The Psychological Impacts and Message Features of Health Misinformation

"This review revealed that exposure to health misinformation can damage relevant psychological antecedents of behaviors such as knowledge, attitudes, or behavioral intentions."

In recent years, health authorities have become increasingly concerned about the impact of misinformation on health behaviours. However, the mere presence of health misinformation on social media is not an indicator of its impact, and correlational studies provide limited evidence on whether unhealthy behaviours or their antecedents are causally due to misinformation. This article provides a systematic review of 45 articles containing 64 randomised controlled trials (RCTs; N = 37,552) on the impact of health misinformation on behaviours and their psychological antecedents regarding infectious and non-infectious diseases, vaccination, medication, nutrition, tobacco consumption, and environmental issues such as climate change.

The goals of this systematic review include: (i) synthesising evidence from RCTs that evaluated the impact of exposure to health misinformation on human behaviours or their psychological antecedents; (ii) analysing message features of health misinformation using the FLICC framework, which based on five common rhetorical tactics of science denialism: fake experts, logical fallacies and misrepresentations, impossible expectations, cherry picking, and conspiracy theories - thereby supporting future debunking and inoculation interventions to counter misinformation by uncovering common argumentative tactics; and (iii) mapping potential knowledge gaps in understanding tactics of health misinformation and deriving directions for further research in this area.

The review uses the constructs of an extended theory of planned behaviour (TPB) as a framework to analyse the impact of health and environmental misinformation on human behaviours and their psychological antecedents. According to the TPB, behavioural intention (i.e., the perceived likelihood of engaging in behavior) is the strongest predictor of actual behavior. The primary antecedents of intention are an individual's attitude (i.e., favourable or unfavourable beliefs about a behaviour), perceived norm (i.e., belief in what significant others do and motivation to comply), and perceived behavioural control (i.e., belief in the ease or difficulty to perform a behaviour). Further, feelings and emotions toward a behaviour, trust, and knowledge are common extensions of the TPB that are also mentioned as relevant outcome variables in research on the impact of misinformation.

The systematic review was limited to RCTs that compare the impact of exposure to health misinformation (i.e., experimental condition) with exposure to either no information or unrelated information (i.e., control condition). RCTs that only compare exposure to misinformation with exposure to interventions (e.g., debunking or factual information) were not included because the effects of those RCTs may result from either the impact of misinformation or the impact of the intervention.

The literature search was conducted on May 23 2022, using the following databases: Medline via PubMed, PsychInfo, and Scopus. The 45 identified articles cover a range of 41 years of 64 RCTs on the impact of misinformation, with the oldest article being published in 1981 and the peak of publications in 2021 (12/45).

The analysis found that in 49% of the cases, exposure to health misinformation damaged the psychological antecedents of behaviours. That is, many studies reported that exposure to misinformation decreased the intention to perform healthy or pro-environmental behaviours, decreased positive attitudes toward these behaviours, decreased the perception that health behaviours or pro-environmental beliefs are the norm, decreased trust in advocates of science, decreased knowledge, or increased negative feelings toward public health measures. For example, seven evaluations revealed that exposure to misinformation caused lower intentions to perform healthy behaviours; among them, exposure to vaccination misinformation decreased the intention to get a fictitious vaccine for a fictitious child among United States (US) adults and decreased the intention to get a human papillomavirus (HPV) vaccine among young Chinese adults.

No RCTs evaluated the impact of exposure to misinformation on direct measures of health or pro-environmental behaviours (e.g., actual vaccination), and few studies explored the impact of misinformation on feelings, social norms, and trust. As an example of the latter: Exposure to vaccination misinformation decreased confidence in news organisations and credibility of journalists for US adults with low prior belief in vaccine conspiracies; trust was increased when the media shared misinformation that was in line with the prior beliefs of a vaccine-hesitant audience.

Other studies did not find any evidence for the impact of misinformation on any of the psychological antecedents or found mixed results or unintended backfire effects (42/82). The lack of persuasiveness of misinformation may reflect the fact that the persuasiveness of any message is largely context specific, varying across characteristics of the sender, the receiver, and the message.

Following the second goal of this review, the researchers analysed the message features of misinformation for 57 unique misinformation stimuli from various health domains using the FLICC framework. Most misinformation was based on logical fallacies, conspiracy theories, or fake experts. RCTs evaluating the impact of impossible expectations (expecting the impossible from science) and cherry-picking are scarce.

In terms of the third goal of this review, the researchers suggest that more experimental analyses of context variables, such as characteristics of the receiver or the sender, are needed to understand when health misinformation causes harm. Furthermore, more conceptual and theoretical work is needed on the causal pathways through which misinformation influences people's beliefs and behaviours. The negative influence of misinformation may often not be direct, but indirect, and the role of potential mediators such as feelings and emotions, social norms, and trust is not well understood. Moreover, there is a lack of diversity in health misinformation research because most RCTs focus on healthy adult US populations and are conducted online. Finally, RCTs that test the impact of deceptive tactics such as impossible expectations and cherry-picking are needed.

These findings "may help researchers who want to test the persuasiveness of different message features of health misinformation and may guide the design of future interventions aimed at detecting deceptive tactics."