Change in Threads on Twitter Regarding Influenza, Vaccines, and Vaccination During the COVID-19 Pandemic: Artificial Intelligence–Based Infodemiology Study
Bibliographic record
Abstract
BACKGROUND: Discussions of health issues on social media are a crucial information source reflecting real-world responses regarding events and opinions. They are often important in public health care, since these are influencing pathways that affect vaccination decision-making by hesitant individuals. Artificial intelligence methodologies based on internet search engine queries have been suggested to detect disease outbreaks and population behavior. Among social media, Twitter is a common platform of choice to search and share opinions and (mis)information about health care issues, including vaccination and vaccines. OBJECTIVE: Our primary objective was to support the design and implementation of future eHealth strategies and interventions on social media to increase the quality of targeted communication campaigns and therefore increase influenza vaccination rates. Our goal was to define an artificial intelligence-based approach to elucidate how threads in Twitter on influenza vaccination changed during the COVID-19 pandemic. Such findings may support adapted vaccination campaigns and could be generalized to other health-related mass communications. METHODS: The study comprised the following 5 stages: (1) collecting tweets from Twitter related to influenza, vaccines, and vaccination in the United States; (2) data cleansing and storage using machine learning techniques; (3) identifying terms, hashtags, and topics related to influenza, vaccines, and vaccination; (4) building a dynamic folksonomy of the previously defined vocabulary (terms and topics) to support the understanding of its trends; and (5) labeling and evaluating the folksonomy. RESULTS: We collected and analyzed 2,782,720 tweets of 420,617 unique users between December 30, 2019, and April 30, 2021. These tweets were in English, were from the United States, and included at least one of the following terms: "flu," "influenza," "vaccination," "vaccine," and "vaxx." We noticed that the prevalence of the terms vaccine and vaccination increased over 2020, and that "flu" and "covid" occurrences were inversely correlated as "flu" disappeared over time from the tweets. By combining word embedding and clustering, we then identified a folksonomy built around the following 3 topics dominating the content of the collected tweets: "health and medicine (biological and clinical aspects)," "protection and responsibility," and "politics." By analyzing terms frequently appearing together, we noticed that the tweets were related mainly to COVID-19 pandemic events. CONCLUSIONS: This study focused initially on vaccination against influenza and moved to vaccination against COVID-19. Infoveillance supported by machine learning on Twitter and other social media about topics related to vaccines and vaccination against communicable diseases and their trends can lead to the design of personalized messages encouraging targeted subpopulations' engagement in vaccination. A greater likelihood that a targeted population receives a personalized message is associated with higher response, engagement, and proactiveness of the target population for the vaccination process.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.002 | 0.011 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.000 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.002 | 0.001 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".