Characterizing Public Sentiments and Drug Interactions during COVID-19: A Pretrained Language Model and Network Analysis of Social Media Discourse (Preprint)

Abstract

BACKGROUND While COVID-19 pandemic has induced massive discussion of available medications on social media, traditional studies focused only on limited aspects such as public opinions and suffer from reporting biases, inefficiency and long collection times. OBJECTIVE Harnessing drug-related data posted on social media in real time can offer insights into how the pandemic impacts drug use and monitor misinformation. This study developed a natural language processing (NLP) pipeline tailored for the analysis of social media discourse on COVID-19 related drugs. METHODS This study constructed a full pipeline for COVID-19 related drug tweet analysis, utilizing pre-trained language model-based NLP techniques as the backbone. This pipeline is architecturally composed of four core modules: named entity recognition (NER) and normalization to identify medical entities from relevant tweets and standardize them to uniform medication names, target sentiment analysis (TSA) to reveal sentiment polarities associated with the entities, topic modeling to understand underlying themes discussed by the population, and drug network analysis to potential adverse drug reactions (ADR) and drug-drug interactions (DDI). The pipeline was deployed to analyze tweets related to COVID-19 and drug therapies between February 1, 2020, and April 30, 2022. RESULTS From a dataset comprising 2,124,757 relevant tweets sourced from 1,800,372 unique users, our NER model identified the top five most-discussed drugs: Ivermectin, Hydroxychloroquine, Remdesivir, Zinc, and Vitamin D. Sentiment and topic analysis revealed that public perception was predominantly shaped by celebrity endorsements, media hotspots, and governmental directives rather than empirical evidence of drug efficacy. Co-occurrence matrices and complex network analysis further identified emerging patterns of DDI and ADR that could be critical for public health surveillance like better safeguarding public safety in medicines use. CONCLUSIONS This study evidences that an NLP-based pipeline can be a robust tool for large-scale public health monitoring and can offer valuable supplementary data for traditional epidemiological studies concerning DDI and ADR. The framework presented here aspires to serve as a cornerstone for future social media-based public health analytics.