Public Health, Twitter Chats, and Measuring Participant Engagement: A Case Study

Theoretical Background and research questions/hypothesis:  The use of social media channels, such as Twitter, is gaining traction in public health practice. However, evidence suggests that Twitter use by public health entities is largely uni-directional and is primarily used to share information. Despite a growing body of literature, there remains a paucity of research on the use of Twitter as a public health communication and dissemination tool. The purpose of this study is to measure participant engagement during the inaugural #LiveFitNOLA Twitter chat using methods that can be applied by other public health professionals.

Methods:  We conducted a case-study of a Twitter chat, #LiveFitNOLA, hosted by the Tulane University Prevention Research Center (@TulanePRC) and the City of New Orleans Health Department (@FitNOLA), as part of Let’s Move! – a nationwide childhood obesity reduction initiative. In addition, Health Fitness Magazine (@HealthFitMag) served as a guest host for the inaugural #LiveFitNOLA Twitter chat. Data were obtained by abstracting tweets from the Symplur Healthcare Hashtag Project website. Tweets that included the hashtag #LiveFitNOLA during a 75-minute period on March 5, 2015, from 1:00 PM to 2:15 PM Eastern Time were included in the analytic database. The study population included all individuals and organizations that participated in the #LiveFitNOLA Twitter chat during this 75-minute period. There were 744 tweets and 66 participants. Using a mixed method approach, we first conducted a social network analysis to measure and visualize the engagement level between Twitter chat participants. Participants were stratified by entity level (individuals and organizations), and network influencers were identified. In addition, we estimated select network-level metrics, such as network centrality, number of communities within the network, and network diameter. Second, deductive, thematic content analysis was conducted. The coding schema consisted of Information-Community-Action framework elements.

Results:  The #LiveFitNOLA network was composed of 143 nodes, 476 edges, and 4 communities. A node represents a Twitter user that participated in the chat or was mentioned by participants during the chat. When one user mentions another in a tweet, an edge is formed – representing a relationship between Twitter users. The #LiveFitNOLA network degree, the number of edges a node has, ranged from 1 to 101. Regardless of direction, the average network degree was 7.18 (S.D = 13.5), which indicated a tight knit network composed of bi-directional conversations. Of note, host Twitter handles sent and received the most mentions. Results from deductive, thematic content analysis using the Information-Community-Action framework will be completed and available prior to the National Conference on Health Communication, Marketing, and Media.

Conclusions:  Our findings suggest that Twitter chats foster bi-directional communication and catalyze audience engagement.

Implications for research and/or practice:  Twitter chats can serve as low-cost tools for public health entities to effectively and efficiently increase bi-directional communication and dissemination. Capacity building efforts to improve Twitter chat moderation and impact measurement may help public health entities reach audiences with real-time public health messages and increase audience participation in focused dialogue. The methodology described in this study can be used by other public health practitioners to measure such engagement.