'Unprecedented' appropriately described 2020. The public thought so – the term was awarded the People’s Choice 2020 Word of the Year by Dictionary.com. Its impact became diluted as 2020 ploughed on with yet more unfathomable shocks. The year in science, however, was certainly without precedent.
In January 2020, the virus behind a cluster of pneumonia-like illnesses in Wuhan, China, was identified: a ‘novel coronavirus’ belonging to the infamous group of single-stranded RNA viruses that had already wreaked havoc during the SARS pandemic of 2003 and the MERS outbreak of 2015. Scientists, aware of the implications of the term ‘novel coronavirus’, shuddered, then engaged.
By 11 January, a Chinese-Australian team had posted SARS-CoV-2’s genetic code online. Within a day, immunologists Barney Graham and Jason McLellan had deciphered the sequence of the virus’s ‘spike’ protein and handed it to Moderna – a small biotech firm that specialises in RNA vaccines, the likes of which had never been brought to market.
By mid-March, as researchers were still learning that virus-laden aerosols could allow human–human transmission, candidate vaccines were already being marshalled into early-phase clinical trials. At the start of the Northern Hemisphere summer, there were more than 270 candidate Covid-19 treatments in clinical trials within the USA alone. By mid-November, news emerged that Pfizer/BioNTech’s vaccine showed above 90 per cent efficacy against Covid-19. Twelve days later, Moderna announced the same.
Never before have the gears of science shifted so rapidly; and never before have our daily lives hinged more upon the determinations and conclusions of scientists. Whatever emerges from this mystifying chapter, science is unlikely to be the same again.
A REVOLUTION UNFOLDING
Writer Ed Yong describes science as a ‘stumble towards ever less uncertainty’. As SARS-CoV-2 began charting its way across the world, pressure mounted on scientists to quickly elucidate its biological, epidemiological and medical characteristics. Uncertainty was no option.
In 2020, around four per cent of the world’s research output was devoted to SARS-CoV-2. Usually, the bulk of this output would have been published in scientific journals – a process that features multiple rounds of peer review. Many scientific papers are released months or even years after the initial research is conducted and is often accessible only to those with subscriptions to the journals in which they’re published. In the urgent times of the outbreak, scientists turned to ‘preprints’ – a way of publishing research that short-circuits the lengthy peer-review processes of scientific journals, quickly delivering insights to the public domain. ‘With conventional publishing, peer review comes first, distribution second,’ explains John Inglis, executive director of Cold Spring Harbor Laboratory Press (CSHL), one of the chief publishers of Covid-19 research. ‘With preprints, it’s distribution first. Preprints can then still go through the process of peer review, but in the intervening time, the community can build on the work.’
In 2013, Inglis helped to co-found bioRxiv (pronounced ‘bioarchive’), the top preprint server for biological research, and then, in 2019, medRxiv, the medical equivalent. The servers are run by a collaboration among CSHL, Yale University and the British Medical Journal (BMJ). In the first six months of the pandemic, more than 7,000 preprints related to Covid-19 were released, nearly 80 per cent on bioRxiv and medRxiv.
This represents a significant acceleration. The number of preprints posted on medRxiv jumped from 240 in January to 2,400 in May; the site’s web traffic increased from one million monthly users to 15 million in the space of a month. ‘The scientific evidence accumulated far faster than would have been possible with the traditional publication system,’ says Richard Sever, another of bioRxiv and medRxiv’s co-founders. ‘People are alive today who wouldn’t have been, had the rapid dissemination of medical data not been possible,’ he adds, pointing to the preprint from the RECOVERY trial in June, which showed that dexamethasone (an existing medication) could reduce mortality in patients with severe Covid-19. ‘Medics were able to quickly see that data, because of the preprint servers, and say, “Okay, we should immediately switch to having this as standard of care”.’
As preprints allowed results to be shared quickly and openly, international institutions were better able to collaborate. ‘In a pandemic, researchers in so many geographical locations are trying to solve the same problems. They need to compare experience, particularly in a clinical context. The ability to do that quickly is paramount and preprint servers are made for such moments,’ says Inglis. On 27 March, a study by more than 80 researchers at the University of California, San Francisco emerged as one of the first to describe the method that SARS-CoV-2 uses to enter human cells. The work was published as a preprint on bioRxiv and simultaneously submitted to the journal Nature. In the two months it took for Nature to accept the study, the authors had already sent out genetic material to 500 research groups around the world, allowing more scientists to replicate and build on the initial findings.
‘If there wasn’t a public health crisis, those authors may have just sent the data to respected scientists within their network, reinforcing the cabal of researchers who have a leg up on everything, see the latest results and get ahead,’ says Jeffrey Shaman, professor of environmental health sciences at Columbia University and one of the most published researchers on Covid-19’s spread. ‘Publishing the research as preprints, they’re agnostic; they’re allowing anybody with the skills to come in and build on the science.’
Preprints aren’t new – they’ve been used by physicists and mathematicians for decades. It’s their adoption in the infectious-disease space during the Covid-19 pandemic that’s significant. By contrast, just five per cent of published papers were shared as preprints during the 2014 Ebola and 2015 Zika outbreaks.
‘I think the preprint acceleration can and will continue,’ says Sever. Surveys are encouraging. A November 2020 poll conducted by Frontiers in Public Health – the largest academic survey in existence – found that, since the outbreak, 29 per cent of researchers are more likely to deposit work on a preprint server.
In a letter released on 29 January, more than 500 researchers called for the publication of all SARSCoV- 2 genomes on a triad of databases that would allow completely unrestricted sharing of genetic data. Many hope that the preprints and open-source data needed to advance Covid-19 research will help to liberate data and resources in other areas. Some 44 per cent of scientists from diverse fields of research now favour open-access journals as a result of the pandemic; 45 per cent will share their data in the future. ‘There have been a lot of voices saying that if we can do this during a pandemic, why should it be any different when we’re looking for Alzheimer’s or cancer cures – even in global warming research?’ Sever adds.
A shift towards open-access publishing is already manifesting in research funding. The EU’s revamped Horizon Europe programme – which will make available €95.5 billion to scientists over the next seven years – now mandates that recipients publish their work either as a preprint or through an open-access journal. However, Theodora Bloom, executive editor of the BMJ – a key publisher during the pandemic – is sceptical that the principle will hold for industry-sponsored research. ‘It was in everyone’s interest to collaborate and solve Covid-19 together,’ she says. ‘Realistically, 50 per cent of medical research is done by pharmaceutical companies and large amounts of geological research is sponsored by oil companies. I’m not sure that they will see it in their best financial interest to continue cooperating to the extent that we’ve seen during Covid-19.’
THE LAYERING OF THE TRUTH
The scientific community’s embrace of preprints has meant that the public has been kept up to date with the latest research, often in parallel with scientists themselves. It was the year that laid bare the mechanics of the scientific process for all to see. ‘With the pandemic, scientists have been speaking from the position of best understanding, which evolves over time,’ explains Inglis. ‘So this year, the public have a greater understanding of how science progresses in small increments.’
Perhaps the most perplexing of SARS-CoV-2’s numerous biological quirks is its ability to spread asymptomatically – a fact that became clear gradually as many studies, data sets and analyses became available. A German study, later described as ‘flawed’, first hinted at the possibility of asymptomatic transmission in early February, but the evidence was deemed insufficient. Later in February, University of Texas scientists reported that the virus could have a ‘negative serial interval’, meaning that some infected people showed symptoms before the person who had given it to them. At the time, the virus’s clandestine behaviour was perplexing and a full scientific consensus was only reached much later: case reports of healthcare workers transmitting the disease without symptoms emerged in April and robust models of asymptomatic spread only emerged in May. Early estimations of asymptomatic spread were as high as 80 per cent of infections, but have since been revised down significantly to between 17 and 20 per cent.
Another salient example, says Bloom, is the vaccine data. ‘The first vaccine results showed that they prevent severe disease in most who are vaccinated, but didn’t tell us anything about whether they prevented transmission.’ On 3 February 2021, Astra-Zeneca announced that its vaccine could reduce transmission by two-thirds – months after interim efficacy data of their phase III trial were released. ‘Now, we’re getting closer to that information. It has really been little steps at a time,’ says Bloom.
These gradual steps have sometimes caused frustration. Preprints may have accelerated research, sparked collaborations and liberated data, but in the periods between clinical trials, public patience hasn’t always been forthcoming. ‘The problem is that during a public health emergency, people want to cut to the chase and know how to respond – unfortunately, there’s no way of doing that,’ says Shaman.
Overall, however, Bloom thinks that the revelation that scientific knowledge is, by its nature, accumulative will be beneficial for public engagement. ‘On the whole, I think it can only be good that people have a more realistic sense of how science works, and what it means when scientists assert something from the position of best understanding,’ she says.
To overcome the pandemic’s various challenges, we’ve solicited the expertise of public health experts on testing regimes; immunologists on vaccine development; and supply-chain experts on vaccine delivery. This layering and combination of expertise has been essential. ‘One of the dangerous myths that I hope the pandemic will dispel is that of the lone maverick scientist who has the answer,’ Bloom says. ‘The answer is almost always that no one person has it completely right; research is like building a large edifice brick by brick.’
THE DANGERS OF MISINTERPRETATION
There’s a darker side to this gradual process – one that has seen scientific deliberations and self-corrections leapt on by opportunists looking to advance alternative theories. The origins of the virus, public health measures such as mask mandates and potential treatments (who could forget the infamous hydroxychloroquine?) have all been debated and sometimes politicised.
‘People from partisan media outlets find this stuff and use a single study as a cudgel to beat the other side,’ said epidemiologist Carl Bengstrom in an interview that appeared in The Atlantic. A study from Stanford University published on medRxiv in April serves as an example. After testing 3,300 volunteers from Santa Clara County in California for antibodies, the researchers concluded that SARS-CoV-2 was less deadly than expected and that stringent lockdowns were overreactions. Statisticians took to social media to point out the study’s flaws, but it was picked up and promulgated by anti-lockdown groups.
‘Scientists like me like to take time to be more methodical and ensure the scientific consensus is correct. But when you have such attention on the results of the scientific process, it can be very challenging, particularly when set against the backdrop of a political milieu,’ says Shaman.
One preprinted study, released early in the outbreak, quickly became embroiled in public hysteria. Published on 31 January 2020, the study speculated that portions of the genetic code of SARS-CoV-2 bore an ‘uncanny’ similarity to HIV, which was latched upon by conspiracists pedalling theories that the virus was engineered at the Wuhan Institute of Virology. Scientists quickly pointed out flaws in the study’s methods and conclusions through social media, which led to the preprint’s retraction within 48 hours.
Advocates of the preprint acceleration argue that it enables scientists to better collaborate – that a new form of scientific communication is emerging; one that breaks disciplinary boundaries and broadens the audience of scientists able to critique or corroborate a study’s findings, more so than through traditional journals. Preprint sceptics, however, are concerned that by accelerating the release of research into the public domain, it can more quickly warp public opinion.
Screening processes by preprint servers are in place as the first line of defence against bad science. But, given the degree of misinformation deliberately peddled during Covid-19, many have argued that the preprint acceleration needs to be paralleled by public awareness efforts, helping the public and journalists to understand what preprints are and how to interpret them. New best practices for scientific communication by journalists is one suggestion; rapid-response review interfaces – platforms that bring journalists into contact with independent scientists to discuss research findings before they are communicated – are another. With such measures, the benefits of expedited research publishing might be afforded without the risk of media distortion. But until then, we may all need to be more critical of our unregulated cyberspace. ‘What we really need now is clarification about the sources to listen to – those to trust and those not to,’ says Inglis. ‘In reality, preprint servers and academic journals alike host both bad and brilliant science.’ To him and many others, preprints aren’t the problem; rather it’s those who deliberately misrepresent their findings. ‘It’s important to distinguish between the scientific process being conducted in a more public way, which is an expression of honest inquiry and the nature of uncovering new knowledge, and the deliberate manipulation of people’s beliefs.’
LED BY SCIENCE
If 2020 highlighted the fact that no lone scientist has the answer – that the entire scientific community must be called upon in times of crisis – it also uncovered the perils of overlooking the scientific consensus that emerges.
Shaman and his colleagues at Columbia University developed a simulation of the early outbreak in the USA between February 2020 and early May. Their results, posted as a preprint on 21 May, found that 35,000 lives could have been saved in the country if it had acted earlier on scientific evidence, which would have more than halved the death toll as of 3 May. Nationwide, the team found that 56.5 per cent of reported infections and 54 per cent of reported deaths could have been avoided had control measures been implemented just one week earlier, as many scientists had advised.
Meanwhile, then US President Donald Trump busied himself with dismissing public health officials as ‘idiots’. ‘One day, it’s like a miracle, it [the virus] will disappear,’ Trump told the American people on 27 February. In a recorded interview on 19 March, Trump was heard admitting to suppressing information: ‘I still like playing [the outbreak] down because I don’t want to create a panic.’ The same day, the US Centers for Disease Control confirmed 4,937 more cases, bringing the count to 13,133. Later, Trump suggested that Covid-19 might be treatable with disinfectant or sunlight; he called the preprint showing the ineffectiveness of hydroxychloroquine a ‘Trump enemy statement’ and used the malapropism ‘herd mentality’ when speaking highly of the Great Barrington Declaration, a proposal to pursue ‘herd immunity’ by allowing SARS-CoV-2 to circulate freely – an approach that a group of top epidemiologists writing in The Lancet called ‘a dangerous fallacy unsupported by scientific evidence’.
In a survey of more than 25,000 scientists across the world conducted by Frontiers in Public Health, just 18 per cent of US scientists agreed that scientific advice had been used to guide decision-making during the pandemic. The UK didn’t perform much better. Only 24 per cent of scientists agreed that their advice has been adequately heeded by policymakers. ‘In the UK, the government claims to have been following the science, but has the worst death rate in the world per 100,000, and has consistently been late and often downright wrong,’ says Bloom. Led by epidemiologist Neil Ferguson, a team at Imperial College London developed a model showing that, if no action was taken, the spread of SARS-CoV-2 would result in 500,000 deaths in the UK and that the R number (the average number of secondary infections produced by a single infected person) at the time was between 2 and 2.6. Yet, the UK government, which would have known about the report in advance of publication, waited a week after its release before enforcing a nationwide lockdown on 16 March.
‘Other countries that were successful, like those in the Pacific Rim, had to deal with SARS and MERS. They had science-based playbooks for how to control respiratory-disease outbreaks,’ says Shaman, pointing to Thailand, Taiwan, South Korea and New Zealand as examples. In New Zealand, Jacinda Ardern’s government closed the country’s borders and imposed a national lockdown and stringent quarantine measures early in March 2020, and promptly again in February 2021, after three cases of the UK variant were detected. New Zealand appears at the top of the Frontiers in Public Health survey. South Korea, thanks to free, rapid testing and heavy investment in contact tracing, shut down chains of infection by following scientific playbooks established after the 2015 MERS outbreak.
REFORMING SCIENTIFIC INTEGRITY
When it comes to the relationship between scientists and governments, it’s the USA that provides the most interesting case study, partly because the presidential election of November 2020 took place in the context of the pandemic – when the relationship between the administration and its top scientific advisers was fraying.
In October, weeks before the election, a coronavirus sub-committee within the US House of Representatives listed 47 instances in which government scientists had been sidelined or had their recommendations altered during the pandemic. ‘If science doesn’t fit a political goal, then we often see decision-makers try to manipulate it,’ says Jacob Carter, a public health researcher who specialises in scientific integrity in democracies at the nonprofit, Union of Concerned Scientists.
Many issues were at stake at the ballot box, but it’s widely believed that Trump’s handling of the pandemic contributed to his downfall. ‘The anti-science stance of the prior administration absolutely led to Biden’s arrival in office,’ says Shaman. Polling surveys have indicated the same. An October survey by the Pew Research Center found that 82 per cent of registered voters who supported Biden said that the outbreak would be ‘very important’ to their vote, compared to 23 per cent of registered Trump voters. Just 14 per cent of Biden voters thought that the pandemic had been blown out of proportion, compared with 63 per cent of Trump voters.
‘We’re already seeing the Biden administration take good steps for science,’ says Anita Delsikan, another public health researcher at the Union of Concerned Scientists. On 27 January 2021, Biden initiated a review of scientific integrity policies in government. It will document instances in which ‘improper political interference’ has suppressed or distorted research and data, and give agencies the chance to re-form technical advisory panels axed under Trump. ‘We had so many political interferences, so this is a great move for scientific integrity,’ says Carter.
Many hope that the political shift in the USA will ensure that greater respect for science’s value to society will be a legacy of Covid-19. Shaman, who identifies himself as a ‘realist’, is hopeful that the events will spur more evidence-based thinking. ‘There’s certainly an imprinting and a sensitivity to pandemics that’s going to persist for some time,’ he says. ‘I think that one effect is that people will want to pay more attention to the evidence on challenges such as climate change.’
For all its horrors, the pandemic has been a remarkable chapter for science. With research accelerating, knowledge of SARS-CoV-2 has grown at breakneck speed. Vaccine development has been a triumph. The public has witnessed science in motion. Many have spoken up for scientific integrity, as epidemiologists and infectious disease experts have been lionised. But perhaps the greatest change to science is yet to come. Perhaps those who’ve watched scientists grapple with this disease have been inspired by science’s finest hour. Perhaps a new generation will feel more willing to trust in the scientific enterprise to guide us through the challenges to come.