Sweden had it right all along

adelaide wrote:

(1) Interesting. I would be interested in seeing the direction of the correlations. Also, are those correlations with total fatalities or per 100,000? It seems like the later would be more meaningful.

(2) Finally, did you break your analysis out separately by region by any chance? Europe is much further along in the cycle than some of the tropical countries like India and Brazil. And the numbers in East Asia are vastly different from the main European countries so pooling them all might mask some interesting patterns.

(1) Here is a list of R-squared values for fatalities per million with a list of other indicators, generated about 3 weeks ago:

I've listed R-valued (not R-squared) for cross-correlations with two of the indicators, to try to get a sense of which indicators are associated with the same things. There you can see which things are negatively (or positively) correlated.

Here are a few graphs showing the correlations between some of these indicators and fatalities (not cases) per million:

The correlations with cases are slightly different than those with fatalities, but not meaningfully so for the purposes of this illustration.

(2) I did not try to subdivide by region, but some of the key indicators are markedly different by region, so you can draw some inferences. I might expect to see the correlations shange as the current outbreaks in India area, Africa and South/Central America start to dominate the world trends. My graphs update daily as I pull in new data, and the general trends have not changed (so far) as the outbreak has migrated around the world. But we will see when all is said and done...

damn sorry, the tabulated correlations are between total CASES per million and the listed indicators, as of about three weeks ago. The plots show FATALITIES per million versus a few other metrics.

It's hard to say anything for certain I think Sweden was in it for the long run hoping that everyone would get immuned to the virus and then they would not have to deal with it anymore what do they turn out in the long run to be right or wrong doesn't matter they tried the way they thought would be best and that's that.

People are hard on them because they did something different and it cost more life at the beginning maybe it will continue to cost more lives or maybe they'll all be immune doesn't matter though they did it their way that's it.

If anyone has poked around the table and graphs I shared (three posts up), you'll see some apparently bizarre results. First, fatality rates are higher in wealthier (higher GDP, lower poverty) countries. Clearly, wealth does not cause people to be more likely to die, so there are confounding factors. My belief is that the important causative factor is age, which correlates with GDP (wealthier countries have longer life expectancy, older populations, more vulnerable people).

Another odd outcome is that female smoking correlates strongly with fatality rate, but male smoking does not; female smoking has a strong positive correlation with GDP (which links to age, which may be the underlying factor), but male smoking correlates negatively with GDP.

Finally, fatality rates are LOWER in countries with higher air pollution. This, again, strikes me as likely relating to wealth and life expectancy. But a bizarre result nonetheless...

interested reader wrote:

If anyone has poked around the table and graphs I shared (three posts up), you'll see some apparently bizarre results. First, fatality rates are higher in wealthier (higher GDP, lower poverty) countries. Clearly, wealth does not cause people to be more likely to die, so there are confounding factors. My belief is that the important causative factor is age, which correlates with GDP (wealthier countries have longer life expectancy, older populations, more vulnerable people).

Another odd outcome is that female smoking correlates strongly with fatality rate, but male smoking does not; female smoking has a strong positive correlation with GDP (which links to age, which may be the underlying factor), but male smoking correlates negatively with GDP.

Finally, fatality rates are LOWER in countries with higher air pollution. This, again, strikes me as likely relating to wealth and life expectancy. But a bizarre result nonetheless...

I think it's really interesting that stuff.

I think example potential cofounding factors you mentioned in the poorer countries, could be things like the likelihood of reporting a death by CV. For example, in poorer countries, there is less access to hospitals, in some regions the poor will just die without ever getting tested/treated, maybe hitting some low tech medical facility, if even that ?

Obesity would be an interesting one (as this links into some vit-d hypothesis along with latitude) ?

Air pollution I think is tricky, as it seems to vary a lot depending on whose stats you look at, and again often a lot of places with suggest high air pollution are poorer countries (not all) which are likely not ever seeing a lot of the actual cases they are (as they have very limited healthcare) ?

dense is as dense does wrote:

ex-runner wrote:

Population density. Honestly Americans have no idea of the difference between say Sweden and the UK.

Sweden's population density is 63 per sq. mile.

Italy 520

UKs is 671!

63 per mile is deserted compared to the UK.

1/4 of sweden lives in one metro area, much of it is sparsely populated driving down the countries overall density. you understand how that works?

The number of people in the entire population of Sweden lives in London alone. Do you understand how that works?

According to Wikipedia, the Stockholm Metropolitan area is home to 2.4 million people (1/4 of Sweden) and is roughly 6500 km2.

London is home to a population of 9+ million in a 1500km2 area.

Please tell me how population density doesn't matter? If a country had a population of 1000 and each lived 1 mile apart from each other do you really think the spread would be as fast as if they all lived in the same sq mile?

ex-runner wrote:Please tell me how population density doesn't matter?

I won't attempt to explain WHY it doesn't matter, but I will assert with confidence that population density, on its own, has no relationship with total cases per capita, total fatalities per capita, or case mortality rate (fatalities per case), as shown in these graphs which are current up to yesterday's world data from Johns Hopkins University:

There are other population metrics that show a correlation with cases and fatalities (but not case mortality), for example proportion of population living in urban centres:

Some will be interested to note a much stronger association between urban population and CASES, and weaker between urban population and FATALITIES, suggesting (sensibly I think) that high density (urban) living creates higher chances for exposure and infection, but is not associated with vulnerability (tendency to die from the disease), which relates to other factors (most strongly seemingly age).

My hunch is that wealthier countries are hit harder due to various factors, including more mature healthcare systems, which enable more detailed reporting, and politics/cultural influences.

ex-runner wrote:

Please tell me how population density doesn't matter? If a country had a population of 1000 and each lived 1 mile apart from each other do you really think the spread would be as fast as if they all lived in the same sq mile?

Asian cities are by far the most densely populated in the planet.

Yet they are far from the most Covid infested.

How does that fit into your narrative?

WWRD wrote:My hunch is that wealthier countries are hit harder due to various factors, including more mature healthcare systems, which enablemore detailed reporting, and politics/cultural influences.

My instinct was to rebuff you out of hand, but then I slowed my roll, and decided to plot per capita GDP against tests per capita:

That's among the strongest of all the correlations I've seen, nice call!

The Asian blooded Scandinavians have the same genes as China and Korea which is averse to COVID-19.

Asians are the Best wrote:

The Asian blooded Scandinavians have the same genes as China and Korea which is averse to COVID-19.

I don’t know how serious your post is but I do think ethnicity genetics plays a role in mortality rates. We’ve seen Asian populated countries on average having much lower mortality rates than many European and American countries despite the virus probably originating there.

I’d be interested in seeing the US death count broken down by race. I’ve heard via a few media outlets claiming the African American community has been hit particularly hard by COVID-19 but couldn’t find any hard statistics.

Ask and ye shall receive:

https://www.apmresearchlab.org/covid/deaths-by-race

In general, people with higher melanin content in their skin have lower vitamin D levels. Does one correlate to the other, or the other way around? If it was purely due to economic issues, then one would expect Asian Americans to have the lowest rate, as they have the highest per capita income.

joedirt wrote:

Ask and ye shall receive:

https://www.apmresearchlab.org/covid/deaths-by-race

In general, people with higher melanin content in their skin have lower vitamin D levels. Does one correlate to the other, or the other way around? If it was purely due to economic issues, then one would expect Asian Americans to have the lowest rate, as they have the highest per capita income.

Thanks. That is interesting. Latino, Asians and whites have very similar mortality rates while blacks are WAY off in death count. It could be purely economical but the fact that Hispanics, Whites and Asian death counts are pretty close to identical while black death counts are more than double any of those three races should raise some eyebrows. I would think there has to be some kind of link between a race’s genetics and mortality rates.

Thanks for sharing your data! You have obviously spent some time on this. Percent urbanized is really just a indicator or level of development rather than population density. And as your data clearly show, there is a fairly strong relationship between a country's level of development, as measured by per capita GDP and cases/fatalities per capita . A country, such as Bangladesh can be densely populated by not highly developed or highly urbanized.

Here are some countries by percent urbanized using World Bank data:

Sweden - 87%

Denmark - 88%

Norway - 82%

UK - 83%

US - 82%

Japan - 92%

China - 59%

India - 34%

Niger - 16%

To me the biggest unanswered question is why some Asian countries don't fit the pattern. Japan has the oldest population in the world (percent 65+) and the advanced economies such as S Korea and Taiwan are similar to the counties of Western Europe in terms of measures such as life expectancy, percent in poverty, etc. I suspect if you removed Japan, S Korea, Taiwan, Hong Kong, and Singapore your correlations would be even higher.

interested reader wrote:

WWRD wrote:My hunch is that wealthier countries are hit harder due to various factors, including more mature healthcare systems, which enablemore detailed reporting, and politics/cultural influences.

My instinct was to rebuff you out of hand, but then I slowed my roll, and decided to plot per capita GDP against tests per capita:

https://ibb.co/Vtx1bX6

That's among the strongest of all the correlations I've seen, nice call!

The problem with using test per capita as a metric is that it is dependent on a country's testing capabilities and strategy. Deaths per capita might be better but it still assumes countries test people who die in a consistent way.

There may be other factors at work to explain why high GDP countries have more cases/fatalities, in addition to age:

1. Those countries with high numbers of cases were hit with the virus first so they are just further along in the cycle. Now we are seeing countries like Brazil and Mexico with patterns very similar to those in Europe.

2. The virus seems to spread mostly inside and people in Europe were spending most of their time indoors when the virus was spreading rapidly. In tropical countries, especially outside of urban areas, people don't use/have air conditioning and have their houses open to airflow 24 hours. They buy their food in a outdoor market rather than a grocery store. So there may be some lifestyle differences that inhibit the spread of the virus in poor, tropical countries.

Sweden has a death rate 12 times that of neighboring Norway yet they are doing the right thing? Crazy how people can disregard facts in order to get to the conclusion they want.

Giles Corey wrote:

ex-runner wrote:

Please tell me how population density doesn't matter? If a country had a population of 1000 and each lived 1 mile apart from each other do you really think the spread would be as fast as if they all lived in the same sq mile?

Asian cities are by far the most densely populated in the planet.

Yet they are far from the most Covid infested.

How does that fit into your narrative?

Extremely strict lockdowns or strong early controls.

China lockdown speaks for itself. South Korea tested something like 40,000 a day and selectively isolating cases and tracing contacts before anyone in Europe had started testing their medical staff.

Found this article today anyway:

You described Brazil in no. 2.