1 Introduction

Liberal democracies rely on an open debate on the threats facing society for appropriate risk assessment and mitigation (Beck 1996; Rosa et al. 2015). Transparent and accountable communication may increase trust in difficult decisions involving complex trade-offs that may require economic and qualitative social sciences (Norheim et al. 2021; Whitty 2015). In an ideal world, accurate evidence assessment from multiple fields is the base for policies with optimal cost-effectiveness. However, even with available evidence, policy will not always follow it (Macdonald 2015). The question of whether scientists should advocate is central to current debates, balancing between “the public duty to engage versus the need to protect science” (Oliver and Cairney 2019). Advocacy and lobbying groups tend to highlight particular data for policy-making, which may help to raise awareness of emerging risks and potential solutions but may also carry bias towards special interests and group-think (Jessani et al. 2022; Schippers and Rus 2021; Hutchings and Stenseth 2016).

During the COVID-19 pandemic, advocacy groups argued for vastly different mitigation policies (Lenzer 2020). Some argued for society-driven, voluntary, or “focused” mitigation of vulnerable groups while leaving children, for example, mostly free of mitigation efforts, such as the “Great Barrington Declaration” (Angeli et al. 2021; Burki 2021). Some argued for public health tradeoffs including consideration of the adverse effects of policies (Schippers and Ioannidis 2022). Conversely, others argued for stricter government-enforced policies, for example the John-Snow memorandum (Alwan et al. 2020), or even stricter virus-elimination strategies (“zero-covid”) (Bar-Yam et al. 2021a; Olio-Barton et al. 2021) as long pursued in China. Some argued for a vaccines-plus strategy using mass mitigations in schools, for example, to still control spread in 2022 after arrival of the omicron variant (Greenhalgh et al. 2022). Mitigations such as school closures were advocated but also warned against by UNICEF (Reuge et al. 2021) and pediatric and education experts, highlighting the contested nature of these debates (Buonsenso et al. 2021).

Social media platforms were extensively used by governments, scientists, and advocacy groups during the pandemic to disseminate information and encourage adherence to vast array of safety protocols (Tsao et al. 2021; Katz and Nandi 2021). The World Health Organization (WHO) raised concerns over misinformation early in the pandemic (Zarocostas 2020), which may lead to conspiracy theory beliefs (van Mulukom et al. 2022) and reduced vaccine uptake (Pullan and Dey 2021; Borges do Nascimento et al. 2022; Van der Linden 2022), as a few examples. In addition to producing erroneous public health behavior if followed, misinformation may cause substantial emotional stress, anxiety, even panic, depression, and fatigue when having a strong appeal to fear (Rocha et al 2021; Krygsman et al. 2023). National governments worldwide addressed misinformation differently, from increasing access to what was deemed accurate information to criminalization of public criticism of government response (Pomeranz and Schwid 2021). A major conceptual difficulty is that the term ‘misinformation’ is poorly defined (El Mikati et al. 2023), and that risks and benefits of fighting this phenomenon are complex (Altay et al. 2023).

Research on scientist-driven advocacy indicates a complex picture with unclear impact and potential risks to trust in science and democratic institutions (Geiger 2022; Grant and Smith 2021; Palthiel et al. 2021), perhaps especially when communicated in scientific journals (Zhang 2023). Some studies have analyzed scientist social media networks and discussed broadly some advocacy groups (Durazzi et al. 2021; Ioannidis 2022). A study from Finland revealed three major clusters of pseudo-anonymous social media activism on X/Twitter during the pandemic, with the “zero-covid” cluster often using person-directed strategies to criticize government policies (Heikkilä 2022). In Sweden, the group Media Watchdogs of Sweden (MEWAS), a group with 200 members aiming to influence how the Swedish government handled the pandemic, used coordinated person-directed strategies (trolling or “swarming”) to influence scientists in the debate (Pamment 2021).

Aside from these Nordic studies, research on zero-covid advocacy remains limited. Considering its relevance for scholarly analysis, this advocacy for what we classify as a “maximalist risk mitigation position” of the advocacy spectrum warrants more detailed study. To this end, we analyzed advocacy associated with the World Health Network (WHN), a group formed in 2020 (Bar-Yam 2020) and promoted in a letter in The Lancet in 2021 (Bar-Yam et al. 2021b), and from this analysis, discuss the need for better advocacy guidelines.

2 Methods

We conducted a thematic analysis (Kiger and Varpio 2020) of the most visible advocacy positions on Twitter/X by co-signatories of the WHN advocacy letter published in the Lancet (Bar-Yam et al. 2021b). The appendix of the letter provides the total list of 132 signatures. Many of these were also associated with various WHN partner groups listed in the appendix of the Lancet letter, such as independent SAGE (indieSAGE) from United Kingdom, ZeroCovid USA, and Endcoronavirus.org. Local groups from specific Western European countries were also noted in the list of associated groups in the appendix of the WHN paper. Although the advocacy also had presence in traditional media that could be studied as well, we focused on Twitter/X because of the rich data, search options, and ability to measure impact. Follow-up studies may investigate the studied advocacy groups in other contexts such as in scientific journals and in media outlets. The methods are described below in detail, to enable replication of our findings.

To capture the most visible advocacy, we identified twenty from the total list of 132 signatories who met all the three or the following inclusion criteria: (1) more than 10,000 X followers, (2) 5,000 tweets or more in total on any topic, and (3) tweets being categorized within each of the ten studied themes (see below; enforced to 198 out of 200 possible tweets, with two instances for Mpox not yielding data). The twenty advocates who met these criteria were analyzed on all ten themes presented in English on X, acknowledging that this is not representative of all zero-covid advocacy but only some of the most impactful in English language. These advocates are referred to as the WHN-advocates below.

Tweets covering 2020–2023 were collected and categorized in ten specific themes: Six (1–6) presenting data on specific health-related topics, and four (7–10) representing discursive strategies used by the advocates when presenting their views (advocacy methods): (1) Views on zero-covid (search terms “zero”, “elimin”); (2) SARS-CoV-2 epidemiological data and new variants (“variant”; “omicron”; “virus”; “fit”, “infect”, “transmit”, “SARS”, “COVID”); (3) post-acute sequelae of COVID-19 and immune dysfunction due to COVID-19 (“long”, “immun”, “infect”, “HIV”, “hepatitis”, “Alzheimer”, “brain”, “cognitive”, “chronic”, “heart”, “diabet”); (4) views on vaccines in comparison to other mitigations (“vaccin” OR “vax” AND “protect”, “plus”, “test”, “air”, “view”, “prevent”, “use”, “N95”, “breakthrough”, “mask”, “alternative”); (5) views on mitigation of children and schools (“child”, “kid”, “school”); (6) views on Mpox (“pox”); (7) personal health advice and advertising remedies (“mask”, “air”, “HEPA”, “ventil”, “test”, “trac”, “need”, “should”, “pursue”, “call”, “sign”, “buy”, “stop”, “respirator”, “monitor”, “policy”); (8) views on other scientists (“he”, “she”, “they”, “you”, “grift”, “minimi”, “misinfo”, “disinfo”, “denier”, “mass infect”, “we”); (9) use of anecdotal evidence (“girl”, “man”, “woman”, “boy”, “baby”, “child”, “kid”, “parent”, “father”, “mother”, “daughter”, “son”, “person”, “people”, “family”, “friend”, “effect”, “I”, “my”, “story”, “anecdot”); and (10) claims of intent (“plan”, “deliberate”, “coordinate”, “conspir”, “intent”, “hid”, “delet”, “active”, “allow”, “infect”, “adviser”, “government”, “eugenic”, “mass”, “kill”, “force”, “cleansing”, “misinfo”, “disinfo”, “bias”, “ideol”). Mpox was included because the advocates were very active on this topic, reflecting a broader perspective on disease beyond COVID-19.

To ensure a balanced representation, and avoid the problem of subjective selection of data even when using independent raters, we included one tweet from each advocate on each of the ten themes that had the highest number of likes among all search terms in the analysis section. Searching was conducted through a new X profile without settings or preferences. The search terms were varied by extending roots where relevant (thus “elimin” covers elimination, eliminating, eliminate”; “immun” covers immunity, immunization, immune). We used the “from: profilename” with each search term above, and the top hits were manually checked for misclassification to ensure that they correctly reflected each theme; the tweet with most likes for each advocate for each theme was then collected and analyzed.

Since views likely change over time relative to the pandemic crisis, the timing was collected, to identify progression in views, and listed in the main figures. This yielded 198 top tweets, with two missing for the Mpox theme (given that 18/20 had views on this topic also and the perspective it affords on another disease as comparison, the inclusion criteria was relaxed slightly in this case). Eight of these tweets were from 2020, 66 from 2021, 88 tweets from 2022 (including all the Mpox tweets), and 36 from 2023, indicating a reasonable spread over 2021–2023, with only few tweets representing the acute first year of the pandemic (Supporting Information).

To estimate dissemination, we additionally analyzed the total number of likes received to the top-listed tweets as of December 2023, and the corresponding combined number of retweets and quote tweets. Twitter likes are dynamic and change over time, and are thus only used as indicative of total impact, i.e., at the level of “hundreds” or “thousands”. To establish relevant context, we also collected additional tweets as perspective on the top tweet, especially for views on other scientists. Collected tweets are quoted in machine-readable format with archived links in Tables S1–S10 and impact metrics given. If more than one tweet is listed, the first one represents the top tweet according to the search method. Whereas this systematic collection of tweets is the main body of data, to save space and focus, the Results section selects a subset of tweets based on number of likes or alternatively, in a few cases, relevance to the following thematic discussion.

3 Results

3.1 Views on zero-covid

All 20 advocates used the term zero-covid. Some views are summarized in Fig. 1, with full quotes and impact metrics in Table S1. Zero-covid was claimed to be a strategy that was more ethical and provided normality to society: “If you want your life back sooner like New Zealanders, choose #ZeroCovid” or “What’s not desirable about having a drink with your mates, about hugging your loved ones?” and “COVID Zero is the only strategy that makes sense & the only humane, ethical strategy”. Though not always clear, the term zero-covid was, when clarified or defined, used to refer to SARS-CoV-2 “elimination” strategies (as opposed to complete eradication), which may have caused some confusion regarding the word “zero”.

Fig. 1
figure 1

Statements on zero-covid 2021–2022 (total list of statements in the supplementary information)

Full size image

The alternative to zero-covid was claimed to be worse or even catastrophic, and many of the views posited that there was no middle ground (Table S1): “We either go for #ZeroCovid or we yo-yo for years”; “The alternative to #ZeroCOVID is #ForeverCOVID”; “I always knew that not having zero COVID would lead to max COVID”; “The choice is actually between #Zerocovid and #ZeroHuman / #zeromink”; and “Every opponent to that vision (called it a cult etc.) wanted to “let it rip” including dear “centrists” who didn’t understand” living w covid is exactly the same as letting it rip!”.

A key argument was that virus evolution (leading to potentially worse immune evading variants) could be prevented with zero-covid: “we also need to get as many countries to #zeroCOVID as possible, so as to guard against more variants.” Or as another advocate put it: “Viruses can’t mutate if they can’t replicate. #StopTheTransmission #StopTheSpread #SARS2EliminationNow #ZeroCOVID #EndThePandemic”.

Obtaining zero-covid was seen as a hard struggle that nonetheless would be transformative, as one advocate put it: “Center our feelings in compassion, our will in defeating an enemy, our creativity in innovation, we can transform society for collective action and win”. Some expressed the view that the media were responsible for not allowing zero-covid: “our media are complicit in the mess we’re in. Trying to portray zero COVID as extreme, when it’s the only sensible position”.

Some advocates stopped using the term “zero-covid” in 2021 or 2022, whereas others continued to use it and argued that it was the best strategy even in late 2022: “It seems it is still a target we should aim for (like traffic deaths, plane crashes, cancer, measles, etc.) and be happy if we just make things better. We have better tools than in 2020.”

3.2 Views on SARS-CoV-2 epidemiological data and new variants

Examples of views on SARS-CoV-2 epidemiology are summarized in Fig. 2 (Table S2). As in most other examples listed, the tweets received many hundreds of likes on X. A main message was that SARS-CoV-2 transmission was primarily airborne, rather than using terms such as aerosols and droplets: “Narrator: Covid-19 is PRIMARILY airborne. The WHO’s failure to acknowledge this cost thousands—perhaps millions—of lives”. WHO acknowledged this terminology later in the pandemic but did not state it as the only or primary mode of transmission (Lewis 2022; Lopez and Srigley 2022; WHO 2023). Simultaneously, advocates argued that “SARS2 is avoidable for people who have a full control over the air they breathe. With good policies, SARS2 would be avoidable for everyone”.

Fig. 2
figure 2

Views on SARS-CoV-2 epidemiology, 2021–2023

Full size image

Emphasis was almost completely on negative views. The outlook was summarized by one advocate as follows: “The virus that causes COVID-19 isn’t going to turn into something mild—not in our lifetimes anyway. Don’t let politicians tell you that this is a virus we can learn to live with. It isn’t.” As to how often reinfections would happen, another advocate claimed that “We can get reinfected every few weeks. So can our children. Neither infections nor vaccines prevent infections.”

Accordingly, when new variants arose, views were offered on their severity and impact on the trajectory of the pandemic. One advocate simply said: “Omicron is not “mild”.” An estimate of the mortality of omicron (called “nu”) in November 2021 expanded on these claims with a “very rough estimate of Nu variant mortality is 8X original variant based upon the current rough estimate of transmissibility”.

One advocate summarized the zero tolerance ideology in a broad policy statement as follows: “We must never again accept seasonal colds and other respiratory viruses as the norm. They are telling us that the ventilation is inadequate, that our indoor air is dirty.” (Fig. 2).

3.3 Views on Long COVID and immune dysfunction due to COVID-19

Examples of views on post-acute sequelae (Long COVID) and immune effects of SARS-CoV-2 infection are summarized in Fig. 3 (full data in Table S3). The main narrative was that Long COVID could lead to many complications and diseases, and according to at least some advocates, a chronic impairment of the immune system, a view that played into hypotheses that compared SARS-CoV-2 to HIV on the internet.

Fig. 3
figure 3

Views on long-term effects of SARS-CoV-2 infection 2022–2023

Full size image

A view was that chronic impairment was predominant, i.e., a narrative of a mass disabling event. As one advocate put it: “My scientific hunch looking at the Covid & brain science is basically most infected persons will have some kind of brain issue at some point. Can we prove this? Of course not. But are will we willing to wait for the destruction of the western human brain before we do something?” Another advocate simply clarified: “COVID is chronic. All COVID. Even mild cases”.

The impact was compared to HIV: “Coronavirus uses same strategy as HIV to hide inside cells.” A narrative was that infections, rather than adding adaptive immunity, made the next infection worse, as one advocate said in a tweet that received more than 2,000 likes: “The very idea that there is some benefit from getting infected, i.e. gaining immunity from future infections, has now been shown to be incorrect. The opposite: an infection makes the next infection worse, not better”.

One advocate summarized it as follows: “All year so far we’ve seen more deaths for heart conditions, diabetes, hypertension, Parkinson’s and dementia and Alzheimer’s. All of these have been linked to prior Covid infection”. A hallmark of these views on long-term effects of infection was a common narrative link to children: “I am curious about how many times folks are willing to allow kids to get infected by a virus that wrecks immune, vascular and neurological systems.”

3.4 Views on COVID-19 vaccines in context of their limitations and other mitigations

Examples of views on COVID-19 vaccines as a pandemic mitigation strategy with emphasis on their limitations vs. other mitigations are summarized in Fig. 4 (full quotes in Table S4). These views notably do not reflect the top view on vaccines alone (some of these had many more likes) but the specific views on vaccines in the context of other mitigation strategies, to understand this balance.

Fig. 4
figure 4

Views on COVID-19 vaccines in context of their limitations and other mitigations 2021–2023

Full size image

The advocates cannot be said to have “anti-vax” views, but the intense advocacy for elimination was associated with reduced emphasis on vaccines as a mitigation strategy and highlighting their limitations. One advocate clarified this: “if I was only allowed to have one—an FFP3 mask or a vaccine, I’d choose the former.” (> 800 likes). Another said: “I view the vaccines the way a fighter pilot views the ejector seat… It is a last resort that may end up saving your life..but you don’t want to have to use it”. A third one stated: “N95s are like brakes on a car, they’re used to prevent accidents. Vaccines are like seatbelts, they result in severity in an accident. Better to avoid accidents than rely on seatbelts to save you.”

Vaccines were seen as means of mitigation protection, but not as an alternative: “The current vaccines are not alternatives to NPIs. They are just an additional layer of protection”. One advocate even argued that “If vaccinated people get together they promote transmission of vaccine evading variants, leading to breakthrough infections and another round of this pandemic.” The zero-covid advocates also criticized vaccination advocates and vaccination scientists, while arguing for a focus on mask mitigation: “Still—some prefer to use their voice to continue advocating for vaccines only while ignoring the increasing NEED to reinstate masks.”

Another advocate went further and described the vaccines as “eroded” and a state of “eugenics”: “Our vaccines are slowly being eroded. Our kids are being sent to hospital in greater numbers than before. Businesses struggling. From the outset a eugenics became widely accepted.”

3.5 Views on pandemic mitigation of children and schools

Examples of views on data and policies relating to schools and children in Fig. 5 (all collected quotes are in Table S5).

Fig. 5
figure 5

Views on pandemic mitigation of schools and children 2020–2023

Full size image

Children and schools were a hallmark interest of the zero-covid advocates, and mitigation of schools was arguably a primary feature of their combined advocacy. A direct example of this was when one advocate stated (June 2020) that “Our children can’t come back to school until we end the pandemic.” Another of the studied advocates who was also affiliated with independent SAGE gave this rationale: “Being off school for a year is damaging. Being dead’s also damaging. Go figure”. These views consistently disregarded, without openly denying, the very low mortality impact of COVID-19 on children, and relevant public health tradeoffs relating to school attendance (Soriano-Arandes et al. 2023; Mazrekaj and De Witte 2023).

The concern over school biosafety was clear: “Schools aren’t safe. Without investment in ventilation and air filtration, kids will get COVID-19 2–3 times per year, damaging their health and leaving them vulnerable to other bugs.” As another advocate put it: “UK schools are really not a safe place for children because DfE, the paeds community, and govt have decided that forced mass infection of children is the way to go.” A third said: “Cholera in Paris in 1832 led to sewers. COVID-19 globally in 2022 led to “get the children back in school, infected, and with brain damage.”

The concern over school safety originated in a very negative view on the potential impacts of infection on the long-term health of children. As the WHN founder said: “Children are aged by covid, we just don’t see it because turning a 3 to a cellular 37 year old isn’t obvious.” The narrative hinted at the possibility of avoiding infection completely in children: “I simply cannot accept this argument that it is better for children to get covid as soon as possible. It is better if children never get covid at all”.

A necessary tool in this policy was masking, as one advocate said: “Of course children need masks. COVID is not a cold. It is a neurotropic virus, can go into the brain and lots of other organs.” Other WHN associates stood ready to sell the masks needed from their companies: “Our manufacturers are standing by to produce the respirators we need to keep children safe & in school.” In this way, narratives and solutions played together in the advocacy and merged with financial interests.

3.6 Views on Monkeypox

As Mpox (monkeypox) spread across Europe and the United States in the summer of 2022, the zero-covid advocates quickly initiated an online discourse about its spread and mitigation (Fig. 6, full quotes and links in Table S6) providing broader insight into their views on disease beyond COVID-19. All these views are from a short period when Mpox was spreading mid 2022; the tweets relating to Mpox reduced quickly when the wave stopped in late summer 2022.

Fig. 6
figure 6

Views on Monkeypox during the outbreak of 2022

Full size image

A characteristic feature was the strong public emphasis that Mpox transmission was airborne, so that masks would be required to contain it. One advocate stated it as follows: “I’m calling it: monkey pox is not a sexually transmitted disease. Sex is the confounder. The left-out variable is sharing air. Try having sex without doing so.” Another advocate simply said: “Monkeypox is AIRBORNE”, an oversimplification at best, and in contrast to the WHO statement that Mpox is transmitted through physical contact and can be prevented by avoiding physical contact (WHO 2023).

Also central was the view that the Mpox spread was a high-risk situation, and a narrative that it was comparable to the COVID-19 pandemic and required its own pandemic declaration. As one said: “Monkeypox is very likely going to be a much bigger s..show than COVID19; it is just going to unfold over a longer timescale.” Exponential projections were made: “As of now Monkeypox cases are doubling in 1.5 days 27X in one week If this growth rate continues the 228 now confirmed will be more than 6,000 in one week 150,000 in two weeks 300 M in one month Everyone in 5 weeks + 2 days. It can’t continue, but what will stop it?”.

These views eventually culminated with WHN declaring a “monkeypox pandemic” (Fig. 6). While danger was seen as imminent, apparent lack of effort was linked to a potential conspiracy, as one advocate noted: “Why no government works seriously to contain monkeypox when it’s still relatively easy to do? Is the plan to spread it to as many people and as many communities as possible?”.

3.7 Personal health advice and advertisement of remedies

The 20 studied WHN advocates were active in proposing specific public health policies, providing personal health advice on the Internet, and promoting specific remedies online. These tweets typically received hundreds of likes. In addition to the advocacy for general zero-covid policies described in Sect. 3.1, examples of such personal health and mitigation advice and advertisements are summarized in Fig. 7 (full quotes in Table S7).

Fig. 7
figure 7

Views on personal protection and advertisement of remedies from 2021, 2022, and 2023

Full size image

Even after broad vaccination and the arrival of omicron, which led many countries to abandon strict mitigation, the advocates emphasized mask use and social distancing even outdoors, for example announcing “#Omicron is so contagious. Make sure your kids use a minimum KN95 or KF94 or FFP2 mask” and “Omicron is so contagious, stay away from people even outdoors and even wearing a mask.” In May 2022, an advocate wrote: “We *need to suppress* with high-grade masks, ventilation, rapid testing, supported isolation—*while* we wait for the next generation of vaccines, and better treatments for long COVID. This will lead to *huge* increases in chronic illness.”

As late as April 2023, one advocate stated: “My various science WhatsApp groups are buzzing. Genetic lineage clips and diagrams flying back and forth. I understand little of the detail but it looks like it’s once again time to MASK UP.”

Some advocates had private companies selling masks, respirators, air monitors (we identified two, http://prescientx.com/ and http://co2radical.com.au) and frequently advertised this, mixed with views on the pandemic, with statements such as “FloMask is one of the key reasons our family has managed to keep safe from school-related infections.”, “Our 7 yr old uses http://Flomask.com “, You can buy this simple biohack from http://co2radical.com.au “ etc.

3.8 Views on other scientists

The advocates had emotion-laden opinions about government agencies and other scientists. Although some advocates kept a less confrontational approach, others frequently quote tweeted, criticized, and name-called scientists doing research on COVID-19, as well as other authorities. Some examples are summarized in Fig. 8 (full data in Table S8). The analysis does not determine if this was more prevalent than among others, including the opposing scientists.

Fig. 8
figure 8

Views on other scientists 2021–2023. The names of the professors have been changed to X, Y, etc.

Full size image

Harsh criticism and strong rhetoric centered on official bodies like CDC, UNICEF, WHO, etc. An advocate with a large X following announced: “The @CDCgov has blood on its hands”. This on you @CDCDirector”, and in another context “The @CDCDirector is directing a disinformation campaign.” Another zero-covid advocate claimed that “@unicef spreading misinformation that surface transmission and handwashing are important for COVID-19”.

Scholarly work that the advocates did not agree with was strongly criticized, a notable example being the Cochrane review on masks (Jefferson et al. 2023), which was described as “the grift that keeps on giving…” The term “minimizer” was on several occasions used to characterize opponents (supposedly someone who is deliberately downplaying COVID-19), with one advocate stating of a well-known professor of epidemiology, “This is X, minimiser in chief”. An epidemiologist and a vaccine expert were quote tweeted with the statement: “This is an information war. Do not be taken captive by ideologues who refuse to engage in proper scientific debate, however senior they are”, implying that the two British professors were driven by ideology as opposed to “honest” science.

University professors with different views were described as “a sick man”, “useless excuse for a researcher”, “he posts garbage constantly”; “she’s a miserable human”; “she’s so incompetent”; “he’s such an idiot”; “propagandist X”, etc. One advocate wrote: “X finds reassurance in mass infection of children” and a third wrote “sure, listen to clinical experience, but not X, Y, and Z, all of whom have been covid minimisers in children- and partly responsible for where we are.” More examples, including claims of being “idiots”, doing “pseudoscience”, etc. can be found Table S8.

3.9 Use of anecdotal evidence

All the studied advocates used anecdotes to make messages stronger and emotionally appealing, with some examples shown in Fig. 9, and all quotes in Table S9. Sometimes the anecdotes were just listed, whereas other times, conclusions were drawn from them suggesting policy implications, as a form of anecdotal evidence.

Fig. 9
figure 9

Examples of use of anecdotal evidence

Full size image

During summer 2020, when the first wave passed in Europe, anecdotes from South Africa’s wave were discussed: “A person with #COVID19 went to a hospital in South Africa. 39 patients and 80 staff were subsequently infected and 15 patients died. We run this risk every day in Australia and will continue to do so unless we pursue elimination like New Zealand.” Anecdotes with children implying consequences of poor mitigation were particularly impactful, such as for example: “A 10-year-old Suffolk girl who died of COVID had been assigned at school to walk with sick students to nurse’s office”.

Concerning stories about unexpected consequences of infection were impactful; one wrote: “One man in his 30s revealed he had struggled to get an erection after infection and, as a result, his penis had shrunk.” Personal experiences were noted, such as “I’ve been struggling with breathlessness, chest pain and brain fog post-COVID”, a statement that received more than 3000 likes. Another story, “We have just received the sad news that a seven year old girl, who is a friend of one of our kids, lost her mother to COVID-19. Her mother caught COVID-19 from her and the child is painfully aware of that. “Fully” vaccinated. Words fail me”, received almost 3000 likes.

Anecdotal evidence was used during the Mpox spread, indicating community (airborne, as opposed to close contact) transmission. One advocate spread a claim of a man having Mpox in a Spanish train that received thousands of likes but was later debunked by Spanish media as fake news, the person in fact having neurofibromatosis rather than Mpox (Daunt and Khatsenkova 2022). Another advocate announced: “Baby caught monkeypox. Probably that “close vigorous contact with MSM usually during sex at a club” (Table S9).

3.10 Claims of intent or conspiracy

In addition to the criticism of official bodies noted above, the advocates also implied or directly accused authorities of ill intent or conspiracy to hide data, enforce mass infection, sometimes hinting at “eugenics” strategies. Examples of these tweets spanning 2021–2023 are summarized in Fig. 10, with all text quotes in Table S10.

Fig. 10
figure 10

Examples of claims of intent or conspiracy

Full size image

Regarding misinformation and data deletion, one advocate simply stated: “Misinformation provided by government = disinformation”. One wrote: “This is important. Our government is rewriting history.” In the US, another announced: “MUZZLING—CDC’s social media has been surreptitiously deleting/omitting key info that #LongCovid can last for “YEARS”.”

One advocate claimed that advisers had been “hand-picked”: “Tweeps, these advisers have been hand picked. I have no inside track.” Another said that “The government’s calculated indifference even appears to be part of a deliberate plan to let the virus rip.” Other views included “The reality: “pandemic cleansing” “; “The eugenics movement is deeply rooted within the UK”; “More evidence of the toll the mass disabling event we’re in. As always, remember the names of those who helped governments fabricate your consent.”

Some of the claims related to notable persons, such as royalty: “They killed the last Queen by infecting her with COVID. ‘To lose one Queen may be regarded as a misfortune, to lose two looks like carelessness.’“ Others hinted at major ignorance and denial at government level, such as: “@wef & @WHO had a global planning session to prepare the world to defend against a worldwide airborne respiratory coronavirus with asymptomatic transmission 3 months before COVID? Then ignored the plan? And denied airborne transmission? You can’t make this stuff up”.

4 Discussion

4.1 Claims made

The following discussion is offered as a perspective, but we stress that the quotes collected should stand on their own. The statements showed a high degree of thematic convergence and coherence, with advocates frequently retweeting other advocates, including others than those studied in the present work, to reinforce views (please see links in Tables S1–S10).

In evidence-based medicine and public health more broadly, the hierarchy of evidence in complex tradeoffs guides policy proposals. Achieving consensus requires systematic evaluation of the available data across diverse disciplines, typically published in systematic reviews and meta-analyses. Due to the multidisciplinary complexity of public health tradeoffs and associated policies (Norheim 2021) and the need for transparent high-level evidence, scientists should not draw strong conclusions and suggest policy on limited observational data. Similarly, scientists and even medical doctors usually do not give health advice on social media, as this requires personal consultation with an appropriately experienced and licensed medical doctor. In addition to the risk of direct harm from erratic behavior or treatment, strong online claims and advice may cause net mistrust in the medical profession (Graso et al. 2022).

The style of advocacy studied here did not follow any of these rules of communication: In Pielke’s model, this type of advocacy narrows options down to one main issue (the disease threat), disregarding other threats, adverse effects, and options. In so doing, it departs from an “honest broker” model of advocacy, which attempts to offer a balanced discussion of data, options, and policy scenarios for public scrutiny (Pielke Jr. 2007). Although some participants tried to role-model a “science arbiter” advocacy approach (analyzing data in greater detail and communicating results of the science to the benefit of readers), the analyses were always in relation to the single issue, the danger of COVID-19, and mostly lacked discussion of alternative interpretations, uncertainties, and other policy options than those favored by the advocates.

Although such views may exist, none of the tweets obtained by our search strategy resulted in optimistic statements about the course of the pandemic. The data collected reinforced a belief in a continued threat to society, even in 2023. Some claims were directly shown to be false, such as the one on Mpox in a Spanish subway. Alleged misinformation by zero-covid advocates was recently described in another research study in relation to Mpox (Knudsen et al 2024). Some advocates were also associated with alleged misinformation relating to Denmark’s reopening that involved Danish Statens Serum Institut actively responding on social media to one of the studied advocates (Delmelle 2022).

Several other claims were misleading or lacked nuance: Stating that 3-year-olds become like 33-year-olds due to infection-induced biological aging; that most people will have some form of brain damage from COVID-19; that all COVID-19 is chronic, even mild cases; that omicron (called “Nu” in the tweet) would be eight times more deadly than the original variant; that HIV and SARS-CoV-2 share features of immune damage mechanisms, and that Mpox is airborne without mentioning physical contact transmission. Many views involved far-reaching policy proposals with complex tradeoffs advocated without systematic evidence analysis, and several narratives, such as “the destruction of the western human brain”, came without field expertise on the topic of the claim, in addition to the curious qualification of the “brain” at risk as specifically “Western”.

Another narrative consistent with the zero-covid concept was that infection was entirely avoidable. Statements like “It is better if children never get covid at all” and “We must never again accept seasonal colds and other respiratory viruses as the norm […] our indoor air is dirty” illustrate a totalizing narrative on respiratory viruses and inherently dirty indoor air that fits with the need for zero-covid. The claims that Mpox is airborne and thus needs to be suppressed with masks and ventilation fed the same “dirty air” narrative. This narrative of “dirt and pollution” has been amply documented in the moral and racial logics informing the class-inflicted projects of the Victorian and other hygiene movements (Porter 1991).

4.2 Social media strategy

The advocates were highly connected and coordinated by quoting and retweeting each other (Supplementary Information). Some advocates were involved in other groups, notably indieSAGE in the UK, OzSAGE in Australia, and the John-Snow Project (JSP). The advocates commonly amplified each other and other similar-minded advocates by retweets and quote tweets, with examples given in the archived links of the supporting information. This coordination to increase impact may have made the advocates appear stronger and more notable, as indicated by their significant followings. The advocacy was intense, as the number of tweets by some advocates show (tens of thousands in several cases), providing a constant stream of information.

Whereas some UK advisors criticized the use of fear in government public health messaging in May 2021 (Rayner 2021), fear remained a consistent feature of the studied advocacy well into 2023 (Fig. 3), with advocates tending to exaggerate risks, as documented above. During our search, we did not find any tweet examples where the advocates indicated low risk of COVID-19, and no tweets were unambiguously optimistic, even after the data for the much lower infection fatality rate with the omicron variant in vaccinated populations (Erikstrup et al. 2022) with emphasis instead moving more to low-evidence but potentially high-risk long-term consequences of infection. Although such tweets may surely exist in the larger advocacy group, they were not captured by our search criteria, whereas many negative views emerged, indicating a strong bias in narratives already without the direct misinformation discussed above.

The claimed risk (“COVID is chronic. All COVID. Even mild cases”; “the destruction of the western human brain”) was coupled to persistent calls for stringent policies (“We *need to suppress* with high-grade masks, ventilation, rapid testing, supported isolation”, “Our children can’t come back to school until we end the pandemic”). In all the cases we examined, these policy proposals lacked core expertise (such as pediatrics and education), clarification of limitations in supporting data despite often being anecdotal, or consideration of tradeoffs and complexities.

As illustrated in many themes, including the claimed need for zero-covid itself (Fig. 1), many statements were value-driven or emotionally appealing (“if you want your life back”), semantically ominous (“The only thing that brings this deadly virus to our shores is human bodies”) or indicating conspiracies (“Is the plan to spread it to as many people and as many communities as possible?”).

The WHN was confused with the official WHO in some circumstances, and the WHN applied not just the similar name but also some of the entrusted actions of WHO, by for example declaring a Mpox “pandemic”. On at least one occasion, BBC confused WHN and WHO (BBC 2022), much like indie-SAGE was on several occasions confused with official SAGE in British media (Dalton 2023; Sample 2023).

The pandemic was associated with substantial hostility, blaming, or stigmatization in the public discourse (Chang et al. 2020), and many physicians and biomedical scientists experienced social media harassment (Royan et al. 2023). Official bodies and experts became contested actors in their field of expertise (Väliverronen et al. 2020). Our study confirms these tendencies with many specific examples, including challenging The CDC, WHO and UNICEF to the point of calling them spreaders of misinformation. The advocates strongly criticized authorities and other scientists, giving the impression of a deep divide in the scientific interpretation of the data, and sometimes even claiming evil intent of opponents. The person-directed methods of advocacy identified in previous research (Heikkilä 2022) was confirmed in our study, with other scientists being called sick, pseudoscientists, ideologues, minimizers, idiots, or propagandists by various of the studied advocates.

As one advocate said, “our will in defeating an enemy, our creativity in innovation, we can transform society for collective action and win”. Such totalizing war-like rhetoric was also reflected in attacks on authorities implying ill intent, hiding of data, etc., and accusing other scientists of biases or incompetence in public. This suggests “groupthink” and a discursive universe that far exceeded COVID itself. Such groupthink may reduce exploration of alternative explanations of data and delay time required to achieve consensus (Akhmad et al. 2020).

Overall, the strategies used constant, coordinated communication of danger or fear, uncertainty and confusion, war analogies and group identity, name-calling opponents, emotional appeal, selective use of data in contrast to evidence-based scientific methods of systematic review, and in all this, strong and direct proposals of potentially intrusive policy on highly complex matters. These methods seem consistent with those used to construct “moral panics” and arguably could even share some features with propaganda (Hier and Hier 2011; Ungar 2001; Wright 2015).

4.3 Promotion by X/Twitter

The selection of tweets in this study is a small sample of the total number of tweets made by the advocates, as seen from their total tweet numbers, which were in tens of thousands in some cases. The number of likes obtained by the tweets (in some cases, many thousands; the misinformation from Spain which was retweeted by one of the advocates received more than 70,000 likes) shows that X was very important in ensuring that the messages and claims reached many thousands of people. Many such examples are given in the Supplementary information.

In many cases, retweeting and quote tweeting of other advocates led to a mutual amplification (Supporting Information Tables S1-S10), boosted by very strong follower networks of anonymous activist accounts liking and retweeting the advocates with tags such as #zerocovid, #COVIDisAIRBORNE, #airborneHIV, #LONGCOVID, #COVIDisNotOver. We did not identify any tweets in our study that were removed or otherwise sanctioned by X. All the attacks on official bodies and other scientists (calling them idiots, minimizers, propagandists, pseudoscientists, etc.) were fully available at the time of the study, showing that X did not enforce any noticeable policy to reduce the harm done by the advocacy to other users and scientists.

The success of the studied advocacy on X meant that some advocates became “celebrities”, gaining in one case more than 700,000 followers in 2023. The reach achieved via X was so large that not only other scientists (Kupferschmidt 2022), but also authorities like the Danish Statens Serum Institut (SSI) (Delmelle 2022) and WHO (Ridlo 2022) had to correct misleading claims directly on X. The fame they obtained on X was also invested in appearances in mainstream media, such as an opinion in The Washington Post arguing that Mpox was a pandemic (Feigl-Ding et al. 2022), or arguing for maintenance of restrictions in the UK in the Guardian (Gurdasani 2021).

4.4 Declaration of conflicts of interests

Advocacy is often motivated by interests such as patient concerns, ideological views, or industry lobbying (Wouters 2020; Salooiee and Dagli 2000). Lack of declared conflicts of interests is inappropriate in science as it undermines efforts to evaluate interests and potential biases (Marshall et al. 2020).

We identified two companies led by some of the studied advocates, selling remedies against COVID-19 (CO2Radical, Prescientx), implying a financial conflict of interest that in a context of policy advice and proposals arguably should be declared. Several advocates were associated with the World Socialist Web Site (WSWS), which also argued for zero-covid but in a context of criticizing capitalism and imperialism. Of the seven experts on the WSWS pandemic panel, six were directly associated with WHN, four of whom are among the 20 advocates studied here (World Socialist Web Site 2021). Table S11 lists 15 examples where the advocates featured in interviews or as advisors to WSWS.

In addition to the impact on X and in some media, WHN views were promoted by The Lancet, which published many viewpoints by the studied advocates, such as advocacy to reduce SARS-CoV-2 infection (Priesemann et al. 2021), combat “mass infection” (Gurdasani et al. 2021a), argue for stricter school policies (Gurdasani et al. 2021b), and the full WHN advocacy letter advertising the group (Bar-Yam et al. 2021b). The advocates who advised WSWS did not disclose this even though The Lancet published the WHN paper within a week after the WSWS pandemic panel webinar (Oct 24, 2021)—we note that the chief Editor of the Lancet also praised indieSAGE upon its formation in May 2020 (Horton 2020) and wrote an editorial discussing the case for zero-covid or “No-Covid”, a German version of the elimination strategy, in January 2021 (Horton 2021).

WHN advocates were also strongly recruited into the Delphi consensus panel on COVID-19 published in Nature in November 2022 (Lazarus et al. 2022), together with other advocates associated with European zero-covid (Bar-Yam et al. 2021a). Ten of the 20 advocates studied here were panelists on the Delphi consensus paper, and the core group selecting other experts for the panel consisted of at least 35% active advocates of zero-covid or related advocacy in Europe and North America such as WHN (where the feasibility is arguably much smaller than in some other regions like New Zealand) (Kepp et al. 2024).

The studied advocates often did not report their WHN-advocacy as a conflict of interest: In fact, even in the high-impact Delphi consensus paper in Nature, none of dozens of WHN-affiliated panelists declared their association with WHN, despite Nature’s guidelines on disclosing membership of advocacy groups, as seen from the declaration section of the paper. This suggests that the advocates did not consider the advocacy a potential conflict of interest, with the consequence that readers of Nature were not alerted to the major presence of zero-covid advocates in the panel.

Interestingly, in a letter in BMJ Evidence Based Medicine (Baker et al. 2024), three WHN advocates called a paper on long COVID (Høeg et al. 2024) a “Trojan horse” and accused the authors of association with ideological positions and organisations and evading declaration of interests, while themselves not declaring any interests, neither their zero-covid advocacy or associations with WHN or WSWS (in the case of the first author).

4.5 Implications and moving forward

Scientists and other experts’ intervention in policy making is complicated by the mixture of values and expertise, and given its importance in modern democracies, this phenomenon is in urgent need of robust analytical frameworks (Bandola-Gill 2021). In the phenomenon studied here, public health advocacy may be seen as distributing along two axes, a “risk axis” describing how severe a new risk is, and a “remedy axis” that describes the efficacy of interventions. One end of the threat axis then represents denial of the public health threat (in the pandemic: COVID-19 denial), whereas the other extreme sees the threat as dominating all others issues such as dementia, mental health, cardiovascular disease, cancer, education, individual autonomy, climate, poverty, etc. On the remedy axis, one extreme could be understood as representing the “anti-medicine” or “anti-intervention” perspective. The other represents an obsession with control dressed in a language of science, or scientism (Haack 2012), which can create opportunities for lobbying for totalizing interventions without societal checks and balances that may risk distortion by profit-driven interests.

At its best, public health policy interventions balance the various threats and considers mitigation effects and relevant tradeoffs in appropriate balance (Schippers 2020). For example, education and school-associated health co-variables of children which were disturbed during the pandemic (Soriano-Andres et al. 2023; Vaillancourt et al. 2021) and cancer care deteriorated in some places, disproportionately impacting people with disadvantaged socioeconomic backgrounds (Muka et al. 2023). In contrast, group-interest driven advocacy and lobbyism may evade accountability systems, and rarely consider tradeoffs when proposing policies, leading to net harms to science and society (Hutchings and Stenseth 2016).

Our study maps the views of a side of the advocacy spectrum on the high end of the risk-axis, and on the end of the remedy-axis favoring stricter-than-mainstream policies. This first manifested itself as advocacy for a zero-covid policy, and later for sustained mass mitigation still in the omicron era, when many European countries had already eased restrictions after infection fatality rates had substantially decreased and vaccinations had been offered (Erikstrup et al. 2022). Interestingly, groups such as WHN spurring governments to be more active seemingly had lots of space on social media and in scientific publications, while comments on unintended and often negative effects of government interventions were possibly given less room; this seems of importance and invites further analysis and corrective mechanisms (Schippers and Ioannidis 2022).

Our analysis highlights the concerns raised previously that “bias, increasingly manifest as science-based advocacy, can hinder communication, diminish credibility, and distort scientific evidence” (Hutchings and Stenseth 2016). We show that science-based advocacy can be very intense, aggressive, coordinated and long-term, and that it exists along both axes described above. In other words, we have identified a notable case of a politicized advocacy network and mapped a set of polarizing forces within the scientific community, which may echo current forms of polarization present in the wider society. These findings raise a wider debate on the ethics at two levels: 1) the ethics of the zero-covid policy itself (such as its consequences for lower-risk groups such as children), a question we welcome zero-covid proponents to argue; and 2) at a higher level, our findings raise complex questions on how science advocacy is conducted ethically, including our interactions with scientist peers and use of available evidence.

Although meeting the challenges posed by the mass spread of misinformation is important, censorship of erroneous information and advocacy is a complicated and a potentially harmful response (Niemiec 2020; Shadmehr and Bernhardt 2015). However, in the social media space, some advocates effectively become instruments of censorship themselves, by intimidating colleagues away from public debate: For example, in Sweden, a professor at Karolinska Institute left the debate after online abuse that involved some advocates studied here (Trysell 2021). This method of public intimidation to remove opposition goes against widely-held scientific principles and may distort the public’s view of both scientists, institutions, and the state of evidence on important topics. In addition to direct personal harm, problematic advocacy tactics risk polarizing the scientific community, creating echo chambers and “death spirals” that ultimately harm both democracy and science, not unlike the most problematic aspects of political populism (Schippers et al. 2023).

Academic freedom is vital and must be unconditional within the boundaries of legality. However, like all freedoms, academic freedom of speech can be subject to abuse, such as scientists being harassed because of their viewpoints. Also, opinions and recommendations should be supported by evidence and conveyed in appropriate balance, particularly when involving policy proposals. Our work outlines the importance of socially accountable evidence-based medicine in policy advocacy (Kerridge 2010; Mercuri and Upshur 2022; Baron 2018). Lockdowns and school closures are ultimately political decisions which science can inform on but not answer, despite the ‘certainties’ held by strong belief systems. There is a need for public health debate to be less polarizing (Cooper et al. 2023; Korczak et al. 2022). Therefore, whilst independent activism is desirable as a counterweight to official bodies in a democracy, checks and balances must not be abandoned, even (or especially) in times of crisis.

Based on our analysis, we propose guidelines to ensure that: (1) conflicts of interests, including association with advocacy groups (membership and signatures, but not social media activity which would be very intrusive), are transparent in journals, and in print and digital media; (2) journals limit endorsement of advocacy and lobbyism—as this endorsement may create the false impression of certainty; (3) academic leaders and institutions protect and support scientists against censorship or abuse, regardless of their views; and (4) new guidelines are developed in the field of meta science to analyze and help guide scientists communicating science to the public. These guidelines should consider multi-disciplinary expertise, including the philosophical, social, political, and practical issues of data gathering, use and interpretation. Such guidelines could promote balanced assessments of evidence with societal cost–benefit analyses, ensure acknowledgement of uncertainties and technical and interpretative confusions, and curb the use of affective language, in particular fear, in the articulation of policy recommendations.

4.6 Limitations

The main limitation of this work is that no study of this type—neither qualitative nor quantitative—can avoid selection bias. Yet the large majority of impactful tweets on X, in the sense of having more than hundred likes, often thousands, supports our search and inclusion protocol in terms of relevance. We stress that the quotes should be seen primarily in themselves, with the themes intended to help the reader understand the advocacy studied. Our analysis of the themes, while aspiring to be systematic, may still reflect subjective interpretation and be incomplete, similarly to commentaries in a narrative review. We maintain deep respect for advocacy and the advocates involved, noting that some authors of our paper are also committed to advocacy. Thus, this paper does not reject science-based advocacy, but raises discussion on how to best conduct it, to serve science and democracy best.

The study emphasized the science of the views, rather than the organization, networks, and funding, which the study also does not inform about. The compiled views are not representative of views on other topics by these or related advocates and should be considered in a specific context of the time period and social media site Twitter/X. We also note the organizations mentioned are not responsible for the views of those advocating for them. There are many other zero-covid groups that could have been studied, and the study says nothing about their methods and views but invites further study to understand this part of the advocacy spectrum, and thereby science-based advocacy in general, better.

5 Conclusions

Scientist participation in advocacy is important to democracies but can also have adverse effects. Our study offers a novel insight into a pervasive form of “science-informed” advocacy that was highly visible and impactful during the COVID-19 pandemic. This style of advocacy used high-risk imagery, war-like rhetoric, and affective-based criticism of colleagues holding different views and authorities. It drew heavily from anecdotes with value-based and emotional appeal rather than systematic evidence to back up the need for mass behavioral policies directed at, for example, schools and towards children. Views commonly placed less emphasis on vaccines and much more emphasis on testing, (infection) control, and tools like masks and air monitors.

The primary finding of our analysis is that polarization on social media has large impact even on scientists themselves, i.e., science itself becomes subject to the polarization forces present in the larger society. Our findings raise questions on how to best communicate science, evaluate evidence, comment outside one’s field expertise, declare conflicts of interests, and perhaps especially engage with views of other scientists. We suggest that science-based advocacy needs clearer guidelines for public engagement and science dissemination efforts. In order to do so, researchers and practitioners in the health sciences and the larger science, technology, engineering, and math (STEM) fields need familiarity with the vast body of research in the humanities and social sciences mapping the often unacknowledged political and moral implications of science dissemination. In parallel, stronger coupling between live systematic evidence synthesis and policy advisory as previously proposed (Elliott et al. 2021) may help reduce the sensitivity of democracies to the imperfections of science-based advocacy.

Finally, we emphasize that this study should not be misconstrued as personal criticism directed at the advocates, or at any individuals or groups directly or indirectly referred to in this paper. Rather, it should be read and interpreted as a case study in science advocacy and polarization in the scientific community, a cautionary tale from which we encourage drawing lessons for the future. Although it is regrettably not possible to do this entirely without being candid and specific, we believe that generosity and magnanimity towards all individuals is a cornerstone of any healthy science and policy debate.