COVID-19 Megathread 5: The Trumps catch COVID-19

[Fmr. Gov. NickG]

Cumulative European case and death graphs (5-day weighted averages).  I will caution that Mondays are consistently the day when these averages appear the lowest.

[Fmr. Gov. NickG]

If the goal of “flatten the curve” were to control the pacing of cases as we move toward herd immunity, we wouldn’t be trying to reduce cases everywhere.  We would be trying to maintain a certain equilibrium number of infections across geography and time such that the health care system is never overwhelmed, but we are steadily increasing the proportion of the population who is infected, and eventually recovered and immune.

This would mean adopting a sort of Keynesian economic approach where the government would enforce restrictions in areas where the outbreak of cases was particularly high to discourage infection, but would also adopt policies to encourage infection where the outbreak was low and the healthcare system is below capacity.

We did a OK job of the first half of this in NYC, where we implemented strong restrictions just in time to avoid a complete health care collapse.  And now we are seeing steadily decreasing cases in NYC as the city clearly moves toward herd immunity.

But we aren’t doing the second half of this at all.  In areas with few infections, we are still trying to reduce them, not increase them.  This is just prolonging the lockdowns and delaying the infection surge that is going to happen in these places regardless.  They are mostly on a pace where there aren’t enough infections to get us to herd immunity within the next year, but also too many infections to declare the virus eradicated.

An additional advantage to encouraging infections in low-prevalence areas is that we could better engineer who gets infected.  The death rate of the virus will end up being much, much lower if we encourage infection among the young and health while quarantining the vulnerable than if we just equally expose everyone to the virus as we are doing now.

If a random 70% of the population ends up infected, probably over a million Americans die of the virus.  But if only the healthiest 70% are infected, this could drop as low as 100,000 deaths.

[Fmr. Gov. NickG]

I keep reading here that flattening the curve was not about reducing the number of cases/deaths, just spreading them out over time. But as far as I remember, all of the models showed *many* fewer cases and deaths (like multiples) with a flattened curve than in a business as usual scenario.

The reduced cases/deaths are a result of a normalized bell curve under the healthcare system encumbrance line. Once a health system collapses, the numbers of cases/deaths also skyrocket because that support structure is now gone.

The loss of life and out of control spread that happens once that scenario takes place is a result of the system collapsing more than the projected COVID-19 cases in a vacuum. In other words, those higher numbers also included lives that would have otherwise been saved from COVID-19 and other diseases/conditions as well.

Those nuanced stats are not discussed in public because that kind of detail is usually not necessary to get the point across, but it does generate confusion if people start scrutinizing the model against totals.

So how many people do you estimate will die from the virus if 200 million people are infected, but they are spread out in such a way that the health care system does not collapse?

[Fmr. Gov. NickG]

I keep reading here that flattening the curve was not about reducing the number of cases/deaths, just spreading them out over time. But as far as I remember, all of the models showed *many* fewer cases and deaths (like multiples) with a flattened curve than in a business as usual scenario.

The reduced cases/deaths are a result of a normalized bell curve under the healthcare system encumbrance line. Once a health system collapses, the numbers of cases/deaths also skyrocket because that support structure is now gone.

The loss of life and out of control spread that happens once that scenario takes place is a result of the system collapsing more than the projected COVID-19 cases in a vacuum. In other words, those higher numbers also included lives that would have otherwise been saved from COVID-19 and other diseases/conditions as well.

Those nuanced stats are not discussed in public because that kind of detail is usually not necessary to get the point across, but it does generate confusion if people start scrutinizing the model against totals.

So how many people do you estimate will die from the virus if 200 million people are infected, but they are spread out in such a way that the health care system does not collapse?

I neither have the access to the necessary data nor the experience to make those estimates. I can tell you that they would vary greatly with the state of the healthcare system, who's infected, where, and how many at the same time.

I’m stipulating that the state of healthcare systems is good.  Assume no shortage of necessary medical care or equipment.  As for where, if we’re talking about herd immunity, you’d have to assume the infected are pretty evenly spread out around the county.  You can make any assumption you wish about who gets infected and when under our current stay-at-home regime.

For reference, the current IMHE model is estimating that 74,000 will die.

[Fmr. Gov. NickG]

I keep reading here that flattening the curve was not about reducing the number of cases/deaths, just spreading them out over time. But as far as I remember, all of the models showed *many* fewer cases and deaths (like multiples) with a flattened curve than in a business as usual scenario.

The reduced cases/deaths are a result of a normalized bell curve under the healthcare system encumbrance line. Once a health system collapses, the numbers of cases/deaths also skyrocket because that support structure is now gone.

The loss of life and out of control spread that happens once that scenario takes place is a result of the system collapsing more than the projected COVID-19 cases in a vacuum. In other words, those higher numbers also included lives that would have otherwise been saved from COVID-19 and other diseases/conditions as well.

Those nuanced stats are not discussed in public because that kind of detail is usually not necessary to get the point across, but it does generate confusion if people start scrutinizing the model against totals.

So how many people do you estimate will die from the virus if 200 million people are infected, but they are spread out in such a way that the health care system does not collapse?

I neither have the access to the necessary data nor the experience to make those estimates. I can tell you that they would vary greatly with the state of the healthcare system, who's infected, where, and how many at the same time.

I’m stipulating that the state of healthcare systems is good.  Assume no shortage of necessary medical care or equipment.  As for where, if we’re talking about herd immunity, you’d have to assume the infected are pretty evenly spread out around the county.  You can make any assumption you wish about who gets infected and when under our current stay-at-home regime.

For reference, the current IMHE model is estimating that 74,000 will die.

I don't know why you keep talking about herd immunity when we're still not even certain about whether or not someone who's had COVID-19 can be reinfected.

The “flatten the curve” model assumes that people have immunity once they recover as well.  Otherwise, how do you explain either of the curves in the “flatten the curve” figure ever declining?

[Fmr. Gov. NickG]

None of the supporters of the current blanket stay-at-home regime seem to be willing to acknowledge its consequences: at least a million people will die if we keep doing what we are doing.

We can get that fatality rate down much, much lower, but we need to radically rethink our strategy.  And this strategy will require young and healthy people to make some sacrifices to protect the most vulnerable.

[Fmr. Gov. NickG]

I think this Washington Post article that was just posted is a good summary of growing consensus across scientific studies of an overall IFR of around 0.6%:

Antibody tests support what’s been obvious: Covid-19 is much more lethal than the flu

[Fmr. Gov. NickG]

Now they’re estimating that 36 ppl got COVID at the WI primaries

Given that the projections estimate about 42,000 people in Wisconsin have been infected in total, this seems awfully low.

[Fmr. Gov. NickG]

IHME model now revising the death toll up to 74k by August 1st and that most states should not begin to reopen until late June

The model is still projecting a rapid and symmetrical decline in deaths....starting now!  

We've already significantly exceeded it's projected deaths for today, and the experience of every country going through this now shows that the fundamental assumptions of the model are dead wrong.

Why is this something we still care about?

If we keep doing what we're doing, we're going to have 10,000 covid deaths every week for the next year.  We need to acknowledge that the failure of our models has led to the failure of our basic policy and move on from it.

[Fmr. Gov. NickG]

Reopening schools should be much more of a priority than returning to work right now.

Seems like simply more evidence, along with climate change denial, that people really don’t care very much about the welfare of future generations.

[Fmr. Gov. NickG]

This is good news if true:

https://twitter.com/NateSilver538/status/1255550542295764996

I think this was pretty obvious to everyone who really dug into and thought through this story when it originally came out.

Unfortunately, so much damage has already been done with respect to the public understanding of the virus due to people who misinterpreted the headline and/or disseminated inaccurate information derived from this to suit their own agendas.

[Fmr. Gov. NickG]

I thought California was supposed to be a success story.  How is it a success if six weeks later, we have to crack down even more?

It seems like the more evidence we have that our strategy is failing, the more we are doubling down on it.

[Fmr. Gov. NickG]

https://mobile.twitter.com/CBSLA/status/1255699561399619584

Here come the fun police.

I don't think it would be much fun for those people at the beaches, their family members or other random people they pass the virus onto to be hooked up on ventilators.

What new evidence do we have today that we didn’t have last week (or six weeks ago) suggesting that closing parks and beaches will reduce serious virus cases? 

If anything, the new evidence from studies in the past week or two indicate that outdoor transmission is NOT a significant source of infection (on top of all the evidence that has accumulated suggesting the lockdowns in general are a failure).

[Fmr. Gov. NickG]

https://mobile.twitter.com/CBSLA/status/1255699561399619584

Here come the fun police.

I don't think it would be much fun for those people at the beaches, their family members or other random people they pass the virus onto to be hooked up on ventilators.

What new evidence do we have today that we didn’t have last week (or six weeks ago) suggesting that closing parks and beaches will reduce serious virus cases?  

If anything, the new evidence from studies in the past week or two indicate that outdoor transmission is NOT a significant source of infection (on top of all the evidence that has accumulated suggesting the lockdowns in general are a failure).

Wow, I wasted a lot of effort explaining the whole thing about stay at home orders to you so that you would acknowledge it and then turn around the next day and say the same thing here in bold again. Noted.

I hope that others benefited from the exchange at least.

IIRC, your explanation was that the stay at home orders were not intended to reduce the total number of infections, but merely to spread those infections out over a longer period of time.  

I agree the restrictions will do that.  And if you are satisfied with 200 million Americans being infected and more than a million American deaths, I suppose you could say the restrictions are working.  I just don’t think much of the public would agree with you.

[Fmr. Gov. NickG]

Long and interesting report based on a national survey of 22,000+ from April 16-26:

http://www.kateto.net/COVID19%20CONSORTIUM%20REPORT%20April%202020.pdf

Really fascinating survey results, thanks for the link!

A couple especially interesting bits of analysis from the first few pages:
- Only 18% of people support “reopening the economy” within the next 2 weeks, but only 26% are willing to wait longer than 8 weeks (p. 6)
- Trump’s approval rating is below that of EVERY governor (p. 9)
- The state where the most people think their state government has overreacted to the outbreak is Wisconsin (25%), followed by Michigan (24%).  
- The state where the most people think their government has not taken the virus seriously enough is South Dakota (35%), followed by Florida and Iowa (32%) (p. 13)

[Fmr. Gov. NickG]

Latest cumulative European case & death graphs (5-day weighted average).  Continued steady progress in Italy and Spain, but no sign that America has significantly declined from its peak death numbers.

[Fmr. Gov. NickG]

https://twitter.com/vote4robgill/status/1255922426120146947

How on earth on guns allowed inside a state capitol?

[Fmr. Gov. NickG]

On a different note, here is some xkcd.

The other big difference is that a lot of the questions our public officials are being asked to address are not questions of science or medicine, but questions of how we should balance competing values. 

In a representative democracy, it is very reasonable to expect that while scientific questions should be answered by experts within the bureaucracy, disputes over competing values should be resolved through democratic deliberation, in which we would ideally like the public to be thoroughly informed.

[Fmr. Gov. NickG]

Oh please. Quit being such an apologist for thugs. Oh goodie! They didn't actually cock and point their guns at legislate arguing in favor of keeping the shut down! Nothing to see here! Totally business-as-usual compatible with the open expression and exchange of ideas in the legislature why nonviolent needs!

Some of us carry guns. Yes, all the time. Just because you don't like guns doesn't mean carrying guns is a threat ... let alone terrorism. Or is the terrorism that they opposed the suspension of our fundamental rights and breathed in public without wearing a mask. I've seen so much hateful invective directed at people who support the Bill of Rights these past months its kind of hard to tell why we should "muh killz the terrist protesters!" Quit being an apologist for a comment clearly calling for violence against protesters. What do you think "treat them like terrorists" means in this context?

When you say you carry a gun "all the time"...do you carry a gun to the airport?  To a courthouse?  
I am completely shocked that anyone is allowed to take a gun into any state capitol.

[Fmr. Gov. NickG]

2nd highest number of cases today.

Deaths are down some though.

The new case numbers are not very important.  Since we are only identifying a small fraction of actual infections, this is mostly a function of testing, and we had a record number of new tests today.

Deaths are a much more important stat, especially deaths outside of New York, since the decline in NYC can be attributed to progress toward herd immunity.

[Fmr. Gov. NickG]

2nd highest number of cases today.

Deaths are down some though.

The new case numbers are not very important.  Since we are only identifying a small fraction of actual infections, this is mostly a function of testing, and we had a record number of new tests today.

Deaths are a much more important stat, especially deaths outside of New York, since the decline in NYC can be attributed to progress toward herd immunity.

We are not trying to get herd immunity through infections! I am not sure why you keep repeating this line. Any effort to accomplish herd immunity would be hugely destructive even if we could somehow limit infection just to people we suspect are more resilient. We are trying to stall transmission until there's a vaccine so we don't overwhelm hospitals. That's it.

The decline in New-York-City-area deaths, which you are attributing to herd immunity, also just so happens to coincide with a massive and successful lockdown. Are you saying that's a coincidence? What do you think the true death rate of this disease actually is?

New York City has had a similar lockdown to almost every other major metropolitan area in the country.  And yet they are seeing steep declines in cases and deaths while other cities are seeing continued increases.

I think over the past two weeks we have established a much better understanding of the true death rate of the virus through the serology studies.  The consensus estimate is around 0.6%.

(Also, the post of mine you quoted doesn’t say anything about trying to accomplish herd immunity through infection, so I’m not sure why you say I keep repeating it.)

[Fmr. Gov. NickG]

What does "the true death rate" mean?

I mean the IFR for a representative cross section of the population.
Obviously is it much lower for younger people and much higher for older people or people with certain health conditions.

[Fmr. Gov. NickG]

That sounds subjective.

How do you mean?  We have estimates for total infections based on serology, and estimates for number of deaths, and you just divide one by the other.  You can critique the method for deriving either number, but I wouldn’t describe that as “subjective”, merely inaccurate.  And a lot of studies seems to be converging on the same estimate using different methodologies, so the estimate should be pretty robust to a flaw in the data of any one study.

[Fmr. Gov. NickG]

Some days ago you made a similar point, that lockdowns don't work and progress in NYC is due to the large number of immune people slowing down spread. I asked you the following questions then, in a post you either missed or ignored. Given that you're still beating on the same drum, I'm just going to copy paste that post. I hope you think about the answers this time.

Let me get this straight. You are doubting that keeping people away from one another reduces the transmission of the virus? What mechanism of transmission do you think the virus has, that apparently doesn't need physical access to new hosts?

You are right that immunity does reduce transmission. But again I am curious by what mechanism 25% immune managed to reduce transmission by more than 50%.

Your herd immunity, not distancing hypothesis also fails to account completely for the mass of other countries that managed to push R below 1 without the massive rate of infected NY has.

Ah, OK, I must have missed that.
It is very plausible that a 25% infection rate might reduce transmission by 50%, depending on how fast the transmission is.  It’s basically the same logic behind how a less than 100% infection rate can reduce the virus to 0.

Let’s say the virus has a baseline R0 of 2 (each person infects 2 new people).  If 25% of the population is infected, the virus has an R0 of 1.5 (since 25% of the new people who would have been infected can no longer be infect).

With no one immune (R0=2), a single infected person infects 2 people.  In the next round, those 2 people infect 4 people.  In the next round, those 4 people infect 8 people.  So over 3 infection cycles, 14 new people have been infected for every 1 person infected at the start.

Now with 25% immune (R0=1.5), a since infected person infects 1.5 people.  Those 1.5 people infect 2.25 in the next round, and those 2.25 people infect 3.375 people in the third round.  Some each infected person infects 7.125 people (compared with 14 people when no one is immune).

If you imagine a 5-day infection cycle, a 25% immunity will reduce transmission by 50% about every 15 days.

Obviously these numbers will change if you alter the length of the infection cycle, but I do find a 50% reduction over some time period given 25% immunity to be totally believable.

[Fmr. Gov. NickG]

Some days ago you made a similar point, that lockdowns don't work and progress in NYC is due to the large number of immune people slowing down spread. I asked you the following questions then, in a post you either missed or ignored. Given that you're still beating on the same drum, I'm just going to copy paste that post. I hope you think about the answers this time.

Let me get this straight. You are doubting that keeping people away from one another reduces the transmission of the virus? What mechanism of transmission do you think the virus has, that apparently doesn't need physical access to new hosts?

You are right that immunity does reduce transmission. But again I am curious by what mechanism 25% immune managed to reduce transmission by more than 50%.

Your herd immunity, not distancing hypothesis also fails to account completely for the mass of other countries that managed to push R below 1 without the massive rate of infected NY has.

Ah, OK, I must have missed that.
It is very plausible that a 25% infection rate might reduce transmission by 50%, depending on how fast the transmission is.  It’s basically the same logic behind how a less than 100% infection rate can reduce the virus to 0.

Let’s say the virus has a baseline R0 of 2 (each person infects 2 new people).  If 25% of the population is infected, the virus has an R0 of 1.5 (since 25% of the new people who would have been infected can no longer be infect).

With no one immune (R0=2), a single infected person infects 2 people.  In the next round, those 2 people infect 4 people.  In the next round, those 4 people infect 8 people.  So over 3 infection cycles, 14 new people have been infected for every 1 person infected at the start.

Now with 25% immune (R0=1.5), a since infected person infects 1.5 people.  Those 1.5 people infect 2.25 in the next round, and those 2.25 people infect 3.375 people in the third round.  Some each infected person infects 7.125 people (compared with 14 people when no one is immune).

If you imagine a 5-day infection cycle, a 25% immunity will reduce transmission by 50% about every 15 days.

Obviously these numbers will change if you alter the length of the infection cycle, but I do find a 50% reduction over some time period given 25% immunity to be totally believable.

I see I used an ambiguous term with "transmission". I meant "R", as jn, how would 25% immunity reduce R by 50%. (Indeed, assuming R(0) is 2 or more, we need a reduction of R by at least  50% to get a decline below 1, which is required to get a decreasing number of cases in NYC.)

It is likely that neither the 25% herd immunity nor the lockdowns alone have reduced R0 significantly below 1.  But they have worked in combination to reduce it below 1. 

It does seem like a lot of cities without significant herd immunity are no longer exponentially growing in infection, but either slightly increasing or slightly decreasing.  So we might imagine that the lockdowns have themselves reduced the R0 to around 1.

Stack the lockdown effect with a 25% herd immunity effect and you get an R0=.75. 
With R0=1, each new infection will infect 3 new people over the course of 3 cycles of the virus.
With R0=.75, each new infection will infect .75+.5625+.422=1.734 people over 3 cycles of the virus.

Which is why you could see a very stable number of new infections in most cities, but an almost 50% reduction in new infections in a city with similar restrictions but a 25% herd immunity.