Infodemics and health misinformation: a systematic review of reviews

1During crises, such as infectious disease outbreaks and disasters, the overproduction of data from multiple sources, the quality of the information and the speed at which new information is disseminated create social and health-related impacts.– This phenomenon, called an infodemic, involves a torrent of online information containing either false and misleading information or accurate content.

2To tackle the production of misinformation (that is, false or inaccurate information deliberately intended to deceive) and disinformation (that is, deliberately misleading or biased information; manipulated narrative or facts; and propaganda) during recent pandemics or health emergencies, research on infodemics has increased. This research focuses on understanding the general effect of infodemics on society, dissemination patterns and delineating appropriate countermeasures policies.– Several studies are analysing the effects of infodemics and misinformation and how societal behaviours are affected by that information.– Particularly, evaluating infodemic-related concepts, such as impact on humans’ lives and communities, frequency and most common sources to widespread unreliable data, using comprehensive and evidence-based criteria, has gained more attention. Therefore, assessing how infodemics and health misinformation affect public health and identifying the availability and quality of evidence-based infodemic characteristics is timely and pertinent to inform appropriate management of its potential harms and support the development of monitoring guidelines.

3We conducted a systematic review of reviews to collate, compare and summarize the evidence from the recent infodemics. To improve and guide the infodemic management, we designed our study to identify current opportunities, knowledge gaps and challenges in addressing the negative effects of the dissemination of health misinformation on public health.

4We registered our systematic review in PROSPERO (CRD42021276755). The review adheres to the Preferred Reporting Items for Systematic reviews and Meta-Analyses 2020 and the Quality of Reporting of Meta-analyses statement.,

5We explored the following research questions: (i) To what extent are evidence-based studies addressing peculiarities and singularities associated with infodemics? (ii) What type of information on the topic of infodemics are published in systematic reviews? (iii) What main challenges, opportunities and recommendations addressing infodemics did systematic review authors highlight? and (iv) What is the methodological and reporting quality of published systematic reviews conveying research questions related to infodemic?

6We used a published definition of systematic reviews and included a systematic review or mini-reviews if: (i) the search strategy was conducted at least in two databases; (ii) the study had at least two authors; and (iii) the study comprehensively presented a methods section or description of inclusion and exclusion criteria. We only included systematic reviews that directly analysed the available evidence correlated to infodemics, misinformation, disinformation, health communication, information overload and fake news (defined as: purposefully crafted, sensational, emotionally charged, misleading or totally fabricated information that mimics the form of mainstream news). We excluded preprints, unpublished data and narrative or literature reviews.

7With an information specialist, we designed the search strategy using medical subject headings and specific keywords (Box 1). We had no restriction on publication date or languages. We searched five databases (MEDLINE®, Embase®, Cochrane Library of Systematic Reviews, Scopus and Epistemonikos), explored the reference lists of the included studies and searched for potential review protocols registered on PROSPERO. We first conducted the search on 4 November 2021 and we re-ran the search on 6 May 2022.

8#1 Communication OR consumer health information OR information dissemination OR health literacy

9#2 (infodemic* OR misinformation OR disinformation OR disinformation OR information dissemination OR information sharing* OR information overload) OR (fake new* OR influencer* OR conspirac* OR hate* OR infoxication) OR ((viral AND (news OR social media OR media)) OR (consumer health information OR health literacy OR health information literacy)

10#3 - #1 OR #2

11#4 Systematic review as topic OR PT Systematic review OR AB “systematic review” OR TI “systematic review”

12#5 - #3 AND #5

13After removing duplicates, two authors independently screened title, abstract and full-text of articles and included eligible articles for evaluation. An independent third author resolved any disagreements. We performed the screening process in Covidence (Covidence, Melbourne, Australia).

14Two independent researchers extracted the general characteristics of each study and classified them into six major categories: (i) reviews evaluating negative effects of misinformation; (ii) reviews assessing the sources of health misinformation and the most used platforms; (iii) reviews evaluating the proportion of health-related misinformation on social media; (iv) reviews evaluating the beneficial features of social media use; (v) reviews associated with corrective measures against health misinformation; and (vi) reviews evaluating characteristics associated with studies’ quality. We clustered systematic reviews based on similar properties associated with the stated objective and the reported outcomes. Although infodemics were primarily defined as the overabundance of information, usually with a negative connotation, we decided to report data from systematic reviews which also described the potential beneficial effects of the massive circulation of information and knowledge during health emergencies. We summarized challenges and opportunities associated with infodemics and misinformation. A third author verified the retrieved data and another author resolved any inter-reviewer disagreement.

15Two authors independently appraised the quality of included systematic reviews using the AMSTAR 2 tool, containing 16 domains. We rated each categorical domain using the online platform and obtained an overall score of critical and non-critical domains. Inter-rater discrepancies were resolved through discussion. We calculated inter-rater reliability with a Cohen’s κ and we classified reliability as adequate if κ > 0.85.

16We synthesized the characteristics of included reviews, reporting their primary outcomes categorized by the similarity of the review question or results. Additionally, we created summary tables showing current evidence and knowledge gaps. We rated the certainty of the evidence through an adapted version of the Grading of Recommendations Assessment, Development and Evaluation approach for the defined primary outcomes.,

17We identified 9008 records and after removing 443 duplicates, we screened 8565 studies of which 111 were eligible for full-text assessment. Of these, we excluded 80 studies (available in the data repository). We included 31 systematic reviews, of which 17 studies were published between 2018 and 2022,– three awaiting classification (we were unable to retrieve full text during our review)– and 11 ongoing reviews (Fig. 1).– Inter-rater reliability was high (κ = 0.9867).

18Out of 17 published systematic reviews, 14 were published after the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) outbreak.– The published reviews included 1034 primary studies covering 12 infectious diseases and three major topics (vaccination hesitancy, disaster communication and disease outbreaks) related to infodemics, misinformation, disinformation, fake news or any other variation of these terms (Table 1). The included reviews covered 19 official scientific databases.

19The main outcomes, categorized in six themes, are summarized in Box 2 and by study in Table 2. Below we describe the outcomes, by theme, in more detail.

20Reduce patients’ willingness to vaccinate

21Obstruct measures to contain disease outbreaks

22Instigate the physical interruption of access to health care

23Amplify and promote discord to enhance political crisis

24Increase social fear, panic, stress and mental disorders

25Enhance misallocation of resources

26Weak and slow countermeasures interventions

27Exacerbate poor quality content creation

28Social media platforms are associated as a potential source of promotion of anecdotal evidence, rumours, fake news and general misinformation

29Twitter, Facebook, Instagram and blogs play an important role in spreading rumours and speculating on health-related content during pandemics

30Digital influencers or well-positioned individuals acts as distractors or judges in social networks

31Closed communication within online communities can be used to propagate and reverberate unreliable health information

32Misinformation can be derived from poor quality scientific knowledge

33Health misinformation in posts on social media is common (1–51% on posts associated with vaccine, 0.2–28.8% on posts associated with COVID-19 and 4–60% for pandemics)

34Approximately 20–30% of the YouTube videos about emerging infectious diseases contain inaccurate or misleading information

35Social media platforms and traditional media might be useful during crisis communication and during emerging infectious disease pandemics, regardless of the geographical settings

36Using social media properly, infosurveillance can be highly functional in tracking disease outbreaks

37Social media can improve knowledge acquisition, awareness, compliance and positive behaviour towards adherence to clinical infection protocols and behaviours

38Correcting misinformation delivered by health professionals is harder than information delivered by health agencies

39Misinformation corrected by experts is more effective than when corrected by non-experts

40The effectiveness of correcting misinformation using text or images is similar

41Use of refutational messages, directing the user to evidenced-based information platforms, creation of legislative councils to battle fake news and increase health literacy are shown to be effective countermeasures

42Most studies published during an infodemic are of low methodological quality

43There is a substantial overlap of published studies addressing the same research questions during an infodemic

44Ten systematic reviews presented evidence of the negative effects associated with the dissemination of misinformation during an infodemic.,,,,,– Several of the consequences were linked to altering people’s attitude towards the situation: (i) distorting the interpretation of scientific evidence; (ii) opinion polarization and echo chamber effects (that is, the formation of groups of like-minded users framing and reinforcing a shared narrative); (iii) offering non-specialists’ opinions to counter accurate information; (iv) promoting fear and panic; (v) increasing mental and physical fatigue of population; and (vi) decreasing credibility of circulating information on different platforms during unforeseen circumstances. Infodemics could also decrease trust in governments and public health systems as well as in the government’s response and accuracy of the official health messaging. Other societal consequences could be amplifying and promoting discord to create a hostile political environment, increasing violence against ethnic and minority groups and affecting the global economy. Within the health system, infodemics could lead to (i) misallocation of resources and increasing stress among medical providers; (ii) decreased access to health care; (iii) increased vaccine hesitancy and conspiracy beliefs; (iv) increased illegal promotion of the sale of controlled substances; and (v) delayed delivery of high-quality care and proper treatment to patients, which could further have a negative effect on public health-care systems.

45Six reviews reported potential links between misinformation and sources of misinformation.,,,,, All reviews emphasized that mass media can propagate poor-quality health information during public health emergencies, particularly through social media. Authors of the systematic reviews highlighted that health misinformation can be quickly propagated through media posts and videos, usually circulated among closed online groups, significantly influencing individuals with low health literacy and elderly patients.,, Similarly, two reviews found that social media networks were often identified as a source of illegal or inappropriate promotion of health misinformation, including the sale of controlled substances., One review tracked the main sources of health-related misinformation spreading on social media during infectious disease outbreaks worldwide, noting that the primary sources of misinformation are groups against immunization, online communication groups (such as WhatsApp groups and Facebook communities) and pharmaceutical and marketing industries, who could favour conspiracy theories.

46Four reviews evaluated the proportion of health misinformation on different social media platforms.,,, In a meta-analysis, the proportion ranged from 0.2% (413/212 846) to 28.8% (194/673) of posts. Similarly, a review identified that the proportion of the literature containing health misinformation is dependent on the topic, which were articulated in six categories (vaccines had the highest proportion, 32%; 22/69, whereas medical treatments had the lowest, 7%; 5/69). One review identified 47 mechanisms driving misinformation spread. The authors also argued that misconceptions about vaccine administration in general and about infectious diseases (45 studies out of 57) and chronic noncommunicable diseases (8 studies out of 57) are highly prevalent on social media; however, the review lacks comprehensive presentation of epidemiologically-relevant data. Additionally, authors of a review estimated around 20% to 30% of YouTube videos about emerging infectious diseases contained inaccurate or misleading information.

47Although infodemics are often associated with negative impacts, eight reviews reported positive outcomes related to infodemics on social media during a pandemic.,,,– Social media can be used for crisis communication and management during emerging infectious disease pandemics, regardless of geographical location of the recipient of information. Furthermore, reviews found that dissemination of information on several social media platforms had significantly improved knowledge awareness and compliance to health recommendations among users.,, Notably, some authors also stressed the fact that social media created a positive health-related behaviour among general users compared with classic dissemination models.,, In particular, these platforms can be used for education, suggesting that social media could outrank traditional communication channels., Also, content created by professionals and published on social networks, especially YouTube, might serve the community with online health-related content for self-care or health-care training. Three reviews evaluated the effectiveness of social media platforms as vehicles for information dissemination during health-related disasters, including pandemics, as well as a tool to promote vaccination awareness. The reviews evidenced the effectiveness of social media platforms as an information surveillance approach, as these platforms could provide complementary knowledge by assessing online trends of disease occurrences, collecting and processing data obtained from digital communication platforms.,, Twitter and Facebook emerged as beneficial tools for crisis communication for government agencies, implementing partners, first responders and for the public to exchange information, create situational awareness, decrease rumours and facilitate care provision. Furthermore, these authors also argued that social media is a viable platform to spread verified information and eliminate unfiltered and unreliable information through crowd-sourced, peer-based rumour control (that is, technologies that network users can collaboratively implement for more effective rumour control).,,

48Interestingly, one study suggested that the use of social media to mitigate misinformation associated with health-related data might result not only from the prioritization of strategies taken by governmental and health authorities, but also from the economy sector, which also includes the information technology market, the media and knowledge-based services. Also, citizens’ intention to spread misinformation by using real information would ultimately serve as a natural controlling system.

49One systematic review evaluated the knowledge levels and media sources of information about coronavirus disease 2019 (COVID-19) in African countries and found that 40% (4/10) of studies reported that the participants used social media as their source to acquire information about COVID-19. Likewise, traditional communication channels (such as television and radio stations), family members and places of worship were also used to receive information about the disease.

50Four reviews evaluated the impact and effectiveness of social media interventions created to correct health-related misinformation.,– In general, eliminating health-related misinformation delivered by family or colleagues is more challenging than eliminating misinformation from organizations. Furthermore, evidence shows that a greater corrective effect occurs when content experts correct misconceptions compared with non-experts. In addition, authors of three reviews suggested redirecting users to evidence-based or well-founded websites, besides computer-based algorithms for rumour correction, as countermeasures to limiting the circulation of unreliable information.– Early communication from health authorities and international health organizations plays an important role in providing misinformation mitigation.,–

51Three reviews reported results on methodological quality of included studies.,, Generally, studies related to SARS-CoV-2 and infodemics showed critical quality flaws. For example, 49% (138/280) of eligible studies critically appraised the quality of original records. In comparison, only 33.0% (29/88) of the studies reported the registration of a scientific protocol before the beginning of the study. Several systematic reviews did not consider in their final analysis and conclusion statements the limitations of each included study’s design.,–,,, One study concluded that the spread of misinformation has been frequently derived from poor-quality investigations. Lastly, a large number of editorials, commentaries, viewpoints and perspectives were published since the onset of the COVID-19 pandemic; these types of articles are fast-tracked publications not based on new experimental and analytical data.

52When appraised using the AMSTAR 2 critical domains, 16 reviews (94.1%) scored as having critically low quality across most major domains.,– Only one review showed a moderate risk of bias for most domains (Table 3). Meta-analysis was conducted in only two reviews, which used appropriate statistical methods and considered the potential impact of risk of bias in each of the primary studies., The overall quality of the evidence is shown in Table 4. All themes had low quality, except the proportion of health-related misinformation which had very low quality of evidence.

53We evaluated reported data on current opportunities and challenges associated with infodemics and misinformation that may impact society worldwide. A summary of the main opportunities for future research and challenges is presented in Box 3.

54Future investigations should be performed to provide different aspects of the impact and reliability of SARS-CoV-2-related or any other health emergency information.

55There is a need to balance the gold standard systematic reviews with faster pragmatic studies.

56Studies need to evaluate effective methods to precisely combat the determinants of health misinformation during pandemics and subsequent infodemics across different social media platforms.

57Novel investigations could focus on creating a basis to conduct future studies (especially randomized trials) comparing the use of social media interventions with traditional methods in the dissemination of clinical practice guidelines.

58Future studies should assess the potential of social media use on the recovery and preparation phases of emergency events.

59Researchers could analyse communication patterns between citizens and frontline workers in the public health context, which may be useful to design counter-misinformation campaigns and awareness interventions.

60A multidisciplinary specialist team could concentrate on the analysis of governmental and organizational interventions to control misinformation at the level of policies, regulatory mechanisms and communication strategies.

61Studies should address the impact of fake news on social media and its influence on mental health and overall health.

62Future studies should examine how social media users process the emerging infectious diseases-related information they receive.

63Focus should be given to how users evaluate the validity and accuracy of such information and how they decide whether they will share the information with their social media contacts.

64Further interdisciplinary research should be warranted to identify effective and tailored interventions to counter the spread of health-related misinformation online.

65Overlap of studies covering the same topic.

66Overall low quality of studies and the excessive and inordinate media attention given to these studies.

67Creation and use of reliable health-related information and scientific evidence considering real-time updates.

68Inadequate orientation of the population and medical providers into wrong pharmacological and non-pharmacological interventions.

69New trends in personal content creation are constantly emerging, such as TikTok, which represent new challenges for regulation.

70Further understanding the economic impact of misinformation, the difference in distribution of health misinformation in low- and high-income countries and the real impact of antivaccine activism groups.

71Decisive and pro-active actions are required from government authorities and social media developers to avoid the destruction of positive achievements that social media has already promoted.

72The difficulty of characterizing and evaluating the quality of the information on social media.

73In our study, most systematic reviews evaluating the social-, economic- and health-related repercussions of misinformation on social media noted a negative effect, either an increase in erroneous interpretation of scientific knowledge, opinion polarization, escalating fear and panic or decreased access to health care. Furthermore, studies reported that social media has been increasingly propagating poor-quality, health-related information during pandemics, humanitarian crises and health emergencies. Such spreading of unreliable evidence on health topics amplifies vaccine hesitancy and promotes unproven treatments. Moreover, reviews evidenced the low quality of studies during infodemics, mostly related to overlap between studies and minimal methodological criteria.

74The increased spread of health-related misinformation in social and traditional media and within specific communities during a health emergency is accelerated by easy access to online content, especially on smartphones. This increased access and the rapid spreading of health-related misinformation through social media networks could have a negative effect on mental health.–

75Although the number of studies evaluating variables associated with infodemics has risen, some variables still require further scientific exploration. For instance, some studies described the need for better methods of detecting health-related misinformation and disinformation, as the propagation methods are constantly evolving.– Different initiatives have been used by individuals to reveal untrustworthy content, including website review, lateral reading approaches (that is, verifying content while you read) and emotion check analysis.– However, no consensus exists on which method is more effective in battling unreliable content. Moreover, the techniques to build social media content that convey misinformation vary across different social media platforms and over time, even for the same platform.– This variation implies the need to employ various multilingual detection and eradication techniques, which should be frequently updated to keep up with misinformation patterns. Evidence-based studies could evaluate the effectiveness of different misinformation detection models by comparing performance metrics and prediction values.– Further priorities include recognizing methods to decrease the high-speed dissemination of misinformation and understanding the role social media plays in individuals’ real life after obtaining a certain content, information or knowledge from these platforms.

76Only one review recommended legal measures against the publication of false or manipulated health information. Indeed, the discussion of this topic in the literature is controversial and polemical and is limited by the diversity of national legislative processes. For several jurists, the criminalization of intentionally sharing health misinformation acknowledges the wrongful violation of the right to life and liberty.– Furthermore, proper attention must be paid to predatory journals that publish articles without minimum quality checks., For anti-criminalization supporters, creating policies controlling health misinformation and disinformation goes against freedom of speech and a free flow of information., Countermeasures not involving legal actions against health-related misinformation can be awareness campaigns for patients and health-care professionals, the creation and dissemination of easy-to-navigate platforms with evidence-based data, the improvement of health-related content in mass media by using high-quality scientific evidence, the increase of high-quality online health information and improved media literacy. Promoting and disseminating trustworthy health information is crucial for governments, health authorities, researchers and clinicians to outweigh false or misleading health information disseminated in social media. Another option is to use social media channels to counter the false or misleading information, which may require further studies to evaluate the best format for this outreach and which channels work best for different populations in different geographical and cultural settings.

77This review has some limitations. First, we did not search for grey literature because studies have suggested that the searching efficiency and replicability of searches depends on geographical location and users’ content profile. Second, we assessed only systematic reviews and may have overlooked helpful non-systematic reviews. Nevertheless, by incorporating reviews published in scientific journals indexed in relevant databases, we obtained a comprehensive snapshot of the literature and we summarized reported gaps and implications for future research. Third, overviews of systematic reviews by nature depend on other researchers regarding inclusion criteria and methods of synthesizing data or outcomes. Thus, our conclusions may have been affected by by the bias that any systematic review author is potentially affected by. We took steps to minimize this bias, through creating a research protocol, assessing records by two authors and evaluating the quality of the evidence. Fourth, the quality of most included reviews were rated critically low due to non-adherence to important methodological features, a known issue of systematic reviews., Therefore, we advocate that researchers comply with reporting and executing guidelines for systematic reviews, which increases the completeness of reporting and assists with transparency and reproducibility of the study. Likewise, journals’ editors and reviewers should put into practice endorsed reporting guidelines which, although commonly displayed at journals’ interfaces, are not systematically employed during the evaluation process. However, we considered the low quality of included reports when interpreting and discussing the results.

78Based on the available evidence, people are feeling mental, social, political and/or economic distress due to misleading and false health-related content on social media during pandemics, health emergencies and humanitarian crises. Although the literature exponentially increases during health emergencies, the quality of publications remains critically low. Future studies need improved study design and reporting. Local, national and international efforts should seek effective counteractive measures against the production of misinformative materials on social media. Future research should investigate the effectiveness and safety of computer-driven corrective and interventional measures against health misinformation, disinformation and fake news and tailor ways to share health-related content on social media platforms without distorted messaging.