capability of large language models (LLMs). For each ICD code, we generate: (1) hierarchy-augmented (HA) and (2) LLM-generated (LG) descriptions to capture rich semantic nuances, addressing linguistic variation. Furthermore, we introduced a prompting framework (PR) that leverages LLM reasoning to handle granularity mismatches, including source-to-parent mappings.
Results
Chapterwise mappings were performed between ICD versions (ICD-9-CM↔ICD-10-CM and ICD-10-AM↔ICD-11) using multiple LLMs. The proposed approach consistently outperformed the baseline across all ICD pairs and chapters. For example, combining HA descriptions with Qwen3-8B-generated descriptions yielded an average top-1 accuracy improvement of 6.5% (0.065) across the mapping cases. A small-scale pilot study further indicated that HA+LG remains effective in more challenging one-to-many mappings.
Conclusions
Our findings demonstrate that integrating the representational power of pretrained encoders with LLM reasoning offers a robust, scalable strategy for automatic ICD mapping.
Schlagwörter
MeSH-Schlagwörter (MEDLINE)
Stichwörter der Autor:innen
Autor:innen und Institutionen
- School of Computer, Data and Mathematical Sciences, Western Sydney University, Sydney, NSW 2196, Australia.
ORCID-Profile: Santosh Purja Pun
Förderung
- Western Sydney University and Australian Consortium for Classification Development Postgraduate Research Scholarship
- Western Sydney University
- Australian Consortium for Classification Development Postgraduate Research Scholarship
Zitieren
APA
Purja Pun, S., Obst, O., Basilakis, J., & Ginige, J. A. (2026). On embedding-based automatic mapping of clinical classification system: handling linguistic variations and granular inconsistencies. Journal of the American Medical Informatics Association. https://doi.org/10.1093/jamia/ocag004
Vancouver
Purja Pun S, Obst O, Basilakis J, Ginige JA. On embedding-based automatic mapping of clinical classification system: handling linguistic variations and granular inconsistencies. Journal of the American Medical Informatics Association. 2026. doi:10.1093/jamia/ocag004
BibTeX
@article{PurjaPun2026,
author = {Purja Pun, Santosh and Obst, Oliver and Basilakis, Jim and Ginige, Jeewani Anupama},
title = {On embedding-based automatic mapping of clinical classification system: handling linguistic variations and granular inconsistencies},
journal = {Journal of the American Medical Informatics Association},
year = {2026},
doi = {10.1093/jamia/ocag004}
}
EUVIMED