We selected 24 studies dealing with health care institution interventions on crisis communication (Table 1), of which 12 (50%) were carried out in the United States, 3 (13%) were carried out in China, and 3 (13%) were carried out in Canada. Engagement was the most represented primary evaluated aspect (17/24, 71%), followed by message framing (4/24, 17%) and campaign effectiveness (3/24, 13%). In this category, the Kmet evaluation score resulted in an average of 0.73 points, with a correlation coefficient of 0.85.

Overall, in this group, studies focused for the most part on the analysis of collections of posts and inquiries (21/24, 88%), whereas the rest (3/24, 13%) focused on the analysis of people as participants. Consequently, the design of these studies included a high percentage of content analyses (14/24, 58%), observational studies (5/24, 21%), and network analyses (4/24, 17%). As for the media channels analyzed in this group, 83% (20/24) of the studies focused on only 1 communication medium, whereas 17% (4/24) dealt with multiple media. Finally, half of the studies (12/24, 50%) referred to a specific communication theory.

In 71% (17/24) of the studies, the primary aim was to assess the success of engagement techniques in web-based communication both on websites and social media.

First, what emerged in the study by MacKay et al [29] was that public health agencies and news media should use guiding principles consistently to increase positive sentiment and build trust among followers.

Another study by Alamoodi et al [12] was focused on TikTok with the aim of determining the factors and influencing mechanisms related to citizen engagement with the TikTok account of the National Health Commission of China during the COVID-19 pandemic. The result of this was that shorter videos are preferred to longer ones, and a positive emotion is better suited than a negative one. Similarly, a study carried out in China [30], this time on the platform Sina Weibo, concluded that posts displaying positive emotions can include more videos or pictures, whereas plain text is more suitable for posts with negative emotions.

The studies by Dimanlig-Cruz et al [31], Hagen et al [32], Lauran et al [33], and Slavik et al [34] dealt with targeting specific population groups. The first and most generalizable study is the one by Lauran et al [33], who stated that deciding on 1 actor and 1 (homogeneous) stakeholder group is not the right strategy. What is advisable is to take the perspectives of the multiple stakeholders into account (and find opinion leaders within those groups) when deciding on the communication strategies to use and to refrain from introducing a new, unrelated issue into the discussion before the original issue is handled. Similarly, according to Hagen et al [32], public health organizations can benefit from understanding the types of content that are transmitted through specific social media platforms and identifying key participants who are authoritative, popular, and connected with disparate communities to efficiently communicate with the public. As for the study by Dimanlig-Cruz et al [31], given the high number of youths on Instagram and YouTube, public health officials may want to consider targeting youths on these sites; similarly, Slavik et al [34] tried to assess tweeting practices during public health crises to improve risk communication and maximize engagement. What emerged was the need for public health agencies to monitor Twitter analytics to understand their audience and leverage whatever Twitter engagement strategies help maximize the shares of their communications.

Creating a community was also a very important point in these studies as coordinating communication efforts by frequently interacting with other organizations to boost one’s network position can facilitate further communication efforts [25]. In particular, what emerged is that organizations should consider retweeting content from health information sources with a high number of Twitter followers if they want to build up their own follower base and that health agencies should coordinate their communication efforts by frequently interacting with each other. This will boost their network position and facilitate further communication efforts. Another key strategy for public health agencies might be to develop a community of trusted users with their own significant base of followers who will pass on tweets from health authorities [35]. In the event of an outbreak, prompt responses from the authorities can be vital in crisis management, as explained by Vos et al [36], who stated that public health officials may want to emphasize the severity of an emerging infectious disease. Efficacy information is an important message element in encouraging an effective response. Precise guidelines have also been proposed in the event of a specific outbreak (COVID-19) or with regard to specific communication channels (Twitter). In the case of COVID-19, Lwin et al [37] focused on the dissemination of posts regarding the COVID-19 pandemic, and their findings showed that the public liked and shared the most in the preoutbreak phase and engaged with posts much less during the outbreak, as well as the fact that the public liked the most the posts that encouraged self-efficacy. Furthermore, in an uncertain environment, public agencies can reach the public—and increase message sharing—with a wide range of practical information regarding the health impacts of COVID-19, protective action measures, and the progress of the pandemic itself. At the same time, some tactics useful in other disasters (such as sentence styles that use exclamatory and interrogative punctuation) were counterproductive during the COVID-19 pandemic [38]. As for studies on Twitter, according to Tang et al [39], the main takeaway was that public health agencies should continue to use Twitter to disseminate information, promote action, and build communities, especially by targeting specific population groups. Similarly, Young et al [40] focused on chats, concluding that this means of communication was effective at answering questions about disease, creating a forum for targeted criticism, and promoting conversation among participants. Government accounts could also take full advantage of social media functions, especially mentions, hashtags, and the number of original posts, and add pictures and text length appropriately to increase interactions with the public and improve the level of engagement [41].

As for the strategies that proposed taking more technical aspects into account, the studies by Guidry et al [42], Renshaw et al [43], and Vos et al [44] offered interesting insights. First, according to Renshaw et al [43], focusing on useful content rather than gimmicks to go viral would be helpful in the long run. Having meaningful content such as relevant images embedded in posts might be crucial for success and, according to Guidry et al [42] and Vos et al [44], organizations should create messages that illustrate information visually and try to include threat and efficacy information in messages. They should also engage social media audiences before public health crises emerge.

This group comprises 17% (4/24) of the studies, all aimed at assessing the way in which a certain message is framed to make communication as effective as possible. To begin with, Pascual-Ferrá et al [45] concluded that the integration of social network analysis is recommended as a best practice in crisis communication on social media. Ort and Fahr [46] conducted a study focused on the interaction between perception of threat and self-efficacy in a crisis situation. Even in this case, health messages promoting people’s self-efficacy perceptions may be preferable to threatening messages. Another study aimed at a specific part of the population—public health officials—carried out by Jang and Baek [47] in South Korea concluded that lower perceived credibility of information from public health officials was associated with a greater tendency to use web-based news, interpersonal networks, and social media. The last study [48] focused on how message construction, style, content, and the textual content of tweets and embedded images affected message retransmission over the course of the pandemic. In particular, the focus was on tweet effectiveness, this time studying hashtags and concluding that public health–oriented hashtag campaigns may help engage individuals to help them feel part of a larger collective body and participate locally by contributing information about their local context.

The last group in this category comprised 8% (2/24) of the studies, aimed at assessing the success of certain campaigns. First, Duong et al [49] investigated the content and format of physical distancing messages directed at Vietnamese youths during the COVID-19 pandemic, concluding that perceived norms and self-efficacy did not fully account for the association between interpersonal communication and behavioral intentions. Second, Harris-Sagaribay et al [50] summarized the lessons learned through an observational retrospective study when it comes to improving information dissemination during a health care crisis. Other than content, the effectiveness of website-based communication was measured through ease of navigation and trust in the information provided by the website.

This section comprises 45% (35/78) of the articles, all concerning studies that deal with the themes of health promotion and disease prevention (Table 2). Of these 35 studies, 19 (54%) were carried out in the United States, 5 (14%) were carried out in Australia, and 4 (11%) were carried out in the United Kingdom. Most studies had to do with engagement (19/35, 54%), 23% (8/35) had to do with campaign effectiveness strategies and, finally, 23% (8/35) had to do with message framing. The Kmet evaluation resulted in an average of 0.77 points, with a correlation coefficient of 0.78.

Studies on these 2 topics were mostly based on the analysis of collections of posts (21/35, 60%) rather than being studies on groups of people (14/35, 40%). The study designs were also diversified in this group, with the most represented being content analyses (16/35, 46%) and observational studies (9/35, 26%). In this group, most studies (27/35, 77%) were conducted on only 1 communication medium, whereas the remaining 23% (8/35) dealt with multiple platforms. As for communication theories, 71% (25/35) of the studies referred to one or more specific communication theories, and 29% (10/35) did not.

In this group, 54% (19/35) of the studies analyzed the engagement of campaigns or interventions aimed at promoting disease prevention or health promotion measures.

One of the main takeaways from these studies is that the message format affects the level of users’ web-based commitment to health organizations via social media [51]. Other studies (2/35, 6%) were tied to the assessment of the effectiveness of different types of posts on Facebook. A study analyzed the extent to which a post can resonate with an indecisive parent when it comes to the human papillomavirus vaccine [52], concluding that designing factual posts so that they include an emotional dimension increased the engagement with these posts, sponsored content can generate more negative comments than organic content, and all people should be addressed in an accommodating manner regardless of their tone. Although Syred et al [53] stated that moderation can help maintain the discussion quality and generate new interest and discussion on a certain topic, Loft et al [52], by contrast, focused more on the technical aspects of Facebook posts by stating that photos with short comments were the most effective in engaging information consumers and greater use of this post type could encourage greater audience engagement. At the same time, professional videos may not be as effective as a mechanism for active audience engagement on social media platforms. Zhang and Zhou [54] analyzed message efficacy, this time with particular attention to fear, and proposed a strategy where the inclusion of more emotional cues such as pictures is emphasized to arouse fear to motivate information dissemination on social media. In the social media context (in this case, Twitter and anonymous web-based discussion platforms), the relative importance of having clear informational content sent from organizations would be much greater for generating highly viewed and shared cancer prevention messages [55].

The studies by Kite et al [56], Lister et al [57], Parackal et al [58], Rabarison et al [59], and Reuter et al [60] measured engagement in and of itself. In particular, the study by Rabarison et al [59] focused on a specific aspect of the social media in question: Twitter chats. Specifically, chats of this kind should be used as an engagement tool with the audience by sharing messages and responding to questions from the public. Focusing on Twitter, Instagram, and Facebook, Reuter et al [60] concluded that engagement with a health message on social media does not indicate user engagement on a website and, therefore, it is recommended that both metrics be taken into account when designing health promotion strategies. It was also suggested to combine organic and advertising messages in health promotion campaigns. More specifically, with regard to Facebook, communication effectiveness could be enhanced using a two-way communication format, which enables the promoter to respond to negative comments [58]. Finally, according to the studies by Kite et al [56] and Lister et al [57], effective engagement through Facebook requires both maximizing the reach of posts through paid boosts and delivering content that users want to engage with and share to capitalize on word-of-mouth marketing.

In total, 6% (2/35) of the studies suggested social media strategies with the aim of improving engagement by comparing the work of institutional social media with that of lifestyle influencers [61] or by investigating the way in which network structures explain retweeting behaviors [62]. More specifically, Klassen et al [63] and Kim et al [62] stated, respectively, that health promotion organizations should try to build relationships with their users in a similar fashion to lifestyle brands and that influential people should be identified and targeted as their messages are more likely to be disseminated.

Regarding more technical aspects, the study by Guidry et al [64] states that, in the case of crisis communication, public health organizations should be present on all major social media platforms, but Instagram may yield the greatest return and user engagement. The study by Chung [65] was aimed at examining whether dialogic messages induced greater risk-preventive behavioral intentions than monologic messages, reaching the conclusion that frequent posting of tweets with images and graphs instead of videos and hyperlinks is beneficial. Similarly, according to Barklamb et al [61], strategies that were associated with higher engagement included the use of hashtags and announcements compared with not prompting engagement strategies. However, imagery should be carefully used as it appeared to be a powerful tool for attracting attention and briefly engaging users (ie, increasing likes) as well as increasing message transmission (ie, increasing shares). However, the use of images with information about illness consequences and control or with messages conveying negative affect could mute responses [66]. In particular, communication effectiveness could also be enhanced by designing factual posts so that they include an emotional dimension that could increase engagement [67]. Moreover, according to Strekalova and Krieger [67], sponsored content can generate more negative comments than organic content, and all people should be addressed in an accommodating manner regardless of their tone. Finally, users were more likely to click, share, comment, or like the content of posts that had photos. Branded, visual content was more effective in facilitating engagement [68].

Finally, 3% (1/35) of the studies focused on the effectiveness of communication theories, in particular that of the health belief model. Analyzing the Instagram accounts of the health departments of the Gulf Cooperation Council, it was found that the health belief model should be included more in internet-based communication [69].

Considerably less studies (8/35, 23%) dealt with the topic of message framing in this category. The first study in this category was by Dockter et al [24], stating that content should be transmitted or retransmitted by well-known, credible sources. On a more specific note relating to content engagement, Yoo et al [70] recommended the use of content-oriented social media when trying to influence risk perception during campaigns, with particular attention to posts with photos as users were more likely to click, share, comment, or like this type of content. Borah and Xiao [28] and Cho et al [71] investigated the effect of health message framing and the moderating effects of social endorsement and source type on credibility perceptions of posts, resulting in a superiority of gain-framed messages to reach a positive campaign outcome. Other studies (4/35, 11%) were tied to the assessment of the effectiveness of different types of posts on Facebook when it comes to engagement [72]. In particular, communication effectiveness could be enhanced using a two-way communication format, which enables the promoter to respond to negative comments. Nobles et al [73] examined the demographic profile in photos concerning HIV prevention and diagnosis, underlining a disparity in the representation of minorities and marginalized communities. Another study by Chung and Lim [74] focused on a long-running campaign on National Breast Cancer Awareness Month and concluded with 2 observations regarding the efficacy of frequent posting and the positive impact of photos and images instead of videos and hyperlinks. Finally, Whitten et al [75] addressed the presence of information targeting low-literacy, racially diverse, non–English-speaking, and age-diverse audiences on breast cancer websites. The results were three-fold: if content were tagged according to ethnicity or language, then this would allow users to browse websites according to the information that is most personally relevant; it would be beneficial for websites containing lower-literacy material to avoid statistical data; and storytelling evidence has demonstrated the ability to serve as a greater motivator for healthy behaviors.

This last group included 23% (8/35) of the studies, which focused on the assessment of the success of a particular campaign. Starting again from a more general framework, the study by Harris et al [76] stated that there are 4 qualities that are key to influencing trust and the subsequent decision to act on the advice given. These are information quality, personalization, perceived impartiality, and design credibility. Delving deeply into the issue of trust, the studies described in this section proposed different strategies to maximize trust from the web users. Social media can also be used effectively in social marketing campaigns and is an essential tool in the promotional mix when targeting young people. According to Potente et al [27], entertaining peer-to-peer messages can be used to engage youths with an important health message for skin cancer prevention. By contrast, Allen et al [77], on the promotion of the human papillomavirus vaccine, recorded no statistically significant change in the intent to be vaccinated in the next 6 or 12 months after the campaign among those who had not yet started or completed vaccination. Focusing on more technical aspects, Frisch et al [78] stated that websites designed for health education should include visual presentations of information such as pictures, charts, or graphs. Similarly, Yoo et al [79] were concerned with more technical aspects of Twitter communication, especially when developing a Twitter campaign. The results included the need to consider incorporating features such as hyperlinks to related websites or live chats with health care providers as well as the creation of tailored messages or edutainment, which may also be considered to engage people in the process of information selection and transmission. Moving forward from just design structure, the way content is presented is also a great source of studies. In this case, Nguyen et al [80] offered useful insights, concluding that mode tailoring may be a tool to reduce or prevent the information overload that may occur when too much information is placed on a nontailored web page at one time. Other than content, the effectiveness of website-based communication is measured through ease of navigation and trust in the information provided by the website. Perrault and Silk [81] used social cognitive theory and media richness theory to prove that the exposure to videos was responsible for the increased engagement in risk-reduction behaviors. Another communication theory is the transtheoretical model. This was used in the study by Pirzadeh et al [82], who stated that the transtheoretical model was the most effective education strategy when it comes to prompting behavior change.

In the category of general health (Table 3), 17% (13/78) of the studies were included. Of these 13 studies, 9 (69%) were carried out in the United States, whereas 3 (23%) were carried out in the United Kingdom. Engagement was, once again, the most represented topic (9/13, 69%), followed by message framing (4/13, 31%). The same proportion holds true when it comes to studies on collections of posts (9/13, 69%) versus studies on human samples (4/13, 31%). The study designs were not overly diversified in this group as 46% (6/13) were observational studies and 38% (5/13) were content analyses. In this group, all studies (13/13, 100%) were conducted on a single communication medium, but only 31% (4/13) were connected with a specific communication theory (the remaining 9/13, 69% were not). The Kmet evaluation resulted in an average of 0.76 points, with a correlation coefficient of 0.82.

In this group, 15% (2/13) of the studies [83,84] focused on the content of health organizations’ Twitter profiles, concluding that the use of hashtags, URLs, visual cues, and user mentions was positively associated with retweets. Another study focusing on retweets and on the type of message brought on by health organizations is the one by Meng et al [85]. They defined a series of recommendations on the matter (ie, designing efficacious information is the key to increasing the aggregated number of retweets, crafting information that can raise risk perception is important to increase the diffusion chain through person-to-person transmission, and tweets that induce negative emotions could be more effective in catching users’ attention and expanding sharing of the information). A similar study by Bhattacharya et al [86], this time on Facebook, also stated that posts containing media or links and expressing positive sentiments correlated with higher or longer engagement. Facebook was also the topic of research of Kite et al [87], who concluded that content providers must encourage engagement and adapt to the Facebook algorithm to maximize message exposure while also ensuring that the content is of high quality. Language also plays an important role in the effectiveness of a post, as explained by Campbell and Rudan [88], who claimed that adjusting the language and presentation can be of more appeal to popular culture. Kim and Kim [89], by contrast, centered their study on the Instagram presence of the Centers for Disease Control and Prevention (CDC), stating that the message design should be different depending on whether the aim is to increase the number of likes and comments or induce a more positive response from the public. The dissemination of health information is also a topical area of research, in particular by Park et al [90], who provided guidelines such as retweeting content from health information sources with a high number of Twitter followers to build up an organization’s follower base. The study by Pang et al [91] was aimed at creating a design for a consumer health website by supporting different health-seeking behaviors. In particular, different types of information-seeking behavior should be supported as access to a dynamic information scope is critical for health information seeking.

This group comprised 4 studies: the studies by Hung and Stones [92], Lazard and Mackert [93], Shim and Jo [94], and Sillence et al [95].

Concerning design and website presentation, Lazard and Mackert [93] stated that high design complexity is often associated with a better perceived comprehensibility, a greater perceived usefulness, a greater message quality, and higher perceptions of visual informativeness. Other than content, the effectiveness of website-based communication is measured through ease of navigation and trust in the information provided by the website. Sillence et al [95] identified the key factors influencing UK and US citizens’ trust and intention to act on advice found on health websites (ie, credibility and impartiality). Moving forward from just design structure, the way content is presented is also a great source of studies. In this case, Hung and Stones [92] offered useful insights, stating that, among other guidelines, to appeal to the public, general terms should substitute professional terms and simplified text-based content should be used. Delving deeply into the issue of trust, Shim and Jo [94] applied the information systems success model, assessing that service quality had a significant association with user satisfaction and that its impact on perceived benefits occurred indirectly to user satisfaction and intention, thus maximizing trust from the web users.