Home // HEALTHINFO 2025, The Tenth International Conference on Informatics and Assistive Technologies for Health-Care, Medical Support and Wellbeing // View article

From Text to Code: Predicting Abbreviated Injury Scale 2015 from Clinical Narratives

Authors:
Chien-Ming Lee
Pei-Ling Lee
Chia-Yeuan Han
Joffrey Hsu
Chuan-Yu Hu

Keywords: abbreviated injury scale; natural language processing.

Abstract:
Accurate coding of traumatic injuries using the Abbreviated Injury Scale (AIS) is very important for trauma registrants. However, manual AIS coding requires trained personnel and is very time-consuming. This study explores the feasibility of using a pre-trained Natural Language Processing (NLP) model to automatically predict complete AIS codes from unstructured diagnostic text entered by emergency physicians. Without additional training or fine-tuning, a publicly available transformer-based model was applied to emergency department narrative data. This preliminary result shows that such models can find clinically relevant information from these free-typing texts and then map to the correct AIS codes. This work highlights the potential of leveraging existing NLP models to assist in injury classification and AIS coding, especially without labeled datasets for training.

Pages: 19 to 20

Copyright: Copyright (c) IARIA, 2025

Publication date: October 26, 2025

Published in: conference

ISSN: 2519-8491

ISBN: 978-1-68558-312-5

Location: Barcelona, Spain

Dates: from October 26, 2025 to October 30, 2025