Health authorities’ health risk communication with the public during pandemics: a rapid scoping review

We included studies of modes of communication concerning pandemic health risk communication from health authorities to the public. Modes of communication in this paper included but were not limited to web-based information, social media, television, newspapers, video, texts, and narratives. We conceptualised health authorities from a broad perspective, including governments, official health experts, healthcare professionals as official spokespersons, health authority officials, health agencies, and official health bureaucrats, at the regional, national, or international level (i.e., the WHO). We excluded studies concerning health communication between individuals, such as a medical doctor and a patient (e.g., e-health, telemedicine), or between healthcare professionals (e.g., digital educational methods, digital solutions).

Pandemics included but were not limited to swine flu (H1N1) and COVID-19. The COVID-19 pandemic is of primary interest, but as it is currently ongoing, searches on this topic could only provide studies on the response phase and use of diverse modes of communication; thus, we included the swine flu pandemic, as it might include knowledge on a long-term perspective on the pandemic. We excluded studies concerning infectious diseases without pandemic potential. Only studies after 2009 were included, which reflects the timeframe of the evidence generated following the last large-scale pandemic (swine flu) and the need for evidence about communication modes to reflect the scale of technological change over the past decade.

We included studies of communication to the public and specific target groups in the public without any predetermined categorisation. Commentaries, reviews, opinion pieces, or other papers not reporting primary empirical research were excluded. English-language articles for both qualitative and quantitative peer-reviewed empirical studies were included.

Fourteen of the included studies used survey methodologies with self-reported use of communication channels, with health authorities among the information sources [36‐51]. These studies found that people receive pandemic health risk information through a broad spectrum of communication channels and information sources and that they are influenced not only by newspapers, television, printed information, governmental websites, scientific articles, radio, and SMS from the government but also by interpersonal and informal sources, such as friends, family, healthcare professionals, and social media [36‐50]. Hence, people are not passive receivers of health risk information but are influenced by the totality of the information they receive from various and multiple communication channels and information sources in addition to health authorities’ health risk communication.

Five of the included studies of health authorities’ health risk communication and traditional mass media (television and newspapers) reported on outcomes related to framing and the use of mass media as a communication channel [52‐56]. A central topic was how framing messages could affect public responses. Framing is defined by Nisbeth [[81], p. 216] as “interpretative story lines that set a specific train of thought in motion, communicating why an issue might be a problem or pose a threat or what might be responsible for it, and what should be done about it”. Studies of media coverage in mass media during the H1N1 pandemic found that the mass media used sensationalist framing of their messages, conflict frames and war metaphors [52‐54, 56]. Tabloid papers used risk-amplifying frames when presenting press releases from the WHO and health authorities across 10 European countries during the H1N1 pandemic. Conflict and damage were emphasised when disseminating the press releases [53]. Hall and Wolf [52] interviewed German public health experts and found that the participants attributed fear and panic in the public to sensationalist media coverage.

One qualitative study examining the perceptions of healthcare professionals, regional public health officers, epidemiologists and public health experts across eight European countries following the H1N1 pandemic found that they all experienced collaboration with the media as poor and that the professionals felt misunderstood [55]. The stakeholders emphasised the importance of establishing good relations between national health authorities and the media and highlighted that society’s trust in public health authorities must be improved long before a pandemic [55].

Overall, the evidence describes framing as an important topic in studies of mass media; however, no studies were identified regarding the effect of framing and public responses. Furthermore, no studies reported on outcomes related to health authorities’ communication with the public through radio.

Six of the included studies on government websites assessed readability using multiple validated indexes, accessibility, and timeliness of the updates on the websites [57‐62]. Cross-sectional Google search studies of the readability of websites found that websites of the UK Health Service government [57] and online educational articles on COVID-19 on websites related to governments, hospitals and health organisations (e.g., WHO) [58] provided information related to COVID-19 that was too difficult for the general population to read. A prospective study of the websites of the U.S. Centers for Disease Control and Prevention during the first months of the H1N1 pandemic classified the documents according to their intended audience and their reading level. The study analysed the material for readability using instruments to assess the suitability of materials, which take into account criteria such as layout, typography, graphics, and surrounding context and its accessibility to an intended audience. The authors concluded that while the webpages were adequately adapted to the reading level of the intended audience, the format and layout (i.e., text-heavy and densely formatted) made the material difficult to comprehend [59]. A study of the accessibility of information on the WHO website found that the information was not accessible for the elderly because the website lacked non-text alternative content for people with vision problems [60]. The lack of timely updated information on webpages has been reported as particularly problematic at the local governmental level in China during the COVID-19 pandemic [62] and during the H1N1 pandemic in the U.S. [61].

The evidence suggests that adaptations for the reading level of the audience alone do not determine information accessibility. The readability, layout, format, accessibility and timeliness of the updates are additional variables that have been studied in assessing the quality of health authorities’ and governmental websites.

Eight of the included studies on the social media platform Twitter and the Chinese microblogging website Sina Weibo reported on message content (i.e., framing), engagement (i.e., message retransmission and number of likes) and misinformation [15, 16, 63‐68]. Sutton et al. [63] conducted an analysis of officials’ communication on Twitter and found that retweeting (i.e., reposting or forwarding) of messages is influenced by message content, message features, the organisational type of the account, the number of followers of an account and the time and day the message was sent. Messages were most likely to be retweeted if they contained content related to surveillance, technical information, self-efficacy, or collective efficacy or symptoms. Messages that included video were retransmitted 63% more often than those without video [63]. In a textual analysis of Twitter messages from U.S. public health agencies’ accounts during the first 60 days of the COVID-19 pandemic, Sutton et al. [15] found that the content changed from focusing on instructive messages in the first phase to motivational messages to sustain action in the long term, which focused on the need to protect vulnerable groups.

Five studies explored factors affecting lack of viewer engagement with health authorities’ social media posts related to COVID-19. Kamiński et al. [64] conducted an infodemiology study (an area of science concerning contributed content on the internet) of Twitter messages published by health authorities across the globe. They found that health authorities often used negative framing in Twitter messages on pandemic information compared to celebrities and politicians, who posted optimistic messages that were associated with higher viewer engagement. However, while more positive posts from health authorities may help to enhance engagement, it may also undermine the perceived seriousness of the message. The authors argued for collaboration between health authorities and opinion leaders to succeed in pandemic risk communication [64]. An information dissemination study of U.S. health authorities’ Twitter messages identified inconsistencies and incongruences in messages over time related to critical topics in the communication of COVID-19 information. However, the study did not report on the effect of message inconsistency and people’s self-protective behaviour [65]. Studies of posts on the Sina Weibo platform found low engagement with COVID-19 posts from the Chinese government [16, 66, 67]. The studies suggested that the lack of engagement was related to a failure to use media richness properly [66] and the use of nonpersonal and nonnarrative content [16, 67].

The evidence suggests that health authorities are faced with low engagement with their posts related to pandemic health risk information on social media. Messages containing narratives and a degree of self-efficacy are among the topics that affect engagement on social media platforms Twitter and Sina Weibo. No studies on health authorities’ use of social media in pandemic health risk communication in relation to effects on self-protective behaviour were identified.

Five of the included studies explored the use of the world’s largest video-sharing platform, YouTube [69‐73]. Complementing traditional television, online video has become increasingly important. However, videos from credible governmental sources and the WHO are highly underrepresented on YouTube, leaving YouTube vulnerable to the spread of misinformation [69‐71]. Li et al. [69] found that 25% of the most popular videos related to COVID-19, which elicited more than 62 million views worldwide, contained misleading content. While governmental videos contained factual information and more accurate COVID-19-specific information, they accounted for only 11% of the videos and 10% of the views [69]. COVID-19-related videos with misleading content were found to have a higher percentage of views on YouTube than those from credible sources [70]. Across nations, the few videos from government and health agencies on COVID-19-related information on YouTube were often found to be credible, but they received a low number of likes and comments [72] and represented a low share of their videos [70]. Two issues that may explain the lack of social media “reach” by health authorities on YouTube are YouTube’s heavy reliance on the English language (videos) and the fact that public agencies tend to have low numbers of subscribers to their YouTube channels [70].

The evidence based on these studies [69‐72] suggests that YouTube is a source of misinformation for COVID-19, and health authorities’ videos on YouTube are underrepresented and have low engagement. No studies on health authorities’ use of YouTube videos in pandemic health risk communication that also reported on their effect on self-protective behaviour were identified.

Ten of the included studies reported data on trust in formal sources, i.e., government sources and health care professionals [37, 41, 46‐48, 50, 51, 56, 68, 75]. Trust in formal governmental sources has been associated with more accurate pandemic risk knowledge and self-protective behaviour [46, 47]. In a survey study of Hong Kong adults, trust in formal sources from the government about influenza was associated with greater reported understanding of the causes of H1N1 and more self-reported hand hygiene among males than among those who trusted informal and interpersonal information sources [46]. In a survey study of the US population, trust in governmental sources was positively associated with accurate disease knowledge related to COVID-19. Younger people expressed higher trust in informal sources, such as the CNN news channel and social networks (e.g., Twitter), than US governmental sources, while the older population expressed higher trust in US governmental sources. The authors argue that younger people’s high level of trust in informal sources makes them vulnerable to misinformation and false news [47].

However, credible sources may differ from the sources that are used most often by people. Jardine et al. [37] found that Canadian people used mass media and friends as information sources about the H1N1 pandemic but found public healthcare professionals to be the most credible sources [37]. King et al. [48] found that despite its lack of trustworthiness, the Australian mass media was the most important source of health risk information for parents during the H1N1 pandemic. Medical doctors, authoritative hospitals and government websites, in contrast, were perceived as most trustworthy.

A survey study of a sample of Slovenian adults by Lep et al. [41] found that medical professionals and scientists were perceived as most credible but that news portals, television news and social media were the most used information sources related to COVID-19.

Two qualitative studies reported on failures in governmental risk communication in China [51, 68]. According to Zhang et al. [51], Chinese authorities failed to disclose their uncertainties during the early Wuhan COVID-19 outbreak, which the authors argue undermined their institutional trust. In another qualitative study, Zhang et al. [68] explored the spread of misinformation related to the effect of a Chinese herb on COVID-19 prevention. A message was delivered by one official health expert through the Chinese authoritative media, which affected the perception of the credibility of the message. Social media shared the message using titles inconsistent with the facts, which eventually led to rumours and hoarding behaviour [68].

The evidence describes both positive and negative behavioural outcomes of health risk information from formal sources. However, while the evidence indicates that trust in formal sources is associated with increased knowledge and self-protective behaviour, the studies lacked controls for education level, which may moderate the relationship between gender and age differences and knowledge outcomes. The evidence suggests a disconnect between the information sources people use and find credible and those they report using for pandemic health risk information.

Corresponding with the literature review by the WHO on risk communication in public health emergencies [12], five of the included studies found that messages related to pandemic health risk were perceived as most trustworthy when they came from a healthcare professional [37, 41, 48, 56, 75]. An RCT assigned Japanese participants to receive intervention messages from different sources. This study found that health care professionals’ written persuasive messages encouraging people to stay home were more effective than messages from the governor, patients or a public health expert or a control message [75]. Luth et al. [56] conducted a content analysis of television news clips and found that one of the reasons why the Canadian government lost control of its message in mass media was related to the choice of governmental spokespersons in press conferences. When Canadian government officials appeared on television in press conferences with a visual identity as bureaucrats and politicians, they decreased the credibility of their message [56].

Two studies suggest that the credibility of governmental spokespersons and governmental trust are highly related to the evaluation of crisis management [50, 51]. A survey study of public perception of the Israeli government’s emergency management regarding COVID-19 found that the higher the public’s trust and evaluation of crisis management were, the greater the public’s compliance with guidelines. Those who perceived that the prime minister was the most credible spokesperson had a high evaluation of crisis management, whereas those with low evaluation of crisis management trusted other sources of information, such as infectious disease specialists, ministry of health websites and scientific articles [50]. Governmental trust was found to be reduced in China due to the perception that officials had concealed information about the COVID-19 outbreak in Wuhan. This led to reduced government credibility and increased spread of information from unofficial sources, conspiracy theories, and rumours [51]. The evidence suggests that while health care professionals are often perceived as a credible source, trust in governmental sources may not be stable over time. However, no longitudinal studies were identified that studied the variability in governmental trust and media types over time. Although the evidence indicates that trust in healthcare professionals is high, independent of media type, no studies were identified that examined the impact of different media types on people’s trust in formal sources.

Health authorities are advised to collaborate with communities to ensure that their concerns and information needs are understood and to tailor advice and messages to address the target groups [12, 13, 18, 82]. Only three of the included studies reported data on modes of communication and ethnic minority groups [78‐80]. One qualitative study with the Latino community in the U.S. found that all respondents wanted more personalised COVID-19 information from researchers or health professionals via communication channels such as personal email, text messaging or a group communication method (e.g., video group call, Facebook group) [80]. A Swedish policy analysis described how the government and decision makers relied on migrant community leaders, representatives of migrant associations, religious leaders, and other influencers to reach out with culturally sensitive information [78]. In a qualitative focus group study of communication priorities for vaccination to indigenous people in Canada, Driedger et al. found [79] that risk messages were transmitted in different dialects and by different formats (radio, television, print, online, community sessions). An additional issue was the need to tailor not only the communication mode but also the message for ethnic minority groups and migrant populations. In a study of indigenous Canadian people, the authors concluded that the communication failed to engender behaviour change because the target group did not understand why they were prioritised for vaccines, instead believing that the government was using them as “guinea pigs” [79].

The evidence from the included studies suggests that ethnic minority groups and migrant populations need personalised information and trusted spokespersons. The review did not identify studies of effective modes of communication with ethnic minority groups.

According to Slovic [83], people tend to consider their feelings to guide their decision making and judgements, and the use of narratives has been explored in the risk communication literature [14]. Six of the included studies explored how the narrative tone of communications plays a role in the way health authorities’ messages are interpreted and acted upon during pandemics [16, 49, 66, 67, 73, 76]. The choice of narratives has been found to affect uncertainty and anxiety in the public [49], pandemic knowledge [73], engagement [16, 66, 67] and self-protective behaviour (stay home messages) [76]. Liu et al. [49] found that messages negatively affected people’s anxiety when hospitals reported a need for monetary donations, as this reflected a shortage of beds in hospitals. People experienced less anxiety when the government reported opening schools in China, as this provided them with hope [49]. An experimental study found that people reported greater knowledge and greater pandemic influenza prevention measures when they viewed non-narrative health videos about pandemic influenza compared to those who viewed narrative videos [73]. Two infodemiology studies of the Chinese social media platform Sina Weibo found that people were more engaged (more comments and likes) with narrative posts than non-narrative posts, both from the government and from personal sources [16, 67]. Posts related to new evidence and a nonnarrative style were strong negative predictors of the number of shares [16].

A focus group study of English residents found that factual, evidence-based messages related to pandemic influenza from health authorities were most convincing, particularly those that included cost-benefit comparisons of H1N1 vaccination safety. Messages that aimed to elicit feelings of anticipated regret for not getting vaccinated were generally perceived as patronising and unprofessional [76]. Another content analysis of the social media platform Sina Weibo found that posts with high media richness (both text and video) and positive emotions increased citizens’ engagement regarding government social media, while posts high in media richness and negative emotions attenuated citizens’ engagement. Consequently, the authors concluded that plain text should be used when sharing posts that elicit negative emotions, while video should be attached to text when reporting content eliciting positive emotions [66].

The evidence indicates that the use of narratives and messages eliciting emotions in health authorities’ pandemic health risk communication leads to engagement with the public, while people learn more from non-narrative messages than narrative messages. The evidence also suggests that negative narratives may negatively affect people’s emotions.

Four of the included studies reported on outcomes related to the use of jargon in health authorities’ health risk communication [57, 59, 68, 77]. In terms of messages and wording in pandemic health risk communication, the WHO recommends avoiding the use of jargon [12]. However, an experimental study on health messages from health agencies in the context of COVID-19 found no negative effect of jargon on how much difficulty people experience in understanding a message (processing fluency), which may relate to people’s greater motivation to process information that includes jargon when urgent and risky concerns are communicated [77]. Two studies of readability emphasised the importance of adjusting written material to the intended audience. Khan et al. [57] found that the use of jargon and long sentences were contributing factors to the purportedly low readability of UK health authority websites. In a prospective study of the U.S. health authorities’ CDC guidance documents, Legasse` et al. [59] found that while the use of scientific jargon was adjusted to the reading level of different audiences, the documents were still difficult to comprehend due to suboptimal layout (i.e., text-heavy and densely formatted) [59]. The use of jargon in a news report from the Chinese government has also been explored as a contributor to the misinformation and development of rumours in a qualitative study [68].

The sparse evidence based on these studies [57, 59, 68, 77] emphasises that the use of jargon in pandemic health risk messages is multifaceted. Studies report on variables related to the reading level of the audience, misunderstandings, and the recipient’s motivation to process the information.

Whether a message is communicated visually through graphics, graphs, moving images, colours, symbols or by the use of text or numbers can have a significant impact on its reach and how the message is received. According to WHO guidelines, engaging and effective risk communication is multimodal and includes visual information [12]. Three included studies explored the impact of graphic and audio-visual means [17, 56, 74]. The understanding of exponential growth, for example, which is a key concept during COVID-19, is interpreted differently when communicated through numbers as opposed to visually. A quasi-experimental study with 1980 participants from 43 countries found that people tend to linearise exponential functions when assessing them intuitively by looking at graphs and underestimate future values based on the current value. The study found that showing people prior data in raw numbers before showing them graphs causally reduced exponential-growth prediction bias, which affected risk perception and was associated with improved safety compliance with the WHO’s COVID-19 recommendations [74].

According to Luth et al. [56], public health agencies also need to be aware of the emotional valence and the effect of colours and symbolism in visual media, as they can be used to encourage positive health behaviours or elicit panic. In a content analysis of Canadian television news of the H1N1 pandemic, Luth et al. [56] found the use of static information screens in conjunction with auditive messages that used colours (red) and symbols (technical, alien objects), which may contribute to negative valence [56]. Luth et al. [56] also emphasised the importance of matching news footage with the public health message delivered. Luth et al. [56] found a mismatch between the visual content and the audio content in the presentation of official priority groups for vaccination. Visual footage showed seemingly healthy nonpriority individuals in line to be vaccinated, but the audio targeted vulnerable groups [56].

The perception of colour may also vary depending on the audience and context. A study of an information leaflet used by the Slovak Ministry of Health in communication about the H1N1 pandemic found that the use of white and red types on a dark blue background was preferred by most primary school students. This combination of colours, however, was perceived negatively by more than half of healthcare professionals, presumably because of differences in ageing and different colour perceptions and preferences in children and adults [17]. The authors argue that reading literacy and age are important factors to take into account when tailoring information in health information leaflets [17]. The limited evidence related to graphic and audio-visual means [17, 56, 74] suggests that different modalities affect the message in a multifaceted way. Visual and multimedia messages are interpreted differently than text- or number-based messages. Symbolism, colour perception and preferences change according to audiences and contexts. While the effect and importance of creative and audio-visual production and means is an important area for pandemic health risk communication, no studies of creative communication choices or evaluations/comparisons of various visualisation techniques relating to graphic and audio-visual communication were identified.

Consistent with other reviews, this review found many studies reporting that people receive information from multiple channels [12, 84]. The studies suggest that people are not passive receivers of health authorities’ information, a wide spectrum of communication channels is used, and meaning making is influenced by both informal and formal sources, as highlighted in the literature [85]. Hence, health authorities are advised to disseminate information through multiple channels, including social media [12, 84]. However, health authorities’ rapid message dissemination on social media suggests inconsistency, which could be a source of confusion for the public [37, 65]. According to Jardine et al. [37], receiving information from multiple sources is only better if the additional information sources contribute to improved informed decision making, less confusion and strengthened credibility. The review findings also emphasise that misinformation is a significant problem across social media sites related to pandemic risk communication. This sometimes results in confusion and even panic for the public [23, 26, 27]. A major challenge for health authorities during pandemics has been to combat the infodemic, the overabundance of information, by identifying and addressing misinformation, rumours, and the contradictory information the public receives from multiple informal sources on social media sites [12, 23, 27, 86]. At the same time, there is a lack of consistency in their own messages that are disseminated rapidly and in parallel on multiple communication channels. As highlighted by Ratzan et al. [20], global health communication demands a communication strategy, particularly on social media, that ensures consistent and congruent messages to the public. Challenges for health authorities across nations are to create engagement in social media, provide the public with webpages that are accessible, updated in a timely way, and tailored towards varying reading levels and with a readable layout to ensure or avoid risk amplification of their messages delivered through mass media.

Social trust is “the willingness to rely on those who have the responsibility for making decisions and taking actions related to the management of technology, the environment, medicine, or other realms of public health and safety” [[87], p. 354]. Social trust in health authorities is value-based and relies on judgements of similarities in intentions and values [88]. Corresponding to the existing literature, our review findings suggest that people tend to have social trust in health care professionals as spokespersons and information sources in public health emergencies [12]. Furthermore, the current review indicates that trust in health authorities as an information source is not a static phenomenon; rather, it is highly dynamic and related to the public perception of health authorities’ crisis management [50, 51]. The evidence in the present review cannot conclude why trust in governmental sources varies. In a critical review of the literature, Siegrist [88] found that social trust varies by hazard and respondent groups, and little is known about the heuristics people rely on when evaluating hazards [88]. Factors related to why people choose informal sources of information may relate both to the respondents (e.g., attitudes, education) and to the government’s crisis management or trust in a particular spokesperson. According to Siegrist [88], there is a need to separate trust and confidence. The type of trust described in the present review is related to the concept of “confidence”. Confidence is affected by past experiences of emerging evidence, indicating that events will occur as expected. While the level of confidence in crisis management may affect risk perception and behavioural outcomes, a total loss of social trust is rarely found in response to governmental risk communication (or the lack of such) [88].

The evidence of the current review also corresponds with the literature indicating that health authorities may need to collaborate with trusted spokespersons and tailor communication methods to reach out to immigrant and ethnic populations [12, 13]. Finally, this study corresponds to the knowledge gained from research conducted on epidemics (notably Ebola, Zika and yellow fever), suggesting that efficient modes of communication vary by location and population. What works best for one population might work poorly for another [28]. Hence, health authorities’ health risk communication needs to be adaptive to the pandemic situation and the multiple receivers in the field, be responsive to its development over time, and choose credible spokespersons in media to maintain public confidence in health authorities as a source of health risk information.

This review identified topics related to the content of the message, such as the use of narratives, emotional valence, framing and jargon, multimodal messages and providing people with messages that support self-efficacy. Our findings correspond with previous literature findings that health authorities should use balanced and factual information based on science and evidence and should motivate self-efficacy and promote specific actions people can realistically take to protect their health [12, 82, 89]. However, while the WHO guidelines propose that messages should be jargon free and not explained in technical terms to reach the general public [12], we identified one study [77] implying that this may not apply to the COVID-19 context. Existing reviews of the literature have emphasised the impact of education and health literacy on protective behaviour during pandemic and infectious disease outbreaks [3‐5] and concluded that individual differences in both attitudes and knowledge about risks suggest that there may be no “one size fits all” approach to risk communication [90]. The use of jargon may be a question of what works for whom.

The evidence to support the choice of message attributes (e.g., narratives, non-narratives, emotional components) in the current review is inconclusive. Although the findings in the present review suggest that people learn more from non-narrative messages, Downs [91] argues that the use of narratives in science communication may help to communicate complex information because it can provide the receiver with a context, capture attention and improve the understanding and processing of the information. According to Balog-Way et al. [14], there is no single dominant formula for risk communication. In a review of the risk communication literature, Balog-Way and colleagues conclude that the effect of single message attributes is typically contingent, indirect and cumulative, and effective risk communication requires a multifaceted approach. Furthermore, it is a matter of what works for whom. The evidence related to jargon and narratives in the current review reflects the complexity health authorities face when designing effective messages for various audiences.