Informing Social Media Analysis for Public Health: A Cross-sectional Survey of Professionals

During the COVID-19 pandemic, analysis of social media conversations about health topics became more common to inform health communication and digital engagement. Understanding the gist, velocity, volume, and underlying drivers of narratives circulating in online communities can help to inform pre-bunking and debunking initiatives, fill information gaps, and inform responses to the "infodemic" (or information overload) accompanying the pandemic. Social media analysis tools have traditionally been developed for commercial purposes, rather than public health. Thus, this research aims to describe the current use and work practice of social media analysis tools and to identify gaps and needs of those working in public health.

The researchers developed a cross-sectional survey and distributed it through global infodemic management networks between December 2022 and February 2023. Questions were structured over four sections related to work-practice and user needs and did not collect any personal details from participants. Descriptive analysis was conducted on the study results. Qualitative analysis was used to categorise and understand answers to open-text questions.

There were 417 participants, 162/417 who completed all survey questions, and 255/417 who completed some, with all responses included in analysis. Respondents came from all global regions and a variety of workplaces. Participants had an average of 4.4 years' experience in the analysis of social media for public health; they had worked across 55 different health issues (though COVID-19 was the most common) and were conducting social media analysis in 56 different languages.

Results reveal a range of training, technical capacity, and support needs. For instance, respondents asked what specific tasks they currently did, wanted to do, or otherwise. For 15/26 individual tasks, over 50% of respondents reported wanting to do the task, but either didn't know how to, or couldn't do it with the tools they currently used. Less than 40% agreed that that social media analysis platform they were using was fully meeting their needs, and over 50% of respondents indicated the need for additional training and support for half of the options given. Respondents were seeking support with advocating for data and platform availability, for support demonstrating value to managers and other partners, and for global coordination and support. Only 57.5% of respondents felt confident using social media analysis tools to their full capacity, and training on platform use was the top training need identified by 69.9% of respondents.

Reflecting on the findings, the researchers argue that more research is needed on how analysis of narratives on social media platforms and more broadly in information ecosystem(s) can be conducted and scaled up, how locally relevant and offline sources can be integrated, how public health workers can be supported to do this, and how better access to data and better analytical platforms can be facilitated. For example, the need for coordination was exemplified by one comment from a survey participant, who described their attempt to create a crawler for Instagram to better enable monitoring of infodemic situations. "Having local teams attempting to create individual solutions is not an effective use of time or capacity. There is a need for more transparent access to data and social media platforms, advocacy to enable this, global collaboration on solutions, and analysis platforms developed in collaboration with those working in public health."

Per the researchers, "broader challenges in metrics and measurement of the circulating concerns, questions, information voids, and narratives, are also required in addition to focusing on social media analytics alone. For example, WHO has piloted a platform for responsible and improved analysis of narratives, concerns and questions expressed by people in digital spaces, while at the same time also explored the opportunities from using digital ethnography and qualitative approaches to understand concerns and narratives circulating in communities. Moreover, WHO and UNICEF have developed a manual and trainings on how to triangulate social media analysis with other health data sources to improve the usability and relevance of infodemic insights."