Saturday, March 21, 2020

Coronavirus stories that seem contradictory, but aren't

First up, the US infection statistics are exploding:
Coronavirus cases in the United States are doubling faster than in any other country and Americans need to reduce contact by 90% to stop the pandemic, a new report by a local nonprofit says. 
“The rate of increase in domestic cases is now estimated to be in the range of 25 to 40% per day near to or eclipsing the alarming rates of case growth occurring in France and Italy,” the report titled “Stopping COVID-19: Short-Term Actions for Long-Term Impact” reads. 
The report was written by the infectious disease analytics team from MITRE, a Bedford-based nonprofit.
Next, the death rate is dropping like a rock:
While coronavirus cases have increased in the past two weeks, the mortality rate has declined, pointed out veteran Fox News analyst Brit Hume.

He retweeted a day-by-day list of the number of cases reported compared to the number of deaths. It showed a drop in the mortality rate from 4.06% on March 8 to 1.84% on March 15.
You could rewrite both stories to say "Early statistics on Coronovirus infection and death rates in the US are crap.  They're still crap, but at least 25% less crappy.  In a couple more weeks we expect the crappiness to have been reduced by 75% from its peak."

As to infection rate, we've only been testing for a week or two at most.  Of course the infection rate is going up, because instead of asking "Have y'all been to China" we're actually, you know, testing people now.  Duh.

And I covered the death rate in some depth here, but this is the key bit:
  • Reported death rates vary from a high of 3-4% to 1% to a low of 0.1%.  Sure, we know the death count with some level of reliability, but since we don't have any idea about how many people contracted the virus, there's no good way to predict how bad (or less bad) things might be.
The good news is that the data is getting better.  It will keep getting better for a couple of weeks.  At that point we can start extrapolating the data and making some predictions that won't be complete garbage.  Until then, keep in mind that we really don't know what's going on right now.


Chris said...

I have been trying to tell my coworkers much the same. But our company's official line is that things are getting apocalyptic. (Large multi-national firm.) Plus the news media is all hype and hysteris. We'll see, as you noted, in a few weeks or so. (BTW, howdy from Catonsville)

Old NFO said...

Lack of data is always the bane of analysis... It will get better as more data comes in. And stop listening to the damn MSM who are trying to sell doom and gloom to keep the panic alive!

LL said...

I have friends in Italy and in the UK who have and are recovering from COVID-19. In Italy they aren't testing anymore, which means that the stats with the exception of the death toll, are unreliable. In the UK, they really never did test much and aren't testing.

So the numbers (everywhere) are skewed.

Jess said...

I think the best indication of the illness will be the number of individuals that directly know someone with the illness. In two weeks, if this indication doesn't arrive, many will feel either duped, or believe the efforts were a knee-jerk reaction. How they respond will be interesting.

Beans said...

We would have had a much better reading of the actual infectious rate if the CDC had just asked South Korea for help in developing a US produced test system.

But, no. The CDC stepped on it, big time. After this panic is over, the CDC needs a major shoe-leather colonoscopy.

As to the real infectious rate, looking at South Korea, Japan, Hong Kong and the Flu Cruise, those are the best examples of spread and mortality. Test early and often, isolate and support, fresh air and food.

On the other hand, Italy, Iran, Communist China... well... Lots of lung-compromised people in a non-really modern world and with a health system already totally hosed, and you end up with.. Italy, Iran and ChiCom. Add in North Korea, which reports are beginning to creep out that, duh, Wuhan Flu is smacking the Hermit Kingdom really hard and we have examples of the worst things to do. Lots of sickly people, poor air quality overall and lack of modern medical attention and supplies.

We are, in many places, set up like South Korea in preparation. We are, unfortunately, in many places meeting or exceeding the bad examples of how not to prepare and react, like, oh, say, the LA Basin.

It is up to us if we rise or fall to the occasion.

Aesop said...

"We are, in many places, set up like South Korea in preparation."
Not so much, Beans.

SK has 22 hospital beds per 1000 people.
Japan has 13.4 per 1000.
Italy has 3.2 per 1000.
The US has 2.8 per 1000.

If we had 6.3M hospital beds, instead of 900K, then we'd be as prepared as Korea.
And we couldn't get the staff (doctors, nurses, etc.) for those beds in 30 years of all-out WWII mobilization effort.

We are, in point of fact, set up in preparation more like Italy.

And the current case surge has told us one thing, undeniably, whether it's due to more testing (I agree that it is) or not: the number of cases in the U.S. in the wild is booming, and we didn't do nearly enough to screen it out at any point.

But that's what happens when you do kabuki theater temperature spot checks at the airport, for a virus that incubates at least 2-14 days.
That FakeScreening probably ensured that 99.8% of all infected persons traveling here arrived, and propagated it here, probably back as early as mid-January, when everyone who could do so got out of China ahead of their travel bans.

We were seeing atypical severe pneumonia ICU admits back that far in the hospital, but no one correlated that to Kung Flu (for which there was no testing possible anyway).

And the real CFR will be notable, but secondary; the real key number to know is the rate of infection requiring hospitalization, and the demographics of who gets sick. That will determine how bad this is, and how bad it'll get, long after this pandemic has swept by us.

If we're going to do a shoe-leather colonscopy of CDC, can we please shoot the shoes out of cannons, after filling the shoes with lead foot forms?

Divemedic said...

The increase in the number of deaths is a better indication of the spread of the virus, IMO. The only problem is that number lags the number of infections by about 14 days. The daily increase in the number of deaths has been holding pretty steady at about 24% for the past two weeks.

Aesop said...

Numbers will be going to crap again:
L.A. County, running out of test kits already, tells doctors not to bother testing anyone, since it won't change how they'll treat patients.

So after a spike in cases, they'll drop, neither of which is representative of reality.

The only number anyone will track will be deaths, and that's one of the least-important numbers in this outbreak.

So much for getting a statistical handle on things.
Top. Men.

LSP said...

We're most definitely in uncharted waters and it'd be nice to think the Chinese death toll was accurate.

Let's pray Italy's not the model.

Richard said...


The CDC needed a shoe leather colonoscopy about 30 years ago. It is a highly politicized and deeply incompetent operation. They have spent untold money over those year on political topics rather than medicine. Had they been doing their actual jobs we would have been better prepared for this. A competent CDC wouldn't have prevented it, as it is a natural process, but we certainly could have stockpiled supplies, trained personnel and had a plan in place instead of having to do it on the fly. The politicization was a plus for Clinton and Obama but WTF were the Bushes doing.

Ken said...

All true about the CDC, but you probably don't want to spend too much time thinking about the proportion of everything everywhere that is done on the fly. I can't tell you how often I feel as though I'm trying to put new head gaskets on a small-block Chevy while the car is rolling down the freeway.

Will said...

"... it'd be nice to think the Chinese death toll was accurate."


Aesop said...