Token Classification for Disambiguating Medical Abbreviations

TL;DR

This study investigates token classification methods, especially transformer models like SciBERT, for disambiguating medical abbreviations in clinical texts, showing improved performance over traditional text classification approaches.

Contribution

It demonstrates the effectiveness of token classification models, particularly SciBERT, for medical abbreviation disambiguation, outperforming text classification methods.

Findings

Token classification models outperform text classification in abbreviation disambiguation.

SciBERT shows strong performance on both datasets.

Postprocessing improves text classification results to match token classification.

Abstract

Abbreviations are unavoidable yet critical parts of the medical text. Using abbreviations, especially in clinical patient notes, can save time and space, protect sensitive information, and help avoid repetitions. However, most abbreviations might have multiple senses, and the lack of a standardized mapping system makes disambiguating abbreviations a difficult and time-consuming task. The main objective of this study is to examine the feasibility of token classification methods for medical abbreviation disambiguation. Specifically, we explore the capability of token classification methods to deal with multiple unique abbreviations in a single text. We use two public datasets to compare and contrast the performance of several transformer models pre-trained on different scientific and medical corpora. Our proposed token classification approach outperforms the more commonly used text classification models for the abbreviation disambiguation task. In particular, the SciBERT model shows a strong performance for both token and text classification tasks over the two considered datasets. Furthermore, we find that abbreviation disambiguation performance for the text classification models becomes comparable to that of token classification only when postprocessing is applied to their predictions, which involves filtering possible labels for an abbreviation based on the training data.