November 5, 2020

“Pestilence is in fact very common, but we find it hard to believe in a pestilence when it descends on us,” declares the narrator in Albert Camus’s The Plague. For the second time this year, a pestilence is descending upon us. But, once again, some are finding this fact hard to swallow.

There are important debates taking place about the effectiveness and necessity of restrictions and lockdown measures. Now that the pandemic has taken hold in communities across Europe — and many governments have been incapable of operating effective test and trace systems — there is unlikely to be a ‘perfect’ strategy. 

On one hand, if we entirely reject restrictions, there is a substantial risk of lethal second waves, overwhelmed healthcare systems, and tens to hundreds of thousands of deaths. On the other hand, as ‘lockdown sceptics’ have rightly argued, these measures cause immense economic, educational, social and health damage, and undermine our civil liberties. We are, for now, stuck between a rock and a hard place.

Suggested reading
Parliament's pathetic Covid response

By Paul Embery

There are, however, some people on the lockdown sceptic side of this debate who would deny the existence of this tradeoff; they are flirting with a sort of ‘Covid denialism’. The pandemic, they claim, if it ever existed, is over because ‘herd immunity’ has already been reached. There is no true increase in cases, they insist, just more testing and ‘false positives’

This approach deserves interrogation, we believe. And a key proponent of the herd immunity and false positives theory is the well-credentialed Dr Michael Yeadon, a PhD in respiratory pharmacology and a former leader in pharmaceutical companies. Yeadon surmises, in a post for Lockdown Sceptics, that:

“The susceptible population is now sufficiently depleted (now <40%, perhaps <30%) and the immune population sufficiently large that there will not be another large, national scale outbreak of COVID-19. Limited, regional outbreaks will be self-limiting and the pandemic is effectively over. This matches current evidence, with COVID-19 deaths remaining a fraction of what they were in spring, despite numerous questionable practices, all designed to artificially increase the number of apparent COVID-19 deaths.”

Yeadon’s central assertion is that, by his calculations, a large proportion of the population have already had the virus (32%), have prior immunity to it through exposure to other coronaviruses (30%), or are too young to spread the virus (10%). This, he says, leaves relatively few (28%) susceptible. 

While Yeadon admits in his report that the precise numbers are not “mathematically perfect”, they could, if true, mean we already have herd immunity (typically estimated to require 60%-70% of the population to be immune). We could all get back to life as normal. Everything could reopen; social distancing, testing, and vaccines would be unnecessary; and we could return to our terrible former hygiene habits. 

Sadly, we cannot. Yeadon’s claims are riddled with leaps of logic and illustrated with highly inflated, speculative statistics that most infectious disease experts dispute, and which are being disproved by the explosion of second waves across Europe. To understand why he’s so wrong, we need to address his claims about how many people have already had Covid, the alleged existence of prior immunity, and whether false positives explain the second waves. 

So let’s start by examining how many people have already had the virus. Serological surveys, which assess whether people have been infected by Covid-19 and have developed antibodies to it, provide weak evidence of population immunity, even in some of the worst affected cities. Despite large outbreaks this year, only 23% in New York, 18% in London and 11% of people in Madrid have Covid-19 antibodies. Overall, the REACT-2 survey found just 6% of England had antibodies by July. A US study estimated 9.3% seroprevalence nationally.

On the positive side of the ledger, it is possible that a higher number of people, even some who lack antibodies, are developing longer lasting T cell memory. T cells ‘reactivate’ the immune response when a virus, which has previously been encountered, is encountered once again. They help reduce the severity of symptoms upon reinfection, but are highly unlikely to prevent it — meaning that those who have T-cell memory may still be reinfected and spread the virus again.

The possibility that some patients develop T cells but not antibodies does complicate the calculation of how many people have actually been infected by Covid-19 and are immune. But it is, nevertheless, a leap to claim one-third of the United Kingdom has been exposed and is immune. This is because most people who get infected do develop antibodies: between 91.1% and 100% of cases who tested positive by PCR then developed antibodies; studies show that antibodies last for several months at least. Thus, only a small proportion of people would be missed by serological surveys, by only having a T-cell response to the virus.

Yeadon’s claim that 32% of the population have already been infected derives from his estimate of the chances that someone will die from the disease if they have been infected, i.e. the infection fatality rate (IFR). He estimates the average IFR is 0.2% which, with 43,000 Covid-19 deaths in the United Kingdom, equates to 21.5 million people having been infected (i.e. 32% of the population.) Yeadon admits this “might be a little high”. 

Indeed. It is dependent on a substantial underestimation of the IFR. The WHO has estimated the Covid-19 IFR is around 0.5%-1.0%. A meta-analysis in September put it at 0.68%, and another analysis puts it at 0.53%, while in England specifically, it is estimated to be 1.5%. A Swedish government study put the infection fatality rate at 0.6%, based on a sample of PCR-positive individuals in late March. An IFR between 0.5% and 1%, and 43,000 infections, equates to between 6% and 13% of the population previously being exposed in the United Kingdom and likely to have immunity.

In sum, Yeadon may be right that serological surveys underestimate the extent of previous infection and immunity, but by nowhere near enough to mean we already have herd immunity.

More from this author
When will the Covid-19 vaccine arrive?

By Saloni Dattani

So let’s turn, then, to his claim about prior immunity. Yeadon states a large number (30%) have prior immunity because they have previously encountered other seasonal and endemic “common cold” coronaviruses. It appears that some individuals have T-cells circulating in their blood which are able to react to the novel coronavirus in laboratory conditions, even though they have not been infected by the virus. A study of donor blood specimens in the United States between 2015 and 2018 suggested half displayed various forms of this “T cell reactivity” to SARS-CoV-2, and a German study found reactivity among one-third of donors.

But this reactivity does not necessarily mean, as Yeadon asserts, that these people are immune. First of all, these studies had small numbers of participants, 20 and 37 respectively, meaning they may not be representative of entire countries. Secondly, their T-cells were found to react to virus particles in cell cultures in laboratory conditions specifically; we do not know how they behave in practice, how these individuals would actually respond to an infection by Covid-19. And contrary to Yeadon’s implication, it would be unprecedented for cross-reactive T-cells to prevent an infection entirely, reducing its spread in the population so substantially.

In reality, T-cells reduce the ability of viruses to make copies of themselves over the course of several days and may potentially reduce the severity of disease. And in previous small human challenge trials, participants were exposed to common cold coronaviruses and re-exposed to other similar coronaviruses later, or the same virus one year after, and most were reinfected and developed symptoms. 

For now, there are reasons to believe cross-reactivity may not have much impact on the threshold for herd immunity. There are too many cases in which too large a proportion of a population has been infected to indicate widespread pre-existing immunity to infection. Two large outbreaks on ships this year, for example, resulted in 67.9% and 85.2% of their passengers being infected. By April, 57% of the population of Bergamo in Italy had been infected and developed antibodies to the virus. By July, 54% had in Mumbai, while 55% in Karachi in Pakistan had by September. Clearly, huge swathes of populations are susceptible to infection.

Suggested reading
How the experts messed up on Covid

By Stuart Ritchie and Michael Story

Finally, Yeadon claims that two thirds of those aged 0-11 years old cannot spread the virus (10% of the population). There is, indeed, widespread evidence that young children are not likely to develop disease from Covid-19; but there is scant evidence that they are unable to be infected or spread it. In fact, the evidence is mixed when it comes to how much less likely they are to be infected than adults or how much they contribute to the spread of the disease.

In any case, if people have T-cell immunity from their exposure to other coronaviruses in the past, that is already considered in empirical estimates of the R. The R for the coronavirus has been estimated in a variety of different ways that examine the way the virus is demonstrably spreading in the population — accounting for any pre-existing immunity they might have. If there were isolated groups of people who did not have this pre-existing immunity, however, it would imply that the R0 (and hence the herd immunity threshold) in those groups would be higher than in the populations that scientists have already looked at — in other words, harder to reach.

Besides, the ultimate proof of whether or not we already have herd immunity is being provided by events unravelling right now. The second wave of infections, hospitalisations and deaths all demonstrate that there is a large proportion of people who are still susceptible to being infected.

This is where the second key facet of denialism comes to the fore. The rise in cases, we are told, is a function of greater testing and widespread ‘false positives’, which is when people who are not infected are receive positive results. We are apparently experiencing a ‘casedemic’, not a pandemic. But this claim is crumbling under the weight of new evidence. 

Back in September, Yeadon said that “because of the high false positive rate and the low prevalence, almost every positive test, a so-called case, identified by Pillar 2 [community testing] since May of this year has been a FALSE POSITIVE.” The inaccuracy of tests is so dire that, according to Yeadon last week, we must immediately stop “lethal PCR testing” that is driving fear and restrictions (rather than helpfully spotting cases to prevent outbreaks). 

You could, perhaps more reasonably, have made this claim during the summer, when far fewer people were infected by the coronavirus. But explaining why the UK has uniquely high false positives rates would be difficult, and so would the observation that many other countries have undertaken similar numbers of tests with much fewer positive results.

As Tom Chivers pointed out in September, tests are used more frequently by people who have symptoms, meaning the chances that a person who is tested is actually infected by the virus is higher, which reduces the chances of false positives. Yeadon’s argument would also require ignoring the presence of false negatives (cases where people are infected but test negative), which occurs very frequently in the first days of the illness. 

Suggested reading
Is this the second wave?

By Tom Chivers

In any case, more recent data from the UK dispels the notion that false positives alone explain the rise in cases. This is because the proportion of positive tests has been increasing dramatically across a number of measures, from as low as 0.4% at the start of July to 8% at the start of November. This would mean, even with Yeadon’s claim that 1% are false positives, almost all of 7% are true cases. The increasing proportion of positive tests cannot simply be explained away by false positives, as that would mean testing quality is severely declining.

It is a similar story for the Office for National Statistics infection survey, the most reliable source on community infection in the United Kingdom (because it is a large representative survey of the population). The ONS indicates that the proportion of the population who tested positive during the fortnight they were tested had risen from 0.03% in late June to 0.1% at the start of September, and 1.04% by mid-October. Again, the change is important: even if much of that 0.03% in June were false positives, it is not likely the number of false positives have increased 35 times. 

This can also be triangulated against the ZOE COVID Symptom Study, which uses a large number of daily symptom reports and tests to estimate the total number of cases in the community. They estimate the case numbers have risen from 22,000 at the start of September to over 540,000 by late October, with about 43,000 daily new infections. Imperial College’s latest REACT-1 study, from another large community survey, is more pessimistic, estimating an increase in community prevalence from 0.60% in late September to 1.28% in late October, which would mean about 100,000 new cases a day. 

Over and above the dramatic rises in cases are the rising numbers of hospitalisations and deaths. The time between new cases and a rise in deaths typically takes around four to six weeks: it takes time for the virus to spread from the young to more vulnerable groups, longer for them to develop symptoms, be hospitalised, die, and finally have their death recorded officially. The average time between infections and deaths is around 22.9 days, and deaths are registered and reported in official statistics even later.

In the United Kingdom, we can see that hospitalisations are in fact rising proportionately to new cases, and while deaths have taken over five weeks to start rising, they are certainly on the way up. There are now already over 10,900 people in hospital with confirmed cases of Covid-19 in the United Kingdom. There are also a few hundred people dying a day, and that number is rising. Across Europe, including in Spain, France, Germany, and Belgium, there is a similar story of rising cases, followed by hospitalisations and deaths. Even Sweden is now instigating a new voluntary lockdown in the face of growing case numbers.

Yeadon’s typical response has been to point to the lack of excess deaths in recent months. But this is simply a reflection of delays in reporting. The death registrations from the Office for National Statistics are now showing excess deaths.There were 980 excess deaths in the week ending 23 October, 10% above the 5-year average, after 726 excess deaths the week before. These numbers are near identical to the number of deaths involving Covid (978 and 761, respectively). The ONS have also found nine-in-ten death registrations listing Covid-19 as the underlying (main) cause of death. This is consistent with the 28 day death numbers from Public Health England (which provides daily death count figures), showing deaths picking up from mid-October — again, about five weeks since cases started increasing at a substantial rate in September.

Suggested reading
What we can learn from the Swedish paradox

By Freddie Sayers

The precise relationship between cases, hospitalisations and deaths, as well as the speed of the outbreak, is not the same as it was in March. There are some who have already been infected, and there is ongoing social distancing, improved hygiene, local lockdowns, more frequent use of masks, better testing and tracing, and improved treatments. If we stopped taking precautions, as lockdown sceptics insist, the cases would increase at a faster rate. 

But even though the pandemic is not over, we should not despair. The smartest forecasters are expecting at least one of the dozen promising candidates to produce a workable vaccine by early next year. We have learnt a lot about the virus and how to treat it, with drugs such as dexamethasone, tocilizumab, potentially remdesivir and monoclonal antibodies — making the disease less lethal than it was earlier this year. We also know that the use of continuous positive airway pressure (CPAP) instead of invasive ventilation, anti IL-6, and blood thinning provide benefits to patients. And although treatments would be hard to scale up to everyone who required them, they could make the disease far more benign. 

Meanwhile, effective testing and tracing can prevent outbreaks and save lives — as in Singapore, Hong Kong, South Korea and Taiwan. We are also on the cusp of a large number of cheap and rapid testing technologies — LAMP, antigen strip tests, and, even, breath tests — that could allow life to return more or less to normal even sooner. 

Earlier this year, exceptionalism blinded many Western countries to the coming carnage of Covid-19. Overconfidence and inept bureaucracies led to dramatic failures in border controls and testing, tracing and isolating that could have prevented widespread outbreaks. The result was harsh lockdowns and tens of thousands of deaths.

Europe is in the foothills of a second wave of Covid-19. This has, rightly, led to a renewed debate about the appropriate policy response. It should not lead to a denial of reality.

Join the discussion


  • December 1, 2020
    Life years lost consistently measured at over 10, which is the average. Some obviously lose very little time (as with most diseases), the flip side is that there must be plenty losing a lot. See e.g.: https://www1.racgp.org.au/newsgp/clinical/a... Read more

  • December 1, 2020
    There _is_ some hard(er) evidence coming out on vit d, e.g. https://www.medrxiv.org/content/10.1101/2020.11.27.20239087v1.full.pdf but it is very mixed. Shows _some_ risk reduction for about half the population (women), and shows no effect for men. Read more

  • November 26, 2020
    Interesting, Phil. I look at this a little differently. As I understand it, immunity doesn't mean you don't become infected. It means your body can successfully kill off the virus. Given that 99.5% of people who become infected recover - as many as 80% of them without developing symptoms of any... Read more

To get involved in the discussion and stay up to date, join UnHerd.

It's simple, quick and free.

Sign me up