Covid-19 age demographics and countries

Here is an excellent document, best material I have seen on this topic so far, by the excellent Elad Gil and Shin Kim.


'This post takes the simplistic approach to asking - if age alone is the primary correlate of IFR (Infection Fatality Rate), what should we expect different countries’ fatality rates to be?'

Which is interesting when you replace a region with a country in the following - 'The results of this simple-minded approach suggest that if age were the primary determinant of fatality rate, Italy should have 7-8X the COVID-19 fatality rate of' - wait for it - Germany.

Or is that the counterfactual of their simple approach, as currently, Germany has a death rate almost precisely 1/7 of Italy's, even with both having fairly similar age structures.

Italy has had 1 death per 7.3 officially recognized cases of COVID-19 while Germany has had 1 death per 23.9 cases. On the face of it that makes the plague 3 times deadlier in Italy, but those figures are unlikely to be readily comparable.

For one thing, COVID-19 took off around 10 days earlier in Italy. says Italy 486 deaths per million, and Germany 83 - 6x, not 7x.

Maybe not the best way to compare, but since no one knows how many unknown infections there are, it shows a certain ratio.

Yeah, deaths are harder to ignore. Of course, even there how they are counted can make a big difference. We may need to look up excess deaths to see if the official deaths look right.

That was amazingly inappropriate

Better than boredom.

Er... Prior approval has been wilfully dumb on understanding the unreliability of international comparison of case counts, and their relationship with far larger real infection pool, for roughly two months now. You're rather optimistic if you think he's suddenly going to have a Road to Damascus moment on this now.

You honestly think that Italy and Germany do not have comparable methods of counting their COVID19 dead, even after having it repeatedly pointed out that the head of the RKI says that Germany's methodology for counting is identical to Italy's, more than a month ago, when people thought that the explanation for the difference in fatalities must be a difference in how Italy and Germany were counting the dead? The bodies are fairly easy to count, even if case counts may not be comparable in light of the fact that in both Italy and Germany, the current official estimate is that confirmed cases represents 1/10 of total cases.

Now, if one were using unreliable numbers like those from the UK or France, you would have a much better point.

“Axis of counting”?

“Germany counts deaths in care homes only” if the authorities have bothered to test. Tests will also not be conducted during autopsies.

Any suspected Covid nursing home death in Germany is recorded as something else.

Something is missing from the literature, so far.
I am not sure the covid is harmful, by itself. The harm comes from white cells who cannot figure out how to digest it.

The patients who get antibodies early and survive without symptoms tell us something. It takes about two week for the B cells to try their possibilities. If they do not have a response, the B cell structure breaks down and macrophage takes over, and the macrophage, a type of B cell, is not very discriminant.

There is a piece of missing information, the virus does not seem to bother birds, or does it? What if the virus just lived with us, and the B cells ignored it, likely it might takes months to develop some permanent nerve damage, its the likely result. But then again, it may be no worse than smoking. If w can do something to shut the B cell system completely, including macrophage, say in rats, might be a great experiment to see what exactly is harmful about the thing.

Check the size of a covid:

The authors pointed to earlier research showing particles 0.04 to 0.2 μm "can penetrate surgical masks." For the coronavirus responsible for SARS, particles were estimated to be within that range at 0.08 to 0.14 μm, they said.

Vs incense smoke:

Our data showed that aerosols from incense smoke were spherical droplets with a count median diameter of about 0.13 μm and mass median aerodynamic diameter of 0.28 μm.

Something to remember. The idea that smoke particles compete for surface space with covid particles may mean something later.

How do we know the covid virus has fouled the b cell system. Because of cytokinetic storm. But we all have rhinovirus, the common cold. If the b cell system is fouled, one of those viruses might be the infecting agent, this has not been tested. Further, the b cell network seems to organize itself around one antigen at a time. The path to a covid antibody may be much longer if the b cell system is already fighting rhinovirus.

These data suggest that rhinoviruses and coronaviruses may be associated with serious respiratory illnesses in frail older adults.

Two viruses interfering.. Not much is know, they are just now figuring this out.

So roughly 15% of the the population in the US (age 65 or over) comprise about 85% of all C-19 coded deaths? A lot of provisional deaths too. Are those deaths where no testing was done, but doctor suspects or thinks C-19 contributed to death...or was pressured to record that way so hospital could gain its share of federal funds?

Basically, if you're under age 65, your chances of dying in auto wreck is 100x more likely than dying from coronavirus. No reason for lockdowns. Protect old and frail if you want, everyone else get back to work. Or stay home and collect your unemployment and disability checks.

Update: Updated CDC website May 1: drops total US COVID19 death rate from 64,283 to 37,308.

This is a hoax. The average joe is more level headed than you ivory tower types. Ohhh, scary virus, death lurks everywhere. Just stay home. It's all bullshit.

And now the CDC website is up to 68, 279.

Well that escalated quickly.

That's more than US deaths in Vietnam. It's also more than US combat deaths in WW1. Or not, since, as you say, it's just hoax. I expect one in 750 New Yorkers to jump out from behind the sofa and yells "surprise!" any moment now.

I agree with most of what you say but let's make sure to include the fuller context. The CDC counts death by major cause across age groups for 2018 here:

We see there were 59,120 deaths caused by pneumonia and influenza and 48,344 suicides. Covid-19 will certainly be a leading cause of death in 2020 in the U.S. but probably mostly in the 45+ age brackets.

Number of years lost and over what time interval is probably a better metric to be using to compare tragedies. Any death is tragic but, of course, someone spared today from covid-19 could just as easily have fallen victim to heart disease or cancer which killed almost 1 million elderly Americans in 2018.

I'd say the fuller context is the United States has proven itself unable to defend itself from a deadly cold and needs to turn things around because this failure has been noted by potential future enemies.

"You've been very hysterical and plain stupid about Covid-19 dangers... What should your punishment be when it turns out we have about 1,000 or less total US deaths (and less than 30,000 worldwide) when the virus dies out?" — Alvin, March 20

Now do a search for "millions", as in, MILLIONS WILL DIE

Probably all you.

+1. Doing the Lord's work, Troll. Keeping an honest comments' section is a thankless chore.

Now the surgeon general is dumping Gates and Fauci’s models and bullshit and just done to put fear into everyone. The sad thing most people fall for it.

Did you see that Neil Ferguson of the Imperial College 2.2 million coronavirus deaths had to quit after violating the shutdown rules? The little SOB catches Covid, but decides to bed down with his married mistress during the 14 day isolation window. Stated he "believed himself to be immune." What the hell? I guess rules only apply to sheep.

He is cut from the same cloth as Gates, Newsom and the rest of these sanctimonious autocrats.

I like your frothing.

I didn't;t like it as an epidemiologist, but now I like him much better as a man.

That is hilarious.

By far the best analysis I've yet seen.


what good are complex models if they distract us from basic facts and the simple strategies they suggest?

we should have tried to limit the attack rate in the most vulnerable populations by locking down nursing homes and isolating hospital patients suspected of having C19. focussing all of our efforts on those two aims would have saved many more lives at a tiny fraction of the cost.

I'm not surprised that the data show covid-19 hitting the elderly the hardest. What I found most surprising in this paper is their comorbidity table. Looking at their data source for that table, the rate of diabetes comorbidity is startling. Sure, hypertension is the biggest comorbidity, but only in 55% of the deaths, which is roughly the prevalence of hypertension in the general population. But 37.5% of coronavirus deaths in New York had diabetes as a comorbidity, far above the 10% prevalence of diabetes in the general population and 25% prevalence in the elderly. Dump the vague warnings about "underlying medical conditions" - diabetics need to be told they're especially vulnerable.

45% of American results have hypertension.

The column for “no underlying condition” was astounding.

This is a question seeking a response.

How good is any of the CFR data without conclusive reliable antibody testing?

I mean none of that data has any reliable antibody data built into it?

As I understand it, CFR by definition is the number of deaths attributed to a given disease divided by the diagnosed number of cases of that illness. IFR is the figure you are looking for if you want something that represents the "true" risk of dying from an illness.

Some of the skeptics here were pointing to Iceland and South Korea as proof that covid-19 was vastly less deadly than people were making it out to be. They certainly had reason to be concerned about undercounting the true number of infected people but these countries show that the alarmists were mostly right all along.

Iceland, which did some of the most rigorous testing anywhere in the world, has 10 deaths out of 1,799 cases or a bit more than 0.5% CFR. And, yes, their infected population is relatively young. South Korea, which did an impressive job of tracing and testing soon after they became a hotspot, has a 2.36% fatality rate but less than 1% fatality rate for people under 50.

It’s interesting to look at this data and the asymptotic rates in prisons and meat plants, both of which have younger populations.

Plugging in prison demographics, their model suggests an Expected Fatality Rate of 1/4th the US average, on par with sub-saharan Africa. Deaths attributed to Covid suggest reality is 1/2 the US actual fatality rate.

Is the infection twice as high in prisons as the rest of the US? Confirmed case rate is about 2.5x in prisons. Plenty of news stories about high rates in individual prisons, but I can't find good data on how many prisoners have been tested. If it's similar to the rest of the country, demographics explains the low fatality rate in prisons.

The UCLA data suggests no confirmed cases in 68% of prisons. What hasn't arrived, the virus or the test?

Agreed. Excellent article. Essentially, the younger and less skilled/educated bear most of the cost of the lockdown (unemployment) while older people reap most of the benefit and most higher skilled/educated are simply WFH. The Great Recession negatively impacted recent college graduates and other new labor market entrants. Twelve years later, when they are in their 30s and maybe starting families and carrying all the debt that goes along with it, they may face a long spell of unemployment. That cohort is already very left-leaning and have/have-not and old/young tension has been greater than the norm post-financial crisis/Great Recession. What will the political/social landscape look like over the next 10-20 years?

The proportion is probably similar to deaths caused by standard influenza; but, I'd like to see this data broken down further. What percentage of those elderly were living in nursing homes/retirement centers or were already hospitalized for other reasons? Those populations were de facto already isolated but their shared isolation seems to have been a major contributor to the death toll. Although co-morbidity plus age is predominately the fatal mix, an elderly person is probably safer on main street than confined in those places.

I now see that my wish had already been anticipated--see section 3 of the paper.

The graph "Expected Fatality Rates s vs. Reported CFRs" is the kicker. It is where theory meets practice in this paper.

Sadly, we don't know why some countries blow past this EFR.

Nor do we know why other countries, per this chart, outperform.

sure we do.

populations do not mix randomly and we do not test randomly, so early estimates of the case fatality rate will depend on where in the ''social graph'' Covid-19 started to spread and the criteria used to test/count cases.

importantly, if it spreads disproportionately among the sick and elderly first, your early estimates of the fatality rate will be too high. if it spreads disproportionately among the young, they'll be too low. in the longer run, underlying demographics and anything affecting the expected relative attack rate among the most vulnerable (more nursing homes and churches, etc.) will increase the eventual rate.

i should add here: it's still early for most countries.

probably. not for sure, but probably.

But if COVID-19 managed to "seek out" the elderly in a population, shouldn't that population show "extra" age bias in caseload?

Per this document NY actually has a lower care-home percentage of deaths than less impacted areas.

For what it's worth, I think this might relate to what some epidemiologists call "overshoot."

The disease has some virulence, and can perhaps be kept in check by some social isolation, but once you reach a critical mass things go pear-shaped.

NY should have less age bias than less affected areas.

Up to 1/3 of the city already caught the disease and is immune.

"Up to 1/3 of the city already caught the disease and is immune."

Some people link to data, some people make assertions. But of course anyone raised on the claims of television commercials will spot the "up to" in that claim.

If coronavirus deaths are 0.2% of population, in US that means 700.000 deaths.
If of those deaths 5% are younger than 50 years, that means 35.000 deaths of younger Americans.
Why is such vital data been so much ignored when debating what to do?
Have the older the right to impose lockdowns that will mortgage the younger’s future?

That's funny. I read the first half of your comment, and saw 700,000 and 35,000 has really big numbers.

My presumption was that no one would like those numbers, not even the young and self-interested.

I don't think your assumptions are realistic; but, that probably serves to reinforce your point. If one uses Italy as an example ( a hard hit country), there have been, as per the date of the study, 197 deaths in the under 50 age group attributed to the virus. That is .00005325819096 of the entire population of 60,463,000 persons. Your estimate of .00001 is very, very pessimistic.

On the other hand, I don't think that it is necessarily "the older" who are imposing the lockdown on the younger. From my experience (in France), it is a mix of generations who are for these extreme measures. And it is the media and the blogosphere (including this one) that have succeeded in creating an atmosphere of fear and hysteria. These are not "the older". This study convinces me further that the first priority should be to protect the elderly in nursing homes, long-term care and hospitals from the virus entering and spreading in those places. Thi sis nothing new--the same precautions should be taken for influenza, etc. For the rest, educating the public to take precautions and letting people move freely and businesses operate is the best policy for the near and longer term.

Yes, +1. The generation in power in France is clearly not the "boomers" anymore (which we would call in France may be the "soixante-huitards", from the students revolt of 1968), it is the generation X (my generation). Macrons' administration is made of people from 35 to 50, mostly. And, to my despair, they were in phase with everyone I know to impose and maintain for a long time a strict lock-down.

Contrast with:

"During a private call, Texas GOP Gov. Greg Abbott admitted "every scientific and medical report shows" state reopenings "ipso facto" lead to an increase in coronavirus cases, even as he announced plans that same week to end his state's stay-at-home order."

So, why choose death?

I'm afraid it is because Republicans have tried everything they think Republicans can do. They have exhausted inside-the-box options.

It is the same reason President Trump talks about winding down his coronavirus response team. They have done everything they consider a reasonable for themselves to do.

Who is choosing death? Lockdown orders are forceful; re-openings are voluntary.

That is a huge difference, one is violent, one is not.

This isn't exactly a super deadly disease...

"volent" lol.

No, I'm afraid this is a situation that brings to fore my most pessimistic take on American exceptionalism. It is that we have become exceptionally stubborn and exceptionally stupid.

We look at news like this:

"South Korea went from the largest ex-China COVID outbreak in the world to zero domestic cases in 60 days."

And then we just blink and look away. We go back to hoping that "Sweden will win" by dying instead.

This is an astoundingly ignorant take.

I hope you learn compassion at some point in your life; South Korea and Sweden took shockingly similar responses.

Nope. South Korea did aggressive track, trace, and quarantine.

In contrast:

"This work is taking place at a regional level, since in Sweden it is the country's 21 regions which govern healthcare, with collaboration across regional borders.

"We are having to use a lot of overtime, and we will need more people working with this. The region has a clear plan of how to do this," said Dotevall, who works in Västra Götaland. The region confirmed its first case of the coronavirus, also called COVID-19, on February 26th, and since then has been working to trace people who could have been infected since then.

It's a similar situation in Stockholm, where as of Friday more than half of the country's confirmed cases were located. Around 20 people in each region were working on contact tracing."

Do you think, in retrospect, 20 people was enough?

If anyone thought for a moment anonymous wasn't simply here to fight the culture war and nothing else, here's Exhibit A.

Note that whenever he attacks his tribal out-group he proposes nothing. Because he's not here to debate or discuss policy or approaches.

Keep in mind the plethora of states beginning their reopening process doesn't even cut across partisan lines. Apparently the Democratic Governor of Colorado has decided he's out of Republican box options.

kek anon, top kek

On the other hand your whole deal is that all criticisms are partisan, even the good ones. That's the reason you've been fighting your own peculiar rear-guard action for Donald Trump these last 3 years. "I don't like him" you say "but don't criticize him, that's partisan."

I'd hope that by May 2020 everybody saw through that.

He's at troll. Perhaps not intentionally, but generally he never applies his own standards to his own comments and resorts to all kinds of logical fallacies routinely.

Appeal to Authority - check
Motte and Bailey attack - check
Move the goal posts - check
Ad hominem response - check
Strawman comments - check, check , check

Even the response here was clearly a strawman argument.

And what do I see? Another guy who can't answer the question on it's merit, and has to turn ad hominem.

Basically, if you had a strong answer to why we can't be like South Korea, except because of our political ideologies you'd answer with that.

But you don't.

Sigh, it's clear that you don't know what and ad hominem attack is. An ad hominem attack is one that distracts from the argument.

I was responding to the comment:
"If anyone thought for a moment anonymous wasn't simply here to fight the culture war and nothing else"

Saying you are a troll that engages in logical fallacies is germane to that comment. Furthermore in your response you committed a classic strawman fallacy by claiming that Skeptical has positions that he clearly doesn't.

Then you respond to my post with another logical fallacy, namely you are trying to move the goal posts. from your behavior to "why we can't be like South Korea". This sub thread wasn't even about South Korea. South Korea never had a significant Lock Down. So, it's off topic for the discussion.

"How South Korea Is Beating Coronavirus Without a Lockdown "

However, for the record, I approve of South Korea's response, think we should be doing just what they are doing. IE Remove the lock down in most of the country and resort to Contact tracing.

So, we agree, in that the US should get rid of the Lock Down (at least outside of the NY cluster) and handle this just like South Korea.

You constantly join my discussions and then tell me what my discussions are about.

Buy some self-awareness for a dollar.

And if you approve of SK response, maybe you don't have a fight with me, other than as a "personal issue."

"You constantly join my discussions..."

This was a thread started by Tyler Cowen. And Per Kurowski made a comment to which many people made replies. It was NOT your discussion.

"Buy some self-awareness for a dollar."

The irony is dialed to 11 for that statement.

These are threaded comments. Anyone may begin a thread. Anyone may respond to a thread.

Internet 101.

You obviously joined this thread because you wanted to fight with me about America's Exceptionalism. Except you never really engaged with it. You haven't explained why exactly we can't ask our government to be as good as South Korea.

BTW, if you agree with it, are you demanding it? If not, why not?

To me the only question is why were red state governors so much better equipped to prevent deaths than Democrat governors. Are they likely higher IQ, healthier, and of higher T levels- yes almost certainly. but is that the only reason? Is it just that red state governors are smarter people? I wonder. I tend to think the biggest problem is blue states are full of people who think hygiene is optional.

Maybe it’s also related to how much bigger republican’s dicks are than Democrat’s dicks. Speaking primarily of the non-black dems like you.

To ask you the direct question, JWatts:

"By the time the World Health Organization issued its plea in mid-March for countries to “test, test, test”, South Korea had spent weeks doing just that, quickly developing the capability to test an average of 12,000 people – and sometimes as many as 20,000 – a day at hundreds of drive-through and walk-in testing centres. The mobile centres conducted the tests free of charge within 10 minutes, with the results were sent to people’s phones within 24 hours. By mid-March more than 270,000 people had been tested."

Why can't we demand that of our national government?

"Why can't we demand that of our national government?"

12K tests per day for South Korea is 12K/51M = 0.02% of the population per day

The US is currently averaging about 250K tests per day.
That's 250K/328M = 0.08% of the population.

Congratulations, your demands were answered.

Did you just punt on national trace, track, quarantine and support?

Typical. When presented with facts, did you change your mind? Did you reflect and admit that perhaps your priors were wrong and outdated?

No, you moved the goal posts. This is why I label you a Troll.

I assumed that you read my link.

If you did not, you aren't replying to me, you are playing games.

No, he's saying that we've met and exceeded the percentage of testing the S korea was doing, meeting your demands. I assumed you read his response. If you fit not that you are playing games. Maybe the problem isn't the reading of his response but the actual connecting of the dots of what he was saying, maybe that is beyond your reach. So Ive done it for you.

"And it is the media and the blogosphere (including this one) that have succeeded in creating an atmosphere of fear and hysteria."

On the one hand, we hear claims that people are overreacting and engaging in "hysteria." On the other, we hear claims from some quarters that China behaved wrongly by not alerting the world to the dangers of covid-19 earlier. Which narrative is true? If the former is true, does Wuhan really have enough nursing homes to have contributed to the death toll that even the Chinese government admits to there? All the signs point to the fact that the Chinese government was desperate to believe that the virus was no big deal but people kept on dying and it became impossible for the government to ignore by late January.

It is not as if no one ever thought of the idea of just encouraging people to take modest precautions and otherwise just go out and live their lives. The problem, as Mike Tyson elegantly put it, is that everyone has a plan until they get punched in the face. One by one, China, Italy, the U.K. and the U.S. each tried pretending and hoping this would be no big deal and each got smacked in the face by reality.

There are about 225 million Americans under age 55. If we take the German fatality rate for under-60s as a rough estimate of the fatality rate we can expect for Americans under 55, it is 0.2% (German number is provided in the linked paper).

If you "let it rip," the death toll is merely a function of how many people ultimately get the virus. If 50 million Americans under age 55 get it, we are talking 100,000 deaths and not 35,000. You are assuming that the virus just stops infecting people for some reason -- what would that reason be?

Are you using IFR or CFR to get the .02 number? Because I don’t think the actual fatality rate for under 50’s is anywhere near that number. There’s just no way. How are you getting .02???

I don't know what you mean by "0.02." German case fatality rate is 0.2% for age group 0-60 -- this number is provided on page 3 of the paper. If you want to corroborate that number, you can look at South Korea, which is reporting a CFR of 0.21% for ages 40-49, 0.17% for ages 30-39 and 0 for under 30s.

Basic math *rolls eyes*

If you want to use Germany as the example, fine, but then you need to accept all the data and not just the data you choose. Read the chart--for Germany, the "attack rate" for the under-60 cohort is just 0.1 percent. *Of those*, the fatality rate is 0.2 percent. If 0.1 percent of Americans under age 55 "get it", that's 225,000. The fatality rate for that group being 0.2 percent results in 450 deaths but probably closer to 500 due to rounding. Granted, the game is not over, but I don't think you are in the ballpark.

"Sloppy writing is a symptom of sloppy thinking"

This document discusses CFR's but fails to define it. They however define IFR but don't use it. Actually, they just tell wat the 3 letters stand for but no more.

I usually skip this kind of texts if it has this kind of errors in it especially if the autors pretend to be serious. They are not.

btw: if you this is " an excellent document, best material I have seen on this topic" then your standards are pretty low.

I also searched in vain for a definition of "CFR". I suppose it isn't Code of Federal Regulations. I was also puzzled by the comorbidity "Demantia", I've never heard of that. And where's obesity? (or BMI) You'll also note that China is (glaringly) missing in some of the data graphs/tables. Finally, that older people are at higher risk has probably been known we've had old people and we've known for years that nursing homes are especially problematic. To sum up, I'm not impressed. I should also note that the data quality from Africa is unlikely to be comparable to the the developed world's. And why isn't more being said about the disparity in sex of IFRs, it had some legs in Jan/Feb but has faded away. This blog pointed to an article which contradicted the idea that smoking was the actual culprit, since more men than women smoke. So, why the silence? /. about a mutant version from Europe being more lethal (and now the dominant virus since it's also suggested to be more infectious).

Excellent article indeed. There are some issues with all data, though. We all know about the lack of testing, particularly in the US when compared to some Asian and European countries. However, I don’t read much about the reliability of tests. I have heard many anecdotes from health professionals and patients that the tests are very unreliable, often showing negative results even when patients clearly have COVID symptoms. Why is that not addressed in the literature or on blogs like this one?

Oh boy another COVID paper written by a pair of non-experts one of whom identifies as an "Entrepreneur and Investor." Thank god, because the Silicon Valley community had been pretty silent on the issue up to this point. Surely these people aren't suffering from their own Halo effect, and this would pass peer review from epidemiologists *eye roll*

Watch as the "confirmation bias" crowd ("best material I have seen on this topic so far"? really? In what position are you to judge?) whines about appeals to authority.

1. Sounds like someone throw together a bunch of random jargon with a link to a paper in an attempt to attract attention.

"We probably now have to redo nearly all the vaccine designs and serology assays. "

Because of course you know what a "serology assay" is.

"What I’m guessing is happening is the immune system is noticing ORF3b/ORF8 on viral-infected cells & inducing cytotoxic response. "

Because of course you knows what a "cytotoxic response" is. And what ORF3b is, for that matter.

Comments for this post are closed