sweden wins again

the all cause deaths data is in for the 2020-21 flu year

many have attempted to malign sweden on covid. mostly, this seems political. they did not adopt the draconian policies of lock down, mask up, and panic, panic, panic of the rest of the world.

unfortunately, this tendency was inflamed and enabled by the fact that for some reason i have never managed to really get to the bottom of, the swedes adopted the incredibly inclusive counting policy for covid deaths of “any death for any reason within 30 days of any covid diagnosis, symptomatic or no.”

this led piles of alleged “covid deaths” (with roughly 75% in hospitals and nursing homes) that has little or nothing to do with covid. covid is an effective opportunistic infection and PCR tests are extremely sensitive. when you’re weak and dying anyway, it moves in. and you get called a covid death even for trace virus. but a lot of these people were dying anyway. this is borne out in the numbers.

  • ~90% of swedish deaths were in people over 70 years old.

  • the average death was over 81 years old. (81.75 is my best estimate).

  • life expectancy in sweden is 82.96.

and this is why “all cause mortality” (ACM) has been such a better metric to gauge the swedish covid response.

  1. it’s a clear, sensible definition with no ambiguity or bias

  2. it captures ALL trades offs from covid deaths to deaths of despair from lockdowns to 20 other knock on effects we might not even have considered. it’s the real, no-exclusions butcher bill. it’s the best basis for a full accounting of public health effects on mortality.

  3. everyone counts it the same way. this means countries can be readily compared.

longstanding gatopal™ @HaroldofWorld has been doing lot of great work here. this is his latest.

note that he’s using the CDC flu year of oct-sept. they do this to capture the whole nov-feb peak in one season rather than split it by calendar year. i think this makes more sense. so do most epidemiologists. otherwise, you’re splicing the second half of one flu season together with the first half of the next.

(also note that he’s cheekily calling it the “gato” year in reference to a certain internet feline who has endlessly hectored over the preferability of flu year to calendar for ACM. it’s good to have pals.)

as can be readily seen, 19/20 was no big deal by the standards of the last decade and 20/21 is looking lower still. this is despite19/20 following a very low year (lowest in swedish history) in 18/19 in a series with mean reversion tendencies. (if you’re very old and do not die in one year, your chances of so doing in the next rise)

it’s even more obvious is we zoom in:

some have tried to argue that 18/19 somehow constitutes a “downtrend” that 2020 “broke” but this is simply not in the data. from 13/14 on, this has been pancake flat apart from one oddly low year. there is no downtrend, just a single outlier (in a year with very low ACM all over the world)

neither of the last two years is even noteworthy in comparison to the years before them. 20/21 is shaping up to be the second lowest in swedish history. hardly the stuff of calamities…

sweden looked very much like the other nordics, none of whom had a bad (or even noticeably variant) all cause mortality year.

contrast this to the US data (through aug, not sept as it’s slower to report, but still highly comparable)

sweden was normal in 2020 vs last 5 years and dropping in 2021.

the US was up 22% in 2020 and rising yoy for aug, looking more like up 26% in 2021. these are numbers not seen in decades.

in a reasonable and sane world where one cared about real health and societal outcomes, there would be a 20 foot tall rocky balboa style statue of anders tegnell under construction in front of the nobel museum in stockholm.

but, alas, as such is not the world we inhabit, this by all accounts reasonable and humble man is going to get maligned, misrepresented, and ignored.

but he got it right. we got it wrong.

most of the west abandoned science for talismanic superstition while a brave few stood firm.

well played, anders.

well played.