Question Answering based Clinical Text Structuring Using Pre-trained Language Model

TL;DR

This paper introduces a question answering framework for clinical text structuring that leverages domain-specific features within pre-trained language models, improving performance on Chinese pathology reports.

Contribution

It proposes a unified QA-based task for clinical text structuring and a novel model incorporating clinical domain features into pre-trained language models.

Findings

QA-CTS improves task performance on Chinese pathology reports

The proposed model outperforms baseline models in specific tasks

Unified dataset sharing enhances clinical research efficiency

Abstract

Clinical text structuring is a critical and fundamental task for clinical research. Traditional methods such as taskspecific end-to-end models and pipeline models usually suffer from the lack of dataset and error propagation. In this paper, we present a question answering based clinical text structuring (QA-CTS) task to unify different specific tasks and make dataset shareable. A novel model that aims to introduce domain-specific features (e.g., clinical named entity information) into pre-trained language model is also proposed for QA-CTS task. Experimental results on Chinese pathology reports collected from Ruijing Hospital demonstrate our presented QA-CTS task is very effective to improve the performance on specific tasks. Our proposed model also competes favorably with strong baseline models in specific tasks.