Quantifying the human costs of lockdowns and restrictions

[Street Pharmacist]

Originally Posted by CASe333

Excess deaths is an important statistic because it helps measure the total impact of the pandemic. It avoids mistakes from countries or regions that under reported Covid related deaths as well as accounts for deaths that would have occurred anyways. I do think it is much too soon to start reaching for any kind of conclusions but the data Gullfoss presented is something I think can invoke good discussion.

On that note here is my observation that I am really puzzled about:

I've been looking at excess deaths using:
https://ourworldindata.org/
If there is any reason their data is flawed please let me know.

If you go down to the second graph plotting excess deaths by age and countries there is no surprise that for most countries the 85+ and 65-74 age groups have spikes in the spring when we first locked down. Its lesser pronounced in the fall because data could be incomplete or the fall flu season does contribute a large number of deaths anyways so excess deaths is at a wash but that can be debated.

What I find perplexing is that if you plot ages 15-64 for almost every country it looks like noise between +/-10% excess deaths. The exceptions are Spain, Italy and England who had their healthcare systems overloaded in the spring. Looking at Germany, Finland, Portugal, Canada, Israel or Sweden or any other country they've counted is close to average when taking the running avg over the year. There is no obvious spike for most unless their public health systems were overwhelmed.

The one exception is the US. Their 15-64 excess deaths mirrors their excess deaths for the older age groups throughout the whole pandemic and even exceeds it % wise in June-July. This is completely different than the trends seen in all other countries. Since mid March the excess death rate for 15-64 in the US has consistently exceeded 20% according to this data! This doesn't make sense to me since their health care was said never to be at max.

My hypothesis is that this is one of the best indicators of how broken the American health system is compared to the rest of the world.
My wife suggested obesity is rampant in the US and could be a factor too but I don't see how this explains the ongoing trend.
A third hypothesis is the data is incomplete and could be faulty too. This always is important to consider

Curious to other interpretations.

Looking at raw data for percentages of deaths ignores the difference in composition of those who got sick. Maybe the US had a much higher level of infection in that age group, leading to a higher death toll?

[opendoor]

Interestingly, the latest excess mortality release from Statistics Canada (covering January - October) discussed this. Based on the data, it looks like excess deaths among those under 45 (particularly males) has been a very real thing from May onward. However it is almost solely occurring in BC and Alberta; other provinces don't seem to really be registering excess deaths in that age group:

While, overall, mortality levels since March have largely been driven by the number of deaths due to COVID-19, a different pattern among people under the age of 45 emerged in May and continued into October.

From mid-May to mid-October, an estimated 7,172 deaths were reported among Canadians aged 0 to 44, an excess of 1,385 deaths. Males accounted for 81% of these excess deaths.

This pattern was seen particularly in Alberta and British Columbia, which were the only provinces with evidence of excess mortality among this age group. Alberta accounted for 298 excess deaths for males under the age of 45, while British Columbia accounted for 260.

https://www150.statcan.gc.ca/n1/dail...01224b-eng.htm

The only real caveat would be that they use a somewhat liberal definition of excess deaths, where any deaths over the average expected number of deaths are considered excess (so there were weeks with excess deaths in January for instance). This is in contrast with the US CDC, which generally uses the upper threshold of a 95% confidence interval. The latter is naturally going to be much more conservative and will tend to filter out normal fluctuations from registering as excess.

That said, the amount of mortality and the week-to-week consistency is something that can't really be ignored. Though the regional differences are puzzling, particularly given that BC and Alberta had some of the most liberal restrictions in the country in that period.

[Dan02]

I find the Covid data interesting because of the stories it tells that the official story doesn't.

For example, the first official USA Covid death was recorded February 29th. However, the excess deaths in the USA started rising the week previous, February 23-29. The number if small enough to fit within their 95% bound but it's a clear trend and the balance of probabilities is undiagnosed Coronavirus infections.

[Street Pharmacist]

Originally Posted by opendoor

Interestingly, the latest excess mortality release from Statistics Canada (covering January - October) discussed this. Based on the data, it looks like excess deaths among those under 45 (particularly males) has been a very real thing from May onward. However it is almost solely occurring in BC and Alberta; other provinces don't seem to really be registering excess deaths in that age group:

https://www150.statcan.gc.ca/n1/dail...01224b-eng.htm

The only real caveat would be that they use a somewhat liberal definition of excess deaths, where any deaths over the average expected number of deaths are considered excess (so there were weeks with excess deaths in January for instance). This is in contrast with the US CDC, which generally uses the upper threshold of a 95% confidence interval. The latter is naturally going to be much more conservative and will tend to filter out normal fluctuations from registering as excess.

That said, the amount of mortality and the week-to-week consistency is something that can't really be ignored. Though the regional differences are puzzling, particularly given that BC and Alberta had some of the most liberal restrictions in the country in that period.

I do wonder what the spike in overdose deaths has done here in BC. Not sure what that looks like in Alberta, but that age group (and mostly males) is who we're seeing dying here. By the end of November, 1548 people had died in 2020 of a drug overdose vs 983 all of last year. That would explain a good chunk of BC's deaths