Is the number of covid deaths inflated? Almost certainly, yes. Does this matter? Of course it does. Apart from being an affront to science, it creates and perpetuates a disproportionate sense of fear about covid, and that in turns drives pathological responses to covid, from national lockdowns to individual hysteria. ‘The bigger the fear,’ observed one of the diarists in the recent compelling BBC history series Berlin 1945, ‘the more ridiculous the excuse.’ It is the fear that puts the pop in the fuel of the Milk Curdler’s blow lamp, it is the fear that powers Dylan’s Idiot Wind, blowing through the flowers on your tomb, the curtains in your room, from the Grand Coulee Dam to the Capitol.

How can Dr No be so sure that covid deaths are inflated? There are a number of reasons. Not one of these reasons finally settles the matter, for that would require a long series of covid viral cultures and post mortem studies, but they do, when taken in the round, satisfy at least the civilian test of the balance of probabilities.

The first reason is historical: we have know for decades that the certified cause of death is often incorrect, often widely so. Studies that have compared causes of deaths identified by post mortem with the corresponding causes on the death certificates have revealed a mark lack of correspondence, in as many as around one quarter of all deaths. Bedside death certification is a far from precise business.

This, of itself, could of course lead to under as well as over inflation of deaths attributed to covid, all other things being equal, but all other things are not equal. We are in the grip of a pandemic, with vast numbers — currently just shy of half a million a day — of people being tested for covid. The government’s corona dashboard, which  defines a covid deaths as one where the patient had a first positive covid test in the 28 days leading up to death, regardless of the actual cause of death, has current numbers of deaths for England and Wales are around 520 per day, or let’s say 3,640 per week. It seems, as the old time GPs used to say, that there is a lot of it about. 

But these 3,640 or so deaths are not certified deaths, instead they are just counts of hapless individuals who happened to test positive in the four weeks up to their deaths. To get the number of certified covid deaths — deaths where covid is listed as one of the causes of death, though not necessarily the underlying cause of death — we have to go to ONS. For the latest available week, to 18th December, 2986 certificates ‘mentioned’ covid, but ONS only recorded 2,557 of these deaths as ‘true’ covid deaths, with covid as the underlying cause of death. Already, the dashboard 3,640 or so deaths begins to look rather inflated, perhaps by as much as 40%, when compared with the ONS number.

Could the ONS 2,557 number also be inflated? Bear in mind we are not in normal times. The pandemic remains front page news, millions are tested for covid every week, and daily death tolls assail us relentlessly. Bear in mind those historical studies showing that death certification is a far from precise business. Could it be that the constant covid pressure creates something called hot stuff bias — or hot stiff bias, as Dr No called it a while back — that nudges doctors to ‘err on the side of covid’, and so ‘mention’ it on the death certificate, ‘just to be on the safe side’?

To get a handle on this, we need to look at all cause mortality. What we find, for the week ending 18th December, is that there were 1,463 excess deaths above the 20-15 to 2019 average. Yet ONS’s own figures record 2,557 covid deaths, which should, given that covid is new, show up as excess deaths — but in fact there are only 1,463 excess deaths. This discrepancy can only be explained on one of two ways. The first is that the 2,557 is inflated, in that within normal range deaths from other causes are being wrongly attributed to covid, through hot stuff bias. The second possibility is that there were 2,557 true covid deaths, but just over 1,000 of those deaths were in people who were expected to die anyway.

This is important. If they were going to die anyway, then whether they died from covid or something else matters not one jot to them — they were going to die, and are now dead. The ones that matter are the 1,463 excess deaths, the ones who had their death hastened, even if only by a short while. These, Dr No suggests are the true covid deaths, the deaths that would not have happened were it not for the arrival of covid. That makes the current corona dashboard figure of 3,640 or so weekly deaths a gross exaggeration, an inflationary mark up by a staggering 248%. Who on earth would want to do such a thing?       

Comments

  1. James Robinson Reply

    Along with: why?

    If only journalism was alive and well, we could have a media which collectively would ask pertinent questions like these, to be broadcast at large.

    Lot’s of supporting evidence to hold the government’s feet to the fire, but who’s going to do that? Individuals like Lord Sumption have tried, but tellingly, him and other opposers, are purposely excluded from the mainstream.

    It amounts to the increasing likelihood of a conspiracy of sorts. Although, objectors should not be regarded as ‘conspiracy theorists’ being the lazy claim by the pathological herd — the herd which are easily distracted by case numbers, flip flop decisions making and obsessing over one thing to the next. Until that changes, the ‘who’ and ‘why’ won’t be asked, never mind attempted to be answered.

  2. dearieme Reply

    I’ve argued for months that only excess deaths are any use, even though they too have their difficulties. It’s just that their difficulties seem to be much less than the difficulties of any other measures.

    Away back in, I suppose, March I said on a blog comment thread that I’d back my ignorant guesses against the mathematical models that so frightened the government. When the models were better then me, said I, I’d stop guessing. Are we there yet?

  3. dr-no Reply

    dearieme – your first paragraph is pretty much spot on, although Dr No wouldn’t go quite as far as to say only excess deaths are of any use, because sometimes other mortality data has its uses. But the key point, that excess deaths are the best, or rather least bad, data we have is crucial.

    Sometimes excess mortality data gets manipulated, as in the EuroMOMO reports, where they use something called the Z score. This is supposed to allow like for like comparison, but there is a problem: the actual calculation used is not, at least not to Dr No, transparent and reproducible, and that always raises suspicions (because if you can’t reproduce it, let alone understand what’s being done, it’s not science, its numerology).

    Which brings Dr No to that branch of numerology know as modelling. Modelling is not science, it is making huge what if Excel spreadsheets and producing scenarios, most of which are absurd. If we were to apply even the most basic science to modelling (theory, ie the model, makes a prediction, then run experiment in the real world to see is the theory (model) predictions are correct), what we would find, have found, time and time again, is that the models are – excuse Dr No’s language, but it is apposite – bullsh*t. Rubbish. Nonsense. Even a quick review of Neil Ferguson’s past form is very persuasive on this.

    Instead of using models to second-guess the future, Dr No prefers to use real data. Inevitably that means a lag (because you have to collect the data, so it is not true real time data) but that small price is worth paying because you are looking at real data, not the imaginings of some ghastly numerological apparatus. Excess deaths are of course real data. To turn an old medical saw upside down and inside out, when it comes to epidemiological data, better a dead patient than a live model…

    James – Dr No is not a natural conspiracy theorist (Occam’s razor and all that) but even he at times has fleeting moments of wondering whether it is a conspiracy rather than a cock up. It’s certainly true that the mainstream media continue to be useful idiots for the government. Only the other day, Mishal Husain on the Today programme crumpled miserably in the face if the great Salisbury vs Salisbury debate, and ended up simpering standard government positions.

    One worries greatly about what has happened to Lord Sumpers, an essential voice of reason in the maelstrom, who has all but disappeared from public life. One hopes he in not locked in the broom cupboard, with a large and colourful cravat stuffed in his mouth.

    • Tish Farrell Reply

      James, this is a fantastic link. Talk about putting things into perspective. Here in rural Shropshire we have obviously been less afflicted, but in my area comprising 2 small towns and limited rural habitation in between, since March there appears to have been a total of 8 deaths with Cov19 mentioned on the death certificate. In our next nearest town, Church Stretton, pop 4,400 including large very elderly cohort, but a busy market centre for surrounding communities, there were 5 in the spring, and apparently none since.

  4. dr-no Reply

    This isn’t death inflation, but it is inflation nonetheless, this time of admissions, from the Guardian (18:49):

    “The latest NHS England figures show a total of 3,145 admissions in England were reported for 2 January, passing the previous peak of 3,099 set on 1 April last year.”

    The 3.099 figure appears to be correct, despite the date, but the 3,145 admissions is just plain wrong. The figure comes from here (daily report link at the bottom of the page), and the number is “Total reported admissions to hospital and diagnoses in hospital“. The true number of covid admissions is considerably less, with an estimate given of around 2,583 from the community (which presumably includes cares homes, though these are also listed in their own table). The 2,583 are “…patients admitted in previous 24 hours for the first time with COVID-19 plus the number of patients diagnosed in hospital in previous 24 hours where the test was within 7 days of admission” so it is probably fair to call some of them covid admissions, but others will be non-covid admissions who caught covid in hospital. So lets say 2,500 genuine covid admissions. Rather less than the Grauniad’s reported 3,145 admissions…

    Edit: And again, at 19:55: “Hospitalisations in England have exceeded the record daily levels of the first wave, with 3,145 admissions reported for 2 January, passing the previous peak of 3,099 set on 1 April last year.”

  5. dr-no Reply

    Googling “3,145 admissions 2 January” gets several national and other media/propaganda outlets on the first page of hits: the Guardian, BBC, Independent and a number of regional papers. Page two has the Telegraph, Standard and more regional papers. Add “daily mail” to the search, and the Mail also has 3,145 admissions, ditto for Sky News, the Express and the Financial Times. The only outlets that don’t appear to be as gullible — or perhaps that should be ‘not as keen to let facts get in the way of a good story’ — are the Times and the Sun, be that coincidence or not.

    For the avoidance of doubt, the NHS England file (link in previous comment)/Table which the 3,145 figure comes from is titled [emphasis added]:

    “1. Total reported admissions to hospital and diagnoses in hospital

    and has the following explanatory footnote:

    “Shows the number of patients admitted in previous 24 hours where patient known to have COVID-19 plus patients diagnosed in hospital with COVID-19 in previous 24 hours.

    Some of the patients were already in hospital when they got a covid diagnosis, and so they are not admissions. To count them as admissions is simply wrong and represents scurrilous scaremongering. That whole swathes of our main stream media are playing this sordid shabby game tells us something that most of us hope we would never have to admit: accurate reporting in this country is dead.

  6. Roger A Reply

    People die every day of the year. The three main causes are cancer, cardiovascular and respiratory disease. While the number of deaths averages out to around 11,500 per week for any given year, winter flu does lead to extra ordinary spikes between September and March.
    The year 2020 was different because it is an indisputable fact that from March until the end of May 2020 a Corona virus epidemic indirectly led to an excess mortality for that time of the year of around 60,000.
    From June the death toll returned to the norm, but come late October and we were told that a second wave is due to hit the UK.
    And sure enough, when I check the stats I can see that from early September until week 51(w/e 18th Dec) the UK experienced a mild increase in the weekly death toll.
    But whether this increase is due to a second wave of the virus is open to debate because the Excel spreadsheets available from Gov. Uk’s Office of National Statistics (ONS) do not provide a detailed breakdown of the cause of death, only those relating to C19 and respiratory disease.
    But this is enough to highlight an anomaly. That during the latter part of 2020, deaths attributed to C19 were rising while deaths from all other causes were significantly lower than for the same period in previous years.
    So I compiled a simple table to highlight this scenario. It shows us that in week 51 of 2020 (w/e 18th Dec.) there were 2,492 fewer deaths – from cancer, cardiovascular and respiratory disease – than during the same week in 2017, which is difficult to accept because one would have thought that ‘clearing out’ the hospitals at the start of the epidemic would lead to an increase in deaths from traditional causes later in the year.

    Covid – ONS stats

    All cause mortality
    wk 51 2017 12,517
    C19 deaths 0
    Deaths from all (other) causes 12,517

    All cause mortality
    wk 51 2020 13,011
    C19 deaths 2,986
    Deaths from all other cause 10,025
    Difference = 2,492 fewer deaths from Cancer and CVD etc

    I admit to ‘cherry picking’ the year 2017 because not only did winter of 17/18 have a slightly higher than average* number of deaths from ‘flu but those deaths passed un noticed by politicians and press alike, other than the perennial line or two in the press re ‘hospital corridors full of patients on trolleys’. Further, C19 was not manifest then.
    The scenario above is, however, typical of almost every week from early September until now.(although the difference is not of the same magnitude) This leads me to believe that at the least, this second wave is over-hyped because if deaths from causes other than C19 were more or less the same for both weeks then surely the deaths directly attributable to C19 for week 51 of 2020 is equivalent to the difference between the total deaths for the respective weeks. That is 492, (13,011-12,517) or 70 odd per day. And therefore nothing like the recorded figure of 2,986.

Leave a Reply to James Robinson Cancel reply

Your email address will not be published. Required fields are marked *