Leveraging Deep Learning to Understand Health Beliefs about the Human Papillomavirus Vaccine from Social Media

"This study demonstrates the potential for utilizing social media to better understand HPV vaccine health beliefs."

The Health Belief Model (HBM) is a conceptual framework used in health behaviour research to explain why people adopt behaviours that lead to healthy lives. Studies have found that HBM constructs are associated with human papillomavirus (HPV) vaccine intention and uptake. With the conviction that understanding parental beliefs about the HPV vaccine could guide the development of effective and targeted vaccine promotion strategies, these researchers proposed a deep-learning-based framework to mine health beliefs on the HPV vaccine from Twitter. (Deep learning is a set of advanced computational models used for various tasks in natural language understanding.)

The researchers collected a Twitter data set related to the HPV vaccine from January 1 2014 to December 31 2017. They focus on 4 primary HBM constructs: perceived susceptibility, perceived severity, perceived benefits, and perceived barriers. Table 1 in the paper shows the constructs, definition, sample tweets, and performance (estimated sensitivity, specificity, and accuracy, with their 95% confidence intervals) of the proposed deep-learning model. In short, the model achieved satisfactory results in terms of sensitivity, specificity, and accuracy on testing sets. The model achieved a mean accuracy of 80.50% for identifying HBM-related tweets and between 80.33% and 89.82% for the 4 HBM constructs.

After applying the model to classify the 956,262 un-labeled tweets, the researchers classified 652,252 tweets, obtained from 216,864 unique Twitter user IDs, as HBM related. Temporal analysis of the overall data showed that the prevalence of tweets in the perceived susceptibility/severity constructs increased every year, which may reflect an improved understanding of the prevalence of HPV and HPV-related cancers as well as an increased awareness of the severity of these cancers. Meanwhile, tweets categorised into perceived benefits/barriers decreased, which may reflect a shift in parental assessment of the risk/benefit ratio in accepting the HPV vaccine for their teen.

A significant shift in health beliefs was seen in 2016. The researchers attribute the shift to promotional articles on the HPV vaccine from several influential media sources, such as the New York Times ("HPV Sharply Reduced in Teenage Girls Following Vaccine, Study Says", February 23 2016) and Time ("The HPV Vaccine Is Lowering Infection Rates", February 22 2016).

The researchers saw 2 spikes in barriers in February and July in 2015, identifying corresponding events: The spike in February was due mainly to the Toronto Star's story on the HPV vaccine Gardasil, titled, "A Wonder Drug's Dark Side" (February 5 2015), and the July spike was due mainly to the news that the European Medicines Agency was conducting a review of the HPV vaccine's side effects.

Further analysis of the impact of these types of events could help inform efforts to promote HPV vaccination.

In the future, the researchers plan to develop novel computational algorithms to understand health beliefs on the user level by analysing the historical tweets for each user.