Coronavirus: figures and fictions
Apr 25, 2020
Snopes has published a generalised lament about the epidemic – not the coronavirus epidemic, but the epidemic of false or misleading posts about it. This article also tackles some of them, and there may be a bit of duplication with some of what I write below, as I’ve added this in a couple of days after starting writing this.
Now, I rarely prune someone from my facebook friends list because they post a lot of stuff with which I disagree strongly – I feel that I need to avoid “living in a bubble” of like-minded people and not knowing what is being said elsewhere. I’ve even stayed a member of a group which seems totally dominated by conservative Christians of various flavours, who spend most of their time explaining why the other (equally conservative but from a different denomination) members are heretics. Goodness knows what they’d say if I shared my own theological stance with them – OK, I have occasionally stuck a barbed response in and got a deluge of proof-texting back which actually only serves to show that the Bible is not univocal. But this is not, on the whole, a theological post…
Among those I have deliberately not unfriended are three or four who, at the moment, are posting just about anything which casts doubt on the science behind coronavirus, the actions of governments (no matter what stance they have taken!) and, of course, Bill Gates. All of them are ferociously intelligent, all of them are, or at least have been, left-leaning politically. And yet they are currently posting stuff which makes them sound like coronavirus-denying Trump supporters…
So, I’ve been sticking the odd rebuttal in on their facebook feeds, which have had mixed results. In particular, yesterday I got a full salvo from one of them of how I’m swallowing the “party line” and aiding and abetting the further takeover of everything by “the deep state” and the demolition of all democratic freedoms. The post I was commenting on there was exactly what Snopes mentions in the link above, the suggestion that Covid-19 is no worse than a seasonal ‘flu.
Now, I am not used to being accused of being gullible in the face of official claims. I tend fairly heavily to the sceptical, and in the case of any and all claims about Coronavirus I tend to fact-check extensively and, where possible, look at real scientific findings rather than the frequently sensationalised versions which appear in the media. The science is too complex for the media to present accurately, even assuming the reporters actually understand it, and a huge amount of it is based on statistics.
At this point, a caveat. While I have some claim to be a scientist (I have a degree in Physics and still from time to time do research work in Chemistry) I am not well versed in the biological sciences. However, I was for some years a Civil Defence Scientific Advisor at a County level, and when the thread of nuclear exchanges lessened, the most obvious civil emergency for us to run scenarios on was a pandemic. This is, therefore, not the first time I have thought long and hard about epidemiology, virology and strategies to limit the effects of umpleasant viruses.
So, allegation 1 – the statistics are false and misleading and therefore we shouldn’t believe anything about coronavirus. I read on the day I started writing this a quotation of “there are lies, damned lies and statistics”, attributing it to Mark Twain, which immediately made me doubt the writer, given that while Twain quoted that in 1907, it’s first recorded use, as far as I know, was in 1891, and even that was probably not original. There is obviously some truth in the statement (much quoted, including by Churchill); statistics need to be read extremely carefully, and issues like sample size and testing methods are supremely important.
Let’s start with the number of deaths from the coronavirus which are touted in the media. This is not in most places a very reliable statistic, though it is not therefore something which can be ignored. One major criticism of it is that hospital deaths are being reported as coronavirus deaths if the patient who died tested positive for CV19 or showed clear symptoms of the disease, though the deaths could well have been primarily for other causes. They are therefore a significant overestimate, it is claimed. However, in most countries, deaths outside the hospital system have not been included (though that is improving), whether because testing is far too limited to include anything outside a hospital setting or because they are not centrally collected in the same way. This would lead to the figures being an underestimate.
The truth therefore is rather hard to assess, though in my country, looking at the total deaths this year compared with an average for the last few years would seem to reveal that at the point at which we recorded around 16,000 hospital CV19 related deaths and estimated another 4,000 outside the hospital system, there were actually around 8,000 surplus deaths over those in previous years. The figure of 16,000 given was, therefore, an overestimate, but not so much of an overestimate as to make the figure anything remotely like trivial. and certainly not “like a seasonal ‘flu” as some have claimed (we had, of course, seasonal ‘flu in the years used as a comparator). 8,000 was around a 60% increase in deaths for the period of the last two weeks. However, if you were to calculate it as an increase compared with deaths over the year to date, as that would include 13 weeks during which the death rate was not massively higher, the percentage increase would be a lot lower, around the 10% mark.
In addition, those figures only included around two weeks of really high death rates; the number-per-day is not increasing radically as I write this, but is also not reducing significantly. Thus, most of that additional 8,000 has occurred within around two weeks and not the three and a half months over which the figure run. Here’s a piece written when there was only really one week of massively greater death toll.
Another statistic which is frequently rolled out is that a very high percentage of the people who are dying are either over 65 or have pre-existing health conditions, which is used to suggest that for the younger and fitter majority, the danger is massively overstated – so why don’t they go back to working as normal? However, in the UK around a fifth of the population is over 65, and around 20% of those under 65 have a chronic health condition, considering government statistics which were prepared to assess the needs of the NHS in more normal times. If anything, those statistics are probably an underestimate, as they include only the people who have sought treatment for such a condition. That is a very large proportion of the population who are at moderate-to-high risk…
There is also a lot of confusion about how lethal the virus actually is. Partially, this stems from the fact that very few countries have managed anything remotely like testing for all those who reasonably might be infected. The UK has until very recently really only been testing patients in hospital, and not all of those; the figures resulting from this have indicated a death rate of around 10%. However, that is 10% of those cases who have been considered severe enough to need hospitalisation, not 10% of the number who have the virus – we have no idea how many of those there are, in truth. Even Germany, which has done at least three times more tests and shows a death rate per identified case of around 1.4% has not tested widely enough for that figure to be regarded as accurate. The larger the number of non-tested infected people, the lower the figure is in reality. A recent study by Stanford University has indicated that the true rate in Santa Clara County California may be as low as 0.12-0.2%, as they found far more people testing positive for antibodies than was anticipated. However, the participants were self-selecting from people who had internet, which casts huge doubt on how reliable the figure actually is, quite apart from being from one of the more prosperous areas of the USA, including Silicon Valley and therefore being somewhat unrepresentative – and I could comment that in the UK, a death rate of 0.12% would result in around 78,000 deaths, and one of 0.2% in around 130,000.
At this point I’ll link to this article, which is particularly useful to illustrate the extent to which Covid-19 is contagious, and (inter alia) indicates that the R0 value for the virus is somewhere around 5 (while that for the seasonal ‘flu is more like 2-3). That would agree with the observation, back in the days when we were able to track contacts, test and contain, that one of the first cases in this country actually infected five others; in the Wuhan example, there were possibly 9 from one encounter in a restaurant. This means that the CV19 virus can spread around twice as fast as does ‘flu, and we can recall that the seasonal ‘flu epidemic in 1918/19 killed around 50 million people worldwide out of a then population of 1.8 billion; that is a death rate for the population as a whole of under 3%. Some countries then weren’t significantly affected, though. It’s worth at this point noting that spreading at least twice as fast as seasonal ‘flu already makes the virus far more dangerous than ‘flu, which at least makes a start on negating allegation 2, that it’s just another ‘flu.
So, having got back to the death rate, one further thing we need to recall is that the death statistics do not show us the prevalence of the virus at the moment; the onset of symptoms of any kind can take between 3 days and 14 days, and typically take at least a week after that before they require hospitalisation; once hospitalised one can expect deaths to start occurring around 7 days later than that. Thus the deaths we are now seeing represent people who were infected 17 days to perhaps a month and a half ago. Hospital admission and testing produces a delay of 10-21 days in and of itself, so we should consider “number of recorded cases” as telling us something about how things were three weeks ago, and “number of deaths” as telling us something about how things were four to five weeks ago.
Doubling in numbers of infected people can take, in the absence of social distancing, 2-3 days. Obviously, most of the world started social distancing around four to five weeks ago, and so the number of deaths occurring now might be telling us approximately where we were at the start of social distancing, and we could have as many as four thousand times as many by now, two to the power 12 or so (which would be so close to complete as to make little difference). The thing is, social distancing reduces the numbers infected per case depending on how draconian the quarantine is, and countries vary widely in how draconian their measures actually are. Do we know with any accuracy what our particular quarantining rules have brought the figure down to? Probably not with any confidence, but if the figure is not doubling every 2-3 days there is obviously a significant effect, and if (as seems the case) it isn’t going up significantly at all, the probability is that we’re fairly close to an R0 figure of 1, which will maintain the number of cases but not increase it.
Again, though, that is to look at now, and our deaths are due to prevalence 4-5 weeks ago. Thus, testing in California which indicates a rather large percentage of people have had CV-19 now doesn’t mean that their current death rate can be compared with that number; it would need to be compared with the number who had the virus around 4 weeks ago, and that would probably be significantly fewer. One should also look at who gets tested, because the figures quoted in California (as they are virtually everywhere) were drawn from people who thought they might have CV-19, and that cannot be extrapolated to the general population who have no reason to get tested.
We don’t, therefore, know with any confidence what your chances of dying if you do contract it actually are, though for a healthy young person that would probably be around 1-5 chances in 100. For a 70 year old with a chronic health condition, it might be far higher – some nursing homes, for instance, have had death rates as high as 25% so far, and might have more deaths to come.
Most of us, of course, told that going to the pub gave us a one-in-twenty chance of dying, would probably avoid doing it… that, I think, finally disposes of allegation 2, that it isn’t any more serious than seasonal ‘flu.
Allegation 3 works against the ideas that it is less serious than is represented, and it’s the suggestion that the tests are not reliable, which some people are backing with evidence that the scientist who invented the polymerase chain reaction process which is used by most of the tests to produce a large enough sample to identify is these days saying that the process produces faulty results. Granted, he has come to believe in astrology and denies that HIV causes Aids, which might lead one to question his judgment…
However, there is clear evidence that not all of the tests which have been on offer do work. One of the things which massively slowed down the US testing programme was that the initial test produced in large numbers was faulty; there is evidence that the UK spent £20 million on faulty tests, and the extent of caution with which they were stressing that they had to ensure all the tests worked in early weeks strongly indicates to me that they were trapped exactly the same way. I am not a biochemist, but note that the standard method of testing involves converting the RNA which is the main content of the virus into DNA and then massively multiplying the amount of that DNA so that it can be readily identified; typically tests identify two or three characteristic elements in the resulting DNA. It is obvious that if a sample is contaminated with any amount of Covid 19 then the result will be to multiply that contamination massively, and this was identified as the fault in some of the tests. What is less obvious is that a bad choice of the two or three characteristic elements can lead, for instance, to false positives for people who have a different (and far less worrying) coronavirus.
Am I worried that the tests now being done are faulty this way? Not really; the fact that contamination was detected in some early testing kits and that a lot of time has been spent in the UK verifying that the tests being used here produce the right result with sufficient accuracy strongly indicates to me that current tests work adequately well – though there will inevitably still be false positive and false negatives, because no tests of this kind are completely error-free.
Allegation number 4 is the conspiracy theory. According to this, the virus has been bioengineered, perhaps in Wuhan, perhaps at the behest of the American government, or Bill Gates. I regard this as a completely ridiculous claim, not least because the characteristics of the virus have been traced back to viruses naturally occurring in bats and pangolin – my conspiracy theorist friends would say that obviously those viruses have been harvested and weaponised. Then the release is either deliberate (opinion seems to be divided as to whether the motive is by certain capitalists to further enslave the general population or by socialists intent on bringing down free market capitalism, which rather negates the thesis in the first place) or accidental.
If deliberate, no sensible megalomaniac is going to have done so without first having a vaccine or a cure, or both. I regard the continuing absence of either as proof positive that there wasn’t a deliberate release.
Could there have been an accidental release of an engineered or cultured virus? I suppose so; that has happened from labs in the past. The thing is, this is just not the kind of virus one would engineer as a weapon; see above, it just isn’t deadly enough. It does remain possible that someone, somewhere, is culturing viruses of the kind found in bats and pangolins. They are certainly being studied, as otherwise the fairly swift identification of Covid-19 as closely related to bat and pangolin viruses wouldn’t have occurred. However, I cannot see anyone tinkering with them to produce transmissibility to and between humans, which is the reason why we haven’t seen Covid-19 before this (it is, after all, a novel coronavirus). Again, what’s the point?
Just to extend this point, you’ll recall that I spent time as a Scientific Advisor and, in the process, ran scenarios involving pandemics. As we were designing the parameters with fictional pandemics, we could choose any combination of lethality, transmissibility, period of infectivity and period before symptoms were seen. What we wanted for an engineered pandemic was a fairly long period of infectivity and period before symptoms were seen; if you only became infectious when you were showing symptoms, you were easily quarantined and if you started showing symptoms more or less immediately you didn’t have enough time to infect others; we also wanted maximum transmissibility. Codiv-19 shows all of those – but the other things we wanted for a weaponised virus were high lethality, preferably similar to Ebola, and a fairly short period during which the virus could endure without a human host (so it didn’t end up contaminating our putative invader), and Covid-19 doesn’t show either of those – it seems to be able to contaminate some surfaces for an alarmingly long time, for instance.
However, if we were designing something to give the political people headaches in working out what the best course of action would be, we wanted a fairly low death rate (as if the death rate is high, stringent quarantine is a no-brainer) and the potential to leave spaces contaminated for quite a while (because that soaked up effort in decontaminating places). Another plus was something which left people needing hospitalisation for quite a long time, because that also put more strain on resources. Covid-19 ticks all the boxes for that – it is remarkably similar to fictional viruses which we were imagining back in the 1980s.
And so the fact that the conspiracy theorists latch on to the fact that various people were running disaster scenarios with very Covid-19-like viruses and conclude that they knew what was coming (and therefore, if you’re a conspiracy theorist, caused it) is a complete red herring. They were using something looking like Covid-19 because it gave the emergency planning people the maximum number of headaches to contend with, just as we were doing forty years ago. The biggest of those was, of course, whether to quarantine (and to what extent) or to let the thing run it’s course and concentrate on ramping up medical facilities. Both of those strategies have been tried with this coronavirus. It remains to be seen what the best course of action was, because this is a real virus, not one where a set of scientists in a back room are giggling while they design a new twist into the scenario to make the politicians sweat a bit more.
Finally, allegation 5 runs something like this: early projections (such as that produced by Imperial College London) were inaccurate, so we can’t believe any of the projections we see, plus, as they were much worse than what we are actually seeing, they were designed to frighten us (and, of course, if they were designed to frighten us, there must be a nefarious underlying motive such as spreading communism or cementing the power of the capitalist elite).
Projections are informed guesses (generally very well informed guesses by very bright people, but still guesses); Imperial only had very early data from China on which to base their initial projections, and much of that has been significantly refined since they published – there are at the date of writing 16 reports published from their coronavirus research team. Those which tend to be quoted are the earlier ones, and yes, they are now outdated. The thing is, these were effectively threat assessments, and the basic principle in emergency planning is “prepare for the worst, hope for the best”. Of course they were going to say “it could be this bad, we should prepare for that” or words to that effect. That is how we should read any report which attempts to foresee what will happen in the future.
—————————————————————————————-
A note here; the assumption behind not quarantining or relaxing social distancing is that having the virus gives you immunity for some period of time, which is a reasonable assumption given that other coronaviruses do produce immunity, though sometimes only for a couple of months. However, there is no clear evidence that this is the case with Covid-19. This is definitely not a feature we would have imagined for our fictional viruses, because again it removes the element of choice from the politicians – if you can catch the virus again immediately after recovering from it, any thoughts of “herd immunity” go straight out of the window. It’s worth noting that the lethality also goes up markedly – the assumption behind the death rates is that you can only catch it once, and if you can catch it two or more times, clearly that figure is going to go up. If you could catch it, say, three times in a year, that might intuitively triple the lethality, but that probably isn’t a true picture; it would perhaps triple the lethality for those who were vulnerable in the first place, but probably not for those who had mild or no symptoms previously. However, that would mean that the only valid strategies would be either trace-and-quarantine for all new cases, which would mean that basically everyone would need to be tested regularly, or a permanent social distancing for vulnerable people – and as I’ve mentioned above, that could well be 25% or more of the population.
If Covid-19 doesn’t produce antibodies which give immunity, there’s one further unfortunate consequence – ‘flu vaccines work by stimulating the production of antibodies which will then fight a ‘flu virus, and if there aren’t any which are specific to Covid-19, production of a vaccine may be impossible. Happily, the most recent information at the point of writing offers hope that this won’t prove to be the case.
A second note is that there are a shedload of other viruses in wildlife which could conceivably mutate into something communicable to humans. We weren’t at all prepared for this one in the UK, and we will virtually certainly see others in the future. If the next one has Ebola-like lethality, unless everywhere in the world is able to identify, contact trace and quarantine (as, in fact, happened with Ebola), we could be looking at reducing world population by half or more. Some historic plagues are estimated to have produced casualties in those kinds of percentages…
So defunding the American Centre for Disease Control and the World Health Organisation, which are the most prominent organisations which actually seek to have the capacity to respond to such a threat is of a level of stupidity which I find it hard to credit.
A third note is that some of those I have deliberately not unfriended are likely to look at this, if they bother to read it, and claim that I am accepting everything Bill Gates says and that I don’t realise that he’s a megalomaniac bent on global domination (or something of the sort). Not so – I loathe and detest the business practices which have led to Microsoft being such a dominant player in information technology, and those are squarely the fault of Gates. I don’t buy the suggestion that his turn to philanthropy excuses the past and present business behaviour of Microsoft. But I do note that he seems to have accepted that he has, in effect, won at the capitalism game and has decided to play a different one, and that that is broadly a beneficial one, and I thoroughly approve of the fact that most of his philanthropic activities involve grants to organisations he doesn’t control – such as the WHO. Which is not “controlled by Gates” (as has often been suggested) any more than, before Trump pulled the US funding of it, it was “controlled by Trump”, even though the US withdrawal actually leaves the Gates foundation as the single largest funder of it.
Sigh. The billionaires of this world have no need to conceal their motivations in doing things – they just need to spend enough money that no-one argues with them. The Waltons and Kochs, for instance, don’t go to much trouble to hide what they are doing…
My last note is that I am far more forgiving of what are widely pilloried as governmental failings in the case of any government which didn’t immediately close its borders and institute a functional programme of tracking contacts and isolating those as well as initial cases than are many of my friends. As I said before, CV-19 is pretty close to the ideal configuration we designed as emergency planners to give our politicians the largest possible headache in making choices. The shock experienced in both the US and the UK of hearing that the tests available didn’t work is just the sort of curve-ball we were coming up with, and that fact probably negated the one obvious “contact trace and isolate” strategy for both, but in any event, the clear danger of a lot of additional deaths (and no, I don’t think there’s any serious chance that this is only about as dangerous as the seasonal ‘flu) could be outweighed by other factors. It’s quite clear, for instance, that concentration on CV-19 in medical systems is prejudicing the health of a lot of people who don’t have the virus, notably cancer patients (who are not getting diagnosed early, which is vital) and heart attack patients (who are not seeking treatment which could prevent another, more serious, heart attack). I could add those with serious kidney complaints and anyone on the transplant list.
Governments could also be excused, in a world which has had an epidemic of libertarianism alongside that of neoliberalism, of worrying that severe measures (particularly at a time when the public were not seeing many cases and even fewer deaths) would just provoke a popular revolt, leading to them losing control of the situation completely. Even in the USA, the population has seemed to have been far more willing to bear with measures which infringe their liberty very considerably than many might have predicted (certainly, after I had thought that Mrs. Thatcher had effectively killed off communitarian spirit here, with her “there is no such thing as society” approach, I have been agreeably surprised by how few people are disobeying the instructions). Yes, if they were thinking that, they were wrong – but that would have been very difficult to predict in February…
All that is without considering what the damage attendant on crashing an economy could be – and that isn’t just in terms of money, it is also likely to cost lives in its own right in the future. The Conservatives’ austerity programme from 2010 onwards, for instance, caused a very significant number of deaths in the ensuing five years (and, to a lesser extent, since then), and there is a strong probability that recovering from the astonishing amount of deficit spending in the last month or two and rebuilding after demolishing a very large number of businesses will need at least that amount of austerity, or an abandonment of neoliberal policies completely. That doesn’t seem particularly likely, though I live in hope. I’ve seen economists argue that recovery from the 1918-19 ‘flu pandemic took until around 1940, for instance, and that indicates what “just another seasonal ‘flu” can do, without a vaccine. And there is currently no viable vaccine for Covid-19.
As a tailpiece, I link this, which frankly makes most of what I’ve said above redundant.