Sentiments towards COVID-19 vaccinations, whether positive or negative, are a sign of future vaccination rates, according to a study of related Twitter posts. These findings provide new insights into the impact of social media on public health initiatives.
Researchers at the Courant Institute of Mathematical Sciences and the NYU Grossman School of Medicine conducted a survey that showed positive sentiment on Twitter about vaccinations was followed a week later by increase in vaccination rates in the same geographical area, while negative sentiment was followed a week later.
The study employed a real-time big data analytics framework based on sentiment analysis and natural language processing (NLP) algorithms. The system takes real-time tweets and identifies tweets related to vaccinations, identifies them by certain themes and provides sentiment analysis, cataloging tweets as positive, negative, or neutral.
We need to understand how vaccine hesitancy and social media impacts on developing and spreading it, according to Mmegan Coffee, MD, PhD, and a clinical assistant professor in the Department of Medicine at the University of New York, the author of the paper. This is the first step towards achieving a barometer to track sentiment and themes about vaccine hesitancy.
As the COVID epidemic puts more people in front of computers and vaccination hesitancy has shaped the epidemic, we need tools like this one to monitor and understand social media''s impact on vaccine hesitancy for this epidemic and for future epidemics, according to an author.
The authors claim that vaccination could help prevent continued outbreaks and new forms of the COVID epidemic. Despite vaccination hesitancy, vaccinations affect individual and collectively. These are also the role of social media, which helps both information and misinformation about vaccination, putting concern about the impact of these technologies on vaccination rates.
The authors designed a large data analytics platform based on Natural Language Processing (NLP), Sentiment Analysis (SA) and Amazon Web Services (AWS).
These topics and related phrases allowed the researchers to keep track of several vaccine-related topics. Topics included: conspiracy, fear, heath freedom, natural alternatives, side effects, safety, trust, drug companies, and hesitancy. These topics and sentiment scores allowed them to identify instinct, negative, or neutral.
The researchers studied daily COVID vaccination data from the Institute of Electrical and Electronic Engineers (IEEE) Dataport dataset.
According to analysis, once vaccinations were available for all adults around mid-April 2021, a spike in positive sentiment in certain regions of the United States was followed by a decrease in vaccination rates a week later. By contrast, in regions where sentiment was falling, a decrease in vaccination rates was observed a week later.
Notably, the big data analytics framework revealed that in the first several months of the epidemic and before the vaccination rollout began at the end of 2020, positive and negative sentiment toward vaccinations was similar, with a slight increase of positive sentiment. By contrast, negative sentiment tweets exceeded positive ones.
Because vaccination rates were found to track regionally with Twitter vaccination sentiment, a more advanced analytics tool might potentially influence vaccine uptake or guide the development of targeted social media campaigns and vaccination strategies, according to Bari, who is the CEO of the Courant Institutes Predictive Analytics and AI Research Lab.
Coffee points out that this method allows us to begin to investigate patterns in vaccine hesitancy over time and place. But it can only monitor, and not influence, vaccination hesitancy, which is constantly changing. More work is required to build trust in life-saving vaccines and rectify vaccine negative effects.
The Courant Institutes Predictive Analytics and AI Research Lab featured Madeline DiLorenzo from the NYU Grossman School of Medicine, as well as Matthias Heymann, Ryan Cohen, Robin Zhao, Levente Szabo, Shailesh Apas Vasandani, Aashish Khubchandani, and Alankrith Krishnanresearchers.
A grant from Amazon AI secured the funds for the project in part.