COVID-19 Megathread 4: Grandma Got Run Over by the Dow Jones

[emailking]

Sure it might be flattened on the chart of infections/day but it's not something we can know.

[Devout Centrist]

I'm concerned by the lack of long term preparations being made. Quarantining the population will flatten the curve...but it doesn't change the area under the curve. The same number of people will get infected over a longer timespan. What's more: we can't remain in lockdown indefinitely; we should be preparing to test everyone and quarantine the infected.

We're not using this time wisely. We need to be ramping up production of masks, PPEs, and ventilators by magnitudes more than what's currently being done. We need to be testing the entire population and yet, we're not mass producing tests in the quantities needed to do that.

No, instead we're relying on overseas shipments of masks (some of which are of dubious quality!), our medical stockpiles are nearly depleted, and the efforts to produce equipment in the US have stalled! Our Federal Government has left the states to fend for themselves.

This simply isn't sustainable. Sooner or later, this ersatz system going to break down, and millions of innocent people will pay the price for this gross incompetence.

[forgotten manatee]

https://twitter.com/NateSilver538/status/1246909791680770050

Flattened? It's gotten steeper. Yes it's not accelerating as much, that's not flattening. I hope he just had a brain fart here.

"Not accelerating as much" is flattening.  You'll never see decreasing numbers of total cases/deaths, only flattening.

I hope you just had a brain fart here.

Flattening would be fewer cases per day. That wasn't happening in recent days. Cases/day had an increasing slope.

Testing capacity is still ramping-up, so we should expect more positive tests as a result of expanded testing.  As Nate Silver says:

https://twitter.com/NateSilver538/status/1246919689416122369

Only 6k more new cases over last Sunday, and that's in-spite of a more than 40% increase in testing.

Any chance that this means that there is a light at the end of the tunnel by earlier than mid-June?

[Del Tachi]

https://twitter.com/NateSilver538/status/1246909791680770050

Flattened? It's gotten steeper. Yes it's not accelerating as much, that's not flattening. I hope he just had a brain fart here.

"Not accelerating as much" is flattening.  You'll never see decreasing numbers of total cases/deaths, only flattening.

I hope you just had a brain fart here.

Flattening would be fewer cases per day. That wasn't happening in recent days. Cases/day had an increasing slope.

Testing capacity is still ramping-up, so we should expect more positive tests as a result of expanded testing.  As Nate Silver says:

https://twitter.com/NateSilver538/status/1246919689416122369

Only 6k more new cases over last Sunday, and that's in-spite of a more than 40% increase in testing.

Any chance that this means that there is a light at the end of the tunnel by earlier than mid-June?

Yes.  We’re definitely not seeing the exponential growth the UW model thought we’d be seeing.  Even New York is looking better in terms of # of hospitalizations/ICU patients then initially predicted. 

[💥💥 brandon bro (he/him/his)]

https://twitter.com/NateSilver538/status/1246909791680770050

Flattened? It's gotten steeper. Yes it's not accelerating as much, that's not flattening. I hope he just had a brain fart here.

"Not accelerating as much" is flattening.  You'll never see decreasing numbers of total cases/deaths, only flattening.

I hope you just had a brain fart here.

Flattening would be fewer cases per day. That wasn't happening in recent days. Cases/day had an increasing slope.

Testing capacity is still ramping-up, so we should expect more positive tests as a result of expanded testing.  As Nate Silver says:

https://twitter.com/NateSilver538/status/1246919689416122369

Only 6k more new cases over last Sunday, and that's in-spite of a more than 40% increase in testing.

Any chance that this means that there is a light at the end of the tunnel by earlier than mid-June?

I need more evidence to be convinced that we're actually seeing a deceleration (people were saying the last thing last Sunday and Monday was gruesome), but assuming it is true, if there are no efforts taken to stop transmission then the end of shelter in place will lead to more outbreaks.

Silver attributes (without evidence) the trend to social distancing. If deceleration is happening due to social distancing, and you release susceptible people into public life without any sort of mass-testing, contact tracing, or serious quarantining, then June will look just like late March.

[forgotten manatee]

https://twitter.com/NateSilver538/status/1246909791680770050

Flattened? It's gotten steeper. Yes it's not accelerating as much, that's not flattening. I hope he just had a brain fart here.

"Not accelerating as much" is flattening.  You'll never see decreasing numbers of total cases/deaths, only flattening.

I hope you just had a brain fart here.

Flattening would be fewer cases per day. That wasn't happening in recent days. Cases/day had an increasing slope.

Testing capacity is still ramping-up, so we should expect more positive tests as a result of expanded testing.  As Nate Silver says:

https://twitter.com/NateSilver538/status/1246919689416122369

Only 6k more new cases over last Sunday, and that's in-spite of a more than 40% increase in testing.

Any chance that this means that there is a light at the end of the tunnel by earlier than mid-June?

Yes.  We’re definitely not seeing the exponential growth the UW model thought we’d be seeing.  Even New York is looking better in terms of # of hospitalizations/ICU patients then initially predicted. 

That would be wonderful, I hope you're right.

However, I think it is a fair point that we aren't doing enough to prepare for a post-flattened curve environment. That lack of preparation, combined with parts of the country that only recently adopted stay at home orders, could result in this dragging on for a good while longer, no?

[YPestis25]

https://mobile.twitter.com/MarkLevineNYC/status/1246911781198495745
From a member of the NYC Council health committee.

[Del Tachi]

https://mobile.twitter.com/MarkLevineNYC/status/1246911781198495745
From a member of the NYC Council health committee.

Could also be people dying at home because....people are being told to stay home?

Irresponsible to attribute these deaths solely to COVID in an effort to inflate numbers and spread panic. 

[GeorgiaModerate]

https://mobile.twitter.com/MarkLevineNYC/status/1246911781198495745
From a member of the NYC Council health committee.

Could also be people dying at home because....people are being told to stay home?

Irresponsible to attribute these deaths solely to COVID in an effort to inflate numbers and spread panic. 

Or just maybe to get a better idea of how many deaths are actually occurring.  Don't automatically assume a conspiracy.

[Del Tachi]

The updated numbers for COVID-19 in the U.S. are in for 4/5 per: https://www.worldometers.info/coronavirus/country/us/

I'll be keeping track of these updates daily and updating at the end of the day, whenever all states finish reporting for that day. The percentages represent the daily increase from the last report.

3/26:

• Cases: 85,390
• Deaths: ±1,200

3/27:

• Cases: 103,798 (+18,408 | ↑21.56%)
• Deaths: 1,693 (+493 | ↑41.03%)

3/28:

• Cases: 123,428 (+19,630 | ↑18.91%)
• Deaths: 2,211 (+518 | ↑30.60%)

3/29:

• Cases: 142,178 (+18,750 | ↑15.20%)
• Deaths: 2,484 (+273 | ↑12.35%)

3/30:

• Cases: 163,490 (+21,312 | ↑14.99%)
• Deaths: 3,148 (+664 | ↑26.73%)

3/31:

• Cases: 187,917 (+24,427 | ↑14.94%)
• Deaths: 3,867 (+749 | ↑22.84%)

4/1:

• Cases: 215,003 (+27,086 | ↑14.41%)
• Deaths: 5,102 (+1,235 | ↑31.94%)

4/2:

• Cases: 244,433 (+29,430 | ↑13.69%)
• Deaths: 6,070 (+968 | ↑18.97%)

4/3:

• Cases: 276,965 (+32,532 | ↑13.31%)
• Deaths: 7,391 (+1,321 | ↑21.76%)

4/4 (Yesterday):

• Cases: 311,357 (+34,392 | ↑12.42%)
• Deaths: 8,452 (+1,061 | ↑14.36%)

4/5 (Today):

• Cases: 336,327 (+24,970 | ↑8.02%)
• Deaths: 9,605 (+1,153 | ↑13.64%)

So the rate of new diagnoses has slowed everyday since 03/27 despite the steady improvements in testing availability?

[Bandit3 the Worker]

Now would be a good time to prepare for testing, testing, testing.

No excuses or whines this time.

[emailking]

So the rate of new diagnoses has slowed everyday since 03/27 despite the steady improvements in testing availability?

If by that you mean the 3rd derivative of cases vs. days has been negative, yes.

[GeorgiaModerate]

A few thoughts, in no particular order:

1. We all want this to be over and life to return to normal, or as near to normal as it can get.

2. There are some mildly encouraging signs in some of the recent data.  Although we all are looking for hope, don't draw immediate conclusions from them!  Being optimistic is fine, but temper it with realism.  Don't assume that just because things may look a little better for a couple of days, it will all be over by the end of April!  That's wishful thinking leading you into conclusions based on what you hope will happen.

3. Life is NOT going to go back to completely normal until there is an effective vaccine.  That isn't going to happen for quite a few months yet.

4. Don't assume that we're going to get a peak (in May, June, or whenever), and then it's going to die out and everything will be over.  Some of you seem to be expecting this.  Sorry, but that's wishful thinking again.

5. This does not mean that a tight lockdown will need to extend for that entire time.  What we are doing now is designed to slow growth and buy time to put less strain on the medical system, and create more resources such as PPE and tests (both for the virus and for the presence of antibodies).

6. (corollary to 4&5) This is not an all or nothing situation!  As the lockdowns buy time and the peaks pass in various locations, restrictions can be eased.  But there may well need to continue being SOME restrictions for a longer time.

7. This is not happening at a uniform rate all over the country.  NYC will probably hit its peak soon.  Seattle may well be past it.  But there are other parts of the country that have yet to be broadly affected, so their peaks will be much later.

8. The earlier preventive action is taken, the better!  So for those areas that don't think they're badly affected -- you WILL be affected.  Taking action early will save a lot of lives in your area.  Would you rather be in the situation NYC is in, or the situation the SF Bay Area is in?

9. (corollary to 7&8 ) This means that not all places will be able to end restrictions at the same time.  It's going to depend on local conditions.

10. There is a good chance that there will be multiple waves of this, just like there is with the flu.  Even if things go back to relatively normal in a few months, there are probably going to be future waves -- and if they happen before a vaccine is available, we may need to go back into restrictions for a time.

[GP270watch]

https://mobile.twitter.com/MarkLevineNYC/status/1246911781198495745
From a member of the NYC Council health committee.

Could also be people dying at home because....people are being told to stay home?

Irresponsible to attribute these deaths solely to COVID in an effort to inflate numbers and spread panic.  

 Your comment doesn't even make sense. It's a 10 fold increase, people being at home doesn't explain that disparity. It's also in line with initial observations in Italy, Spain, and France about mostly older folks dying at home or nursing homes in much greater numbers. It's not like these elderly are normally zipping about town.

 The charecterization that an attempt to look at data is somebody "inflating numbers and spread panic" is what's really irresposible. We are going through a pandemic, what do you still not get about that?

 

[💥💥 brandon bro (he/him/his)]

Also worth noting that a decrease in Arch's reported percentage increases isn't necessarily informative because the overall number of cases (i.e., denominator used in the calculation) is increasing. You could have a constant number of new cases each day, and that percentage increase would be going down.

Yesterday was a record high number of absolute new cases (34K), and the day before was the first date with 30K new cases. Anybody acting like there is a week-long trend in decreases is either innumerate or dishonest.

[Nyvin]

Also worth noting that a decrease in Arch's reported percentage increases isn't necessarily informative because the overall number of cases (i.e., denominator used in the calculation) is declining. You could have a constant number of new cases each day, and that percentage increase would be going down.

Yesterday was a record high number of absolute new cases (34K), and the day before was the first date with 30K new cases. Anybody acting like there is a week-long trend in decreases is either innumerate or dishonest.

Yeah, if the "percentage" of increase in cases was increasing exponentially (or even linearly) over the long term then we'd really be screwed.

[Okay, maybe Mike Johnson is a competent parliamentarian.]

We're not going to know the real number of cases or the real number of deaths for probably some years to come. That's not to say they're being undercounted by orders of magnitude or anything but they are being undercounted at least somewhat. The idea that the current official stats are identical to the actual numbers of people who have or who have died from COVID-19 is wishful thinking.

[Meclazine for Israel]

You could have a constant number of new cases each day, and that percentage increase would be going down.

Due to a consistent rate of sub-exponential growth in the USA, we have had an increase in case numbers whilst having a drop in case % growth.

Case % growth is therefore not a useful indicator under these circumstances.

[Dr. Arch]

You could have a constant number of new cases each day, and that percentage increase would be going down.

Due to consistent sub-exponential growth, we have an increase in case numbers whilst having a drop in case % growth.

Case % growth is therefore not a useful indicator under these circumstances.

What would be the preferred metric?

[Meclazine for Israel]

You could have a constant number of new cases each day, and that percentage increase would be going down.

Due to consistent sub-exponential growth, we have an increase in case numbers whilst having a drop in case % growth.

Case % growth is therefore not a useful indicator under these circumstances.

What would be the preferred metric?

Just basic mortality numbers are ultimately what history records.

https://en.wikipedia.org/wiki/Spanish_flu#/media/File:Spanish_flu_death_chart.png

[brucejoel99]

https://mobile.twitter.com/MarkLevineNYC/status/1246911781198495745
From a member of the NYC Council health committee.

Are at-home deaths not being tested post-mortem?

[GP270watch]

https://mobile.twitter.com/MarkLevineNYC/status/1246911781198495745
From a member of the NYC Council health committee.

Are at-home deaths not being tested post-mortem?

With what tests?

 

[Grassroots]

The national peak is projected for the 15th. We appear to be on the course for that date being moved earlier.

[Fmr. Gov. NickG]

I don’t think we can convincingly say the worst is over until we see week-over-week decline.  

This Sunday the US had 1165 coronavirus deaths.  
Last Sunday the US had 363 deaths.  

So we definitely aren’t there yet.

[Meclazine for Israel]

OK,

Some great news for people going crazy in isolation. You will be let out earlier than expected.

[tinfoilhat]

In terms of the top of the curve for Australia, Italy and Spain, I am saying that this has now been reached. But the 'official' data does not reflect that yet.

Based on growth rates, mortality, new case numbers, Italy and Spain should have seen their peak by now. But I could not figure out why.

Pulaski highlighted to me yesterday that in Australia, 1,800 recoveries were added suddenly to the Government stats. These were apparently from people who were isolating at home after a positive test, but there was no follow up by medical authorities. Certainly not daily, and nowhere near systematic enough to be used on these graphs effectively.

When I back-added these recoveries to the dataset over the last 7 days, it gave Australia a peak on April 2 2020.



That means Australia has already turned the corner and reached the top of the curve before we expected.

Now taking that thinking about recovered cases over to Europe.

For the UK, they are simply not reporting recovery data anywhere near what logical thought would assume the recoveries should be with a rate proportional to the new cases 10-14 days previous.

For Italy and Spain, I am seeing the same thing. Italy is simply not reporting recoveries anywhere near an appropriate level, and without those numbers, the curve will never 'peak'.

[statadjustment]

Now, assuming Italy and Spain have not been reporting recoveries from positive test patients isolating at home, I have added 4,000 recoveries per day to the Italian dataset and 3,000 recoveries per day to the Spanish dataset from a point in time 10-14 days following the accelerated diagnosis of new cases.

[/statadjustment]

That provides the following graphs:



Italy (Adjusted) Active Cases




Spain (Adjusted) Active Cases

These look more realistic relative to the mortality and daily new case rates coming in. The lag between new cases and mortality showed a correct lag. But the 'top of the curve' is not matching this 7-10 day lag until these adjustments are made.

It is unlikely that a country will have a peak in death rate before the peak in 'Active Cases'. Causative logic prescribes that one precedes the other.

They might be completely wrong, but I am assuming that Italy and Spain are not 'tracing' their positive cases from home following recovery.

If correct, we will see Italy and Spain dump a large number of recoveries in the near future on a single day.

I will plot up Germany tomorrow with adjusted recoveries and show a cumulative growth curve.

[/tinfoilhat]

Any chance that this means that there is a light at the end of the tunnel by earlier than mid-June?

That could be possible. It is unknown at the moment in the USA until the numbers start to calm down and a growth curve can be matched to the data.