October 26, 2020

Masks work? NO.” Scott Atlas, a member of Donald Trump’s coronavirus task force, wrote a tweet the Saturday before last that opened with these words — only to find it deleted by Twitter a day later. In the offending tweet, Atlas had written that the World Health Organisation (WHO) and the Centers for Disease Control and Prevention (CDC) had both recommended against mask-wearing, and had in fact argued that it could cause “many harms”.

This violated Twitter’s policy against Covid-19 misinformation, so it had to go. As the New York Times noted, both the WHO and the CDC do recommend wearing masks to protect against the coronavirus.

The only problem is this: mere months ago, both the WHO and the CDC both had argued — in fact, argued very confidently — against the use of masks, just as Atlas said. It won’t do to ignore this rather inconvenient fact, as the New York Times did, or to pretend the statements were never made. That’s because the Atlas affair provides an object lesson in how overconfident claims by experts, especially on issues as fraught as Covid-19, can come back to bite them.

Throughout the pandemic, experts have been all too willing to make claims about the virus that bordered on the hubristic. It’s easy to forget, for example, that the first months of 2020 saw widespread and confident scepticism of any risk from the novel disease that was sweeping Wuhan. As an infamous (and since updated) BuzzFeed article advised readers not to worry about the coronavirus, but instead to worry about the flu, leading political figures found it difficult to resile from their initial “don’t panic!” stance, even as the pandemic began to take hold.

As late as 3 March, New York mayor Bill de Blasio, supported by his health commissioner, told residents to “go on with your lives + get out on the town despite coronavirus”. This plea for normalcy was made on the day that global coronavirus cases reached 90,000, and a mere month before the total Covid deaths in New York City exceeded those of 9/11. 

It shouldn’t be damning to make a bad call — nobody gets every prediction right; any professional forecaster will tell you that the best you can hope for is to be wrong less often. But the fundamental problem is the high degree of confidence with which the bad calls were expressed. Stating a position without the attendant uncertainty makes it very difficult, should the situation change, to update your views without losing face. In a crisis, slowing down that view-updating process could cost time, money, and even lives. Even after the update is made, those who heard your original, dogmatically-expressed view might have lingering doubts — or might even use it against you.

Few areas better illustrate the pitfalls of expert overconfidence than the question of facemasks.

The initial messaging on masks from the WHO emphasised that masks were needed in medical settings where infected patients were likely to be coughing directly on or near healthcare workers.  The message was echoed by the US Surgeon General on Twitter, who exhorted citizens to “STOP BUYING MASKS!”, since they “are NOT effective in preventing [the] general public from catching #Coronavirus”. Both emphasised the need to preserve mask supplies for healthcare workers.

This might have made sense given the assumption that the coronavirus behaved like influenza — the model disease most countries were using to plan for a pandemic. If that was the case, it was thought, masks would likely be ineffective at preventing transmission outside of hospitals. Indeed, studies of facemask use for influenza had found mixed results. This, coupled with a norm that health organisations should require ultra-solid evidence before making recommendations, somewhat paradoxically meant that the “masks don’t work” message became ever-more-confidently projected — even as the evidence behind it looked shakier and shakier.

In the UK, the public health community embraced the anti-mask message more strongly than most. For example, in care homes, where around half of the UK’s Covid deaths have occurred, staff were informed by the government in February that “face masks do not provide protection from respiratory viruses such as Covid-19 and do not need to be worn by staff” (incidentally, the following statement from the same document — “[i]t remains very unlikely that people receiving care in a care home or the community will be infected” — stands as one of the most tragic of the entire pandemic).

On 4 March, Chief Medical Officer Chris Whitty argued masks would reduce the risks of the non-infected “almost not at all” and said he would not advise wearing them. The medical website patient.info noted that its readers might see people wearing masks out and about. “Don’t worry if you haven’t bought one”, they wrote. “The masks are fairly ineffective for the average person. Only people caring for infected people and the infected people themselves needs [sic] to wear masks.” Even volunteer sewing groups were asked to make morale-boosting patterned scrubs for NHS staff instead of using their skills to boost the mask supply.

The anti-mask campaign soon escalated beyond messaging. In early March, two businesses selling facemasks were banned from producing adverts that claimed their products offered protection from “viruses, bacteria, and other air pollutants”. The adverts, it was said, were “likely to cause fear”. This was on the advice of Public Health England, who didn’t just not recommend masks, but claimed they might raise the risk of transmission, since they were “likely to reduce compliance with good universal hygiene behaviours”. Professor Stephen Powis, the National Medical Director of NHS England, said that the firms selling facemasks who linked their product to the pandemic were “callous” and “outright dangerous”, and that advertising masks in this way had “rightly been banned”.

The campaign to keep British faces uncovered culminated in a video released by the UK Government on 11 March, where the Deputy Chief Medical Officer, Jenny Harries, told the Prime Minister in no uncertain terms that wearing a facemask was “not a good idea and doesn’t help”.

From the vantage point of mid-October, this seems more than a little odd. Masks are now mandated on public transport and in shops, are worn prominently in public by politicians and their public health advisers, and have become a normal part of life. The US Surgeon general who had so vehemently decried the buying of masks on Twitter now has a picture of himself wearing one at the top of his page. Once again, the issue was not necessarily that the original messaging was wrong — as we stated above, the evidence was patchy, and decisions still have to be made.

But the sheer effort undertaken to double down on what could only ever have been an uncertain message helped to narrow the range of options, and likely slowed the eventual change in policy as studies, models and reviews arrived to bolster the case for masks. Overall, whereas the data is far from knockdown, and there hasn’t yet been time to run and publish high-quality randomised trials, the observational and other evidence does point towards a protective effect of masks for this disease.

Had the evidence for the experts’ views been rock-solid, some impatience with public disagreement might have been justified — particularly given the emergency situation. As it was, though, those who even gently queried the evidence quickly became subject to ridicule.

In a difficult-to-watch clip from Jeremy Vine’s Channel 5 show in mid-March, the model and businesswoman Caprice Bourret was upbraided by the show’s medical expert, Dr. Sarah Jarvis, after suggesting that a country-wide lockdown might be helpful. Jarvis perfectly encapsulated the expert attitude with which we’re familiar, telling Caprice that “unless you have read every scientific paper… you cannot argue with me on that. You can have an opinion, but it’s not a fact.” Undeterred, Caprice went on to note that in some East Asian countries that had enjoyed (and still enjoy) a better Covid trajectory than the UK, a large proportion of citizens wore surgical masks. “…which make no difference at all”, Jarvis snapped in rebuttal, laughing at Caprice’s attempts to respond.

Other experts followed suit in suggesting that coronavirus worries — and in particular, disagreements with expert advice — were indications of ignorance, or of low-quality thinking. The American Psychological Association’s Senior Director of Practice, Research and Policy, Lynn Bufka, told Time magazine in March that mask-wearing was a “superstitious behaviour”: “Even if experts are saying it’s really not going to make a difference,” she argued, “a little [part of] people’s brains is thinking, well, it’s not going to hurt.”

One of us has written previously about the overconfidence of experts in the realm of psychology — in particular, how many writers over-applied our shaky, laboratory-based evidence on how people deal with risk to this real-world, unprecedented pandemic situation. Harvard’s Cass Sunstein self-assuredly wrote in February about “probability neglect”, the tendency of people to exaggerate a threat that “triggers strong negative emotions”. This, Sunstein argued, explained people’s “excessive fear” about the coronavirus pandemic — essentially, they just couldn’t properly think it through.

Sunstein, for his part, soon performed an impressive volte-face, writing a month later about how “we can’t be too careful” about coronavirus — without once acknowledging his previous downplaying of the threat. Perhaps if he’d admitted how drastically wrong he’d been — that is, if he’d projected somewhat less confidence — he wouldn’t have subsequently been hired by the WHO to chair their “Technical Advisory Group on Behavioural Insights and Sciences for Health”.

Research that looks at people’s beliefs about Covid often bolsters the narrative of “the experts” versus “the unthinking”. A recent paper from academics at Cambridge surveyed people’s beliefs in Covid misinformation. Some of the beliefs they examined were bonkers conspiracy theories (“the new 5G network may be making us more susceptible to the virus”), and some were just daft (“breathing in hot air… (e.g. from a hair dryer) kills the coronavirus as it can only live in cool places”). They were all taken from the WHO’s Covid “Myth Busters” page, and they are all indeed myths that it would be silly to believe. But it can’t help but feel ironic that, at the time these academics accessed its website, the World Health Organisation was spreading its own confident misinformation: just one click away was the WHO’s page telling people not just that they shouldn’t wear masks if they weren’t directly looking after a Covid patient, but that it might actively be dangerous to do so.

Imagine travelling back in time to various points in 2020, knowing everything we now know about the coronavirus and its spread. In February, as you warned the world of the grimness to come, some experts would have told you that you were suffering from “probability neglect”, and that you should be more worried about the flu. In March, your view that masks would be helpful in controlling the disease would have marked you down as an agent of misinformation — psychologists might have suggested techniques to cope with your superstitious thinking, while Public Health England and the WHO might have said you were increasing the risk of spreading the virus.

Equally, if a time-traveller with the approved public health opinions of March 2020 journeyed just seven months into the future to the present day, their mainstream anti-mask position would be viewed with similar derision: they would perhaps have their thinking explained with regard to their gender, or might even be labelled a likely sociopath.

Which brings us back to Scott Atlas and his “Masks work? NO!” tweet. Although his views might seem absurd with our October-2020 hindsight, it wasn’t difficult for Atlas to find a rich seam of recent, strongly worded expert quotes to back up his anti-mask case. This wasn’t inevitable: although being misinterpreted or misrepresented is an occupational hazard for any expert making public statements, if so many experts had avoided going so far beyond the evidence, Atlas and those like him would’ve had a harder job.

The “expert view” vs “the misinformation” is a false dichotomy. It can be tempting to offer only certainty in times of crisis — but part of being an expert is knowing when to be uncertain, and being strongly aware of the provisional nature of our knowledge. Trustworthiness doesn’t just come from being right, but from communicating the limits of the evidence, and regularly updating one’s view in light of new data and analysis.

Overconfidence from the experts, coupled with a willingness to denigrate and even pathologise those who publicly dissented, might have made it harder for us to change course during the pandemic, costing us precious time that we couldn’t afford. If experts fail to reckon with the inevitable uncertainties of our current times, we risk delaying the next crucial update — or worse, overlooking it altogether.