Diffusing science through social networks: The case of breastfeeding communication on Twitter

Abstract

Breastfeeding is one of many health practices known to support the survival and health of mother and infant, yet low breastfeeding rates persist globally. These rates may be influenced by limited diffusion of evidence-based research and guidelines from the scientific community (SC). As recently highlighted by the National Academy of Sciences, there is a need for the SC to diffuse its findings to the public more effectively online, as means to counteract the spread of misinformation. In response to this call, we gathered data from Twitter for one month from major breastfeeding hashtags resulting in an interconnected social network (n = 3,798 users). We then identified 59 influencers who disproportionately influenced information flow using social network analysis. These influencers were from the SC (e.g. academics, researchers, health care practitioners), as well as interested citizens (IC) and companies. We then conducted an ego-network analysis of influencer networks, developed ego maps, and compared diffusion metrics across the SC, IC and company influencers. We also qualitatively analyzed their tweets (n = 711) to understand the type of information being diffused. SC influencers were the least efficient communicators. Although having the highest tweeting activity (80% of tweets), they did not reach more individuals compared to IC and companies (two-step ego size: 220± 99, 188 ± 124, 169 ± 97 respectively, P = 0.28). Content analysis of tweets suggest IC are more active than the SC in diffusing evidence-based breastfeeding knowledge, with 35% of their tweets around recent research findings compared to only 12% by the SC. Nonetheless, in terms of outreach to the general public, the two-step networks of SC influences were more heterogenous than ICs (55.7 ± 5.07, 50.9 ± 12.0, respectively, P<0.001). Collectively, these findings suggest SC influencers may possess latent potential to diffuse research and evidence-based practices. However, the research suggests specific ways to enhance diffusion.

Real-time Twitter interactions during World Breastfeeding Week: A case study and social network analysis

“…The proposition that the virtual world can be socially mapped and measured to provide insights and action is a quickly evolving space [39,40]. Precedents of effective SNT-based applications for online data have been set, predominately in the field of education and policy and, more recently, in public health [19][20][21]39]. An ample amount of data from these studies suggests social media is the "backbone" of the Internet, where all types of media, links, ideas, opinions, and tools are shared among social networks.…”

“…We used the same data collection approach during the six months prior to the campaign to capture all tweets and associated user data that included breastfeeding-related hashtags (Table 1). These specific hashtags were chosen based on our previous research that identified commonly-used hashtags in this space [19][20][21]. With no previous data available from last year's campaign, we chose a sixmonth period as a conservative period to prevent covering tweets related to the previous year's campaign.…”

“…To understand the background of the identified influencers, we used inductive qualitative coding [34] to categorize profiles into two categories: Scientific Community (SC) (academics, researchers, health care practitioners, and non-governmental agencies), or members of the general public (GP). Coding was completed by two independent researchers after careful examination of influencers' profiles and tweeting history, as well as verifying claimed credentials using websites such as those of academic institutions, governmental and non-governmental agencies, hospitals, and clinics as we have done in other published studies [19][20][21].…”

“…These interactions can be studied using SNT-based methods, which can be used for face-to-face relationships and for those happening online. We have recently applied these methods to understand exchanges in public health areas on Twitter using communications about breastfeeding as an exemplar case [ 19 , 20 ]. We were able to identify unique communities (clusters of users) who discussed breastfeeding around the world and the specific influencers who dominated the conversations.…”

Building Strong Futures: The Feasibility of Using a Targeted Digital Media Campaign to Improve Knowledge About Pregnancy and Low Birthweight Among Black Women

“…Social media influencers have recently been highlighted for their potential to reach audiences in a credible, relatable way that is outside the scope of traditional on-the-ground or digital campaigns (Djafarova et al 2017 ; Freberg et al 2011 ). Research has shown their potential utility in promoting reproductive health-related behaviors such as breastfeeding or HPV vaccination (Moukarzel et al 2020 ; Ortiz et al 2019 ). Influencers were identified manually, by searching local hashtags and business for accounts that appeared influential or attracted engagement from others.…”