At long last, Dr No has found a way of estimating official covid death inflation. By official covid death inflation, he means the extent to which the authorities have deliberately inflated the number of covid deaths to increase, in SPI-B’s infamous words, “the perceived level of personal threat…among those who are complacent”. Let us be absolutely clear: this is the government, and its agencies, deliberately massaging official statistics, to scare the population witless, into submission and compliance. In SPI-B’s own words, this needs “hard-hitting emotional messaging” to be effective, and what could be more hard-hitting and emotional than official accounts of tens of thousands of deaths from the novel coronavirus? It is classical propaganda, “information, especially of a biased or misleading nature, used to promote a political cause or point of view”, a shocking and lamentable abuse of power that has no place in a modern democracy. Yet, for the last eighteen months, this tissue of partial truths, lies and deceits has been relentlessly streamed day after day by the government, via a complicit mainstream media, to the British public. The vast majority of other countries have also fared no better.

First, we can dispose of PHE’s comedy numbers of covid deaths, based as they are on the 28 day rule, a nonsense on a par with saying that if you detect a whiff of tobacco smoke on a person, and the person dies within 28 days, they are automatically counted as a death caused by smoking. Yet this is exactly what death from covid within 28 days of a positive PCR test within 28 days does: it uses a test of exquisite sensitivity to detect a possible whiff of a bit of a virus, and uses that to define a covid death. It is patently absurd, and we can consign it to the bin, and move on. Next slide, please.

The next slide concerns ONS’s numbers for ‘deaths involving covid’. These are orders of magnitude better than PHE’s comedy numbers, but until recently there has been no way of determining how many of these deaths are true covid deaths, where the underlying cause of death as we would normally understand the phrase is covid — where ‘he died from covid’ really does mean ‘he died from covid’ — and how many are conveniently, but misleadingly, added to these numbers, the better to achieve the inflated numbers that will provide the necessary hard-hitting emotional messaging to increase the perceived level of threat.

At this point, we need to review some of ONS’s definitions. Deaths involving covid are simply a count all deaths where covid in any shape or form appears anywhere on the death certificate. This is the headline figure you will normally see or hear, and it is itself already inflated, because it includes both deaths deemed — we shall have a lot more to say about this deeming shortly — by ONS to have covid as the underlying cause of death, that is, the person died ‘from covid’, as well as deaths where the person died ‘with covid’, that is, there was another, non-covid, underlying cause of death. ONS define the underlying cause of death as “the disease or injury which initiated the train of morbid events leading directly to death, or the circumstances of the accident or violence which produced the fatal injury”. In practice, this is a little bit more sophisticated than the man on the Clapham omnibus’s account of his father’s cause of death. The man on the bus might say his father died of a heart attack, but ONS are more likely to code the underlying cause of death as ischaemic heart disease.

ONS do sometimes say what percentage of the deaths involving covid (deaths both from and with covid) had covid as the underlying cause of death (death from covid). Typically this percentage is high, especially during peaks in the death rate. In mid January, for example, of all deaths involving covid, around 90% had covid coded as the underlying cause of death. By mid March, when the total number of deaths had declined greatly, this percentage fell to around 80%. This means that, in the headline figures that we read and hear for covid deaths, somewhere between 10 and 20% are not covid deaths; they are instead, by ONS’s own definitions, deaths with covid. The headline figures for covid deaths, based as they are on ‘deaths involving covid’ are thereby automatically inflated by around 10 to 20%, by simply including the deaths where covid was not the underlying cause.    

It gets far trickier when we try to tease out how many of the deaths deemed by ONS to have covid as the underlying cause are in fact true covid deaths, where the patient really did die from covid. We know there must be some inflation, because of the WHO/ONS requirements for coding deaths due to covid, which boil down to a rule that says if covid appears anywhere on Part I of the MCCD (medical certificate of cause of death), then covid trumps any other potential underlying cause of death, even when the other cause makes better medical sense. More details on this nefarious coviddery can be found in Dr No’s earlier post, Ode to the Death Certificate.

The problem until now has been how to get a handle on the extent of this nefarious coviddery. We know that almost all patients who are deemed to have died from covid will have had ‘pre-existing conditions’, any or all of which are eminently capable of causing the patient’s death. The key question is: how often does covid improperly  snatch death from the jaws of the other equally or even more capable underlying causes of death? In other words, how often do the WHO/ONS coding rules wrongly assign covid as the underlying cause, when by all normal common sense, and indeed medical sense, the other equally or even more plausible cause was indeed the true underlying cause of death.

The answer, it turns out, lies in taking a closer look at those ‘pre-existing conditions’ and considering what would happen under normal, non-covid conditions, and what actually happens, given the current covid conditions. We can do this using ONS’s own data, and get an estimate, if not a precise number, of the extent to which conditions which would, under normal circumstances, become the underlying cause of death, now get trumped by covid as the underlying cause, under the current rules.

The data comes from last year, but they appear to be the most recent data that include  sufficient detail on those elusive pre-existing conditions, and there is no reason to suppose that what is happening now differs in any substantial way from what happened in the early months of last year. The crucial table is Table 5 in the underlying dataset. This details the “Most common main pre-existing conditions in deaths involving COVID-19, all ages and sexes, England and Wales, deaths occurring between March and June 2020”, and the key thing is the definition of main pre-existing condition. It is the “the one pre-existing condition [on the death certificate] that is, on average, most likely to be the underlying cause of death for a person of that age and sex had they not died from COVID-19”. In other words, it is the condition that would most likely have been the underlying cause of death, had covid not trumped it.

What do we find? At the time of the report, there had been a total of 50,335 deaths involving covid, that is, deaths where covid was mentioned anywhere on the death certificate. Of these, 4,476, less than ten percent, had no pre-existing condition, so we can allow those as likely true covid deaths. But what of the other deaths? Over a quarter, 12,869, had ‘Dementia and Alzheimer’s disease’ as the main pre-existing condition – the disease that would have been the underlying cause of death, had covid not reared its ugly head. Let’s now consider first what would have happened pre-covid, and then what appears to happen now, after the arrival of covid, and the new covid death coding rules.

Pre-covid, the MCCD may have just had one entry, Alzheimer’s disease, and that would be coded as the underlying cause of death (UCOD). But often there might be another entry, let’s say bronchopneumonia on the first line of Part I on the MCCD, with Alzheimer’s disease underneath, on the second row. In this case, the MCCD shows a sequence, Alzheimer’s (leading to frailty) leading to bronchopneumonia, and in keeping with the normal definition of the UCOD, Alzheimer’s disease is the  “disease or injury which initiated the train of morbid events leading directly to death”, and so becomes the UCOD.

Now let’s replace bronchopneumonia on the first line with covid. Under normal rules, and by all standards of common and medical sense, exactly the same thing would happen: the sequence is Alzheimer’s (leading to frailty) leading to covid, and in keeping with the normal definition of the UCOD, Alzheimer’s disease is the “disease or injury which initiated the train of morbid events leading directly to death”, and so it  becomes the UCOD. But under the covid rules this does not happen. The medically sensible UCOD gets trumped by covid, and covid is recorded as the UCOD, despite the presence of a pre-existing condition that is, on average, most likely to be the underlying cause of death, had they not died from COVID-19. Occam’s razor hasn’t just been blunted, it’s been twisted back on itself until it snapped in two.

Dementia and Alzheimer’s disease aren’t trivial, they are serious killers in their own right. Yet the covid death trumping rule expunges them from the official record, and replaces them with its own cuckoo, the covid death. A death that would normally, and correctly, have Alzheimer’s as the UCOD now has covid as the UCOD. Let’s be brutal and say that all the 12,869 deaths that had ‘Dementia and Alzheimer’s disease’ as the main (ie it’s present on Part I of the MCCD, and would most likely have killed the person were it not for covid) pre-existing condition were in truth dementia and Alzheimer’s disease deaths, but got hijacked by covid. If so, that means around a quarter of the alleged covid deaths have been wrongly given covid as the UCOD. Or, to put it in inflationary terms, say 32,466 true covid deaths (50,335 less say 5000 ‘deaths with’ that we already know about less 12,869 dementia and Alzheimer’s) have been inflated by around 55% to get 50,335 covid deaths. Cripes. That’s quite some inflation.

And that’s just for dementia and Alzheimer’s disease. The next most common main pre-existing condition (so it was on Part I of the MCCD, and a likely candidate for underlying cause of death) is ischaemic heart disease, with 5,002 entries. We don’t know what was on those 5,002 MCCDs, but what we do know is that even if there was a medically sensible cardiac cause of death, we no longer know about it, because covid trumped it as the UCOD. Let’s be brutal again, and say that all those true cardiac deaths got misleadingly poached by covid, and do the inflation calculation again. Covid death inflation now runs at a staggering 83%. Cripes, and cripes again. Much more of this, and there will be hardly any true covid deaths left.

The third row for main pre-existing condition is Influenza and Pneumonia, with 4,582 entries. This opens up yet another can of worms, or rather chickens and eggs, as in which came first. Did the patient with covid develop pneumonia? That probably is a true covid death. But what about the patient admitted with pneumonia who them succumbed to hospital acquired covid? That is almost certainly not a true covid as UCOD death. Or perhaps it is. Who knows? Dr No’s brain is beginning to hurt, so let’s not go there… 

Clearly, and explicitly, Dr No has made some brutal assumptions. But equally clearly, it is impossible to avoid the conclusion that there has been significant and substantial inflation of the official counts of deaths involving covid, courtesy of the covid death coding rules, leading to a gross exaggeration of the perceived threat from covid. We will never know the true extent of the inflation, because the evidence has been buried and gone up in smoke, and so the necessary post-mortems and viral cultures can never be done. But the inflation, based on the above assessment, is so rampant that even if it is only half that suggested by the assessment, the official account of covid deaths is still wildly in excess of the true account. The government may have achieved its aim of increasing the perceived level of threat by using hard-hitting emotional messaging, but it is a shocking disgrace that it left truth at the front door, when it came inside to do its nefarious coviddery.

* (added 08:48 18th Aug 2021): The title for this post should really be (and was in the draft) Nefarious Coviddery and Covid Death Inflation. But Dr No is rather fastidious, perhaps too fastidious, about keeping post titles to a certain length, so they don’t word-wrap onto a second line. So he removed what he thought was the least important word. This may have led some readers to conclude, perfectly reasonably, that Dr No was talking about overall death inflation (an inflated number of overall deaths from covvidery). This is not the case: the effect described in this post increases the proportion of covid deaths, but the overall number remains the same. (added 09:15 18th Aug 2021): On further reflection, the title for this post should have been Coviddery and Covid Death Inflation. Not only does it have coviddery, which  rhymes conveniently with a certain other wittery, it also has better alliteration. But Dr No was foolishly glued to nefarious, so he dumped the wrong word to get the title within length; and the title will stay as it is, because Dr No is also fastidious about never doing post-hoc editing (apart from the correction of minor typos). As the saying has it, Publish and Be Damned…      


  1. dr-no Reply

    Rick – thanks for the thanks.

    As this is rather a long post, and the central key bit might be a bit lost in the rest, it is perhaps worth going over it again here, with a visual.

    The key thing is that the ONS data set with the pre-existing conditions means we can make a very good stab at reconstructing the actual death certificates (the MCCDs), and then assess how they are coded.

    1. The data set covers ‘deaths involving covid’, so we know covid is somewhere on the MCCD for each death in the dataset.

    2. For each main pre-existing condition, we know this is not only on the MCCD, but that it was in Part I of the MCCD, and was significant enough to mean it would likely have been the underlying cause of death (UCOD) had covid not ‘intervened’. We know this from the ONS definitions in the report.

    3. The most common by far main pre-existing condition is (in aggregate, and also unsurprisingly) Dementia and Alzheimer’s Disease. This makes our task much easier, because they are ‘simple’ (no long and ambiguous chains) sequences, typically just one or two entries on Part I.

    4. Thus we know that those deaths where Dementia and Alzeimer’s Disease was the main pre-existing condition must have both covid and either dementia or Alzheimer’s in Part I of the MCCD. Since no doctor that Dr No knows is likely to believe covid leads to Alzheimer’s, Part I on the MCCD must have looked like this (it may have a third entry, but as we shall see that makes no difference):

    5. In the past, ie before the introduction of the covid death coding rules, there is absolutely no doubt that Alzheimer’s disease would be recorded as the UCOD. This is an Alzheimer’s death. But then the covid rules came along, requiring that covid trumps all, and covid becomes the UCOD. As noted in Dr No’s previous post Ode to the Death Certificate, medical implausibility is no bar to this nonsense.

    6. The effect of the covid death coding rules is to convert what is actually an Alzheimer’s death into a covid death, and in so doing inflate the numbers of covid deaths. Since there were almost 13,000 out of a total of just over 50,000 deaths that had ‘Dementia and Alzheimer’s Disease’ as the ‘main pre-existing condition’, we can be as certain as can be that the effects of this covid trumpery are far from trivial.

  2. dearieme Reply

    Which agrees better with the number of excess deaths – the government’s numbers or yours? Isn’t that probably the best test? If your analysis results in numbers consistent with the excess deaths then you are on to a winner.

    Second point: in a way the interesting question about an Alzheimer’s/Covid death is whether Covid brought forward the Alzheimer’s death and, if so, by how much. I suppose it’s also the interesting question about X/Covid deaths for all diseases X.

  3. dr-no Reply

    dearieme – will look into your first point. On the second point, there is always an implied element of bringing forward – consider the Alzheimer’s / bronchopneumonia example (the ‘old man’s friend’ bringing forward a death) – but the idea in coding is to code the underlying cause (Ic led to I b led to Ia, so Ic is the underlying cause), on the grounds that one is the one that matters most, the usual reason being that it is the one most amenable to intervention, but of course that somewhat falls apart when the underlying cause is (currently) untreatable. But if the UCOD was say ischaemic heart disease, or a treatable cancer, then it makes more sense. It’s also about getting as good a picture of possible of what the true causes of mortality are. Lots die from bronchopneumonia as the immediate cause, but not really very helpful having stats that just show lots die from bronchopneumonia…

  4. dr-no Reply

    dearieme – to answer your first point: there were 58,113 excess deaths (give or take a few) over the period covered by the pre-existing conditions data (1st Mar to 30th June 2020), using the most recent pre-covid five year average (2015-19) as the baseline. This is much closer to the ONS estimate of covid deaths for the same period (50,335) than any revised down estimate based on allowing for covid inflation. Does this mean Dr No has backed the wrong horse?

    At first glance it might appear that it does, but it doesn’t, because his core argument in this post is not that the excess deaths didn’t happen – obviously they did – but that far too many of them were attributed to covid, when under normal coding rules they would not have been attributed to covid.

    The official narrative is there were 58,000 excess deaths, made up of 50,000 covid deaths and 8,000 non-covid deaths. The point made in this post is that this narrative is inevitably (because of the reasoning given in the post, and the second comment above, heavily, and wrongly, skewed towards a huge preponderance of covid excess deaths, when in fact the true numbers are actually very different, perhaps closer to half and half – say 29,000 covid deaths and 29,000 non-covid excess deaths. Which, to put it lightly, is more than a little awkward for the authorities, which is why they stick to the official narrative.

  5. dearieme Reply

    Well, what caused the 29,000 non-Covid excess deaths? Lockdown deaths? It would hardly be traffic accidents if everyone was cowering at home.

    Come to think of it, shouldn’t the government be publishing deaths-and-causes figures for all manner of deaths? Would those figures give a clue as to the causes of non-Covid excess deaths? In other words, which other causes of death leapt up?

  6. dr-no Reply

    “Well, what caused the 29,000 non-Covid excess deaths?” Not forgetting the 29,000 is an estimate – the problem is the evidence has been buried/gone up in smoke so we can’t to post-mortems and viral cultures, so we have to estimate the number using methods such as the one used in this post – indeed that is the question left hanging in the air at the end of the post. Apart from the obvious point made by the post, it is the reason for writing this post in the first place. Dr No is increasingly persuaded that history will show that the collateral damage caused by anti-covid measures actually did more damage than covid ever did, perhaps far far more damage, when we consider the wider ramifications – this post, after all, only considers mortality – of the damage done to health, social well being, education, the arts and the economy. History may well decide that 2020 was the year when the world collectively lost its head, and committed one of the largest blunders short of all out war prosecuted in modern times.

    The government do routinely publish mortality by cause, but there is a long lag before the data becomes available, which is why Dr No despaired of trying to assess the extent of non-covid mortality. Then he found the datasets used in this post and Murder by Decree, which do allow a degree of assessment. The Murder by Decree data had weekly counts by residence for all deaths, and covid deaths, so by simple subtraction we can get weekly non-covid deaths, and look at patterns in all cause overall mortality. The data used in this post had within it counts of main pre-existing conditions, which by definition are in Part I of the MCCD, and would have been the UCOD if covid hadn’t ‘intervened’. The bottom line of this post is it appears very likely the government/ONS found it ‘convenient’ to reassign the UCOD from these main pre-existing conditions to covid, to maintain the narrative that there was a ghastly epidemic raging among us. The truth is more likely to be there was a minor epidemic of a new coronavirus, combined with substantial collateral damage from the anti-covid measures. But this effect has been conveniently hidden, by re-assigning the UCOD for many of the deaths to covid.

    Dr No is not alone in this assessment. A reader kindly emailed a link to a paper that comes to much the same conclusions, and contains additional evidence eg a sharp fall in A&E attendances over the relevant period. The language is rather convoluted and the maths absurdly complicated but its bottom line conclusion, assuming the maths make sense, is that overall, lockdown did more harm than good, chiefly through collateral damage.

    Ah, our old friend alcohol! It’s also by coincidence in the news this morning. Dr No often wrote about alcohol policy on the old blog. It (the policy) is riddled with numerology, coercive healthism, puritanical hysteria and ludicrous public health ideology posing as theory (the invidious sick individuals/sick populations and tail wags the dog dogma). It’s all enough to turn anyone to hard liquor. Alcohol is on Dr No’s list of things to look into, but his current view (without having looked in detail at the data) is that the journey to most alcohol deaths starts many years before the death occurs, and so if there is a lockdown effect, we won’t see it until we are a few years down the road. On the other hand, its entirely possible there has been a rise in alcohol deaths, partly from acute alcohol poisoning (drink till you drop), and a sudden rise is in alcoholic decompensation (decompensation is a rather hideous generic medical term for a precarious chronic medical state suddenly going pear shaped) triggered by increased alcohol intake in people with chronic alcohol related illnesses. Such effects are certainly plausible, but without looking at the data, it impossible to say whether they are true or not.

  7. Tom Welsh Reply

    I would be delighted to see you investigate the topic of alcoholic drink and its effects on health, Dr No.

    You might be interested to look at Tony Edwards’ book “The Good News about Booze”, which I read a few years ago. Since then I have been drinking regularly in (what I consider) moderation. That is a glass or two of wine with dinner, and possibly either an afternoon G&T or an after-dinner drink such as whisky or brandy. In general the obvious guideline applies: if you wake up with a detectable hangover, you drank too much.

    The book can be found at

    Edwards appears to have based his book quite carefully on the literature, and comes to the conclusion that in the case of almost every known serious disease alcohol in moderation significantly improves outcomes. Circulatory disease, cancer, even if I remember aright) infectious diseases.

    Edwards concludes that wine or spirits offer benefits far greater than almost any pharmaceutical drug. (I think beer is out because it contains too much carbohydrate – dry drinks are always better than sweet).

    • dr-no Reply

      Tom – Dr No expects that at some point he will do one or more current posts on alcohol but in the meantime he has put together a quick list of number of posts on the old blog that cover alcohol, mostly nonsensical safe limits and minimum unit pricing (MUP) (note the blog isn’t quite as slick as it might be, because it needed a radical behind the scenes rejig to keep it working on modern platforms, so only the basics are there):

      One current paper of interest might be this one from the Lancet, published earlier this month, which claims to show MUP in Scotland (and Wales) is a Jolly Good Thing. But even on a cursory glance, there are some oddities that might prove interesting on closer inspection…

      • Tom Welsh Reply

        Many thanks for your thoughtful kindness, Dr No! I had not seen any of those posts, as I have been reading your blog for a relatively short time.

        I shall read them with interest.

  8. dearieme Reply

    Thanks for the lengthy reply, doc.

    “history will show that the collateral damage caused by anti-covid measures actually did more damage than covid ever did, perhaps far far more damage”: it was my guess in March 2020, when the lockdowns started, that that would probably prove true. About all we could do, my wife and I, was take Vitamin D and zinc – which we started in the middle of that February – and avoid crowded places where people would sing or shout a lot. We did start disinfecting grocery deliveries but when the government went silent about infection from touching surfaces I inferred that it was unnecessary and we stopped.

    Of all the official medical stupidities the ones I shan’t forgive are those that led to (i) the deaths in care homes, and (ii) to the infection of patients admitted to hospital for other reasons.

    Of all the non-medical follies the ones I shan’t forget are (i) stopping people walking in the park, and (ii) the panic-stricken closing of schools. I’ll grant that the latter was particularly the fault of the teaching unions but the government had ample power to crush them – a wasted opportunity, alas.

    The Swedes got it less wrong.

Leave a Comment

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