Disorientation towards routine immunization, COVID-19 anxiety, and propensity to vaccinate. An analysis based on Twitter data in Italy.

Background Online social media (OSM) play an importart role in our life, for many even indispensable. To speak with and to get in touch with other, people post huge amounts of contents, often personal, and often dealing with sensible topics, as health. OSM and the web are used to share and to acquire information based upon which they often take decision potentially impacting on health and quality of life. There are many positive aspects in the use of social media: reduce physical distance, share information, and retrieve knowledge from new kind of data. However, social media might become harmful e.g., when they become a vehicle for the spread misinformation capable to destabilize the public opinon - especially for controversial topics like vaccination - and generate disorientation amongst people, who can loose awareness of what they read or watch, becoming misinformed and spreaders of misinformation in their turn. Objectives and Methods Using Twitter and Sentiment analysis techniques on tweets written in Italian we attempted at i) characterizing the temporal flow of communication on Twitter about vaccines related topic in the year 2018 and during the COVID-19 Pandemic, ii) identifying the main triggering events, iii) evaluate the opinion and polarity in the case of immunization policy in 2018 and study the effect of COVID-19 on Twitter, iv) investigate if contrasting announcement and decision on immunization policy generated disorientation in public opinion in Italy on 2018 and see if Twitter flow in Italy respond to the pandemic spread at national, regional, and province level. Result Political events originated, in both analyses, major reactions. On sentiment about vaccine and immunization policy we found that 75% are favorable, 11% unfavorable, and 14% undecided, with the first and latter proportions that changed in trend, synchronized with the change of government in Italy, suggesting evidence of long term disorientation in public opinion. We tested for presence of disorientation, in form of instability in polarity proportion, also for short term. Proportion of people involved in vaccination was negligible. For COVID-19 we found a clear a positive correlation between Twitter flow and Covid cases reported, especially in the most hit regions by pandemic. Conclusion Use of social media analysis is useful to estimate and have a better overview of the public opinion for critical health-related topics. Disorientation appears on social media in controversial topic, such as vaccination decision, showing that health topic and healthcare should never be used to raise political consensus. Disorientation may raise also due to lack of presence of public health institutions on social media calling for efforts to contrast misinformation, which needs further analyzed to understand how this will translate in disorientation and future vaccination decision. During the most important health threath of last years, people use social media to espress concern, anxiety, and presence of denialism, it remains to be seen how these sentiments arise and spread during a pandemic and what is the role of social media exposure and misinformation.