The Shape of the Covid19 Curve

[djAdvocate]

Apr 5, 2020 21:19:27 GMT -5 pulmonarymd said:

Apr 5, 2020 20:40:20 GMT -5 djAdvocate said:

so, this was a good day, folks.

this is the first day in TWO WEEKS that we had a 24 hour drop in cases.

of course, there might be trailing totals that have not been registered (the clock just reset) but it is coming in at 25k rather than 35k.

if that holds up, it is the first sign in a good while that we are making NATIONAL progress.

I will circle back in a few hours to check it.

I noticed that too. But since it’s a Sunday, official reporting may lag. Need to see a few days in a row where it is consistently less than predicted. Also need a few more data points to see if curve shifted. Last I saw the deaths appeared to be lower than the covid19.healthdata site predicted, unless a bunch came in after I last looked

today was also better than 2 days ago.

on a darker note, we broke through Italy's "curve" today.  there is a good chance it won't end as well for us.

we also have broken into the top 20 for infections AND deaths per million.

Italy is at 273 deaths per million.  you can do the math.  though I admit it is hard to imagine, right now.

[djAdvocate]

further updates: france is right on the Italy curve, and Spain is ahead of it.
it will take us a while to catch up to Spain. longer than I predicted last Friday.
it won't happen this week.

[pulmonarymd]

Apr 7, 2020 1:00:51 GMT -5 djAdvocate said:

Apr 5, 2020 21:19:27 GMT -5 pulmonarymd said:

I noticed that too. But since it’s a Sunday, official reporting may lag. Need to see a few days in a row where it is consistently less than predicted. Also need a few more data points to see if curve shifted. Last I saw the deaths appeared to be lower than the covid19.healthdata site predicted, unless a bunch came in after I last looked

today was also better than 2 days ago.

on a darker note, we broke through Italy's "curve" today.  there is a good chance it won't end as well for us.

we also have broken into the top 20 for infections AND deaths per million.

Italy is at 273 deaths per million.  you can do the math.  though I admit it is hard to imagine, right now.

That’s how you get 90,000 dead. Still a scary number. My state’s curve is flattening by all appearances

[djAdvocate]

it looks like today we will be "on" the Italy curve.

for those of you that don't know what I am talking about, there are a set of curves of the growth of the CV posted on a NYT website, updated daily. Spain and Italy have the worst curves- which is total deaths by days following the 25th case of the virus. now, keep in mind that this is in absolute terms, and we have a much larger population than Italy. their total cases per million is 2x ours, but they have also done 2x the amount of testing per capita. so I think it is fairly safe to say that we will be in similar circumstances.

I hope I am wrong. but as pmd just pointed out, it is how we end up with 90k dead. and that would be a fairly sunny outlook. as you will recall, I predicted 1M if Trump did nothing, and sent everyone back to work on Easter. well, Trump extended the SOE to 4/30, and the states have picked up the slack (with the exception of 7 red states, which still seem to think it is the flu). the shelter in place order is impacting over 90% of the US population, now. so, there are promising signs.

the problem is that the slope of our curve is still pretty steep. we are not near the apex yet, imo. though the rate of growth is decreasing rapidly now, what we have seen in other places (like Italy and Iran) is that the doubling falls down to about every 2 weeks, which is great, but that still means that by the end of April we would have over 1M cases- and this is not even close to where we are. we are currently doubling every 4 days. so we need to cut that (rate) down by a factor of 3.

[pulmonarymd]

Didn't get a chance t look, but our state curve is more linear now by reports. Still going up, but better, and we might be able to handle it. We look like we may also have a wave progressing across the state, with different peaks in different areas. Also good. But, with that, we won't have the peaks in the state over with until early June. So lockdown would have to extend beyond that. That is what Trump should be messaging. The piecemeal approach is not going to cut it. But, as the science said it should, social distancing works. Interesting question is what percentage of the population will have been infected by summer. That will determine the next step

[djAdvocate]

Apr 7, 2020 12:54:04 GMT -5 pulmonarymd said:

Didn't get a chance t look, but our state curve is more linear now by reports. Still going up, but better, and we might be able to handle it. We look like we may also have a wave progressing across the state, with different peaks in different areas. Also good. But, with that, we won't have the peaks in the state over with until early June. So lockdown would have to extend beyond that. That is what Trump should be messaging. The piecemeal approach is not going to cut it. But, as the science said it should, social distancing works. Interesting question is what percentage of the population will have been infected by summer. That will determine the next step

I think it would be a good exercise for THE PRESIDENT to show the American public what this thing would be like if it WERE like the Flu.

right now, our target is a 1% infection rate nationally. that would put the death toll in the range that Trump mentioned recently.

the typical flu hits 30-70% of the US public.

he should NEVER have compared this to the flu. what he SHOULD have said (and should say now) is that this is 25x worse than the flu, and that in order for it to not kill more people than the flu, we have to have 1/25th the infection rate. that would be strong, accurate messaging. and from that, we could develop a set of NATIONAL policies that would engender that rate.

instead, he has "left it up to the states" which will ensure that this will have much heavier impacts in some places than others.

it is a disastrous policy that started with a very wrong idea of what this virus is capable of.

[pulmonarymd]

Apr 7, 2020 12:59:58 GMT -5 djAdvocate said:

Apr 7, 2020 12:54:04 GMT -5 pulmonarymd said:

Didn't get a chance t look, but our state curve is more linear now by reports. Still going up, but better, and we might be able to handle it. We look like we may also have a wave progressing across the state, with different peaks in different areas. Also good. But, with that, we won't have the peaks in the state over with until early June. So lockdown would have to extend beyond that. That is what Trump should be messaging. The piecemeal approach is not going to cut it. But, as the science said it should, social distancing works. Interesting question is what percentage of the population will have been infected by summer. That will determine the next step

I think it would be a good exercise for THE PRESIDENT to show the American public what this thing would be like if it WERE like the Flu.

right now, our target is a 1% infection rate nationally. that would put the death toll in the range that Trump mentioned recently.

the typical flu hits 30-70% of the US public.

he should NEVER have compared this to the flu. what he SHOULD have said (and should say now) is that this is 25x worse than the flu, and that in order for it to not kill more people than the flu, we have to have 1/25th the infection rate. that would be strong, accurate messaging. and from that, we could develop a set of NATIONAL policies that would engender that rate.

instead, he has "left it up to the states" which will ensure that this will have much heavier impacts in some places than others.

it is a disastrous policy that started with a very wrong idea of what this virus is capable of.

If the states that have not locked down, or have exempted religious services have their usual Easter crowds, it will be interesting to see what they look like in 2 weeks. Louisiana could have a nightmare worse than NY.

[dondub]

Apr 7, 2020 12:59:58 GMT -5 djAdvocate said:

Apr 7, 2020 12:54:04 GMT -5 pulmonarymd said:

Didn't get a chance t look, but our state curve is more linear now by reports. Still going up, but better, and we might be able to handle it. We look like we may also have a wave progressing across the state, with different peaks in different areas. Also good. But, with that, we won't have the peaks in the state over with until early June. So lockdown would have to extend beyond that. That is what Trump should be messaging. The piecemeal approach is not going to cut it. But, as the science said it should, social distancing works. Interesting question is what percentage of the population will have been infected by summer. That will determine the next step

I think it would be a good exercise for THE PRESIDENT to show the American public what this thing would be like if it WERE like the Flu.

right now, our target is a 1% infection rate nationally. that would put the death toll in the range that Trump mentioned recently.

the typical flu hits 30-70% of the US public.

he should NEVER have compared this to the flu. what he SHOULD have said (and should say now) is that this is 25x worse than the flu, and that in order for it to not kill more people than the flu, we have to have 1/25th the infection rate. that would be strong, accurate messaging. and from that, we could develop a set of NATIONAL policies that would engender that rate.

instead, he has "left it up to the states" which will ensure that this will have much heavier impacts in some places than others.

it is a disastrous policy that started with a very wrong idea of what this virus is capable of.

But Obama...

[dondub]

I’m hoping the residents of the 7 red states are like Catholic women and birth control. That 98% are ignoring the edicts from the idiots.

[haapai]

OT:  The answer to this question, if I get one, probably won't affect the shape of the curve.

I've been looking at the historical death data incorporated into the second graph (deaths per day) and it differs from the data that I have been collecting daily when the state issues its numbers. 

Most of the discrepancies are minor.  They differ by one or two deaths each day and the overall pattern is quite consistent. That is, I can look at a data point and recognize it as the number from a prior day plus or minus one or two. 

Then there are the numbers from April 4th and onward.  I just don't recognize those numbers at all.  They bear absolutely no resemblance to the numbers that I have been collecting on paper and are much larger than what the state reported.

What's going on?  Any clues would be helpful. 

[teen persuasion]

Apr 7, 2020 13:52:32 GMT -5 haapai said:

OT:  The answer to this question, if I get one, probably won't affect the shape of the curve.

I've been looking at the historical death data incorporated into the second graph (deaths per day) and it differs from the data that I have been collecting daily when the state issues its numbers. 

Most of the discrepancies are minor.  They differ by one or two deaths each day and the overall pattern is quite consistent. That is, I can look at a data point and recognize it as the number from a prior day plus or minus one or two. 

Then there are the numbers from April 4th and onward.  I just don't recognize those numbers at all.  They bear absolutely no resemblance to the numbers that I have been collecting on paper and are much larger than what the state reported.

What's going on?  Any clues would be helpful. 

Looks like the IHME site has changed their death modeling, averaging numbers over a few days, because the U.S. daily death numbers fluctuate too much.  Definitely true locally - tests and waiting for results can shift things around by quite a few days, depending on whether they even have tests to administer and how long results take.

Based on the now multiple iterations of our COVID-19 death model, we have noticed that, for at least some US states, there are massive fluctuations in the number of COVID-19 deaths reported each day. These substantial day-to-day vacillations are more likely due to an artefact in how statewide deaths are being compiled and then reported each day than actual fluctuations in COVID-19 deaths. As a result, our reported predictions – that you can view and download from the online visualization tool – are now based on averaging the last three rounds of predictions. In other words, what is shown today (April 7) is the average of model predictions from reported COVID-19 death data up to April 4 (model 1), data up to April 5 (model 2), and data up to April 7 (model 3). We view this as an important refinement that helps to strengthen model stability and buffer predictions from data fluctuations less related to observed epidemic patterns and more driven by variable data collection or reporting practices.

Healthdata.org/covid/updates

[haapai]

I'm not sure that I agree with that fix.  

It also doesn't explain the April 4 and April 5 data points, which are not labeled as projections, either by the use of solid or dashed line, or in the pop-up that appears when you move cursor to the right spot. 

Both numbers are over twice what the state has reported publicly and they are presented as actual, not smoothed numbers.  I'm beginning to wonder if these numbers include both the number of confirmed and presumptive/probable Covid-19 deaths. 

[djAdvocate]

we just crossed Italy on the CV curve. at this point in time, Italy was doubling every 2 weeks. we are still doubling every 4 days.

my prediction of 1M by EOW is going to be high again, but 500k appears likely. and we are going to pass Spain in deaths tomorrow. and we broke 400k today.

so, the situation is improving, for sure. but we are not out of the woods, yet.

[haapai]

IHME just updated.  It appears that my state's reported Covid-19 deaths are not going to peak nearly as high as earlier estimations. 

I'm fussing a lot about the under-reporting though.  Wouldn't it be amazing if history showed that the earlier death projections were more accurate than the later ones and as the quality of the data being fed into the model deteriorated, the projections got less and less accurate?

ETA:  A lot of projected peak resource use and peak deaths per day dates changed. 

[oped]

New York City has 100-200 extra dead at homes per day than average... these are presumed covid but not tested and not counted. That’s in NY which has no political motive to obscure numbers and has more tests... what do you think I’d happening
in Florida and Texas?

champ.gothamist.com/champ/gothamist/news/surge-number-new-yorkers-dying-home-officials-suspect-undercount-covid-19-related-deaths?__twitter_impression=true

[haapai]

Apr 8, 2020 5:29:41 GMT -5 oped said:

New York City has 100-200 extra dead at homes per day than average... these are presumed covid but not tested and not counted. That’s in NY which has no political motive to obscure numbers and has more tests... what do you think I’d happening
in Florida and Texas?

champ.gothamist.com/champ/gothamist/news/surge-number-new-yorkers-dying-home-officials-suspect-undercount-covid-19-related-deaths?__twitter_impression=true

You can call me a cockeyed optimist but I'm guessing that Florida and Texas are probably doing more post-mortum testing than NY.  They have enough tests to do that.  New York does not. 

At least that's the conclusion that I'm drawing from the Worldmeters site and dividing the caseloads by the number of tests performed.  The numbers for NY and NJ are gut-emptying. 

[oped]

Apr 8, 2020 6:44:13 GMT -5 haapai said:

Apr 8, 2020 5:29:41 GMT -5 oped said:

New York City has 100-200 extra dead at homes per day than average... these are presumed covid but not tested and not counted. That’s in NY which has no political motive to obscure numbers and has more tests... what do you think I’d happening
in Florida and Texas?

champ.gothamist.com/champ/gothamist/news/surge-number-new-yorkers-dying-home-officials-suspect-undercount-covid-19-related-deaths?__twitter_impression=true

You can call me a cockeyed optimist but I'm guessing that Florida and Texas are probably doing more post-mortum testing than NY.  They have enough tests to do that.  New York does not. 

At least that's the conclusion that I'm drawing from the Worldmeters site and dividing the caseloads by the number of tests performed.  The numbers for NY and NJ are gut-emptying. 

You are going to have to explain that to me better? I'm not sure exactly what you are saying.  What calculations are you doing?
Why would you assume that they are post mortum testing? And what makes you think they have more tests than New York? 

[haapai]

Basically I'm using data from the Worldmeters site and comparing the total caseload of a state to the number of tests performed.  Divide the number of tests administered by the number of cases.  The lower the number is, the more scarce tests have been in that state. 

The implications of a state having performed less than three tests for every confirmed Covid-19 case are very scary. 

Then you have to make a bit of a leap and assume that a states' historical scarcity or abundance of tests is still the case.  I'm not entirely comfortable making this leap.  The capacity to perform the tests (i.e. lab capacity and testing sites) may be there still, but the supplies (i.e. physical stuff, swabs, PPE for the swabber, etc) may not.  So I'm kinda going out on a limb when I assume that the present is similar to the past, 

[oped]

So 42% of NY tests are positive... while 10-11% are positive in Texas and Florida... I'm still not sure why you would assume that translates to Texas and Florida testing the dead at higher rates?  

[deminmaine]

Apr 7, 2020 13:17:22 GMT -5 pulmonarymd said:

Apr 7, 2020 12:59:58 GMT -5 djAdvocate said:

I think it would be a good exercise for THE PRESIDENT to show the American public what this thing would be like if it WERE like the Flu.

right now, our target is a 1% infection rate nationally. that would put the death toll in the range that Trump mentioned recently.

the typical flu hits 30-70% of the US public.

he should NEVER have compared this to the flu. what he SHOULD have said (and should say now) is that this is 25x worse than the flu, and that in order for it to not kill more people than the flu, we have to have 1/25th the infection rate. that would be strong, accurate messaging. and from that, we could develop a set of NATIONAL policies that would engender that rate.

instead, he has "left it up to the states" which will ensure that this will have much heavier impacts in some places than others.

it is a disastrous policy that started with a very wrong idea of what this virus is capable of.

If the states that have not locked down, or have exempted religious services have their usual Easter crowds, it will be interesting to see what they look like in 2 weeks. Louisiana could have a nightmare worse than NY.

My wife has a nephew who works in IT at a hospital in Louisiana.  They are making and distributing cloth face masks for use by hospital personnel.  Hopefully not for use in the ERs.  OK, that last sentence was sarcastic (I think/ hope)  but it is troubling that they even have to do this!  

[haapai]

Apr 8, 2020 8:29:23 GMT -5 oped said:

So 42% of NY tests are positive... while 10-11% are positive in Texas and Florida... I'm still not sure why you would assume that translates to Texas and Florida testing the dead at higher rates?  

Flip it around.  Calculate the other ratio.  Tests performed per confirmed case.  Think about what that means.  

[teen persuasion]

Apr 8, 2020 7:28:41 GMT -5 haapai said:

Basically I'm using data from the Worldmeters site and comparing the total caseload of a state to the number of tests performed.  Divide the number of tests administered by the number of cases.  The lower the number is, the more scarce tests have been in that state. 

The implications of a state having performed less than three tests for every confirmed Covid-19 case are very scary. 

Then you have to make a bit of a leap and assume that a states' historical scarcity or abundance of tests is still the case.  I'm not entirely comfortable making this leap.  The capacity to perform the tests (i.e. lab capacity and testing sites) may be there still, but the supplies (i.e. physical stuff, swabs, PPE for the swabber, etc) may not.  So I'm kinda going out on a limb when I assume that the present is similar to the past, 

Tests have been in very short supply, at least on this end of NY - not sure if they are all getting sucked into NYC's black hole, or its even across the state.  Scratch that, I know that recently the county city #2 is in (near me) had many fewer tests performed than the counties for cities #3, 4, or 5.

Locally, they have had partial weeks with no tests to use.  So when they got tests, they limited who is eligible: healthcare workers, first responders, THEN extremely ill patients who are threats to the first 2 categories and where knowing if they are Covid-19 makes a difference in their care, then group living quarters people showing symptoms (nursing homes and prisons).  Anyone else, not being tested.  County Exec has said he expects there to be at least 11 unconfirmed cases for each confirmed case.

They must be getting more tests, now, because 2 residents  at one nursing home with fevers prompted testing. When they were confirmed, they tested all patients and staff there. Found 41 patients (out of 130) and 25 staff with Covid-19. That's the biggest group testing I've heard of to date.

[oped]

Apr 8, 2020 8:49:40 GMT -5 haapai said:

Apr 8, 2020 8:29:23 GMT -5 oped said:

So 42% of NY tests are positive... while 10-11% are positive in Texas and Florida... I'm still not sure why you would assume that translates to Texas and Florida testing the dead at higher rates?  

Flip it around.  Calculate the other ratio.  Tests performed per confirmed case.  Think about what that means.  

What do YOU think it means? 

[djAdvocate]

Apr 8, 2020 8:49:40 GMT -5 haapai said:

Apr 8, 2020 8:29:23 GMT -5 oped said:

So 42% of NY tests are positive... while 10-11% are positive in Texas and Florida... I'm still not sure why you would assume that translates to Texas and Florida testing the dead at higher rates?  

Flip it around.  Calculate the other ratio.  Tests performed per confirmed case.  Think about what that means.  

it could mean a dozen things. what does it mean to you?

[thyme4change]

New york is short of tests?
Texas and Florida have bad tests?
New York is being stingy on who to test, while Texas and Florida are testing all the rich people with no symptoms?

[haapai]

For me it means that someone can die of Covid-19 (positive test) and their spouse will not be tested.  They'll be asked to self-isolate instead.  The living spouse's contacts will also be asked to self-isolate, but I can't imagine that they'll be nearly as compliant as they would be if there were positive tests to point to. 

 

Is this any way to fight this disease? 

[tallguy]

As another indication of how well Washington is doing, the Army Corps of Engineers built a 250-bed field hospital in Seattle to treat non-Covid-19 patients so that hospitals could save their beds to treat victims of the virus.  Governor Inslee announced that it is going to be broken down and sent somewhere else that has a greater need.  It had not yet seen a single patient here.  Certainly appears that the hospitals are not expecting to be burdened by too many patients.

[Tiny]

Here are some current numbers on the New York "at home" deaths. Some eye opening numbers. It also highlights that the number of deaths we're seeing as being from covid19 aren't the full picture (I'm sure most of us knew that...)

www.npr.org/sections/coronavirus-live-updates/2020/04/08/829506542/after-deaths-at-home-in-nyc-officials-plan-to-count-many-as-covid-19

My bolding of the last two sentences

However, even those numbers failed to include many instances when first responders encountered someone who had already died at home or in other non-hospital settings. That happened 280 times on Monday, according to data from the city's Fire Department.

While not all of those deaths are necessarily caused by COVID-19, it's a sharp increase over the average 25 home deaths per day the city was experiencing before the pandemic swept the five boroughs.

Over the last two weeks, the city's fire officials said more than 2,192 New York City residents died in their homes, compared to 453 during the same time period last year.

As of Tuesday afternoon, a total of 3,555 people was confirmed to have died from the virus in the five boroughs.

[djAdvocate]

our death totals are the same as Spain tonight. we are going to cut right through their curve, since we are growing 2k/day, and they are down to 500. they are in a much later stage.

[oped]

They were still 700 yesterday I think? But still.