It’s not the virus spreading exponentially but the number of tests, study shows

As the number of Covid-19 tests increases worldwide, what are they telling us? If the mainstream media are to be believed they are showing an epidemic that is spreading exponentially. But what if that is an illusion?

It sounds strange, but a German researcher, Dr Richard Capek, has analysed the tests in the United States and found that although the number of confirmed cases of the virus has been increasing exponentially, that’s because the number of tests has been increasing exponentially. If you look only at the proportion of those tested who test positive then that so far has stayed the same. The latest data from the Robert Koch Institute in Germany agrees with this finding.


What this suggests is that the virus itself is not currently spreading exponentially, possibly because, as a number of virologists have argued, it has been around longer than we have been testing for it.

Dr Capek himself describes it as not so much a viral epidemic but an ‘epidemic of tests’.

A German Virology professor, Dr Carsten Scheller from the University of Würzburg, agrees, explaining in a podcast that the apparent exponential spread of Covid-19 as presented in the media is more to do with the increasing number of tests than with the spread of the virus itself. In his view the Wuhan coronavirus is comparable with influenza and so far seems less lethal.

Two Stanford professors of medicine, Dr Eran Bendavid and Dr Jay Bhattacharya, writing in the Wall Street Journal add their voices to this message, arguing that the lethality of Covid-19 is overestimated by several orders of magnitude and is below that of influenza. The reason for the overestimation, they say, is the greatly underestimated number of people already infected but without symptoms.

But if this is the case, what on earth is going on in Italy and Spain, with hospitals overrun and deaths in the hundreds per day – so many they have now begun registering in Italy’s overall death statistics?

Some have suggested this is not just due to the virus but also to poor air quality, low hospital capacity, a panic-induced collapse of the health system and to the lockdown itself. Dr Gerald Gass, President of the German Hospital Association, has argued that ‘the extreme situation in Italy is mainly due to the very low intensive care capacities‘.

There are currently many mysteries about the virus and why some countries are affected so much worse than others, and many theories competing to explain the differences, with a hope of getting clues as to how more effectively to counter it.

A key piece of evidence is how widespread the virus really is and how many have already been infected. This is where testing is moving next, with the World Health Organisation urging countries to conduct ‘serosurveys’ that test the blood of a sample of the population to estimate how widely the virus has spread. Such surveys will finally answer the crucial question of how common non-symptomatic infections are and thus give an accurate indication of the true mortality rate of the virus.

For the moment, however, the priority must be to get the world back to work again, and to bring the level of public concern back into proportion with the real level of threat. This includes encouraging those most at risk to self-isolate during the epidemic as much as possible, those with symptoms staying at home, and boosting health services so they are better able to cope with the increased demand. It also includes restricting travel from heavily infected regions and screening and isolating arrivals (a precaution the UK still has not implemented). But it doesn’t include such costly extreme measures as overriding basic liberties or bringing the economy to a near halt. Such a ‘cure’ is clearly worse than the disease, not least in its impact on people’s health as social distancing hits people’s incomes and mental health (creating suicide risk) and leaves vulnerable people isolated.

Now that even the lead author of the Imperial College study that first spooked the UK government into extreme action has revised down his estimate of deaths from 500,000 to 20,000 (including those who would die this year anyway), it’s time to get things back into proper perspective.

6 thoughts on “It’s not the virus spreading exponentially but the number of tests, study shows

Add yours

  1. The fact most tests continue to come back negative is strong evidence the virus is not massively more prevalent in society than believed and that the death rate is significant.

    Otherwise you have the believe the virus is incredibly infectious and there are already 18 million cases in Europe!

    If that is true then there’s no need to worry about restrictions affecting the economy because the whole of Europe will be infected in 14 days and the whole business will be over by the end of April.

    Liked by 2 people

    1. Hi Stephen.

      I’m not sure about this; lets imagine another scenario. We are testing for a common flu. It’s enedemic in the population but not everyone is going to have it. We start testing in hospitals. We get alarmed – 10% of people with the flu are dying! (this is the usual death rate if you get the flu in hospital; if you’re ill it’s a really dangerous disease) We increase our testing exponentially, and we get an exponential curve of cases! It’s a pandemic!

      But actually we’re just picking up the usual 5% case load. The actual death rate is 0.02. Some young and healthy people get it – very few – and die. It’s tragic – but sometimes that happens. But before we started testing for the new flu we just assumed that the deaths were part of the usual cycle of winter excess mortality and didn’t think much of it.

      Maybe this isn’t what is happening, but it is one reasonable explanation for the data there. If the Oxford simulation is correct (or even one with a lower % of the population infected) we could have a stable population with the disease already.

      Liked by 1 person

    2. @ Stephen Wigmore

      “The fact most tests continue to come back negative is strong evidence the virus is not massively more prevalent in society than believed and that the death rate is significant.”

      Unless, that is, many of those who test negative for this virus (having been tested because they have symptoms due to a different virus) have already had this virus and developed immunity to it, before catching the other virus that causes symptoms.

      Without testing a sample of the population repeatedly (say once a week), for symptoms consistent with the virus, for the virus itself (even if they have no symptoms) and for antibodies to the virus, over a period of months, I don’t see how we can learn enough about the disease the virus causes to become able to infer anything useful at all from the kind of different-priority-criterion-driven testing that is being done haphazardly, ad hoc, in this or that jurisdiction.


  2. You are doing an important service with this info Will. Don’t know if you saw this article from the Spectator

    Let me tell you, lockdown here in South Africa, with a collapsed economy and no provision for self-employed people like me, and the masses of poor people, is serious. Worse still, if you raise objections or dissent from Covid orthodoxy you are shunned as a heretic. Yesterday we were told of the first death here – my wife tells me that it was later revealed by the victim’s husband that she had a heart condition. One suspect death, in an immunosuppressed population with high HIV and TB rates.

    Liked by 2 people

  3. It looks as though there has been an exponential growth in the number of people who met the criteria for being tested. Why? What were those criteria, and did they evolve over the course of the period to which the data relate, or remain the same throughout?

    It looks as though the virus wasn’t the cause of most of that growth in the numbers meeting the criterion for being tested. So, what was?

    We have no data to inform us how many people have had the virus without developing symptoms which caused them to meet the criteria for being tested.

    My impression is that the haphazard manner in which data is being compiled makes it impossible to formulate a unique hypothesis which alone predicts the data we’ve got. Even if it were, that is cheating. One should use statistics to show a correlation that a hypothesis predicts, rather than search with hindsight for a hypothesis that predicts a correlation one has already observed.

    I don’t see how anybody can learn anything worthwhile from the data.



Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Website Powered by

Up ↑