Health misinformation: protocol for a hybrid concept analysis and development

Introduction

Misinformation is increasingly recognised as a serious barrier to implementing public health advice and programming. For example, exposure to misinformation has been identified as contributing to reduced intention to receive vaccines^1–3, engagement in unproven and potentially harmful treatments for cancer⁴, and the decision to delay or avoid interactions with health services and institutions⁵. This has, in turn, raised concerns that health misinformation may negatively impact the capacity of health services and public health programmes to support public well-being.

General interest in misinformation as a problem has surged over recent years: indeed the number of annual research citations containing the word “misinformation” has nearly tripled since 2016⁶. This likely reflects anxiety about other, related shifts in the information landscape. These shifts include, for instance, the mass shift to online information-seeking (and resultant need to adapt literacy skills to a digital environment)⁷; the potential of the internet and social media to accelerate the spread of unreliable claims⁸; and growing concerns about the impacts of “post-truth politics” on public decision-making⁹.

Concerns about misinformation in health contexts have increased since the onset of the coronavirus disease 2019 (COVID-19) pandemic. In March 2020, the World Health Organization (WHO) announced an “infodemic”¹⁰, resulting in significantly scaled-up investment in interventions targeting health-related misinformation^11,12. Proposed interventions to combat misinformation have included, for example, “prebunking” or “psychological inoculations”⁹, warning labels on social media posts¹³, fact-checking programmes¹⁴, and health literacy training targeting functional and critical thinking skills^15,16. While these interventions have shown some promise for increasing resilience against unreliable health claims^9,13,14, much work remains to be done to understand how health misinformation spreads, who is most vulnerable to believing unreliable health claims, the best strategies for supporting individual and group resilience against health misinformation, and the best ways to measure the success of such interventions.

In addition, there remains limited consensus about what types of information constitute “misinformation” in health science. Many discussions of health misinformation have drawn on general, transdisciplinary concepts, such as the Oxford Dictionary definition (“wrong or misleading information”)^17,18 or an early definition from Lewandowsky et al. (2012) (“any piece of information that is initially processed as valid but is subsequently retracted or corrected”)^19,20. However, the relevance of these definitions for use in health science is unclear; and many studies of misinformation do not explain how they have conceptualised misinformation, instead treating it as universally understood²⁰.

Differentiating between “reliable” and “unreliable” health information in practice is complicated, not least because scientific concepts of reliability are intentionally kept flexible, and depend on evolving evidence: what is considered to be “true” about a health topic can (and should be expected to) change, both over time and in line with evolving evidence. In addition, evidence on a given topic may be uncertain, unclear, or conflicting, resulting in disagreement about which information should be regarded as “reliable.” For example, in the early months of the COVID-19 pandemic, when evidence of airborne severe acute respiratory syndrome coronavirus 2 (SARS-COV-2) was limited, many public health authorities discouraged the use of face masks outside of healthcare settings²¹; and social media posts advising people to wear face masks in public were “debunked,” including by major authorities like the World Health Organization, as “incorrect”²². However, as evidence of airborne SARS-COV-2 transmission mounted, and research emerged demonstrating that face masks could reduce the transmission of SARS-COV-2 under laboratory conditions²³, major public health organisations pivoted to recommending masking by the general public²⁴, despite evidentiary and expert disagreement about masks’ potential impact in real-world settings^25–27. This disagreement points to theoretical tensions – as-yet unresolved – around how evidence can (or should) relate to concepts of truth in public health. Resolving these tensions is especially urgent in light of evidence that changing public health advice can have negative consequences for public trust: for example, in a qualitative study by Zhang et al. (2021), many participants reported perceiving the change in masking advice during the early months of the COVID-19 pandemic as an “inconsistency,” and described how this undermined their trust in public health authorities during subsequent phases of the pandemic²⁸.

Furthermore, the processes scientists use to generate and evaluate evidence are not always transparent to the public, resulting in potential uncertainty about how truth has been conceptualised or measured. The term “misinformation” (and proxy terms, such as “fake news” and “alternative facts”) have in some cases been wielded to further particular partisan agendas, or to discredit opposing viewpoints; and the term therefore remains controversial. Concerns remain about how “misinformation” items have come to be classified as such²⁹. In recent years, scholars have turned their attention to clarifying these processes and to identifying the values that inform them. For example, Vraga & Bode (2020) proposed defining misinformation as “information considered incorrect based on the best available evidence from relevant experts at the time”³⁰. As Southwell et al. (2022) note, this definition of misinformation has major strengths for use in science³¹. First, it moderates the expectation that scientific information will contain any inherent quality of “correctness”: it clarifies that information can only be regarded as correct or incorrect, based on the available scientific evidence^30,31. Secondly, it recognises that the “accuracy” of a particular piece of information may change over time³⁰; while, at the same time, affirming the value of scientific evidence in informing public decision-making^30,31.

Southwell et al. (2022)’s subsequent definition of scientific misinformation – “publicly available information that is misleading or deceptive relative to the best scientific evidence or expertise at the time, and that counters statements by actors or institutions who adhere to scientific principles without adding accurate evidence for consideration” – draws heavily on these ideas³¹. Their definition of misinformation is distinguished, however, by its emphasis on the value of scientific rigour (rather than simple expertise or authority) for determining the reliability of information³¹. This is a potentially important evolution in the concept of scientific misinformation because it substantially resolves challenges around expert disagreement³² and the potential of experts to themselves spread misinformation. Significantly, Southwell et al. (2022) also specify that, in order to be classified as misinformation, unreliable content must be publicly accessible and accessed³¹. In this framework, misinformation is distinct from – and larger than – individual misconception or untruth³¹. Rather, it is a “disorder of public science” (or, in other words, information which challenges our ability to achieve scientific consensus and come to “shared understandings”)³¹.

The focus on misinformation as a threat to “shared understanding[ ]”³¹ is echoed in other places as well; but more often it is positioned as a symptom, rather than a cause, of public dissensus. For example, over recent years, academics have described misinformation, as, variously, a symptom of breakdowns in trust between the public and social institutions, a function of increased political polarisation and decreased community cohesion, and a consequence of increased social inequality^19,33–35. It has also been hypothesised that increased social fragmentation may have catalysed a movement against scientific values, often in favour of so-called “alternative epistemologies”³³.

These phenomena may, of course, have serious negative consequences for implementing public health initiatives. However, some scholars have questioned the value of pathologising these “alternative epistemologies.” For example, Furman (2020a) argues that it is reasonable for an individual to reject scientific information, including health information, when the epistemic values which produced it are incompatible with their priorities³⁶. In addition, she suggests that real-world processes for evaluating the reliability of health information may, in some contexts, go beyond the “strictly epistemic”; because health-related decision-making requires people to make “all-things-considered judgements”³⁶. For this reason, information that is communicated in an accessible way^36,37, transmitted by trusted sources³⁷, or identified as relevant to a person’s needs or values³⁷, may reasonably be understood to be reliable, even if it was not generated through scientific processes.

On the other hand, there are academics who dismiss these kinds of evaluative criteria as (for example) epistemically “lazy”³⁸, “emotive”³⁹, “irrational”³⁹, “willfully ignorant”⁴⁰, and “peculiar”^37,41. This risks neglecting the social-ecological and cultural influences (including, for example, perspective, values, contextual relevance, and institutional trust) that impact public understandings of reliability¹⁹. It may also risk creating a situation in which academics are left “talk[ing] passed” [sic] their audiences³⁷ about what it means for information to be “reliable” (or, conversely, “misinformation”). Nonetheless, much of the discourse on misinformation continues to use scientific expertise³⁰ (or its proxy, institutional affiliation³¹) as primary indicators of content reliability. For example, recent studies^42,43 and interventions⁴³ in misinformation have persisted in conceptualising “misinformation” as any item which originates from a list of pre-identified websites deemed “unreliable sources.” This disconnect should be of particular concern for those serving marginalised, underserved, and otherwise vulnerable groups, who may be more likely to have differing perspectives about, approaches to, and needs regarding the delivery of information¹⁸.

One proposed solution to the challenges of determining information accuracy or reliability, outlined above, is to handle information in terms of “normativity” or “disnormativity.” In this framework, normative information represents information that is generally accepted to be true in a given context, and disnormative information is information that deviates from this accepted truth^18,44. For Ruokolainen & Widén (2020), citing Haasio (2015), this approach has the advantage of allowing researchers to examine information processes without making ontological or epistemic commitments about the reliability of information^18,44. It also creates conceptual space to explore how information processing, and evaluations of credibility, may be informed by different “views, norms, and values”⁴⁴. However, this framework may not be intuitive or well-suited to discussions about health misinformation, because health information is often deliberately tied to a scientific epistemology in which subjective factors are meant to be managed and minimised.

Information characteristics such as “accuracy” and “reliability” do important work in conversations about health, and there is an urgent need to develop a model for differentiating between more and less reliable health information. “Misinformation” is a useful designation for information that is not reliable. However, there is an urgent need to clarify: (1) what characteristics are accepted as aspects of health misinformation, (2) how the concept of misinformation should be operationalised for research and intervention purposes, and (3) what consequences of health misinformation have been identified. Clarifying the concept of health misinformation will be essential for developing a coherent evidence base around misinformation processes, establishing shared goals for interventions between researchers and the professionals involved in moderating health information, and ensuring that information interventions are both accessible and acceptable to the populations who will most benefit from them. Therefore, we propose to undertake a hybrid concept analysis to explore the characteristics, measurement, and applications of the concept of health misinformation.

Protocol

Discussion

Conclusion

Exposure to unreliable health claims carries risks for individual health and health-related decision-making; and may have further, indirect consequences for public health programmes, policies, and health systems. “Misinformation” is a potentially useful concept for communicating about the reliability of health claims. However, there remains a lack of clarity – and, in some cases, active disagreement – about how misinformation should be identified, characterised, or addressed. Clarifying the concept of misinformation will allow for more accurate identification of at-risk groups, pathways of vulnerability, and potentially useful goals for intervention. It will also facilitate trust between health authorities and their audiences; and help to ensure that misinformation interventions reflect the information needs, values, and goals of the populations to which they are directed.