Embedding Transfer for Low-Resource Medical Named Entity Recognition: A Case Study on Patient Mobility

TL;DR

This paper explores transfer learning of word embeddings for low-resource medical NER, specifically extracting patient mobility information from electronic health records, demonstrating that domain-adapted embeddings improve extraction performance.

Contribution

It introduces a domain adaptation approach for word embeddings in medical NER, addressing low-resource challenges in extracting patient mobility data from clinical texts.

Findings

Domain-adapted embeddings improve NER precision and recall.

Small in-domain corpora can yield comparable embeddings to large out-of-domain datasets.

Challenges include complex entity structures and linguistic variability in mobility descriptions.

Abstract

Functioning is gaining recognition as an important indicator of global health, but remains under-studied in medical natural language processing research. We present the first analysis of automatically extracting descriptions of patient mobility, using a recently-developed dataset of free text electronic health records. We frame the task as a named entity recognition (NER) problem, and investigate the applicability of NER techniques to mobility extraction. As text corpora focused on patient functioning are scarce, we explore domain adaptation of word embeddings for use in a recurrent neural network NER system. We find that embeddings trained on a small in-domain corpus perform nearly as well as those learned from large out-of-domain corpora, and that domain adaptation techniques yield additional improvements in both precision and recall. Our analysis identifies several significant challenges in extracting descriptions of patient mobility, including the length and complexity of annotated entities and high linguistic variability in mobility descriptions.