Comment by salimmadjd | Hacker News Reader

Comment by salimmadjd | Hacker News Reader

There is a lengthier interview by UnHeard with Professor Johan Giesecke (another experienced epidemiologist) of Sweden on YouTube [0]. It’s worthwhile listening to their perspective.

One of the comments he made was that no lockdown would work in democratic countries and the real test will be fatality rates between Sweden and others in 1 year. He believed it’ll be very little difference between those countries that have implement the lockdown and those that haven’t

I had watched the simulation of covid19 infection by 1blue3brown[1] and some of his modeling appeared to me (maybe I misunderstood them) to confirm Swede’s point of view.

Ultimately the same number of people got infected albeit in a “flatter” rate. However one counter point by Johan Giesecke was that these flattening the curve is motivated by the assumption of a static capacity of ER rooms to deal with this pandemic and he pointed out they have increased capacity. Overall I found listening to his side very informative, at least it gave me a different view of looking at from their view.

[0] https://youtu.be/bfN2JWifLCY

[1] https://youtu.be/gxAaO2rsdIs

You don't have to assume static ER capacity and the like to be logically in favor of flattening the curve. You can assume a 100% increase in COVID treatment capacity or even 300% and still note that without any flattening you are going to have series logistical issues with disposing with all the bodies that drop day after day after day. I agree that roughly the same number of people are going to get it over a year, but the goal is for them to get it over a year instead of 1 to 2 months. That's a huge difference in capacity constraints.

It is also why this approach would be foolish in countries with less developed treatment capacity. Capacity, for example, that is already at it's limit even with social distancing.

So, did Sweden have logistical problems? If not, how is their way wrong?

Inability to dispose with bodies is a reason to put 16 million Americans out of work and cancel school?

With such an enormously high r0 as currently predicted, completely stopping the virus once firmly established is an impossibility, but many examples e.g. Finland and Korea have shown that you can seriously reduce infections and deaths. There's no reason a lockdown followed by extensive testing shouldn't work, and indeed it _has_ worked in countries such as Korea and Taiwan. In fact, it worked in Hong Kong and Canada during the original SARS epidemic.

There has to be a willingness to act early and responsibly, closing your borders, attempting to isolate highly infected areas, and slowing the spread of the disease as you dramatically ramp up testing. It seems as if the Western world has complete amnesia of previous events like SARS and is throwing caution to the wind as they assert that nothing could have been done to prevent this.

It's a policy choice to allow all of the reaction to work through the private sector instead of just deploying a national response strategy. In the U.S., the CDC tried to take the lead on testing (good) and essentially banned any other testing infrastructure (bad) and fell on its face with contamination issues (very bad), and pivoting to opening the testing infrastructure was a good policy choice. But the real answer has always been: DO BOTH. The CDC could be doing a million tests a day right now. Why don't they? It's madness.

Reminds me of "Contact". Why build one, when you can build two for twice the price? :)

The reason the CDC aren't doing a million tests a day right now is because there is a massive global shortage of every single consumable needed to do this and no easy way to expand manufacturing. Even Germany, which has pretty much the best testing capacity for coronavirus of any country out there in per-capita terms, has been stalled at about 50,000 tests per day for a few weeks and it's not for any lack of desire to expand that. Trusted American media institutions like the New York Times and their Editorial Board have been giving people wildly overblown ideas about what would be possible if it wasn't for the hated Trump and his incompetence for stupid, partisan reasons.

There is no doubt that there are shortages of some consumables; namely the swab. But every single? That is utter nonsense. The lysis reagents, extraction reagents, nucleotides, enzymes, primers probes etc. are available today by the metric tonne. The swab shortage could be designed around, but they don't. No one wants to go out on a limb and glue a Q-tip to the end of a thin flexible coffee stirrer, but i bet you it would work. Something thinner in all dimensions and one-piece would be better. Are we really ready to believe that this marvel of swab technology can't be produced at scale in a minimal amount of time from scratch? That we really are screwed because they are largely made in Northern Italy?

Hong Kong had 300 deaths from SARS, they had over 600 suicides above the average during the same period.

This time around they are keeping businesses open

Two points: - Your numbers don't indicate the number of lives saved by shutting down. It could well have been more than 600. - It seems likely that it is possible to reduce the number of suicides in such an event with better planning and support for vulnerable people.

their experts conclude they overacted with how they dealt with SARS

Okay, so mathematically proven that for every life lost to a respiratory disease two more will commit suicide. Impeccable math of yours.

China is the most prominent example of how dramatic an effect quarantines and social distancing can have. There was a massive outbreak in Wuhan, with tens of thousands sick. Two months later, there were almost no new cases.

Extensive testing and contact tracing currently seems to be keeping case numbers under control. We'll see if they eventually have to go back into lockdown mode, but for the moment, it looks like the measures taken turned things around.

In Wuhan, the army was called in. Roads out of the city were not only blocked off but actually torn up in some places. People were strictly limited to their housing blocks and had to await food deliveries. The Chinese state was able to rely on the fact that each apartment block already had its appointed minor Communist Party official who could ensure discipline. The population already had a single government-controlled app for communications, and this could easily be upgraded to show the three-colour health code before citizens were allowed to move anywhere in public.

A lockdown so strict simply would not work in most Western countries. The legislation is simply not there for that, even with the emergency powers that some authorities have, and enacting such legislation would require time and debate. Also, China was deeply concerned with protecting its elderly people due to such ingrained cultural respect for the elderly, but in other countries citizens may well feel that, if the army has to be called in, better that the army help with triaging and disposing of the mainly elderly victims than enforcing a crackdown on the population at large.

> lockdown followed by extensive testing shouldn't work, and indeed it _has_ worked in countries such as Korea and Taiwan

Korea and Taiwan had no "lockdown", they had contact tracing and targeted quarantine. They never closed non-essential businesses, and Korea didn't even ban flights.

That was a good interview, and obviously he's very experienced, but I'm still concerned with a lot of things he said.

First of all, he claimed that the death rate was going to turn out to be 0.1%; but he didn't even give a hand-wavy answer for why he thinks that. Now obviously, he's a professional who's been studying epidemiology for decades, so his "gut feel" is actually worth something; but I'd feel a lot better about that if he at least gave some other evidence that corresponds to that.

Later in the show, when the interviewer asks him how many deaths he thinks the UK might have in total, he says "12,000"; and the interviewer says, "We've already had 13,000". I mean... dude, your model is obviously off by at least a factor of 2 or 3; what if it's a factor of 10?

He claimed that Korea was backtracking because they couldn't contain it; but looking at the numbers just today, they're still above 10,000, which is what they've been at for weeks now. (I think he may have gotten Korea confused with Japan, who do seem to be having hundreds of new cases on a regular basis.)

But, I mean -- he does have a point: suppose you get to the place where Korea and Taiwan are. Are you really going to keep the country hermetically sealed for the next 18 months while we wait for a vaccine? Are you really going to keep people under lockdown that long?

If you don't have a coherent plan to prevent sickness, but only to delay it, then this "controlled burn" approach -- where you try to let your country become infected as fast as your health capacity will safely allow -- is going to lead to exactly the same number of deaths in the end, but will reduce the economic and societal impact greatly.

>but I'd feel a lot better about that if he at least gave some other evidence that corresponds to that.

I can't speak for the man himself but recent evidence from California, the Netherlands and Germany (Heinsberg) suggests that the true number of infections may be underestimated by a factor of 20x-100x, which would drastically push down the fatality rate.

I read a blog analyzing the numbers out of California, which said that they'd made some pretty critical errors; namely, it wasn't clear what the false-positive rate of their test was, and that even a fairly small false-positive rate would put "0" within their error bars.

Germany is running nearly half a million tests per week now against a population of 83 million. I can believe that their true infection number is 2 or maybe even 3 times the official number, but I have a pretty hard time believing 20x.

BTW, here's a link to the blog I mentioned:

https://statmodeling.stat.columbia.edu/2020/04/19/fatal-flaw...

People are grasping at straws. It's so horrendous to believe that the world will be struggling with the pandemic for 18-24 months that any plausible-sounding indication otherwise is taken as truth.

Been seeing it for three or four weeks already, ever since Elon Musk first quoted the hopeful-sounding but incorrect study stating that most of the UK had already had the disease.

20x-100x underestimation seems like fairy-tale land. Even 3x underestimation is hard for me to believe, but 100x ??? That would mean 84M Americans have it, and that is _very_ hard to believe.

why is that hard to believe? It's clear that many (most) experience no symptoms, and if the size of the infected population doubles every ~3 days, well, you do the math...

Well, we'll know in two weeks. Bet you it's wrong.

The Heinsberg study specifically said they shouldn't extrapolate to the entire country since the epidemic there had some unique characteristics. E.g: they had an event which boosted spread a lot.

Anything above a factor 20x is wishful thinking. Limiting it to the studies that have been done in europe show 10-20x which is much more realistic

well when I look at Sweden https://www.worldometers.info/coronavirus/#countries it has 16,004 total cases, 13,517 active cases, 550 total recovered, 515 serious/critical, total deaths 1,937

and Denmark has 7,912 total cases, 2,441 active cases, 5,087 recovered, 80 serious critical, total death 384

So what does that tell me?

1. They have slightly over 12% death rate currently and Denmark has slightly over 4%.

2. They have very many active cases and no recovered cases yet so this makes me think like all these people that have it right have just barely gotten it. So it is a little bit early to talk about how successful it is (otherwise it is time to talk about why is it taking so long to recover in Sweden if they have had it a long time)

If indeed my supposition that they have just barely gotten all these cases and have not had the time to let anyone recover so that means they don't know how many of the people they have right now will actually recover or how many will get critical and have life long complications or die and they are already at a 12% death rate with another 3.2% currently critical === HOLY HELL WTF!!

on edit: of course percentage death rate only for symptomatic cases we know have it, there should be much larger portion of population that has it.

>They have very many active cases and no recovered cases yet so this makes me think like all these people that have it right have just barely gotten it. So it is a little bit early to talk about how successful it is (otherwise it is time to talk about why is it taking so long to recover in Sweden if they have had it a long time)

I'm struggling to find a good source on this, ([1] is the best I could find, though it's in Swedish) but Sweden doesn't track recovered cases so that number is meaningless.

[1] https://www.mabra.com/halsa/friskforklarad-fran-coronavirus/

Ok, then less holy hell wtf. Although I think though most of the cases they have are still relatively recent because I could swear I remember Sweden having less numbers of people than Denmark with it about a month ago.

When I look at this chart of death rates they had 451 dead on April 2 https://www.statista.com/statistics/1105753/cumulative-coron...

which argues a much lower number of cases back then (haven't been able to find how many active cases they had then though, so I may be wrong)

I think this article is right on target

https://www.project-syndicate.org/commentary/swedish-coronav...

on edit: added "then Denmark"

Perhaps the icu stats provides some insights https://www.icuregswe.org/data--resultat/covid-19-i-svensk-i...

> So what does that tell me?

That you look at very unreliable data that is even more difficult to compare across countries?

No it tells me I was looking to get a really smart HN comment by somebody more competent than I in the relevant subject that would explain to me patiently why I was wrong - don't worry though - I shall continue to look.

on edit: I see some other people have replied with more cogent explanations as to why my worries are probably wrong.

A highly relevant XKCD: https://xkcd.com/2295/

Comparing death rates across countries is highly misguided because is depends directly on the testing strategy. The sickness is not different in Sweden than in Denmark and neither country's health care system is overburdened (even in Stockholm, worst hit region of Sweden) which would otherwise be a strong contributing factor to a higher death rate. So there is little reasons to believe that there should be a difference, let alone such a stark one.

Total case count isn't very interesting on its own because it depends highly on testing and policies related to testing. This is one reason why the reported death rates are total crap and why there is such interest in anti-body testing. What is less affected by testing policy and availability is the number of hospitalizations. The number of people currently in hospital presents a pretty good idea of the progress of the disease (it lags a bit.. typically 10 days).

Just growing the infection denominator alone should not change policy very much though without also considering how infectious COVID-19 appears to be. NYC, Italy, etc... have clearly had a lot of deaths from the virus regardless of the final death rate ends up being when this all over.

Where antibody testing does come into play (and assuming antibodies at levels those are tested for do give some immunity) is when things can start to go back to normal.

> 1. They have slightly over 12% death rate currently and Denmark has slightly over 4%.

This site isn't a great way to compare death rates. There's some baked in assumptions. If we compare statistical results we need 1 to 1 comparisons, or an understanding of the limitations of the comparisons.

The lion's share of testing is going to those that are sick. We've already established that <20% of covid cases go to the hospital. It is frequently recommended that one does not go to the hospital unless one is seriously sick. We also know that covid is highly asymptomatic. These people aren't ever going to be tested unless they are in a study specifically that is looking at random but representative samples of the population. Additionally, these countries have different median ages.This is important when we're talking about a disease that affects people dramatically differently by age. (e.g. Italy would be expected to be hit hard because it has an old population that frequently smokes). Also, countries are not uniform in their travel structure, i.e. how dense the country is and strength of nodal connections between dense areas. Finally, strategies were applied at different times in their curves.

So the issues are:

- We don't have 100% testing coverage and this data isn't a representative sample.

- Countries aren't uniform in their age or habits or pollution

- Countries aren't uniform in density and nodal connections.

- Countries applied different strategies at different times.

You should also be aware that the way Sweden is counting deaths by CoViD-19 is by attributing ALL deaths which tested positive for the virus to the disease.

This is very different from other countries statistics and one of the reasons why Sweden statistics are showing that there is very little undercounting of unexplained excess deaths when seasonally adjusted.

actually that seems to be the default method of counting deaths. I find it questionable to attribute (100% of) a death to covid when 4 other conditions existed and contributed to the fatality but there you have it. The undercounting issue has been a result of deaths of untested patients, e.g people found dead at home in NY.

Belgium on the other hand reports all deaths that could plausibly be covid related.

Reply on news.ycombinator.com