Examining Public Messaging on Influenza Vaccine over Social Media: Unsupervised Deep Learning of 235,261 Twitter Posts from 2017 to 2023

Abstract

Although influenza vaccines are safe and efficacious, vaccination rates have remained low globally. Today, with the advent of new media, many individuals turn to social media for personal health questions and information. However, misinformation may be rife, and health communications may be suboptimal. This study, therefore, aimed to investigate the public messaging related to influenza vaccines by organizations over Twitter, which may have a far-reaching influence. The theoretical framework of the COM-B (capacity, opportunity, and motivation component of behavior) model was used to interpret the findings to aid the design of messaging strategies. Employing search terms such as “flu jab”, “flu vaccine”, “influenza vaccine”, and ‘“ influenza jab”, tweets posted in English and by organizations from 1 January 2017 to 1 March 2023 were extracted and analyzed. Using topic modeling, a total of 235,261 tweets by organizations over Twitter were grouped into four main topics: publicizing campaigns to encourage influenza vaccination, public education on the safety of influenza vaccine during pregnancy, public education on the appropriate age to receive influenza vaccine, and public education on the importance of influenza vaccine during pregnancy. Although there were no glaring pieces of misinformation or misconceptions, the current public messaging covered a rather limited scope. Further information could be provided about influenza and the benefits of vaccination (capability), promoting community, pharmacist-led influenza vaccination, and other avenues (opportunity), and providing greater incentivization and support for vaccination (motivation).