I’m not arguing that they didn’t. I’m merely pointing out (as I have many times before) that those peaks are time-shifted. Deaths that might have occurred later have been shuffled together into peaks. Some deaths will have been shifted a lot; some, a little. There is no other way that “excess deaths” can occur - you can’t magic dead people out of nowhere.
COVID-19 doesn’t kill healthy people. There is no debate about this. The inescapable conclusion is that almost all of those people who died in 2020 were going to die fairly soon. The only question here is “how soon?”. If you see a large “excess” in a given 12-month period, the implication is that some deaths have been brought forwards by more than 12 months.
Dying ‘of COVID’ implies that your life was drastically foreshortened. Dying ‘with COVID’ implies that it was not.
It’s also important to remember that other things besides COVID would have killed (vulnerable) people in 2020. We know, for example, that many more people died of heart attacks simply because hospitals were shut down. Their deaths will appear as “excess”. But they were not killed by COVID.
I’ll reiterate that the concept of an “expected deaths” is a fiction, predicated on the idea that what happened last year ought also to happen this year … and if it doesn’t happen as projected, then something has gone horribly wrong. Have a think about that; it’s a pretty strange philosophical position, because it implies that our duty in life is to endlessly predict, observe, and correct reality to match our predictions. That is, our predictions come to define what is (or ought to be) real, however foolish those predictions might be.
Bottom line is that “expected deaths” is not a number. It cannot be ‘accurate’ or ‘inaccurate’. Whether you think any given figure is plausible depends on your assumptions and biases.
Anyway.
Have a look at the raw data for the year: PHE are claiming 98,000 excess deaths over the March20-March21 period, and 127,000 people dying “of or with” COVID-19.
Well, that in itself looks pretty odd. How can excess deaths be lower than the number who died “of or with” COVID? One possible explanation is that something happened in 2020 that caused the natural death rate to be much lower than it was in previous years, ie., the real figure for expected deaths (which we cannot know directly) was lower than the estimated/modelled figure. COVID-19 then killed an additional 127k people. This is the simplistic media presentation.
But that’s probably not what happened. Some events in 2020 would have resulted in a lower-than-usual natural death rate (eg., fewer flu infections). Some would have resulted in a higher-than-usual death rate (eg., restricted access to therapy for treatable disases). We cannot model any of this stuff accurately: as I said earlier, all we can do is bear in mind that ‘expected deaths’ for any given week is a probability distribution, not a scalar.
Now, pull up the PHE data and total up ‘Deaths with COVID-19 as the Underlying Cause’. You get 248k, ie., twice the number of declared COVID deaths. And people can die of many more things than the diseases listed here. The reason for the discrepancy is obvious: when people die, they die with multiple problems. Usually. There’s a lot of double-counting in these figures.
So when you look at the diabetes data and see “28265 excess deaths” alongside “23678 COVID-19 deaths” you might conclude that an abnormal number of diabetics died and that they were (nearly) all killed by COVID-19. And you’d be wrong.
Someone who dies with diabetes usually doesn’t die of diabetes, nor does he only have diabetes. Each of those diabetics who died would have had several contributory causes listed on their death certificates - in other words, his existence may register in any of the other disease rows, and the fact of his death may also register in any of the other rows. In fact you can see this in the PHE charts: diabetics who die invariably have ‘other causes’ or ‘COVID’ listed as the primary cause of death.
I’m struggling to find a way to tease out all this overlap in an intuitively-meaningful way which is still “true”. It’s a kind of inverse problem, and it’s impossible to compute an accurate backprojection from the given data.
Perhaps we can consider diabetics in three categories:
- People who have advanced, chronic, untreated metabolic syndrome (which would include diabetes) and have a shortened life expectancy;
- The same as (1) but an older age group, with a very short life expectancy;
- People who are diabetic but do not have any serious complications and are not at any immediate risk of death.
People in category (2) have a very high probability of dying in a 12-month period. People in (3) have a very low probability of dying. People in (1) are somewhere in between.
-
If someone in category (3) has COVID on his death certificate, it is entirely reasonable to state that he died ‘of COVID’. His death was completely unexpected.
-
If someone in category (2) has COVID on his death certificate, the fact that he had COVID is purely incidental. He probably would have died if he’d caught a cold (yes, that actually happens).
-
If someone in category (1) has COVID on his death certificate, it’s reasonable to suggest that he died ‘with COVID’.
The PHE data strongly suggests that a small minority died ‘of COVID’: look at the orange sections, and consider the aforementioned overlap in these various diseases:
