Study Objectives
During a public health emergency such as the COVID-19 pandemic, emergency physicians may use Twitter to learn about innovations in care. However, knowledge is limited about how information on innovations spreads on Twitter and what the implications are for quality. Studying the spread of information about ivermectin, a medication not recommended to prevent or treat COVID-19, may yield insights into how communication can be improved in the future. Our objective was to describe the dissemination of information about ivermectin during the pandemic through analysis of mentions on Twitter.
Methods
We collected Twitter posts and related metadata using Brandwatch, a leading platform for aggregating and analyzing social media data. Posts were included in our sample if they: 1) occurred between March and December 2020; 2) included the word “ivermectin”; and 3) were written in English. We collected text from posts; metadata describing post date, author location, and profession; and social media network data on user interactions. To analyze the data, we used embedded capabilities in Brandwatch (eg, graphing of mentions of ivermectin over time). We paired data on ivermectin mentions with COVID-19 case counts from the New York Times to improve contextual understanding. Finally, we used RAND-Lex, a suite of text analytics and machine learning tools at RAND, to perform social media network and corpus analyses for identifying: communities of users discussing ivermectin; and non-ivermectin topics and hashtags trending within communities.
Results
We identified 165,159 posts meeting eligibility criteria from 68,497 unique users. Authors of tweets were most commonly located in the United States, India, the United Kingdom, and Canada. Professions of tweet authors were most commonly categorized as health practitioner, scientist/researcher, and executive. The number of ivermectin mentions spiked in March (week 6; max 11,097 posts/day), August (week 27; max 6,486 posts/day), and December of 2020 (week 43; max 21,024 posts/day). [See Figure for a graph of ivermectin mentions per week, which is paired with US national COVID-19 case counts for contextual understanding.] Common trending topics among users in the sample included “Pierre Kory”, “early treatment”, and “brave and keep fighting against COVID19”. We identified 3 large-scale communication communities: A) users interacting with the “COVID-19 Critical Care Alliance” (a non-profit led by Dr. Pierre Kory), who commonly tagged posts with #followthescience, #vitamind, and #realnews; B) users opposed to police enforcement of lockdowns for COVID-19 in India, who commonly tagged #justiceforjayarajandfenix, #idtwitter, and #beatntds (ntds = neglected tropical diseases); and C) users proponing hydroxychloroquine and opposing Dr. Anthony Fauci, who commonly tagged #fraudulentfauci, #newsmax, and #hcqworks.
Conclusion
During 2020, ivermectin was mentioned as many as 11,097 times per day on Twitter. We identified 3 large- scale communication communities who also discussed other hot topics related to health care and popular news, such as emerging treatments, fairness of lockdowns, and veracity of the news. Our findings underscore the importance of tailoring dissemination strategies during a public health emergency so that important consumers of information, such as frontline emergency physicians, can learn about key innovations and ascertain the quality of evidence that supports or refutes their use.
No, authors do not have interests to disclose