COVID-19 Update; 3/14/2020. A Message From Concerned Physicians

Howard Luks MD ^

[P.O.E.]

Unfortunately, evidence is often slow, methodical, and boring and has a tough time against clicky headlines and exaggeration

That's why I stocked up over time months ago, I'm taking the "better safe than sorry" route for now. I maintain my right to reserve judgement until more facts come in.

[Lagmeister]

You are nitpicking semantics. Corona appears to have a dry cough with deep lung in the elderly or those who have weak immune systems. H1N1 was overall head, sinus and lung. Both lead to pneumonia, but the H1N1 was more deadly because it attacked with ferocity the young. There was vaccine available for H1N1 and it still killed up to 18,000 in this country.

Actually, I just recently read that 22,000 have died this year from flu... so, there's that.

[DoodleBob]

Here is today's update, using the JH US totals as of now. My forecast is under by -5.79%:

Date	Forecast Cases	Forecast Percent Growth	Actual Cases
3/14/20	2,999	38.0%	2717
3/15/20	3,552	18.4%	3478
3/16/20	4,376	23.2%	4645
3/17/20	6,837	56.3%	.
3/18/20	8,271	21.0%	.
3/19/20	10,202	23.3%	.
3/20/20	12,029	17.9%	.
3/21/20	13,526	12.4%	.
3/22/20	15,264	12.8%	.
3/23/20	16,577	8.6%	.
3/24/20	18,125	9.3%	.
3/25/20	19,408	7.1%	.
3/26/20	20,410	5.2%	.
3/27/20	20,410	0.0%	.
3/28/20	29,488	44.5%	.
3/29/20	33,280	12.9%	.
3/30/20	34,408	3.4%	.
3/31/20	35,590	3.4%	.
4/1/20	36,696	3.1%	.
4/2/20	37,731	2.8%	.
4/3/20	37,945	0.6%	.
4/4/20	38,196	0.7%	.
4/5/20	38,331	0.4%	.
4/6/20	39,201	2.3%	.
4/7/20	39,201	0.0%	.
4/8/20	39,325	0.3%	.
4/9/20	39,630	0.8%	.
4/10/20	39,875	0.6%	.
4/11/20	40,125	0.6%	.
4/12/20	40,320	0.5%	.
4/13/20	40,579	0.6%	.
4/14/20	40,927	0.9%	.
4/15/20	41,047	0.3%	.

[gas_dr]

This is pretty awesome. I will follow you closely — please keep me up to date. I love your model. Tomorrow looks bad then hopefully we see the flattening. The moment of truth is at hand.....

[Alas Babylon!]

The vaccines are even now being tested.

Hopefully, we’ll have a good vaccine and medicines to directly treat it.

The key is now, not some future date.

Help stop the spread.

[DoughtyOne]

This document is trying to help you to understand the situation
at hand and not to terrify you.

We want to make sure you understand the facts and understand what is
at stake. This is a Pearl Harbor moment for our country.

Seriously...

[ClearCase_guy]

"Was it over when the Chinese bombed Pearl Harbor??"

"Chinese?"

"Forget it, he's rolling ..."

[DoughtyOne]

Was that intended for me?

[ClearCase_guy]

Your post contained a quote about Pearl Harbor. I fell back on the dialogue from “Animal House” (”Was it over when the Germans bombed Pearl Harbor?”). I just made it Chinese in honor of Wuhan.

[DoughtyOne]

Ah okay.

I get it. Thanks.

[cherry]

I still think isolating the most vulnerable group would have made more sense than closing down the whole country.....

we should have asked for voluntary quarantine for people 65 and older plus no visitors at all to senior facilities...and we should have done this 4 weeks ago....

[DoodleBob]

I'm introducing a new forecast but retaining the old one, in part because I'm a dweeb but also in light of gas_dr's most excellent analysis.

I also read this very sobering comment this morning, with which I agree (though my memory isn't as long this FReeper's):We are in the midst of the worst panic I’ve ever seen in my life, and that includes the Cuban Missile Crisis, the 1987 stock market crash, SARS, Ebola, and 9/11. I'm not a sappy person, but honestly, it moves my heart to see so many scared people. The Truth, therefore, needs to be unleashed: as Jefferson said, To prove this, let Facts be submitted to a candid world.

I apologize at the outset for anything I'm posting that you've seen before.

The Hubei pattern of cumulative Coronavirus cases isn't exponential, bro: it's logistic .

This is nothing new: this logistic pattern is what we saw with SARS in most nations where this hit in 2003.

While there is nothing magical about a virus' spread and a logistic function, they seem to both reflect one of two things (or both): 1) that a virus MAY infect early the people most susceptible (i.e. The elderly, the infirm) which is the steep portion of the curve but as the quantum of susceptible people shrinks, the spread isn't as great and you get the flattening, 2) the virus spreads like wildfire early but then outside controls like hand washing and self-quarantining takes hold and the curve flattens.

My original forecast - which I'm retaining - was a simple projection of US cases as a function of Hubei's daily case growth starting in 1/31/2020. I picked that date because the US' pre-3/13 and Hubei's pre-1/31 growth rates looked similar. Hubei went into lockdown on 1/23 so perhaps I am off by a week from a "similarity of controls" perspective. However, I chose to be guided by the data more than the news, and the Hubei data may be faked for all I know.

However, there is a nasty jump continuity in Hubei's data in Feb 11-13 (look at the graph above - it's clearly visible). To address this, and after posting (ad nauseum, some may say) about the logistic function, I decided to take a crack at modeling the Hubei data in that manner.

Ironically, I took the same approach that gas_dr did without reading his post: I modeled an index from 0 to 1 that follow's Hubei's cumulative cases as a function of the day of the epidemic, then multiplied that index by my estimated maximum (or thereabouts) number of cases we'll see in the US of 41,500.

Here is my model:

Overall Model Fit...

Chi Square= 32.6367; df=1; p= 0.0000

Variable	Coefficient	Standard Error	p-value
Intercept	(3.2392)	0.0108	0.0026
Day of the Epidemic	0.1988	0.0567	0.0005

The "Day of the Epidemic" is an integer, reflecting the number of days under which the epidemic has been ongoing. For this model, I set Day = 1 on March 10. Thus, for today, March 17, you would multiply the coefficient of 0.1988 by "7" and tomorrow it'd be "8" and so on.

Now, you may be asking: why bother? Well, notwithstanding my own personal view of this saga, I'm watching my fellow Americans be literally scared for their lives. And yea, I know...'they're sheep and let them worry blah blah.' At some point, though, it pays to see someone put things into really, reasonable terms. As such, I believe Wayne Allen Root put it best:

I have many great friends and guests on my national TV and radio shows who are medical experts. Half believe this is the pandemic to end all pandemics. They quote Centers for Disease Control and Prevention models that report as many as 1.7 million Americans could die. So people are rightfully scared out of their minds. American business is shutting down. But the other half of my medical friends and expert guests say this is an overreaction. They predict fewer Americans will die than during the flu season of 2017-18 that killed about 80,000 people. They don't believe we need to close down American business and lock ourselves in our homes. The problem is we won't know who's right until it's over.

I want this to be over. NOW. And while I'm just one, dopey Deplorable posting on FR, ya gotta believe. Therefore, to the extent we can bring some facts to the hyperbole, maybe we can calm our fearful brethren that the curve can (and will) 'flatten' without taking her advice...

...and we can bring this flipping crisis to a halt, FAST.

Date	Forecast Cases: Hubei Pattern	Forecast Percent Growth	Actual Cases	% diff: Actual-Forecast Cases	Logistic Forecast: 1/(1+exp(-( -3.2392+0.1988*day of the epidemic)))*41500	% diff: Actual-Logistic Forecast Cases
3/14/20	2,999	38.0%	2717	-9%	3,315	-18%
3/15/20	3,552	18.4%	3478	-2%	3,974	-12%
3/16/20	4,376	23.2%	4645	6%	4,748	-2%
3/17/20	6,837	56.3%	6362	-7%	5,650	13%
3/18/20	8,271	21.0%	.	.	6,693	.
3/19/20	10,202	23.3%	.	.	7,885	.
3/20/20	12,029	17.9%	.	.	9,233	.
3/21/20	13,526	12.4%	.	.	10,739	.
3/22/20	15,264	12.8%	.	.	12,395	.
3/23/20	16,577	8.6%	.	.	14,189	.
3/24/20	18,125	9.3%	.	.	16,099	.
3/25/20	19,408	7.1%	.	.	18,096	.
3/26/20	20,410	5.2%	.	.	20,144	.
3/27/20	20,410	0.0%	.	.	22,204	.
3/28/20	29,488	44.5%	.	.	24,236	.
3/29/20	33,280	12.9%	.	.	26,201	.
3/30/20	34,408	3.4%	.	.	28,066	.
3/31/20	35,590	3.4%	.	.	29,806	.
4/1/20	36,696	3.1%	.	.	31,401	.
4/2/20	37,731	2.8%	.	.	32,842	.
4/3/20	37,945	0.6%	.	.	34,125	.
4/4/20	38,196	0.7%	.	.	35,255	.
4/5/20	38,331	0.4%	.	.	36,238	.
4/6/20	39,201	2.3%	.	.	37,086	.
4/7/20	39,201	0.0%	.	.	37,811	.
4/8/20	39,325	0.3%	.	.	38,427	.
4/9/20	39,630	0.8%	.	.	38,947	.
4/10/20	39,875	0.6%	.	.	39,384	.
4/11/20	40,125	0.6%	.	.	39,749	.
4/12/20	40,320	0.5%	.	.	40,054	.
4/13/20	40,579	0.6%	.	.	40,307	.
4/14/20	40,927	0.9%	.	.	40,517	.
4/15/20	41,047	0.3%	.	.	40,691	.

[Vermont Lt]

The jump you make reference to was when the CCP started reported cases that were presumptive as the result of a CT scan, but not a test.

It caused an increase of about 14,000 cases in one day. They later “backed them out.”

What the forecast misses is the political and logistical aspects of the Chinese experience. They had a meeting sometime after the 2/14 jump where the political leadership in Hubei was replaced. The new leaders indicated they were going back to confirmation through testing only. Once they had tested the backlog, it was announced they were done and there would be no more cases in Wuhan. And, as if by miracle, there were not.

During the exponential part of the curve the Chinese only reported cases and deaths that were in hospitals. They would not let you into the hospital without being confirmed by a test. But they literally only had a maximum of 400 tests a day.

During that time if you died outside a hospital without being tested it was deemed to be “flu related pneumonia.”

Without taking those factors ( the physical constraints AND the political pressure) using those numbers is a crapshoot. How you would adjust for the changing of counting methods is what stopped me from counting.

Just some thoughts. Not criticisms.

At this point, I think doing the statistics on this an academic exercise. But I also think it’s like counting raindrops in the middle of a storm. Counting the raindrops as they hit the ground is about as lagging an indicator as you can get.

[Robert DeLong]

One thing I see is that you cannot update the day you are on until the last update has occurred for the day, to make the actual number accurate. For example the actual currently sits at 6,423 infected. But I have to be honest, you understand math much better than I. But your projection numbers are certainly in the ballpark as far as I can see. 8>)

The problem with the panic is that they are making this even worse than it is, and it may end being devastating. But so far I do not see any rational reason for the panic. The press is responsible for this, and they do it to try and assist the Democrats in regaining power. This is not the free press the founders envisioned, nor is it the press we thought we had as little as 10 years ago, even though we saw the bias creeping in more and more.

Sooner or later I fear this country will explode or implode. Not sure I want to be around for either event. Sometimes its good to be old. 8>)

[DoodleBob]

Thank you, those are all good and fair points. I recall the data gap in mid Feb and, as often happens in statistics, homogeneity and consistency in the data was likely lost - if it was even there in the first place.

I've dealt often with messy and imperfect data....my kingdom for a clean dataset! My hope is that even with likely faked and fraudulent data, that the fakery was consistent. Thus, maybe the level is off but the trend may be right, in which case my model isn't totally screwed.

I won't be surprised if I'm off by a few thousand cases....maybe even 10k. Forecasting isn't easy. But the absolutely wild and asinine estimates of hundereds of thousands or millions of cases in America is borderline criminal negligence.

If my long-winded post can calm down one person, it shall not have been in vain.

[Pollster1]

My hope is that even with likely faked and fraudulent data, that the fakery was consistent.

That is why I assume no one will bother with the Iranian data.

[DoodleBob]

Here is the updated CoronaDoodle Forecast - the logistic model seems to be off by a day while the Hubei-driven forecast is tracking more closely to the actual numbers. Stay tuned, and pray.

Date	Forecast Cases: Hubei Pattern	Forecast Percent Growth	Actual Cases	% diff: Actual-Forecast Cases	Logistic Forecast: 1/(1+exp(-( -3.2392+0.1988*day of the epidemic)))*41500	% diff: Actual-Logistic Forecast Cases
3/14/20	2,999	38.0%	2717	-9%	3,315	-18%
3/15/20	3,552	18.4%	3478	-2%	3,974	-12%
3/16/20	4,376	23.2%	4645	6%	4,748	-2%
3/17/20	6,837	56.3%	6362	-7%	5,650	13%
3/18/20	8,271	21.0%	7769	-6%	6,693	16%
3/19/20	10,202	23.3%	.	.	7,885	.
3/20/20	12,029	17.9%	.	.	9,233	.
3/21/20	13,526	12.4%	.	.	10,739	.
3/22/20	15,264	12.8%	.	.	12,395	.
3/23/20	16,577	8.6%	.	.	14,189	.
3/24/20	18,125	9.3%	.	.	16,099	.
3/25/20	19,408	7.1%	.	.	18,096	.
3/26/20	20,410	5.2%	.	.	20,144	.
3/27/20	20,410	0.0%	.	.	22,204	.
3/28/20	29,488	44.5%	.	.	24,236	.
3/29/20	33,280	12.9%	.	.	26,201	.
3/30/20	34,408	3.4%	.	.	28,066	.
3/31/20	35,590	3.4%	.	.	29,806	.
4/1/20	36,696	3.1%	.	.	31,401	.
4/2/20	37,731	2.8%	.	.	32,842	.
4/3/20	37,945	0.6%	.	.	34,125	.
4/4/20	38,196	0.7%	.	.	35,255	.
4/5/20	38,331	0.4%	.	.	36,238	.
4/6/20	39,201	2.3%	.	.	37,086	.
4/7/20	39,201	0.0%	.	.	37,811	.
4/8/20	39,325	0.3%	.	.	38,427	.
4/9/20	39,630	0.8%	.	.	38,947	.
4/10/20	39,875	0.6%	.	.	39,384	.
4/11/20	40,125	0.6%	.	.	39,749	.
4/12/20	40,320	0.5%	.	.	40,054	.
4/13/20	40,579	0.6%	.	.	40,307	.
4/14/20	40,927	0.9%	.	.	40,517	.
4/15/20	41,047	0.3%	.	.	40,691	.

[Disestablishmentarian]

Thanks for the great analysis.

I just discovered your work here on the Kung Flu. It is quite convincing.

I had finally noticed the significance of the flattening out in Wuhan for projections in the U.S., but had no idea about there being an off-the-shelf function for modeling it.

Not to pry too much, hopefully ... but is a lot of this work totally original, or is it a professional expertise ... or are you basing from some other published material? Just curious.

Please add me to your geek ping list.

[DoodleBob]

Thank you for your kind words.

My work is original but the concept of the coronavirus following a logistic vs exponential pattern was well-developed by others with SARS. There are several sites online that can give you logistic model parameters

The Hubei-based forecast is nothing special-anyone with Excel can do that.

If you did well in Statistics 301 then what I did isn't too extraordinary...

[aMorePerfectUnion]

What data source are you using?

https://www.worldometers.info/coronavirus/

Shows 9,477 US cases today.