Improving Adverse Drug Event Extraction with SpanBERT on Different Text Typologies

TL;DR

This paper demonstrates that SpanBERT combined with a CRF significantly improves adverse drug event extraction from social media and blogs, outperforming previous models across different text types.

Contribution

It introduces the novel application of SpanBERT for ADE extraction and validates its effectiveness across diverse text typologies.

Findings

SpanBERT with CRF outperforms all competitors on SMM4H and CADEC datasets.

SpanBERT's multi-token span capabilities enhance ADE extraction accuracy.

The approach is effective across social media and blog text formats.

Abstract

In recent years, Internet users are reporting Adverse Drug Events (ADE) on social media, blogs and health forums. Because of the large volume of reports, pharmacovigilance is seeking to resort to NLP to monitor these outlets. We propose for the first time the use of the SpanBERT architecture for the task of ADE extraction: this new version of the popular BERT transformer showed improved capabilities with multi-token text spans. We validate our hypothesis with experiments on two datasets (SMM4H and CADEC) with different text typologies (tweets and blog posts), finding that SpanBERT combined with a CRF outperforms all the competitors on both of them.