D&D and the rising pandemic

[Celebrim]

Interesting article on why the higher morality rates are probably wrong.

The WHO Estimated COVID-19 Mortality at 3.4%. That Doesn't Tell the Whole Story

Why the real number may be much lower

time.com

Yes, well that article is wrong. There is a ton of reasons why, but one big clue is they don't understand the difference between mortality rate and transient case fatality ratio, and so they say really stupid things like:

"The mortality rate in South Korea, where more than 1,100 tests have been administered per million residents, comes out to just 0.6%, for example."

No, at the time they wrote that the mortality rate in South Korea - defined by the number of deaths divided by the number of closed cases (deaths + recoveries) stood at like 24%. The number they use there, the infamous "0.6%" quote represented the transient case fatality ratio. CFR is useful for determining whether you have found most of the cases that exist, but has nothing to do at all with how many people will die. The easiest way to realize that is to realize that if you have a CFR of like "0.6%" but you still have 7000 people who are sick, then actually "0.6%" represents the absolute floor possible of your mortality rate since it is based on the unspoken assumption that all 7000 people who are still sick will not die.

Not surprisingly, South Koreas transient CFR has been rising steadily all through the period were people were using the number to claim - without a shred of epidemiological training and no understanding what they were saying - that the overall mortality rate is low.

Let's get this settled on this thread once and for all. The observed mortality rate globally is not 3.4% - right now it's 7.2%. The 3.4% number builds into it an assumption that many mild cases don't get counted to the total. Right now, to get to 3.4% you more or less have to assume about half the cases are so mild they don't get counted. If that assumption is wrong, then it's highly possible Covid-19 is more deadly than is being reported. 3.4% is not a sensational number based on hype. I'd calculated 3.6% 10 days before WHO came out with their estimate. 3.4% is a low ball let's not start a panic number.

 

[Dannyalcatraz]

Science does provides a better path. Wuhan has to all appearances defeated the Covid-19 bug. Even outside of Wuhan, where it was less epidemic, the cases are being choked off. We can detect it, and it is slow enough to spread that even an active epidemic can be choked off.

There is what appears to be a path forward that isn't 10 years in house arrest.

The predictive stats say what happens if we choose other paths. We go with 10 years+ house arrest where 1% of the world dies to this, we go with 5%+ mortality of everyone, or we pick a costly, annoying, hard path and beat it.

Do you vote for 10+ years of house arrest, 2 million dead in USA, and 200 million dead worldwide? Do you vote for 350 million dead worldwide and no more house arrest anywhere? Or do you vote for doing something difficult?

Or do you vote for "I hope we get lucky with medical science"?

Knowing what the alternatives are, it makes paying the price easier.

That’s still just public health measures.

Chinese science still has not produced a vaccine, cure, or treatment.

 

[Zardnaar]

That’s still just public health measures.

Chinese science still has not produced a vaccine, cure, or treatment.

Go to YouTube and type in China wetmarket.

Our health inspectors would have kittens. In addition to things you expect such as hosing the floor down into a public street they put baskets of vegetables on the wet concrete.

That's on top of the bats, frogs and whatever else gets slaughtered without basic hygiene standards.

Worked in a fish factory years ago and a farm. Completely differently. If a fish touched the ground it got binned. Everything gets washed down and sanitized between shifts.

 

[Acolyte of Zothique]

I quite enjoy abstract philosophical discussions, even ones which challenge basic assumptions that are usually taken as axiomatic. In the abstract, if a friend were to come up with me and ask "so why DO we value human life?", I'd be happy to engage in the navel-gazing. Especially after like, beer number 4.

If someone is lying bleeding in the street and someone looks over at me and asks, "hey, have you ever thought about how many problems overpopulation causes?" then quite simply I'm not going to have patience or sympathy for their self-indulgent BS. When lives and wellbeings are at stake, I feel no particular compulsion to humor that sort of solipsism.

So no, I'm not bothered by that poster's questioning of philosophical concepts. I am appalled at the concrete behaviors that they are espousing and apparently practicing, and their stated disregard for the risk that said actions run, and moreso for their stated rationalization of "if my actions cause harm to others it's their fault". I don't find that point of view worthy of debate, no, except to simply denounce it.

There's a time and a place for philosophy - this isn't it. This is a time to be part of the solution, otherwise you are part of the problem.

 

[FrogReaver]

Let's get this settled on this thread once and for all. The observed mortality rate globally is not 3.4% - right now it's 7.2%. The 3.4% number builds into it an assumption that many mild cases don't get counted to the total. Right now, to get to 3.4% you more or less have to assume about half the cases are so mild they don't get counted. If that assumption is wrong, then it's highly possible Covid-19 is more deadly than is being reported. 3.4% is not a sensational number based on hype. I'd calculated 3.6% 10 days before WHO came out with their estimate. 3.4% is a low ball let's not start a panic number.

I get it - it's building into the numbers an unknowable assumption - an assumption about the mortality rate of the virus from an unknown population of infected but not tested people. The problem is they could be way overestimating or way underestimating that number in their calculation. If this virus spreads as easy as it seems to then it's reasonable to conclude that way more than half the people with it don't know they have it, which tends to start significantly decreasing that 3.4%. Tell me where that's wrong?

 

[Acolyte of Zothique]

Guess Im not eating that day.

It's EVERY day. You'll get used to it.

 

[Lem23]

Dude, just stop eating at mcdonalds.

 

[FrogReaver]

No, at the time they wrote that the mortality rate in South Korea - defined by the number of deaths divided by the number of closed cases (deaths + recoveries) stood at like 24%. The number they use there, the infamous "0.6%" quote represented the transient case fatality ratio. CFR is useful for determining whether you have found most of the cases that exist, but has nothing to do at all with how many people will die. The easiest way to realize that is to realize that if you have a CFR of like "0.6%" but you still have 7000 people who are sick, then actually "0.6%" represents the absolute floor possible of your mortality rate since it is based on the unspoken assumption that all 7000 people who are still sick will not die.

It's not even a floor because we don't know how many people with the virus aren't in this ratio. If 5 times the people have it than are confirmed as having it (presumably most of those cases would be milder) then you have a much lower mortality rate.

 

[Celebrim]

Tell me where that's wrong?

The South Korea data set combined with other evidence like the doubling rate being between 6 and 7 days.

What SK's low CFR actually proved was that they had found a very high percentage of the cases that were 'in the wild'. The chart in that article showing that the more testing you did the lower your CFR was almost a tautology. It proved something, but not what the people who created the chart understood it to prove. It proved only the obvious, that initial CFR will be lower the higher percentage of the cases you have documented.

Think about the reverse situation. If you don't test widely, then the first cases that you'll notice are almost always anomalous deaths and severe cases requiring hospitalization. So without extensive testing, you'll almost certainly initially only document the worst cases, giving you a very high documented CFR simply because that's all you know about.

And SK's low CFR combined with the high percentage of tests that they do that come back negative in SK proves that there are not a lot of carriers with mild cases. (Now there is some weirdness here with asymptomatic cases that could complicate this discussion, but lets for now ignore all that because it just ends up with a bunch of question marks around things we don't understand yet.)

The other piece of information is the observed doubling number seems to be under uncontrolled conditions 6-7 days. If you see faster doubling than that,then it means your documentation is catching up to reality. If you see slower doubling, then it means you have some sort of quarantine that is slowing the rate of spread. However this very tightly constrains how many cases can exist after a certain period of time. Exponential series are very sensitive to the base number being raised to a power - in this case time. Change base number even slightly and you get massive differences in results.

The high R0 low mortality rate model you are describing was a popular model a lot of people believed in or hoped for early in the epidemic, but the things that have happened since make that model very unlikely. If the R0 was that high, we should be seeing even faster spread than we do. We should be finding all those mild cases. And we just aren't.

 

[seebs]

This isn't "flattening the curve". "Flattening the curve" is accepting that everyone will get sick, and trying to get them to get sick slower to reduce mortality. This is using social distancing to buy time to eliminate the epidemic via tracking and containment.

No, that's not what "flattening the curve" means.

Flattening the curve means slowing down the spread so that fewer people get sick at any time ever, not just so that it's spread out more. The point of the flatter curve isn't just that it's spread out more to even load on the health care system; it's that fewer people get sick at all.

If there exists a curve, at all, if anyone gets sick and then eventually people stop getting sick, you're still in the territory covered by "flatten the curve".

Eliminating the epidemic, by whatever means, is the tail end of the curve, but there is always a curve. The point is that it kills a lot fewer people if it's flat -- and part of that is because of load on the health care system, but part of that is because fewer people ever get it at all.

So, yes, social isolation, distancing, quarantining, all good things, and could conceivably keep it from hitting absolutely everyone, but there's no chance of having it completely contained from day one, especially not in the US, where we have over 50 distinct governments in charge of setting these policies and a lot of them are run by people who explicitly believe that all the claims that we should do something quickly are just excuses for attacking the President. So it's not gonna happen; we're not in a position to fully contain the thing. But we might be able to contain it enough that it never really hits the majority of the population. If we get super lucky.

But it turns out that we're arguing for the exact same course of action, here, I'm just pointing out that part of the benefit of it is that it reduces deaths no matter how successful it is, and that I don't think we have the resources or social tools necessary to eradicate an epidemic that effectively.