Mining vaccine adverse events mentions from social media using Twitter as a source (Preprint)

Abstract

Traditional monitoring for Adverse Events Following Immunisation (AEFI) relies on various established reporting systems, where there is inevitably a lag between an AEFI occurring and its potential reporting, and subsequent processing of reports. AEFI safety signal detection strives to detect AEFI as early as possible, ideally close to real-time. Monitoring social media data holds promise as a resource for this.

1) To investigate the utility of monitoring social media for gaining early insights into vaccine safety issues, by extracting vaccine adverse event mentions (VAEM) from Twitter using natural language processing (NLP) techniques. 2) To document the NLP processes used and identify the most effective of them for successively identifying tweets that contain VAEM, with a view to defining an approach that might be applicable to other similar social media surveillance tasks.

A VAEM-Mine method was developed that combines topic modelling with classification techniques to extract maximal VAEM posts from a vaccine-related Twitter stream, with a high degree of confidence. The approach does not require a targeted search for specific vaccine reactions, but instead identifies any VAEM post within many unrelated posts.

The VAEM-Mine method successively isolates vaccine adverse event mentions from the massive amount of other vaccine-related Twitter posts, achieving an F1-Score of 0.91 in the classification phase.

Social media can assist with detection of vaccine safety signals as a valuable complementary source for monitoring mentions of vaccine adverse events. A social media based VAEM data stream can be assessed for changes to detect possible emerging vaccine safety signals, helping to address the well-recognised limitations of passive reporting systems, including timeliness and under-reporting.