COVID-19 Data Model Updated Daily

[Admin]

...and we know that the data for this new virus has all of the issues that IMHE noted can result in larger uncertainty intervals.  I thought that the IMHE definition and graphic was great but I certainly appreciate the additions.

Every death is tragic, but in the big picture we're having a relatively mild Influenza season and must be doing something right to be keeping our CoVID19 numbers so low. 

The alternate perspective would be that even with a country shut down for 2.5 months, the IMHE is estimating 69,000 deaths on top of whatever the yet to be determined flu #'s end up being.

[Chan]

Today alone, the total national deaths will surpass the highest weekly deaths during last years flu season.  Many of the models, and most experts, are predicting a steep drop soon but adding to the uncertainty is the fundamental challenge of predicting human behavior.

[Admin]

A few interesting updates from https://www.worldometers.info/coronavirus/country/us/ .    Someone had posted a link to this site and it is really interesting for watching the management of some of these issues in near real time.

April 15th (incomplete)

23310 new cases and 1819 new deaths in the United States
New York Governor Cuomo: "we will begin reporting all categories of fatalities pursuant to new CDC guidelines and are contacting facilities to get updated numbers" There may be additional people who died that have not been counted because not in a hospital

April 14th

26945 new cases and 2407 new deaths in the United States
New York City today has reported 3,778 additional deaths that have occurred since March 11 and have been classified as "probable," defined as follows: “decedent [...] had no known positive laboratory test for SARS-CoV-2 (COVID-19) but the death certificate lists as a cause of death “COVID-19” or an equivalent" [https://www.worldometers.info/coronavirus/country/us/]. We will add these to the New York State total as soon as it is determined whether the historical distribution can be obtained

[LaPerouseBay]

Thanks Admin. 

Interesting article from CIDRAP this morning. 
https://www.cidrap.umn.edu/news-perspective/2020/04/antibody-tests-may-hold-clues-covid-19-exposure-immunity-its-complicated

I copied/pasted and plugged their numbers into Dr. Attia's spreadsheet.  Interesting to me that the PPV rises with the prevalence.  I inserted 5, 15 and 30%. 

Dr. Attia's serologic test he's been using (for the past 4 weeks, he deleted the video. (?)), are upthread.  80 and 99%. 
Copy/pasted from the CIDRAP article:
----------------------------------------------------------
According to Cellex, the test was 93.8% sensitive in testing on 128 samples from Chinese COVID-19 patients confirmed by reverse transcription polymerase chain reaction (RT-PCR), and 96.4% specific.

Those impressive-looking numbers, though still leave the possibility of a significant number of people having false-positive and false-negative tests. For example, if 5% of the US population actually had the virus, a test with 95% sensitivity and 95% specificity conducted in a million people would correctly detect 47,500 cases, along with 2,500 false-negatives (those who were infected but were missed).

But it would also produce 47,500 false-positives. That means that if you were to test positive, there would be only a 50% chance that you actually had the virus and have some immunity.
---------------------------------------------------------------------------------

[LaPerouseBay]

I went to this guys website (Peter Attia) and downloaded the spreadsheet.  I have to admit it is making my hair hurt.   

I feel like my wife and kids when they say

"Dad I have no idea what your are talking about . .  no one understands your graphs and you're just saying blah, blah, blah . . $&^%^&&^"

I respond "But it is so obvious." 

Clearly not obvious.

As a math novice, it's comforting that the spreadsheet may have elicited a bit of uncertainty on your part.  I'll just keep plugging other peoples numbers into other peoples spreadsheets...  Go science!   

[Tom]

OK, you guys are way over my head, but you may find this data model interesting from the CDC, take a look.

https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm?fbclid=IwAR065pbyShNfBLvyH0bjzdCjQtmmbXwjuCBCE0VPdOaf1FX3Hg4TwzPJGVY

[Beasho]

OK, you guys are way over my head, but you may find this data model interesting from the CDC, take a look.

https://www.cdc.gov/nchs/nvss/vsrr/COVID19/index.htm?fbclid=IwAR065pbyShNfBLvyH0bjzdCjQtmmbXwjuCBCE0VPdOaf1FX3Hg4TwzPJGVY

OK This is Pretty Earth Shattering!!!!

I am still sharpening my pencil but according to this CDC Website:  Since Feb 1st 73,000 Fewer People have died in the US vs. Expected Numbers (I assume Expected come from Forecasted Deaths by Month). 

Checking Numbers: 631K / 11 Weeks = 57.4K per Week * 52 Weeks = 2.98MM per year.

As of today there were 26,819 Deaths from Covid-19.  This means that 99,880 fewer people died that WOULD have otherwise died because of COVID-19.

Net Deaths DOWN by 73,000 Year Over Year because of Lock-down ().  Almost 100,000 people otherwise SAVED because of Shelter in Place. 

Please check my numbers. 

[Beasho]

I will circle back on this BUT if true the result would be better than my ZERO net deaths theory. 

It is NEGATIVE net deaths from Covid-19.  26K people Died of Covid-19 but 100K were otherwise spared.

NEGATIVE NET DEATHS.  For every 1 person that dies of Covid-19 4 People were otherwise saved.

Can this be true?!

[justsomeguy]

I will circle back on this BUT if true the result would be better than my ZERO net deaths theory. 

It is NEGATIVE net deaths from Covid-19.  26K people Died of Covid-19 but 100K were otherwise spared.

NEGATIVE NET DEATHS.  For every 1 person that dies of Covid-19 4 People were otherwise saved.

Can this be true?!

Data is good, but it's purpose is to help us make decisions ... So lets say it is true ... maybe it's not 4 to 1, maybe it's only 2 to 1... either way, lets run with that for a bit ... starting with, so what decisions should our leadership make based on that?

Do we keep our foot on the brake and continue a conservative path and drag out re-starting our economy (already in a very tenious position)? ... or do we ease off on the brakes a bit and initiate a faster re-start of the nation by smartly re-opening businesses, sending people back to work, re-starting national air travel ... basically let businesses and individuals get that stimulus money moving, while only applying the brakes enough to prevent sliding off the road and ending up in the ditch again.

Of course the medical/health/science aspects also need to be urgently pursued (testing, vaccine, prevention), but it could be a 2 pronged parallel effort vs. a serial delayed effort that could leave all of the living in dire straights.

[peterp]

Murder and traffic deaths are down by over 30% in South Africa due to lockdown. That translates into 40+ lives saved everyday (=840 after 21 days of lockdown). Covid-19 has claimed 34 deaths TOTAL as of this morning......Covid-19 is saving lives (+800 so far).

[Admin]

It is valuable to compare to where we would be without mitigation.  Here is a graph that was presented by the Trump administration earlier this month:

[Admin]

Concerning current status and mitigation efforts by country:

The US has 22% of the world's Covid-19 deaths with 4% of the world's population.

The US currently has 31,000 deaths and the deaths per day is still increasing.  The deaths from the data error (from a few posts up) are now being added back in to the past few days which had signaled a possible top.  It is also worth noting that most of those deaths have been lumped onto one day while they actually were spread more evenly among days since the 11th.  These kind of errors are where smoothing (rolling averages) becomes important.

New York is now at 14,000 deaths.  This is higher than all but two countries. 

[Beasho]

I will circle back on this BUT if true the result would be better than my ZERO net deaths theory. 

It is NEGATIVE net deaths from Covid-19.  26K people Died of Covid-19 but 100K were otherwise spared.

NEGATIVE NET DEATHS.  For every 1 person that dies of Covid-19 4 People were otherwise saved. [/b]

Can this be true?!

I couldn't sleep on this one.  And woke up reading fine print.  CDC Data is rolling on Deaths it will keep getting revised so we'll have to keep watching it.  There may be some or even more than all truth to it.   But Admin's posts also woke me up. 

Yes single day death toll highest ever.  We had been tracking at ~ 2,000 per day and then dropping to 1,500.  But Today SPIKE.  This is materially bad. 

Bad for the summer.   Bad for the proverbial Hurricane forecast cone. 

[Beasho]

US Death Spike.  Admin also referred to this above. 

Data catch-up or Horror Show? 

Italy shown in Second Chart.  The US had been following this exact same course until today.

US Accounts for 3.6% of all Deaths for All Time in One Day from Covid-19. 

Full Model:
https://app.powerbi.com/view?r=eyJrIjoiNzExYmM2ZWQtZWVmMS00OTI1LWE1NzYtMTU2OTk3ZGE4ZjMyIiwidCI6IjQzODE2ODE2LTY0OTYtNDU4My1iMmVmLWU5Y2U3MTg1NjI4MCIsImMiOjF9

[mrbig]

It looks like the countries with the highest test rates for entire populations are faring better especially when combined with restrictions on movement.

Scared that opening for business before June at the earliest is going to be horrible.

Hope I am wrong.

Beasho - thanks so much for your model and observations..