How risk assessment informs what to do about health-harming narratives

When I find myself in conversations with colleagues over monitoring health-harming narratives and what one can do about it, I often refer back to how the public health principles and functions of risk assessment translate into the context of the information ecosystem.

Elisabeth Wilhelm and I started developing ideas around this with other colleagues at the beginning of the pandemic. We wanted to introduce public health approaches to risk assessment and response into what used to be a single analyst's subjective assessment, often based on a few dashboards that monitored specific topics in the information environment, but not considering the context and information landscape. We described the basics of risk assessment and the rationale for it in the WHO/UNICEF infodemic insights report manual for field practitioners.

The challenge of such approaches in the public health context is that often such practices monitor the wrong thing, measure the wrong thing, and produce descriptive summaries that are not easily translated into the work cadence and objectives of an emergency response team, or if insights are delivered for routine programming, into the design and adaptation of the delivery of health services, products or guidance.

When producing an insights report in the public health context, our insights are only going to be actionable if they are translated through the lens of decision-makers who may be used to triangulating data and making decisions on insights from epidemiological and socio-behavioral data sources, but have less experience and training on how insights from the information environment can inform their actions.

How one defines harm to health matters in prioritizing what should be done to address it

There are many types of harms that can come from circulating health narratives. This includes stigmatization of groups of people, lowered risk perception, avoidance of health guidance or health information, using non-recommended treatments or prophylaxis and encouragement of others to do the same, advocacy for harassment and violence against health workers, and continuing and/or propagation of inaccurate health information.

If one focuses analysis in the context of health outcomes, as above, then the research questions, choice of data sources to triangulate, analysis, and risk assessment all are guided by the objectives of the analysis.

It also becomes less important to describe and count reactions of the loudest on one social media platform, and the focus shifts to what is people's health-information seeking and consumption-related behaviors, and what may be the impacts of these on their health behaviors and downstream on the health service utilization and health system impact.

Analogously, how we understand the public health harms we aim to mitigate will also inform the recommended intervention areas. If one only sees a hammer as a tool in the toolbox, everything will look like a nail.

In analyzing and understanding people's expressions of questions and concerns, and information- and health behaviors, we often way too quickly think only about how to respond through communications or engagement actions. However, if a community is expressing concerns over a lack of access to testing or services, no amount of messages or engagement is going to resolve those concerns. If one truly is open to addressing the findings from analysis of community concerns, the intervention areas emerge easily.

Assess the risk with public health action in sharp focus

Risk assessment is not an academic exercise detached from the work objectives it's supposed to inform.

Snazzy reports that cannot be turned into public health action rapidly, especially in a health emergency, will fall like a tree in a dense forest that no one will hear or pay notice to.

Assessing risk from circulating health narratives helps determine which ones are the highest priority to respond to. Things to consider in the risk assessment:

●       If you want the insights you generate to be used, they need to talk to the policy and operational context for which they are being produced

It’s impossible to respond to all health-harming narratives quickly enough, and even if we wanted to, we cannot fix the internet to do so.

When considering what determines the thresholds of risk in a localized context, this must be informed not only by the risk tolerances of the organization at hand but also by the areas of intervention that their mandate offers for action - or the ability of the organization to pass the insights on to a partner organization for action.

When delivering insights on the circulating narratives and misinformation, analysts must consider that if they present summaries that the public health organization is unable to respond to, the created tension within the organization that something must be done against mis- and disinformation can result in actions not conducive to public health action.

A report that describes risks that cannot lead to an immediate public health action can drive misunderstandings within the organization on how it can address health mis- or disinformation.

●       A risk of a harmful health narrative is not directly proportional to its spread.

Just because a mis- or disinformation narrative is widespread, this does not make it the most important one to address. It may, in fact, be more important to address a narrative that is circulating in a vulnerable community that might already be dealing with low trust in government and limited access to health services where mis- and disinformation may more easily gain traction.

●       Risk level is also a function of risk to localized harm to a population and action that harm health disproportionately.

There are relatively lower-risk narratives, such as those that are limited in platform or geographical scope or advocating for actions that are misguided but are not harmful to health (for example, limiting farting to reduce air pollution). Narratives that are higher risk to health would include those who can be observed on different platforms and communities and advocate for actions that can cause direct harm or target a specific vulnerable population (for example, mpox vaccine turns you gay).

● Risk to the reputation of the public health organization must not factor in the assessment of risk to the health of the populations it serves.

A public health approach to assessing risks in the information environment focuses on the harm to people and communities, and not on the harm to the reputation of the public health organization that is performing it.

How we understand harm and how we measure it determines the actions that we will take to mitigate it. A public health risk assessment takes into account also the organization’s ability to take action on the insight, not merely a qualitative assessment of the signal.

While the reputation of a public health organization is important for its ability to maintain the status of a trusted messenger, mixing of the reputation management and public health objectives and actions in communications can backfire.