In Belgium, 1/2000th of the Population Has Died from Coronavirus

Belgium's number, which just passed 500 deaths per million, is the highest of any substantially sized country. (I set aside tiny San Marino, in which the rate is more than double that.) In the U.S., this would be equivalent of about 160,000 deaths, rather than the 40,000 we've suffered so far, though of course we should keep in mind the possibility that different countries' numbers are hard to compare because of different reporting practices. And while it looks like the Belgian death rate is not increasing any more, and may even have begun to decline, there still seem to likely be many more deaths to come.

Comments

[Á àß äẞç ãþÇđ âÞ¢Đæ ǎB€Ðëf ảhf]

I get the impression there is huge variability in what counts as a coronavirus death -- some count everyone with coronavirus no matter how slight, some count only deaths in a hospital, etc. It will be interesting when someone publishes a comparison of all this.

[MatthewSlyfield]

" some count only deaths in a hospital, etc."

I've seen a suggestion that the Italian government was counting all deaths in a hospital even if that hospital had only one corona patient.

[Dr. Ed]

I'm reminded of the "body counts" in Vietnam and the so-called "5 O'Clock Follies."

[David Nieporent]

I've seen a suggestion that Covid is a hoax arranged by Bill Gates to implant microchips in everyone. Given how many nuts there are in the world, one can see a suggestion of just about anything. Not sure why you thought it was useful to relay such an obviously false suggestion.

[librarian]

Maybe this is too early to suggest, but I wonder if the death rate will be lower than average for some period of time after the virus has passed.

[Dr. Ed]

It statistically has to be as it is reducing the number of people with underlying comorbidities and hence those people won't be around to die from those comorbidities in the next 18-36 months.

It's like what Obama's Cash-for-Clunkers did to the supply of used cars.

[Life of Brian]

I wonder if the death rate will be lower than average for some period of time after the virus has passed

Apparently we're already there. According to the CDC's historical data (available here, and updated every week), all-cause mortality for this flu year to date (starting on week 40 of the prior year) is the second-lowest in the past 5 years.

And that's just a comparison of the raw numbers. The contrast becomes even more extreme when you correct for the expected year-over-year baseline increase (conservatively around 0.5% per year just from population growth, but probably more like 1% as we work through the Boomer bubble).

Yes, I understand the numbers for the last week or two aren't completely reported, and yes, I snipped them off. Download and graph it yourself -- it's quite illuminating.

[mydisplayname]

Yes, "different countries' numbers are hard to compare because of different reporting practices," as are the numbers from different US states. But my fear is that our tolerance for disease and death might be unreasonably low: if a generally non-fatal illness (about 80.70% of CoViD-19 sufferers recover without hospitalization) causes this much drama, what would a single puff of highly contagious, highly toxic smallpox do? What would we be doing if CoViD-19 was as widespread and as deadly as HIV, where 1 in 25 Africans are infected?

[NToJ]

It's possible that 1 in 25 Americans have already been infected by the virus that causes COVID.

[Armchair Lawyer]

Certainly the Stanford study would indicate that is likely.

[wnoise]

You can't conclude anything from the Stanford study -- it got results entirely consistent with everyone of those positives being false positives. And it wasn't an actual random sample. And... well, read this:

https://statmodeling.stat.columbia.edu/2020/04/19/fatal-flaws-in-stanford-study-of-coronavirus-prevalence/

[Armchair Lawyer]

"You can't conclude anything from the Stanford study"

I wouldn't go that far. Let's break down the two major issues.

1. False positive rate.
-If the false positive rate is really that high, then the false positive rate in the standard testing should also be that high. That of course is extremely useful knowledge to have, and it would mean that the number of cases we're seeing is being overestimated by a large margin. Now, there may be discrepancies in the types of testing (antibody versus PCR), but as the antibody-based testing is moving forward as it's higher throughput...extremely valuable knowledge.

2. Non-random sampling

Selection bias is a real problem in any type of survey procedures, and many do their best to adjust. That bias can go pro (more people want to be tested) or anti (people don't want to be tested). However, the other more limited US studies, if anything show more dramatic rates of infection. In Manhattan, for example, two maternity wards decided to test every single mother coming through (215 in all) for the Covid-19. There's no self-selection bias there, and a 15% infection rate was found. A homeless shelter in Boston tested all their residents, and the infection rate was even higher (35%). 200 Chelsea residents (MA) who were on the street were randomly tested by researchers, and a 30% infection rate was found.

I would expect the infection rates are actually a lot higher than initially anticipated, believed on this data, and that the non-symptomatic rate is much higher as well. The Stanford study may very well be a lower bound on the infection rate.

[Dr. Ed]

HIV clearly isn't as deadly as it was in the 1980's, it isn't even full-blown AIDS -- and how deadly is it today?

[Armchair Lawyer]

HIV is far, far, more deadly than the coronavirus. It is only held at bay (not cured, held at bay) with extensive pharmacological treatment, developed through decades of research. And if you get HIV, you either get that treatment, or you die.

With COVID-19, the mortality rate is much lower. Most people who get COVID-19 survive without any treatment.

This isn't meant to understate the severity of COVID-19. COVID-19 is far more easily transmitted than HIV. But it is far, far, less lethal.

[Dr. Ed]

And who paid for these decades of research that created the drugs that are keeping the Africans alive? That's not often mentioned, is it?

[Sarcastr0]

Because people generally think of it as research that is keeping people alive.

[AmosArch]

It can't be that serious since California practically decriminalized spreading it intentionally.

[Dr. Ed]

1:2000 is 0.05%
Average annual on-the-job death rate for loggers: 0.1%

Let's keep these numbers in perspective....

[loki13]

"Let’s keep these numbers in perspective…."

Sure, let's!

66 loggers die every year (on average- 2006-2015, source BLS).
We have over 40,000 deaths so far this year from Covid-19.

See, statistics are fun!!!!

Here's another one-

99% of things Dr. Ed says are wrong.

[Armchair Lawyer]

Here's a good question for you Loki.

How many Americans die every year?

[Dr. Ed]

See Chart #4 and remember that this is FTE and not headcount.
https://www.bls.gov/news.release/pdf/cfoi.pdf

I really don't make things up.

[Lawification]

More logging stats here:

https://www.bls.gov/opub/btn/volume-7/the-facts-of-the-faller-occupational-injuries-illnesses-and-fatalities-to-loggers-2006-2015.htm

Very interesting. Thanks Dr. Ed.

[jubulent]

Or about the same as on-the-job deaths of roofing and garbage men.

[Armchair Lawyer]

I would take the Belgian numbers with a large grain of salt.

54% of the deaths Belgium is reporting as a coronavirus death are from nursing homes, and of those, most importantly, only 7.8% actually have a positive coronavirus test associated with them. The other 92% are "suspected" deaths.

At the end of the day, you'll need to compare the deaths in the nursing homes against a baseline level of expected deaths to have a real appreciation for the actual numbers.

[Leo Marvin]

The other 92% are “suspected” deaths.

For those already buried I'll go out on a limb and say they're probably dead; cremated, almost certainly.

[Armchair Lawyer]

Sigh. Suspected coronavirus deaths

[NoVaNick]

It’s like the 440,000 deaths from smoking/tobacco in the US we hear so much about. We’ve been hearing this for 20 years at least, even as the number of smokers now is half what it was then. So they must count every death from anyone who has ever smoked or lived/worked with a smoker as smoking-related. My parents never smoked, but both worked with smokers from the 1950s -90s and died of strokes in their late 70s (they had underlying health issues too). So their deaths would surely be among the 440k. Except now, if they also had CoV-2 in their bodies, it would be Covid-19 that killed them.

[NToJ]

"We’ve been hearing this for 20 years at least, even as the number of smokers now is half what it was then."

I don't have a dog in this hunt, but cigarettes don't kill you when you smoke. They kill you decades later.

[Dr. Ed]

Unless you smoke in bed...

[MatthewSlyfield]

Or an easy chair... Basically if you fall asleep while smoking anywhere near flammable furniture.

[Rossami]

Assuming the rate of deaths was at steady-state before the decline in the number of smokers, even an average lag of decades should still have shown a decline in deaths by now.

I concur that the lag means we should not expect the 440k to have dropped to 220k (proportional to the ratio of smokers) but it should definitely be well below 400k by now. It is implausible that the actual number is precisely unchanged across this time.

[NToJ]

The population has increased in that time frame as well.

[Dr. Ed]

Another aspect of this is the decline of college attendance in the fall.
Vermont's State System is considering closing down three colleges for the fall -- which is freaking out the state. I can't help but wonder how many private colleges will go under as well.

See: https://vtdigger.org/2020/04/17/vermont-state-colleges-chancellor-to-recommend-closing-three-campuses/

[Bob from Ohio]

168,000 if it was the US.

But I have been assured our response was the worst in the world.

[Jason Cavanaugh]

Let's play a game and see how long it takes you to figure out the primary error of your argument.

And yes, our response has been terrible. No surprise given who's at the top of the totem pole right now and his propensity to be idiotic and selfish.

[Bob from Ohio]

Our response has been fine. Except in NYC which has 10 times the cases and deaths it should have based on population and seeded many of the other bad spots.

"idiotic and selfish"

Cuomo kept the subways open and then allowed them to reduce service, which meant crowded cars. Not surprising since brother "Fredo" went out in the community while actually sick with it.

[Bill Bowser]

New York State, which has about 46% more people than Belgium, has lost 1,100 people per million. More than twice the rate in Belgium. The death rate in New Jersey is slightly greater than Belgium also.

[Martinned]

More evidence just came out on this "differences in reporting" point: https://twitter.com/jdportes/status/1252640438122680320/photo/1

(Table taken from the NY Times.)

TL;DR, while most countries have "excess deaths" in the low double-digits, i.e. around 10%-30% more deaths than usual in this period, there are large difference in how that excess deaths number compares to the number of reported Covid deaths. In England & Wales, for example, there are 16,700 excess deaths compared to 10,335 reported Covid deaths (at the time), leaving a difference of almost 40%. In Belgium and Sweden, on the other hand, that difference is essentially zero.

[NOVA Lawyer]

I am re-posting without links because with links is awaiting moderation. It is easy to google the linked pages.

Life of Brian,

It’s not just the numbers from the last week or two that aren’t completely reported, the revisions keep happening for up to two months, as far as I can tell. I originally looked for the data to see if all-cause mortality increased and by how much. (The UK, the Netherlands, Italy, and Spain all seem to have more up-to-date numbers that show huge spikes in all-cause mortality far above historical and predicted values.)

And if you compare the numbers from the graph you shared with today’s numbers from the CDC, there are changes in all the numbers from April and March (all revised upwards), such that the numbers for week 14 (which is the week ended April 4, 2020 on the CDC website) are, as of my comment, at 101% of expected (56,195 compared to your chart’s 48,292 to give an idea of the size of the revisions even three weeks out). The highest numbers this year had been about 96% of expected (i.e., a somewhat low death year). Obviously, late March and early April is when C-19 deaths started to take off.

And those numbers from week 14 will be revised upwards again (from 48k, now to 56k, and by next week something more than 56k) if practice over the next few weeks follows the updates I have observed over the last few weeks. I fully expect the remaining numbers for April, at least, to be well in excess of 100% of expected when fully tallied and would be somewhat surprised if actual numbers didn’t exceed 120% of expected in at least one of the weeks ending April 11, 18, or 25.

Of course, even with all that caveat. What you have correctly identified is that the 2019-20 flu season was slightly below the predicted value based on prior years, which says nothing about the impact of the coronavirus as flu season was essentially over (though March can be significant) by the time coronavirus starting having a statistically significant impact in the US.

I think you’ll find the story told by UK’s Weekly All-Cause Mortality Surveillance in Week 16 (Week 15 data) more up-to-date and compelling regarding the effect of C-19 on overall mortality. These numbers are also subject to revision, but they show predicted deaths to be about 10,000 for week 15, upper 10-z limit (the highest bound) to be about 14,000, the highest peak in the awful 2017-2018 flu season to be about 14,000, and an actual spike from a normal year beginning in about week 13 of 2020 to something more than 20,000 all-cause deaths in the UK in week 15 of 2020. That is, all-cause mortality appears from this preliminary data to be well over 50% higher than normal for week 15.

C-19 is having a huge impact on all-cause mortality.