Detecting Contradictory COVID-19 Drug Efficacy Claims from Biomedical Literature

TL;DR

This paper develops NLP models to automatically identify contradictory claims about COVID-19 drug efficacy in biomedical literature, aiding experts in understanding conflicting scientific evidence during the pandemic.

Contribution

It introduces a new NLI dataset created by domain experts and demonstrates how NLP models can assist in analyzing contradictory COVID-19 drug efficacy claims.

Findings

Models effectively identify contradictions in drug efficacy claims.

The NLI dataset improves model training and accuracy.

Case study shows practical utility for experts analyzing remdesivir and hydroxychloroquine.

Abstract

The COVID-19 pandemic created a deluge of questionable and contradictory scientific claims about drug efficacy -- an "infodemic" with lasting consequences for science and society. In this work, we argue that NLP models can help domain experts distill and understand the literature in this complex, high-stakes area. Our task is to automatically identify contradictory claims about COVID-19 drug efficacy. We frame this as a natural language inference problem and offer a new NLI dataset created by domain experts. The NLI framing allows us to create curricula combining existing datasets and our own. The resulting models are useful investigative tools. We provide a case study of how these models help a domain expert summarize and assess evidence concerning remdisivir and hydroxychloroquine.