Comparing Human Papillomavirus Vaccine Concerns on Twitter: A Cross-Sectional Study of Users in Australia, Canada and the UK

"International networking on Twitter suggests that vaccine related controversies in one country could reverberate around the world and impact vaccine coverage."

This study conducted an international comparison investigating how human papillomavirus (HPV) vaccination concerns are expressed on Twitter. Many individuals use the internet and social media to access health information; however, these sources have been described as having the potential to rapidly amplify unbalanced, distorted, or inaccurate information about vaccines. The goal was to examine the domestic and international network connectedness of HPV vaccine concern expression.

Tweets related to HPV vaccines were collected from January 2014 to April 2016 in Australia, Canada, and the United Kingdom (UK). These countries were selected because they are English-speaking countries, share a similar history and commonwealth membership, and are similar in their administration of the HPV vaccination in schools. Data captured included information about users' locations, the text of the tweets, and information about social connections. To enable the classification of a large number of tweets, 2 stages of machine learning classifiers were constructed from a sample of tweets that were manually coded by 2 investigators.

The Health Belief Model (HBM) was used as the basis for coding the types of HPV vaccine concerns expressed on Twitter. The HBM has been used previously to evaluate the determinants of HPV vaccination and non-compliance by identifying perceived susceptibility to HPV, perceived severity of HPV, perceived benefits of HPV vaccination, perceived barriers of HPV vaccination (including tangible barriers such as logistical challenges and psychological barriers such as perceived harms of receiving the HPV vaccine), and cues to action (e.g., influences prompting HPV vaccine uptake such as information from healthcare providers, family, or friends). To account for additional prominent concerns that were not captured by the model, the constructs were also informed by previous content analyses of media and social media related to the HPV vaccine, as well as literature on vaccine hesitancy. The coding scheme was therefore extended to include mistrust, undermining of religious principles, undermining of civil liberties, additional concerns (not otherwise specified), and ambiguous tweets. The coding scheme was used to code 12 types of concerns expressed in a second sample of 1,000 tweets. Combining categories was done based on conceptual similarity and trying to remain as true to the HBM as possible while attaining accuracy of the classifier.

The network of social connections formed by users who posted tweets about HPV vaccines was used to compare the proportions of local (within a country) and international (across countries) followers. The group of users for whom at least half of their relevant tweets were expressing concerns was assigned to one group (concern), and all other users were assigned to another (non-concern). The users were then also split by country, and the proportions of local and international followers were compared across groups.

The study population was 16,789 Twitter users who posted 43,852 tweets about HPV vaccines. Tweets expressing concerns about HPV vaccines made up 14.9% of tweets in Canada, 19.4% in Australia, and 22.6% in the UK. The types of concerns expressed were similar across the 3 countries, with concerns related to "perceived barriers" (e.g., logistical challenges and psychological barriers such as perceived harms of receiving the HPV vaccine) being the most common. Users expressing concerns about HPV vaccines in each of the 3 countries had a relatively high proportion of international followers also expressing concerns, creating clusters of users by country. That is to say, the results indicate that users who mostly expressed concerns were disproportionately well-connected to international users discussing HPV vaccines.

"This research provides public health practitioners and policy-makers with evidence that concerns about 'perceived barriers' on Twitter are widespread; effective communication campaigns could be designed and implemented to target this concern in locations where it is likely to have the greatest impact. However, it is important for further research to analyse results by type of sender. It would also be critical for future research to design and evaluate appropriate messaging of such a campaign so that this intervention does not 'backfire' and increase hesitancy."

The researchers suggest that, "[g]iven the international connection between Twitter users who express concerns, public health organisations seeking to improve the uptake of HPV vaccines may benefit from tools that help them monitor the impact of vaccine scares on social media locally as well as in other countries in order to pre-empt and respond to misinformation. Greater research would be helpful to further investigate how public health organisations can monitor and intervene [perhaps by working together in a coordinated fashion] to address vaccine concerns. Such support could have been beneficial for Japan and Colombia when the media had a detrimental impact on HPV vaccine coverage."