So many questions have arisen about how to deal with Covid19. A lot of us have been asking who did it best? The lockdowns, the border closures… the list goes on. Problem is we are comparing apples with oranges when we compare countries at the moment.
First a little background. The recording of deaths is generally done internationally in accordance with the World Health Organisation’s ICD-11 for Mortality and Morbidity Statistics (ICD-11 MMS) – to be implemented in 2022, but the ICD-10 is still in action. Problem is, you won’t see COVID19 in that list of causes. That’s because this is a new virus, and a new potential cause of death. So it doesn’t appear in the ICD as a cause of death. Health statisticians have not yet had time to come together (social distancing and no international travel and other priorities and things) to arrive at a clear coding and definitional agreement on what constitutes a COVID19 death.
So when you see countries compared against other countries, particularly by lazy journalists with no idea what they are talking about, don’t trust it as far as you can throw it. We won’t have a clear picture until the health statisticians come together to agree on a definition and singular method for identifying COVID19 as a cause of death.
Let us start with the basics. The typical death certificate used by the majority of countries in the world looks like this:
ICD-10, which ICD-11 references to provide guidance on how to code cause of death states:
Hence, “underlying cause of death”, it is meant to be the last mentioned, or “underlying” reason for death, i.e. the last line (with exceptions, which skilled codes are trained to apply). So if you die of pneumonia, and you happened to have a suppressed immune system as a consequence of AIDS, the first reason would be Pneumonia, and the second (as a consequence) reason would be AIDS. You’d be an AIDS death.
Except that does not happen for COVID19 in many countries. COVID19 is outside the remit of the ICD-11, because it is so new. So instead, under pressure from the public and the media to report “COVID Deaths” most countries have made up their own new rules outside of the ICD-10 and ICD-11.
So for starters, in many countries statisticians started off down the path of “probable” COVID19 deaths. This happened at time when it was brand new and when we knew little about the disease. The major tests used in the first few months were antibody or serology tests, which the CDC appropriately pointed out: “there is a chance that a positive result means you have antibodies from an infection with a different virus from the same family of viruses (called coronaviruses)” e.g. the common flu.
New tests were subsequently developed that ensured only COVID19 was detected, such as the RT-PCR test. This was also rolled out widely into the community, including testing asymptomatic members of the community. However that test, according to the Office for National Statistics, has a false positive rate of between 0.8% and 4.3%. The rate for false negatives is also in the same range let’s call both an average of 5% as per the UK paper. While this sounds low, think about it for a minute.
The statistics show that between 13% and 18% of COVID sufferers present with symptoms. Let’s assume a community of 100,000 people. Estimates on the infectiousness rate in the UK differ between 1 in 640 people and 1 in 1570 people. Call it 1 in 1000 for the sake of argument.
Our imaginary community therefore has 1000 who actually have COVID19. With hospital only testing, only those with significant symptoms would present at a hospital. Let’s say a third of the“13–18%” Call it 5%. So 50 people would turn up at hospital with major complaints, be tested and diagnosed and be included in the COVID stats.
Then we introduced RT-PCR tests, with an error 5% plus or minus. Widespread community testing comes into play, and we know that not everyone would be tested, but let’s say for this example the NHS got out there and tested 30% of the community.
So they would presumably pick up the 30% of actual sufferers, 300, minus 5% for the false negative rate, say 285. But they would also pick up plus 5% of people without COVID of 1,500. So 1,785 total “cases”. This means community testing should never be used for statistical purposes, the errors are far too high. It should be used to detect outbreaks, but never reported on publicly.
But it gets worse. Remember back to the whole ICD-10 and ICD-11 thing? No COVID coding in the data, all the coders making up their own rules. So here’s the WHO’s snap shot of how different countries are determining “COVID” deaths.
So many of these countries are coding “COVID” deaths to whether a person died “with” COVID rather than as the underlying cause of death.
So the next line of defence for most COVIDIOTS (pardon the use of a pejorative) is that we should look at “Excess Deaths”. Okay, let’s try that for my home country of Scotland.
What hypothesis can we draw from this graph? On first look, it appears COVID is a massive increase on the last non-COVID year (2018). But then, remember, in Scotland, if you died “with” COVID as opposed to your underlying cause of death – say impaled by a leprechaun – you are still recorded as a COVID death. So we can assume that a certain proportion of those “COVID” deaths, if they were properly coded against the rules of the ICD-10 and ICD-11 would be lower.
Let’s take the first statistically significant deviation from 2018, “Mental Health and behavioural disorders”. They appear down. Great news! But a deeper dive into the stats show that this is really out of character with ordinary trends, and that almost all of these deaths used to be coded to “Dementia”. We all know that care homes were hit hard by COVID. So ordinarily someone might have died of “Dementia” but because they tested positive for “COVID” the coders reallocated the death to a COVID death.
Similarly, Respiratory illnesses through the floor, most of it pneumonia. Recoded to COVID perhaps because of the new rules?
Another example? Heart attacks up significantly. Okay, that makes sense, we’ve all been told to keep away from medical services. People get a heart pain, they go “Oh, yeah, protect the NHS, I’ll get over it”. So we see a sharp jump in heart disease deaths.
Reduction in “External causes of morbidity and mortality”, that’s pretty much accidents and suicide. Makes sense accidents are down, no one is going anywhere in their cars, unless you get killed by an aerially fatal baguette falling from your pantry, no one is dying of accidents. But dig into the stats a little more and you find out that suicides are up by 200 over a pretty stable trend.
If you really want to have fun with this, you can look into mental health waiting lists… I think I have one handy.
But that really is another kettle of worms entirely.
Edit: Bottom line is that “Excess Deaths” is a poor measure as well. Some of the additional deaths can be attributed to government lock down policies, leading to higher heart related deaths and suicides. COVID deaths are likely overstated due to a change in statistical rules in contradiction to the previously well adhered to WHO standards.
Whichever way you cut it, COVID is no plague. It has modestly increased death rates, largely among very old people at end of life. Nevertheless, Long COVID is a thing, but we aren’t really directing resources at it yet.
Edit 2: For fullness, not every country uses the ICD set, 117 of 195 nations do. So that is a further complication in comparing like-for-like stats with countries.
TLDR: You can’t compare countries because they aren’t using the same rules like they used to. Also, a significant proportion of COVID deaths are likely recoding from other underlying illnesses that should have been recorded as the primary cause of death instead of COVID.