Extracting factors associated with vaccination from Twitter data and mapping to behavioral models

Area of Science:

• Public Health
• Social Media Analysis
• Behavioral Science

Background:

• Social media platforms like Twitter offer valuable insights into public opinion on vaccines.
• Established behavioral models, including the 5C psychological antecedents model and the Health Belief Model (HBM), provide frameworks for understanding vaccine-related behavior shifts.

Purpose of the Study:

• To investigate if Twitter data can yield insights into public attitudes toward vaccines within the Arab population.
• To determine if these insights can be mapped to the 5C model and HBM.

Main Methods:

• A two-year Twitter data search (Jan 2020-Jan 2022) was conducted using Arabic vaccine-related keywords.
• BERTopic modeling was applied to extract topics, which were then manually mapped to the factors of the 5C model and HBM.
• Analysis included 1,068,466 unique users and 3,368,258 tweets, with demographic data on gender and region (MENA).

Main Results:

• BERTopic modeling identified 25 topics, successfully mapped to the 15 factors of the 5C model and HBM.
• Twitter users showed increased vaccine acceptance when trusting safety and effectiveness.
• Vaccine hesitancy was significantly linked to conspiracy theories and misinformation.

Conclusions:

• Twitter data is a rich, diverse source for understanding and predicting vaccine-related behavioral changes in the Arab population.
• Mapping topics to behavioral models like the 5C and HBM demonstrates the utility of social media data for public health.
• This approach enables more timely and effective interventions for vaccination programs compared to traditional methods.