Twitter Conversations and English News Media Reports on Poliomyelitis in Five Different Countries, January 2014 to April 2015

"Social media data analysis, as demonstrated by this study and others, can assist public health professionals to better monitor the communication environment and thus better communicate health information to the public."

Prior research has identified peaks in polio-related Twitter data corresponding to dates of news articles such as the Taliban banning the polio vaccine, suggesting Twitter could amplify polio-related news dissemination. This study was an effort to obtain a better understanding of Twitter conversations about polio and what inspires them, in the hope it could enable public health professionals to better communicate about polio vaccination efforts and to galvanise greater public support for polio eradication efforts. The hypothesis was that themes of polio-related tweets and media articles would differ by location of interest (hashtag of country name mentioned in the tweet; country name mentioned in media articles) but would be similar to each other (tweets and media articles) for each location of interest.

The researchers retrospectively examined a 40% random sample of Twitter data containing the hashtag #polio from January 1 2014 to April 30 2015 (N = 79,333), from which they extracted 5 subcorpora each with a co-occurring hashtag #India (n = 5,027), #Iraq (n = 1,238), #Nigeria (n = 1,364), #Pakistan (n = 11,427), and #Syria (n = 2,952). (This study did not compare polio-related Twitter content tweeted by users in these 5 countries; instead, it compared Twitter content tweeted by users globally about these 5 countries.) Various summary statistics are provided, including the top 10 hashtags, keywords, and mentions, along with the distribution of languages of the tweets and the number of tweets over time (for the entire sample, as well as stratified by country co-occurring hashtag). The researchers conducted descriptive analyses and unsupervised machine learning (latent Dirichlet allocation (LDA) modeling) on the 5 Twitter subcorpora.

As hypothesised, polio-specific Twitter conversation themes differed by the location of interest. Examples:

• In the #India subcorpus, the majority of tweets centred around India achieving a polio-free status, represented with a theme of celebrations/polio-free (n = 4,115, 83.54%).
• Most of the tweets from the #Iraq subcorpus were in regard to current eradication/vaccination efforts (n = 566, 46.36%) and cases/outbreaks/incidence (n = 469, 38.41%).
• Most #Nigeria subcorpus tweets were categorised as celebrations/polio-free (n = 691, 52.03%) - perhaps in expectation that the country would soon be declared polio free - followed by current eradication/vaccination efforts (n = 476, 35.84%).
• 44.21% (n = 4890) of the #Pakistan subcorpus tweets were categorised as current eradication/vaccination efforts, whereas 36.67% (n = 4,056) were categorised as violence/conflicts with polio efforts.
• In the #Syria subcorpus, 38.07% (n = 1,096) of the tweets were categorised as current eradication/vaccination efforts, and 32.16% (n = 926) as cases/outbreaks/incidence.

Reviewing the topics for each of the 5 countries, one can see some similarities and differences in what is important in creating or maintaining progress toward the eradication of polio. For instance, since the death of Osama Bin Laden, radical Islamist militants had increased their hostility and violence against community health workers who provided polio vaccinations in Pakistan. Concerns about this political instability are reflected in the fact that 7 in 10 tweets with co-occurring hashtag #Pakistan were pertinent to violence/conflicts or current eradication/vaccination efforts.

Notably, there were a large number of mentions for either @UNICEF or @UNICEFpolio, and a large portion of tweets in the subcorpora of the co-occurring hashtags #Syria and #Iraq contained such mentions. Because polio eradication efforts rely on international donors who, in turn, rely on global public support, the researchers speculate that communicating information regarding the plight of children in conflict-ridden countries could help maintain the attention of global Twitter users and solicit their support for the polio eradication efforts in those countries.

The researchers also retrieved and categorised 73 polio-related English-language news stories from within the same time frame, using: online databases from 2 newspapers, The New York Times and The Independent; 2 television news stations, Cable News Network (CNN) and Fox News; and 2 radio news stations, National Public Radio and Public Radio International. Each article pertaining to polio was then manually categorised into 1 of 5 general themes (celebrations/polio-free, current eradication efforts, cases/outbreaks/incidence, violence/conflicts, and other) and tagged with the specific country (or countries) of the article's concern.

The results show that 6 of the 8 articles about India were celebrating India's achievement of being polio free. In contrast, most traditional news articles on polio in Iraq, Pakistan, and Syria were focused on violence/conflicts with polio efforts (63.64%, 59.46%, and 58.33%, respectively). In the case of Nigeria, one-third of the traditional news articles were categorised as celebrations/polio-free, and another one-third were in regard to new cases, outbreaks, or the incidence of polio.

Contrary to the researchers' hypothesis, this evidence suggests Twitter content differs significantly from English mass media content. Themes with positive undertones such as celebrations/polio-free and current eradication/vaccination efforts were more common in the Twitter sample than in the English-language media article sample for each country of interest. "Negative topics revolving around polio eradication efforts reported by English-language media may hinder public support of these efforts. Therefore, social media may be a better source for soliciting global health support compared with English-language media."

In conclusion: "Because differences in polio-specific themes between social and mass media were identified, further research may elucidate the relationships between the 2 forms of communication. Social media data can assist public health professionals in pinpointing the general concerns or needs of the public during infectious disease events to create location-specific campaigns....Similarly, social media could serve as a positive environment where support can be harnessed for these campaigns."