Cause Exploration Prizes: Misinformation Gone Viral: Education As Pandemic Prevention

By Open Philanthropy @ 2022-07-12T07:38 (+8)

This essay was submitted to Open Philanthropy's Cause Exploration Prizes contest and posted to the Forum with the author's permission.

Importance

There is an increasing risk of deadly pandemics. With the risk of biological threats, zoonotic transmission, dangerous laboratory pathogens, and biological weapons, pandemics will become more probable and more severe (WHO, 2020). As the world becomes increasingly interconnected through travel and commerce, novel diseases can spread rapidly around the globe. Traditional approaches to public health will no longer be effective. At a critical point, pathogens may kill the entire host population, in what is referred to as a “transition to extinction” (Bar-Yam, 2018). 

Mortality and Morbidity

Besides existential risk, pandemics bring about a wide range of issues. COVID-19 has deepened the socioeconomic crisis, pushed 70-100 million people into extreme poverty, disrupted education, and negatively impacted the mental health of millions (WHO, 2020). In addition to the 15 million deaths recorded in the first two years of the pandemic, COVID-19 also caused significant disability, often becoming one of the top three most burdensome diseases in many countries. In the Netherlands, 1570 DALYs per 100,000 people are attributed to COVID-19, and 1980 DALYs per 100,000 people in Scotland – and this is almost certainly an underestimate, as new sequelae of COVID-19 are constantly being discovered (Wyper et al, 2022). The impending risk of pandemics illuminates the need for novel approaches to public health.

Financial cost

Countries have varied widely in their COVID spending. In the Netherlands, spending in the first 1.5 years was $82 billion euros, or $172 euros/DALY. The United States has spent $4.3 trillion dollars and the UK has spent £368bn (Appleby, 2022). The World Bank has provided over $200 billion in aid to developing countries for COVID relief.

Identifying the problem 

The success of any public health measure depends heavily on public trust and cooperation. Public trust may be the most critical measure in pandemic prevention. Higher government trust is significantly associated with adoption of health behaviors and prosocial behaviors such as prioritizing long-term benefits and altruism. Trust in government is associated with a government’s perceived ability as organized, clear in messages and knowledge, and fair (Han et al, 2021). 

In the most promising study on this subject, recently published in the journal Lancet, researchers analyzed the effects of pandemic prevention indices, which measure health security, and other factors including trust. Surprisingly, pandemic prevention indices were poor predictors of  infection rates. Instead, the level of public trust – trust in government, trust in community, and low corruption levels – had the highest associations with low infection rates and higher vaccine uptake.

Compare citizens in the United States, which ranks in the 25th percentile for trust in government and community trust, with Denmark, which ranks in the 75th percentile for both. If the United States could improve its trust in government to match Denmark’s, then infections would be reduced by 12.9% (for government trust) and 40.3% (for community trust) (COVID-19 National Preparedness Collaborators, 2022). Bolstering confidence and trust in public health and safety recommendations could be an integral piece of pandemic prevention. 

During the COVID-19 pandemic, there has been a significant decline in public trust. The first line of defense after an outbreak depends on the will of the public to adhere to recommended safety measures including mask wearing, quarantine, social distancing and vaccination. These individual actions require public trust in the government and public health recommendations. In November 2020, Pew Research polling found that 84% of US adults reported at least a fair amount of trust in scientists to act in the interests of the public. However, by late 2021, studies conducted by General Social Survey, Pew, and others have documented the declining trust in government institutions, including the Centers for Disease Control and Prevention (CDC), Food and Drug Administration (FDA), and the World Health Organization (WHO). 1 in 4 adults, for instance, reported that the pandemic may be planned by “powerful people.” (Cross, 2021). During a time when public trust is needed most, there has been a reported decline.

Identifying causes 

1. Motives in mass media

The erosion of public trust can be largely traced to problems with the media, leading to unfounded information being spread. Social media was predicted to produce 84.9% of misinformation surrounding COVID-19, and the internet as a whole was responsible for 90.5%. Several key actors including former President Trump were single handedly responsible for a significant portion of the misinformation (Al-Zaman, 2022). 

1.1 Social media algorithms

Algorithms on social media platforms are designed to maximize the user’s time spent on the platform, enticing users to fall into rabbit holes of misinformation. The algorithms may incentivize and accelerate the spread of misinformation. Social media systematically evaluates an individual’s susceptibility to falsified information and likelihood to share it. The most susceptible individuals are the elderly, young people, people active on social media, and people with less formal education. 

1.2 Political polarization

Certain media outlets have provided politically-biased instead of science-based reporting. Throughout the pandemic, Fox News, the leading cable news network in the United States, claimed hundreds of falsehoods, including that the pandemic was a hoax; that masks didn’t work; that Covid vaccines killed thousands of people; and that hydroxychloroquine and ivermectin cured Covid. A Kaiser Family Foundation poll found that Fox News viewers were three times more likely to believe Covid-19 misinformation compared to those who watched news elsewhere (KFF, October 2021). 

Politicization led to the ineffective public health measures in the US. Political extremism leads people to believe information confirming their biases. Exposure to politicians and digital media is associated with a higher degree of belief in conspiracies and misinformation surrounding the pandemic (De Coninck, 2021). Political ideology also predicted the perception of the risk of COVID-19. Conservatives were more likely to perceive less vulnerability to the virus, and imagine that the virus was less severe. They were more likely to believe that the media overstated the impacts of the pandemic. When asked to discern between true and false information related to COVID-19, conservatives had lower truth discernment. These results suggested to researchers that political ideology and perception of risks depends on the political lens of the media (Calvillo, 2020).

2. Psychology and misinformation 

Other research has pointed to psychological effects as the main driver in spreading misinformation. Social media is designed to provide validation from numbers of likes, even if the post is inaccurate or misleading. This leads to emotional, reactive thinking, instead of careful analysis of the credibility of sources and accuracy of information. The result is viral spreading of emotionally charged content that can mislead or sway people from the truth. 

With this lens, an effective intervention to fight misinformation is for the media platforms themselves to emphasize the importance of accuracy. Social media platforms could periodically ask users to rate the accuracy of certain headlines. A trial of this intervention found that it nearly tripled participants’ truth discernment surrounding COVID-19 headlines. To test the implications of this intervention, the researchers found that the participants with higher truth discernment were more likely to spread true rather than false information. On the other hand, the control group participants were equally as likely to share true and false information (Pennycook et al, 2020). The two aforementioned studies involved 1,600 participants matched with quotas to fit the demographics of the United States along the lines of age, ethnicity, gender and geographic region.

3. Ineffective risk communication

Another contributing factor is a lack of effective public information, or risk communication. Risk communication is “the exchange of real-time information, advice and opinions between experts and people facing threats to their health, economic or social well-being.” Risk communication is modeled through both objective and social constructionist lenses, as risk perception may be connected to sociocultural context. The public relies heavily on social media to inform risk perception, which makes an essential part of risk communication imparting on the public the ability to identify and manage misinformation. 

Since risk perception differs within a population, with consequences for how different groups respond to messages, communicators must take on a risk communication perspective. Ineffective risk communication along with heightened risk perception has led to antisocial behaviors such as hoarding personal protective equipment including masks (Abrams & Greenhawt, 2020). In the future, risk communication must avoid these issues.

Solutions 

Combating infodemics includes several steps. Facilitating accurate knowledge and translation is one major goal, since knowledge translation processes between public health policy recommendations and the media must mitigate the interpretation of “facts” by influencing factors including politics. Knowledge refinement, through filtering and fact checking, ensures that people have access to reliable and accurate information. Building electronic health (eHealth) literacy may ensure that the end user can consume all types of information and responsibly find trustworthy information. Increasing monitoring, infodemiology, infoveillance, and social listening is a research goal (Eysenbach, 2020). Different methods to achieve these goals are outlined below.

1. Supporting community-led risk communication

Effective risk communication and community engagement could break chains of transmission until vaccines and other biomedical tools are available. Social science research may provide tools for efforts such as involving community health workers to convey public health information through trusted communications channels (Sullivan & Lander, 2021). Utilizing socio-behavioral data, risk communication reduces the spread of pandemics by understanding people’s perceptions, attitudes and motivations to adapt to the recommended behaviors by public health experts. 

Utilizing this information, community-led, data-driven, collaborative, and local approaches could be catered towards the target community. These approaches are more likely to build trust and thus mitigate the effects of pandemics. Collaborative efforts should involve multi-stakeholder groups, institutions engaging communities, coordination groups, and multi-sectoral engagement (WHO, 2020).

2. Developing early alert systems

Increasing confidence in public health and improved defenses should focus on communicating effectively during disease outbreaks. Government leaders should provide the public with information in an outbreak early alert system. Certain countries in Southeast Asia have had effective models of a similar system which informs the public of respiratory diseases, many of which were mitigated with preventative measures such as screening, temperature checks and mask wearing. During COVID-19, early alert systems have been strained, defunded and politicized throughout the globe. Countries should recognize that real-time information sharing is a part of collective defense, and work to improve the efficacy of early alert systems. Pre-investing in security infrastructure and meaningful ways to break chains of transmission could be more cost-effective than the $5.7 trillion dollars provided in direct relief efforts. The issue is that in global defense spending, the security industrial complex leads to a lack of investment in mitigating pandemics (Disparte, 2022). In the security industrial complex, fear and “otherness” is used as a tool by politicians that leads to the prioritization of securitization and building military defense instead of focusing on real societal challenges including pandemic risk and education (Disparte & Dichter, 2018). Developing early alert systems will face a structural challenge and require political will.

3. Strengthening evidence-based communication

Governments and healthcare providers can also motivate vaccination and other safety measures through evidence-based communication to increase public trust. Instead of immediately correcting misinformation with accurate information, effective communicators should focus on communicating about the disease to make individuals perceive the risk and believe in taking effective action to prevent disease. 

Experts should use frequent, presumptive communication, motivational interviewing, use empathy, and tailor communication to the individual (Limaye & Sauer, 2021). Social media companies should censor misinformation and disinformation, redesign algorithms and direct users to COVID-19 resources. Scientists and physicians should engage with mass media by making public corrections strategically and respectfully (Gisondi et al, 2022). Social media interventions to combat misinformation were statistically most effective when misinformation was debunked by experts (Waltar, 2020). 

4. Researching infodemiology 

Infodemiology is the study of information and managing information. Infodemiology involves assessing sources, reading beyond headlines, identifying the author, checking the date and current context, checking personal biases, and double checking with fact-checkers (WHO, n.d.). 

Infodemiology may help predict public health events within a population, and should be explored further to inform public policy and health by studying the relationship of information supply and demand with population health. Information may be collected to analyze how people use the Internet to collect, communicate and share health information to predict their behavior. This may include analyzing queries in Internet searches to predict an upcoming disease, monitoring social media sites like Twitter, and gaining more health information. Automated tools may track information spread and effectiveness of health marketing. Infodemiology is a promising emerging research area to gain insight (Eysenbach, 2009). 

Infodemiology shows promise in its infancy and can help find solutions to the limitations of current problems and interventions. A survey of 338 infodemiology studies revealed that in health informatics, the unique benefits of infodemiology stem from its use of web-based sources, opening up a previously inaccessible source of information. Web-based data provides real-time assessments, whereas data from the health sector traditionally may take years to process (Mavragani, 2020). 

One limitation to infodemiology is that it does not inform researchers of the social psychology that might lead users to believe inaccurate information. More research on the social and psychological aspect of persuasion will be useful in addition to infodemiology.

4.1 Developing tools for infodemiology research 

Individuals and infodemic systems alike should develop the tools to responsibly engage with information distributed during a pandemic. The WHO has declared COVID-19 as an “infodemic,” or an information epidemic. In an infodemic, there is a rapidly developing overabundance of information during an epidemic, some of which is inaccurate misinformation or disinformation (Zielinski, 2021). 

COVID-19 misinformation typically includes “distinct language generalization patterns,” which create a negative sentiment and cause sympathy in the consumer. Auto-infodemic systems may be able to identify the discriminant features of misinformation, which would provide humanity with a useful tool to prevent the infodemic (Wan, 2022). 

Comprehensive research has studied the abilities of different machine learning models to identify fake news within a dataset of COVID-19 information. Of the most promising models, language and ensemble models taught to classify misinformation on social media were able to pinpoint misinformation with 97% accuracy. 

Future research projects should utilize datasets including both pictures and texts, and expand the sets to include more languages (Biradar et al, 2022). In addition to technological tools, pattern recognition may also help individuals to navigate the web thoughtfully. 

4.2 Building a benchmark dataset

Fake news diffusion is different from the diffusion of real news. They follow distinct paths and patterns while propagating through social media. A more comprehensive understanding of the social dimensions, life cycle, and identification of spreaders is needed. Recording and modeling the spread of fake news could help to mitigate the spread. A benchmark dataset that understands fake news could then provide early alerts to halt further spread of misinformation (Shu et al, 2017).

5. Facilitating media literacy interventions

Individuals should be empowered to seek out reliable information before making choices. eHealth interventions intended to build health literacy skills in target populations had significant positive or promising outcomes. The most effective mechanisms of intervention for motivating behavioral change have yet to be determined (Jacobs, Lou & Ownby, 2014). 

6. Creating a top-level domain

One strategy to facilitate accurate knowledge during infodemics includes creating a top-level domain for health information. Creation of the domain will aid in developing standards for information and it will be a step towards providing universal health coverage. 

This strategy has been attempted unsuccessfully twice in the past by the WHO. Commercial considerations tended to overwhelm the ethical and altruistic ones. In light of COVID-19, a top-level domain should potentially be attempted again, possibly managed by an international partnership operating independently. 

7. Creating an SDG

Additionally, nations should create a sustainable development goal targeting health information to prevent future infodemics. There is not yet a goal targeting information and knowledge, but the development of the goal would make it easier to facilitate the creation of international targets and secure funding. A goal could also lead to direct planning within the health sector. Each country could develop a knowledge preparedness plan (Zielinski, 2021).

8. Media literacy education

Promoting media literacy education (MLE) is another promising solution. Media literacy is the ability to obtain accurate information by analyzing sources and content. In the age of COVID-19,  the media has become difficult to navigate and interpret. US individuals with higher media literacy are more likely to take the recommended preventative measures from public health experts (Austin et al, 2021). 

Several European nations already have an MLE curriculum in schools for children. Research determining the effectiveness of such programs from a media literacy perspective is a relatively neglected issue. An issue with collecting data on these programs is the lack of standardization in curriculums. Another is the difficulty of measuring the effects of MLE on media literacy through tools and methods. A comparative study on the diverse MLE curriculums should be prioritized (Zhang et al, 2020). These programs tend to focus on the perceived benefits of MLE. More attention should be given to tractability.

The effects of media literacy education were studied in larger populations and found no differing effects from factors including age, setting, country, or topic. A study of 51 media literacy interventions found positive effects on media knowledge, criticism, and behavior, amongst other variables (Jeong et al, 2012). 

Tractability

Media literacy is measured by performance- or competency-based measures that are intended to evaluate an individual’s ability to analyze and criticize media based on its content, format and techniques to capture the viewer’s attention. Data is collected in self-report measures, where individuals rate statements that reflect certain beliefs. 

Limitations to tractability include a potential lack of objectivity. As new methods of data collection for media literacy develop, these might reduce the possible cognitive bias. The other limitation is the unavailability of evidence surrounding MLE curriculums, although there is some evidence on media literacy interventions.

Certain methods of intervention are more likely to be ineffective. One possible risk in advancing media literacy is motivating participants to become overconfident in their abilities to recognize inaccurate information. One method that is criticized is providing participants with a checklist to evaluate websites and sources. The issue with this approach is that unlike the general public, professional fact checkers tend to leave a website to evaluate its credibility instead of lingering. Fact checkers understand how information is presented online, and the reason that certain sources appear before others on search engines. A checklist tends to be the most popular way to teach media literacy because of its tractability; the impact can be easily quantified (Bulger & Davison, 2018).

The greatest problem to measuring quantitative impacts is the lack of long term data on tractability. The majority of available research on media literacy is from cross sectional studies, including a group of participants over a short period of time. An effective investment would be in longitudinal studies. Longitudinal data would likely be collected with the creation of a national body that would act as a storehouse of media literacy information. Routes to create this potential organization are discussed in greater detail in the next section.

Neglectedness and Funding Areas

Addressing the infodemic will involve multidisciplinary efforts because of the diverse causes and interactions between fields. Much of the work that furthers the media literacy field has come from outside of media literacy. Advances in social psychology allow researchers to understand human decision making. Political science can be informative in regards to partisanship and the political climate that may influence the media and end users. Sociology and communications allows health experts to understand how to address and communicate issues effectively. Media literacy should be addressed holistically as the field progresses. The next step forward is to identify gaps in the research, and uncertainty surrounding training methods for teaching media literacy (Bulger & Davison, 2018).

Funding research

Philanthropic organizations could be a locus of change to fund research projects in infodemiology and related fields. Infodemiology is considerably less neglected after the onset of COVID-19 than when Eysenbach introduced the concept with a pioneering study in the early 2000s. However, it remains a novel research area, especially in the US. To date, public funding agencies have led the majority of the available research on trust-building to resolve the aforementioned issues of trust. Addressing the problem of public trust requires a broader effort, and should involve funding by philanthropic organizations. 

There are a considerable number of gaps in the current research base. One that should be addressed involves focusing on clinician’s trust with healthcare organizations, the broader healthcare system, and each other (Cope et al, 2022). Research should focus on how healthcare organizations can build trust between staff and patients. Another involves undertaking long-term research on media literacy. A third research gap is measuring the impact of mass media in supporting the goals of media literacy through campaigns.

Building a national evidence base

Philanthropists should fund the creation of a national literacy evidence base to speed up the generation of available information on infodemiology. Current media literacy initiatives are decentralized and underfunded. Other nations have a more robust effort guided by a national organization, such as Ofcom in the UK. As a media literacy evidence base, Ofcom tracks progress with longitudinal data that would otherwise be difficult to collect. An independent organization should be developed in the US, modeled after Ofcom (Bulger & Davison, 2018).

One organization that may live up to the role is the National Association for Media Literacy Education (NAMLE). NAMLE is a nonprofit bringing together practitioners and advocates to teach media literacy to students, parents, health care providers, political leaders and communities. NAMLE holds biannual conferences and a peer reviewed journal. NAMLE could become a unifying national force for media literacy (Hobbs, 2010). NAMLE is currently being funded by social media platforms including Tiktok.

Supporting health literacy campaigns

Health promotion and literacy campaigns may be an effective way to further the movement. The WHO has introduced and promoted developments combating infodemics. These include a global campaign to increase awareness of misinformation, and increasing the body of research available. The WHO research agenda may help researchers determine the best approaches to manage infodemics.

Campaigns that address misinformation are likely more effective than those intended to change the algorithms of social media. Addressing misinformation outside of adjusting algorithms includes “social correction,” or encouraging citizens and public health experts to correct misinformation by challenging inaccurate information with evidence and reasoning. However, more research needs to be done to determine whether social or algorithmic correction is most effective (Bode & Vraga, 2018). This is a decent preliminary study, but relatively limited in scope. This study used a student sample and focused on Facebook, which may not represent the general public or sources of misinformation in general.

Supporting educational interventions

Other funding areas include supporting organizations intended to increase the awareness of misinformation, and increasing eHealth and other educational interventions. US organizations include the Action Coalition for Media Education, Alliance for a Media Literate America, Center for Media Literacy and The New Mexico Media Literacy Project (Kellner & Share, 2005). Support should be directed towards an online video documentation tool for teachers teaching media literacy.