trustworthy and non-trustworthy
With the end of Corona Virus Disease 2019 (COVID-19) as a global health emergency in May
2023 [1], experts can now meet face-to-face to discuss what people were saying and doing
behind the keyboards during our first global ‘paninfodemic’, the term used to describe the
abundance of information [2], helpful or dangerous, during a disease outbreak. The emerging
field of infodemiology deals with promoting the first and preventing the latter, mixing traditional
tools of surveillance and epidemiology with the new world of social media content and
telecommunications metadata.
Traditional epidemiology is already very complex. The scope includes demographics,
environmental factors, and non-communicable and communicable diseases like COVID-19.
Public health officials periodically check the catalogue of the symptoms people report in clinics
and hospitals, results of laboratory tests, and causes of death. With this in hand, they hope to
localize the outbreak, study the pathogen, enforce quarantines and other measures, and
propose exit strategies, such as vaccination, to ensure a safe return to everyday life.
The more online a population is, the more the data, including location, financial activities, and
communications patterns, can be cross-referenced into big data. Using this data to understand
what is happening within a population is called social listening. It is a growing industry. Public
health, corporate marketing, and public security are among the fields that use social listening to
translate what is dispersed and not accounted for into consolidated, numeric, and visual
information to support decision-making.
With the popularisation of the internet in the 2000s, infodemiology [3] emerged as a research
area focused on content analysis of online medical information. Health information on
webpages was compared with authoritative sources, usually printed books. How do we
distinguish between trustworthy and non-trustworthy content? As we dove deeper into
cyberspace, new tools emerged, including active monitoring of search queries that can
anticipate a surge in disease cases based on information from those who research symptoms
online before visiting a doctor. Additionally, smartwatches and other devices promise to detect
problems before the user becomes aware of them. Not without controversy, big data can relate
specific genetic information with societal behaviors.
Both social listening and infodemiology use social media platforms, mobile carriers, and search
engines to collect data for population studies. An important task is to do public health research
in the digital age while improving ethical and scientific standards. Security officers and
marketing teams may not be concerned about the social determinants of health or medical data
privacy, but health workers are. Infodemiology and social listening professionals can anticipate
outbreaks, understand the demands of a population, and propose and execute health
communications strategies.
With these concerns in mind, experts from around 10 countries or regions attended the
International Conference on Governing Social Listening in the Context of Serious Health Threats
at the University of Hong Kong's Faculty of Law conference hall. From the 22nd to the 24th of
September, 2023, they discussed infodemic management, vaccination, and rules-based
research. In the same way syndromic surveillance has fever as an indicator, infodemiologists
should put forward their own indicators, such as the term “vaccines cause autism”, suggested
Professor David Scales, from Weill Cornell Medicine, in New York. These indicators would be
used on a regular basis, and not only during an event, like a disease outbreak. As Dr. Cherstyn
Hurley, from the United Kingdom Health Security Agency, pointed out: “Meteorologists don't
measure the weather only when there is a typhoon. They measure it every day”.
Professor George Fu Gao, from the Chinese Academy of Sciences and the Chinese Center for
Disease Control And Prevention, introduced new concepts: ‘inforus’ [4], to theorize about the
causative agent of the infodemic, coronaphobia, to describe the uneasiness to discuss COVID-
19, and coronahypognosia, the inability to act.
Memes, news sources, and hashtags are the most common sources of social media analysis.
During the COVID-19 pandemic, they displayed the denial, distrust, and confusion expressed by
ordinary people on social platforms. In the first stage, it was necessary to convey the
seriousness of the situation and persuade those still in denial to use masks and stay at home.
Vaccination was the second stage. In some countries, it was mandatory. In others, it was
voluntary. Some restricted the access of non-vaccinated people to public areas, and others
demanded negative COVID-19 tests.
Once emergency usage vaccines became available, persuading people to get immunized was the
big challenge. Vaccination requires campaigns on radio, community halls, pamphlets, television,
and any network that socializes information. Infodemic managers from different countries
reported a variety of misinformation and disinformation that harmed vaccination strategies:
stories denying the existence of a pathogen, theories of public health officials being part of a
great conspiracy to curb freedom, false studies disproving the immunization effects of vaccines,
or that some readily available drug could wash all difficulties away. Many lives were lost.
Eventually, quarantine, vaccines and masks became a consensus in most places. However,
changing people's expectations and severing economic activity to a minimum required good
governance from trusted authorities. In some countries, influential political figures openly
discredited vaccination and the use of masks, acting too late and hampering the work of health
professionals. Nonetheless, it is worth remembering all the professionals and volunteers who
took time and effort to listen and talk to promote anti-pandemic responses.
According to the experts' presentations at the conference, the fight against the vaccine inforus
was made difficult by platforms’ unwillingness to cooperate. For example, social listeners
searching for disinformation and misinformation online reported difficulty in methodologically
studying datasets from private internet companies. When the data is obtained, the lack of
transparency makes it difficult to ensure the representativeness of the sample and, thus, of the
situation taking place. Moreover, social listening tools widely available in the market could be
expensive for academics or local public health officers.
There are some reasons why the data is challenging to obtain. First, it is difficult to combine
different data sources. Second, platforms are obliged to protect user privacy. Third, some
judicial frameworks do not attribute responsibility to platforms for user-generated content.
Fourth, data is a commodity for companies, so they purposely make it challenging to obtain
freely.
More openness and transparency from social media platforms would be welcomed by the
researchers, especially with data that is already legally commercialized with the intent of profit.
Different countries will have different judicial strategies. In episodes of severe health threats,
internet platforms could move from a position of neutrality to one of normative responsibility,
suggested Marcelo Thompson, from the University of Hong Kong. A question is how this
normative responsibility will occur: will the private companies set their standards, or will they
comply with local requests? When outbreaks happen, who will be the accredited source of
information?
The Platforms need to be held accountable, and so do users. One method discussed at the
School of Law is to move the burden of proof to users trying to challenge the scientific
consensus rather than placing it on public health authorities asking the information to be
removed or marked as untruthful. Of course, the scientific consensus is sometimes inaccurate
and may not even be a consensus. During the first months of the pandemic, authoritative
scientific organizations did not recommend using masks, while others promptly endorsed it.
Infodemiology can also help to set the consensus straight. Epidemiologists require contact
tracing methods to identify who may have been exposed to a pathogen and how the disease
spreads. Now, with the right amount of data, proximity tracking (e.g., the difference in signal
strength of two smartphones) can predict if two users were close enough to infect one another.
Associated with precise pathogen detection tests, proximity tracking is considered to have
opened a new anti-epidemic field, considerably de-risking infection.
Big data powers, big responsibilities. Many countries are working to improve the legal
framework to define data, to whom it belongs, and which rights real persons have over it. Rules
for processing personal information include consent, law-based analysis, and deleting personal
information or other data obtained for a specific purpose and timeframe. Law-based analysis
can require data to be anonymized and encrypted, serving only to the specific purpose for
which it was collected. For example, police officers should refrain from using health-related
location data to make arrests. On the contrary, criminal laws should be used to punish those
who infringe on personal information. Ideally, professionals who handle this data should have
clearance to do so, as biosafety lab staff must. Moreover, the procedure needs to be
overviewed, from time to time, by some form of ethics committee.
Before COVID-19, health scientists had already speculated about a global disease of an airborne
coronavirus, and it happened. Along with it, an infodemic. Now, with more tools and
experience, social listening and infodemic managers can work together to better prepare us for
the next inforus.