Monday, March 9, 2020

COVID-19 # 23 (Still Required Parts are Highlighted)

The WHO has officially declared this a pandemic. The official declaration appears to ... not actually have much significance.

The Trump administration is considering declaring the entire country to be in a state of emergency. That would allow FEMA resources to be used. Except that FEMA closed its Atlanta office today due to infections.

You should probably read this tweet storm down through # 13:
It is time to familiarize yourself with the origins of the medical term triage. Relate this to the spreadsheet that I covered in class today showing that Italy is not doing that well. Also, here's a full-length article, with many links, about what's going on in Italy.


The U.S. has done ... wait for it ... 4,889 tests as of Tuesday night. Total. That is not a daily figure. The CDC is not collecting this data at all. Instead, there is a private project that is doing this. (Part of this has to do with that state/federal breakdown of responsibilities I posted about last week). So, go here for the numbers: 17% of tests have come up positive. This rate is quite high because in many places we're really not testing anyone until they're visibly ill. For example, Utah has done ... wait for it ... 2 tests. Both of them were positive. Except that Utah has this goofy behavior, that as a state our government is free to choose, of not reporting negatives at all. This sort of behavior goofs up any sort of inference we could make about prevalence.

Given that the U.S. is dragging its feet on testing, it might be worthwhile to go back to the New York Times piece entitled "Italy, Mired in Politics Over Virus, Asks How Much Testing Is Too Much". This was published on February 27th. At that time, there was a serious political debate in Italy about whether they were testing too much or not enough. It's been 2 weeks, and it seems like we know the answer.The times keeps their articles behind a firewall. You can access them, in a text-only form, through the SUU library. This one seems apropos, so I saved a copy of it to the G drive.

Coachella has been postponed until the autumn. Parts of Washington are considering banning gatherings of 250 or more people, while the city of San Francisco is discussing banning those over a thousand; Washington D.C. too. Norway banned gatherings of more than 500.

The Minister of Health for the UK has tested positive. And Joe Biden and Bernie Sanders are cancelling campaign rallies.

Good news. Testing capability for the U.S. is up to 15,970 per day. They're not doing that many tests yet, but hopefully soon. If you're curious why it is so hard to get tests done, check out this interview.

For the nerds out there, here's an excellent analogy:
A DDoS attack is the main way that hackers bring down websites (often for ransom). It's short for Distributed Denial of Service. In it, hackers connect robot programs on a large number of compromised computers to simultaneously try to access a particular website. The servers overload and crash.

 Iran. Consider this. China has few announced cases amongst known government officials, mostly healthcare workers. Iran has ... dozens. And they're not low level officials like in China. Either the Chinese are hiding everyone whose face and name are known in public, or the Iranian outbreak is much larger than we suspect. Graeme Wood of The Atlantic runs some numbers based on available tidbits of information, and comes up with 7 estimates: the smallest ones put the number of infected in Iran at 600K. This makes their outbreak ten times the size of China's, and means their numbers are underreported by a factor of one hundred. Keep in mind what I said in class on Monday: Iran has largely "gone dark" like in a zombie movie (and keep in mind that I said "like", so that's clearly a metaphor not a fact). Oh ... not so dark ... another one of their vice presidents has tested positive; hope this one doesn't die like the other one.

Bahrain repatriated some citizens from Iran this morning. Forty percent of the group has tested positive. Is it plausible to interpret this as a random sample from the population of Iran?

Governor Cuomo of New York has instituted a "containment zone" in New Rochelle (northern suburbs of New York City). No one is sure what the difference between a containment zone and a quarantine is. The National Guard is being deployed to the area (remember what I said about governors being able to call out soldiers in uniform if needed).

What follows is a very difficult chart to read:
We pay a lot for healthcare in the U.S., and part of what that buys is a lot of hospital capacity that is unused much of the time (the gray bars). In Wuhan, they started with a lot less; everything they built was immediately filled until about February 14th. Their rate of hospitalization eventually plateaued just below the rate for hospital beds in the U.S. Their rate for severe cases peaked right about our rate of ICU beds. Here's a way to think about this: China needed to have a severe quarantine in Wuhan because they were going to exhaust their more limited facilities more quickly. The U.S., or a subsection of it, can afford to be slower or more limited about a quarantine, but there's every reason to believe the outbreak can fill what the capacity we have.

Here's a tweet storm about the outbreak in Seattle. It's fairly detailed on the epidemiology:
If you're interested, click the box above, and then when a new page opens with this box at the top, click the "Show this thread" link. Basically, the point estimate is 1100 people infected in the Seattle area, and that number is still growing. They are still processing tests from the middle of last week, a few hundred every day. Lots of people show up in healthcare facilities with respiratory problems; the percentage of them associated with SARS-CoV-2 is low, but still rising. This indicates that all the rest are still out in the community infecting others: point estimate for March 5th was 620 infections.

Then there's this:
Data on illnesses in Washington shows that it has gone from below the national average a month ago, to double the national average today.

Xi Jinping has now made a state visit to Wuhan. Funny how that happened a few days after American legacy media noted that he bears more responsibility than what was once thought, yet had not made a public relations tour of affected areas. Oh, and the visit was unannounced, so press, pictures, and videos were not available live.

Of course, soccer is a lot bigger in Europe. Historically, one of the biggest countries for that is Italy. They have now suspended play in their top level league. No word yet on what will happen to the Champions league match involving an Italian team scheduled for next week.

Ireland has cancelled all St. Patrick's Day parades. No decision yet on those in Chicago and Boston.

Here's the killer statistic that Americans should be focused on. We have some awareness that we're just not testing very much, and here's the data:

Do the math: South Korea is doing tests 700 times as fast as we are, they've done 35 times as many in total, and they've gotten a lot more positives. Here's the forecast: expect cases in the U.S. do go up rapidly this week as testing becomes more common.

Also, now that we are actually getting the tests out, the new roadblock is the number of tests that can be evaluated every day. Currently this is estimated to be 7,840. That's a third of what South Korea was doing last week. Since not all tests run will be positive, this puts a ceiling on how fast U.S. numbers can grow.

There is now a list of full and partial college and university closures, with dates, and shutdown details. The big names are Barnard, Columbia, Fordham, Hofstra, Monmouth, NYU, Ohio State, Princeton, Rice, San Jose State, Santa Clara, Standford, The New School for Social Research, Berkeley, Florida, USC, Vanderbilt, Yeshiva, Harvard, and of course the University of Washington. That file has some hidden rows that make it a pain to count, but as of Tuesday morning it went down to row 62.

Here's some environmental irresponsibility dressed up as science (personally, I want to care about environmental issues, but I find the abuse of supporting information offensive). Anyway, this post at G-FEED entitled "COVID-19 reduces economic activity, which reduces air pollution, which saves lives" (sorry, not even worth linking to) comes up with results that look all science-y, but which are implausible on their face. To wit, the reduction in air pollution in China due to the outbreak prevented over 77K deaths. In an earlier post, I discussed rates at which crematoria are used in China: normal rates are about 25 times per day, times the 5 in the city of Wuhan, which has over 11 million people. Obviously some people, maybe a lot, are not cremated. There is no reasonable way to reconcile these numbers, even if all "normal" deaths are due to air pollution.

The UN is currently estimating the hit to global GDP at $1T, with a worst case scenario of $2T. This s 1-2% of the global economy. Do note (and no slam intended) that the UN is not really known for its economic analysis.

Italy will be quarantining the entire country on Tuesday. Main focus is on the transportation system, since so many people fled the partial quarantine put in place the other day. Now, almost all public activities have been shut down in a Wuhan/Hubei like manner, and can be confirmed from videos of Milan readily available on the news. On Tuesday there were reports that the healthcare system in the Lombardy region (the big rich one) has reached capacity; the death rate there is at 8%, which is double what it was in Wuhan (see the paragraph below outlining that the strain in Europe is definitely different, and plausibly worse than the original one).

If you recall, when I first started covering COVID-19 in January, I didn't post directly to many videos of life inside the quarantine area — there was reasonable doubt that some of them were fake ... heck, no one ever imagined bars being welded over apartment doors. We now know that most of them were true. Anyway, there are videos of armored vehicles patrolling streets in the smaller city of Foggia in southeastern Italy.

Also, related to JE's point made in class on Monday: ski resorts in Italy were closed after students were found to be planning ski trips during their newly found time off. This is a classic example of the problem of tragedy of the commons and the insights we get from studying the prisoners' dilemma. The tragedy of the commons aspect is that since the quality of public health isn't owned by anyone in particular, no one has an incentive to defend their property rights. Then the desirable good (public health) gets abused. The prisoners' dilemma aspect is that if your strategies are to obey or disobey a quarantine, the dominant strategy is to disobey, but if both parties disobey they arrive at the poorest outcome (potentially infecting each other). Due note that this outcome is a Nash equilibrium to the prisoners' dilemma game, so we should not be surprised if it shows up ... pretty much every time the game gets "played".

 Excellent chart of progress, and lack thereof, in the global hot spots:
Note the log scale (yeah). A useful addition to this chart would be something showing incidence rates per million people. (Always warning that we may not be able to trust China's numbers due to incentive problems), China has 28 times as many people as South Korea, 23 times as many as Italy, and 18 times as many as Iran. This means the outbreaks in South Korea, Italy, and Iran are probably worse than the original one in China.

I also like this map of Italy showing the old quarantine zone that started Saturday. The purple dots show population, so you get a sense that this is like a satellite view of the region at night, where the important macroeconomic/epidemic detail is the population concentrations, not the political borders.
Also note the pattern that's similar to satellite images of Utah: population is strung out in threads along highways and railroads.

Trump changed his chief of staff last week. Now the new COS is in self-quarantine. Scary thing is, this is related to attendance at the CPAC convention, so he was exposed 12 days ago. Five members of Congress are now in self-quarantine.
 
BD brought up in class that there are different strains of the virus. By strain, they mean something that through mutation is no longer chemically identical, but which is genetically close enough to be the same virus. I remarked that there's a lot more than one. Here's the chart:

As you can see, there are over 100 on here. They are organized into clades. A clade is set of things that are closer to each other than to other clades. Viruses collected in the U.S. are most of the magenta branches. They also mostly belong to the S Clade, found primarily in China. The V clade is predominant in southeast Asia and parts of Europe; it popped up fairly early in Singapore and got transmitted through business travelers. The G clade is worrisome. This is almost exclusively in Europe, and Italy has had higher rates of ICU care than China did. So it is plausible, but unproven, that the G clade is a bug that has evolved to be slightly more dangerous.

Regarding JE's comment from Monday's class:

I will be working on this one up until class time on Wednesday.
Posted by at

No comments:

Post a Comment

Subscribe to: Post Comments (Atom)