[CardinalSagehen]

I'm old enough to remember when a correlation was found between latitude and COVID-19 and everyone thought that it was because of temperature and that it suggested that COVID-19 was going to go away in the summer.

I remember when early in the pandemic people were wondering why Germany's CFR was so low; it has since caught up with much of the rest of the world (excepting countries that were overwhelmed).

I think most of these ecological correlations are likely to turn out spurious.  Numbers of cases and even deaths attributed to COVID-19 are very subject to the testing regimen, which has been very uneven.  In some countries *cough, Russia, cough*, the numbers are manipulated to fit the country's narrative.  Even deaths, which I thought would be harder to get wrong, are frequently under-counted (see, for example, this article: https://www.nytimes.com/interactive/2020...hs-us.html).

I would not trust any analysis that depended on reported COVID-19 cases and deaths.

BC

I agree with that. Case stats are meaningless, so CFRs are meaningless. This one seems to be based on death rates in European counties, which I assume to be a fairly reliable stat. But the correlation will likely prove to be random.  

It's pretty much impossible to know what is really going on in Africa.

[CardinalSagehen]

I'm old enough to remember when a correlation was found between latitude and COVID-19 and everyone thought that it was because of temperature and that it suggested that COVID-19 was going to go away in the summer.

I remember when early in the pandemic people were wondering why Germany's CFR was so low; it has since caught up with much of the rest of the world (excepting countries that were overwhelmed).

I think most of these ecological correlations are likely to turn out spurious.  Numbers of cases and even deaths attributed to COVID-19 are very subject to the testing regimen, which has been very uneven.  In some countries *cough, Russia, cough*, the numbers are manipulated to fit the country's narrative.  Even deaths, which I thought would be harder to get wrong, are frequently under-counted (see, for example, this article: https://www.nytimes.com/interactive/2020...hs-us.html).

I would not trust any analysis that depended on reported COVID-19 cases and deaths.

BC

I agree with that. Case stats are meaningless, so CFRs are meaningless. This one seems to be based on death rates in European counties, which I assume to be a fairly reliable stat. But the correlation will likely prove to be random.  

It's pretty much impossible to know what is really going on in Africa.