14th September 2022
Some of you may remember COVID19. We had an epidemic, or a pandemic, or … choose whatever word you like best. The legacy of it still hangs about in many strange, disconnected actions.
My last flight in late August, on Lufthansa, required me to wear a mask. The connecting flight with Swissair, did not. No mask was required whilst waiting at Munich airport or being transported to and from the planes in a crowded bus. Go figure. I am sure this all makes sense to someone, somewhere.
Anyway, I thought it might be good time to have a catch up and see if we can learn anything more about the pandemic and the drastic actions taken to control it. The first thing to say is that this is a complex task. Mainly because the data surrounding COVID19 are unreliable. To say the least.
How many people have been infected with Sars-Cov2? How many have died? I believe we can only really guess. Worldometer, as of the tenth of September 2022, confidently informed me there have been around six hundred million people infected with COVID19 worldwide (613,234,326). The number of deaths is just over six million (6,514,989)1.
Quite remarkably, this represents infection fatality rate of pretty much bang on one per cent. As predicted by Imperial College London and Professor Neill Ferguson. Take a bow that man? Or perhaps not. How many people think that those figures are remotely accurate? Certainly not me. Just to start with, do we really believe that ninety per cent of people have managed to avoid a Sars-Cov2 infection?
My own belief is that virtually everyone in the world has been exposed to/infected by Sars-Cov2 and at least once. (The concept of what ‘infection’ means has undergone a bit of a transformation, a.k.a. mangling). We already know many people have been infected several times. In fact, as early as the autumn of 2020, doctors were seeing people who had been, proven, to be infected twice. Even then, these cases were considered to be the tip of the iceberg.2
If people were getting infected twice, within six months of the virus arriving, I think we can safely assume almost everyone else has been infected at least once. Maybe a few villagers in the Amazonian rain forest have remained exposure, and infection, free. As for everyone else… unlikely.
And as for the numbers of COVID19 deaths, again, can we know anything for certain? I wrote out four death certificates which included COVID19 as a cause. This was early on, before much testing was possible. I have no idea if they had COVID19, or not. I cannot imagine I was alone in adding to the COVID19 stats, whilst blundering around in the dark.
If we can’t really rely on these, the most basic of facts, then can we learn anything? I think so, I hope so. Indeed, from the very start I tended to focus my attention away from COVID19 specific data, towards data I felt I could trust. Namely, the overall mortality rate.
Although these data do not allow us to be certain who died of COVID19, the numbers are the most robust we have. Someone is either alive, or dead, and it is difficult to get the diagnosis wrong. Yes, there can be some delays in reporting etc. but in general dead is dead and alive is alive, and that is that. One hundred per cent accurate.
Of course, in order to use these figures, I have to make assumption (made by many others), that spikes in overall mortality would be the best way to get a fix on how many people COVID19 was actually killing. A big spike – more deaths from COVID19. No spike, no extra deaths from COVID19 (or very few).
So, did every country show pretty much the same pattern in mortality? Or were there extremes or outliers? I am a believer that it is at the extremes where answers can often be found.
I began by looking for countries – or populations within those countries – that suffered a major increase in overall mortality. Then I looked for matching countries, and populations, that showed no change, or very little change. Because here, maybe, we could find some solid ground to stand on.
The most easily accessed data can be found at EuroMOMO3. This is a resource where data on overall mortality are collated from many different European countries. The site then plots mortality against a (moving) five-year average.
I ended up focussing on European data, not just because it was easy to find, but mainly because most European countries are very similar in many important parameters. Standard of living, health service provision, demographics, life expectancy and suchlike. Which means that you are comparing like with like. Try to compare Norway with, say, Kenya, and you end up with a mess.
On the other hand, you can more reasonably compare Norway and Sweden. There are far fewer differences between them, which should make it simpler to spot the key one(s).
However, to start with I am not going to look at Norway and Sweden. Instead. I want to draw your attention to four countries that are not, in truth, separate countries. They are Scotland, England, Wales and Northern Ireland. Four different ‘parts’ that make up the single entity known as ‘The United Kingdom of Great Britain and Northern Ireland’. Longest country name in the world – good pub quiz question.
It is true these four ‘countries’ did not do precisely the same things during lockdown. But the differences in timings and actions, were small – with a few (look at me, I’m locking down harder than anyone else, vote for me … thank you Nicola) variants. However, you will struggle to find any other four countries that were more alike in their characteristics and actions.
Despite their many similarities, one of these populations showed a hugely significant increase in overall mortality, and the other three did not. This difference can be seen most starkly within the age group of forty-five to sixty-four.
First, a short explanation of what you can see in the graphs below.
The – somewhat difficult to make out – dotted line represents the rate of overall mortality five standard deviations above the norm for the time of year. Mortality is always higher in winter than summer, but these graphs take this into account, and are mathematically flattened out.
The darker, spiky line represents the overall mortality rate. If it rises above the dotted line this is considered to be a ‘statistically significant’ event. Or, to put it another way, something is happening that is killing far more people than we would expect to see, and we need to find out what. This is normally due an infectious disease of some kind, almost always influenza.
The scale on the left -10, 0, 10, 20 is not an absolute figure. It represents the standard deviation from the mean (the z score). If it goes above ten, this is big time trouble. Above twenty, look out, the sky is falling. In general, two standard deviations from the mean is considered ‘statistically significant’ in medical research.
OVERALL MORTALITY RATES AGE 45- 64 IN THE UNITED KINGDOM
2017 TO SEPT 2022
As you can see. England had a three major mortality ‘peaks’. Spring 2020. Winter 2020/21 and a far more diffuse mountain range in autumn and winter 2021. The other three countries showed almost nothing at all.
I will just add in here that the difference is not restricted to this one age group. Below is a graph of the sixty-five to seventy-four-year-olds.
OVERALL MORTALITY RATES AGE 65 to 74 IN THE UNITED KINGDOM
2017 TO SEPT 2022
Pretty much the same pattern emerges. Two massive upticks in overall mortality in England, very little elsewhere. Absolutely nothing to see in Northern Ireland. If COVID19 was killing lots of people in Northern Ireland, it was not showing up.
First question, does England have a worse health service than the other three countries? No, it does not. Is the overall health worse in England? Well, in general, the English have a longer life expectancy than those in the other three countries, rather than the other way around. Which suggests that the English are, in general, healthier 4.
What was the same in these countries
- The health services
- The age of those dying (I matched people for age)
- The lockdowns (very minor differences)
- The treatments given
- The vaccinations given
- The climate
- Overall life expectancy (very minor differences, should be favouring England)
So, what was different?
Over to you. Because if we can work out what caused all these people to die in England, and not in Scotland, Wales and Northern Ireland, we can probably learn something of great value.
Before that – and changing tack for a moment or two, in the early days of COVID19, everyone jumped around claiming that Norway had done things fantastically well, as they had no change in overall mortality, and very few recorded COVID19 deaths. ‘Look at them shutting their borders and enforcing a very tight lock-down. Way to go whoop, whoop.’
No-one bothered to mention Northern Ireland. Which did precisely the same as England. Yet also had no change in overall mortality, as per Norway. You could argue that Northern Ireland did not fit the agreed narrative, whereas Norway did.
Sweden, on the other hand, famously did not lock down, ‘shock-horror, everyone in charge should be fired, or thrown in jail’. Sweden did have significant uptick in overall mortality. Proof that lock-downs were essential?
Possibly … probably not. Many other countries in Europe which did lock down, have had far more COVID19 deaths, and a greater impact on overall mortality, than Sweden.
Here are the European countries that have recorded more COVID19 deaths, per head of population, than Sweden. In descending order1:
- Bosnia and Herzegovina
- North Macedonia
- San Marino
Here, I did use COVID19 deaths, as reported on Worldometer – with all caveats recognised. The reason for using these figures rather than overall mortality, is that they were, initially, used to attack, the Swedish response. [People are a lot quieter about Sweden now] Also, calculating the overall mortality increases in these countries represents a very major task – with complex adjustments to be made. So, I didn’t do it here. I would also point out, for the sake of completeness, that Sweden is reported to have had 1,968 COVID19 deaths per one million of the population. Norway 728. [Two per thousand vs. point seven per thousand]
Lithuania, by the way, like Norway, is very similar to Sweden. For about a hundred years they ruled central Europe together within the Union of Kedainiai. In many ways, they have more in common than Sweden and Norway. It should be noted that Lithuania locked down early, and hard. You may note Lithuania pops up at number ten in the list above. Reported COVID19 death rate 3,528 per million.
You may disagree with my definition of European country … Gibraltar? Listen, I got this from Worldometer, so you can fight with them. However, if anyone wishes to tell me that Sweden suffered a unique catastrophe due to their reluctance to fully lock down, they may struggle to convince me that it was the critical factor. In fact, I may give a hollow laugh, even raise a quizzical eyebrow.
So, what else was different between Norway and Sweden? Something that could reasonably explain the difference in both recorded COVID19 deaths, and overall mortality. I believe there is another clue within the EuroMOMO data. If you choose to look at what you are actually seeing.
Below are the data from Norway from late 2017 (slightly annoyingly, their data only started in late 2017).
OVERALL MORTALITY NORWAY LATE 2017 TO 2022 – all ages
What stands out very clearly is that the Norwegian overall mortality rate has never spiked. At least not since late 2017… on EuroMOMO. This was even the case in the winter of 2018, which was a bad flu season across most of Europe. Something that shows up most clearly in Germany, although the same pattern can also be seen, to a lesser extent, in France, Belgium, Austria, the Netherlands, UK, Portugal, Italy etc.
OVERALL MORTALITY GERMANY 2017 TO 2022 – all ages
Did the Norwegians lock down in 2018. No, they did not. So, what stopped them dying from flu? The answer is … something else. And that something may well be the same thing that stopped them dying of COVID19.
As an aside, why did the Germans not panic in 2018, when more people were dying then, than from COVID19 in 2021? They had a z-score of very nearly twenty. Did anyone even notice? Was it front page headlines? No, of course not. It passed in virtual silence. Compare and contrast, as they say.
Anyway, I hope that I have given you a little puzzle to solve. I have been contemplating this puzzle for some time, and I think I may have identified the key factor that can explain the patterns in the UK, and also between Norway and Sweden. I am interested to see what other people’s thoughts might be.
Before coming back with answers. Remember, these data are age-matched. They compare overall mortality, not the number of recorded deaths from COVID19. They are not the absolute numbers of deaths, but variation from the mean. The z-score.