COVID-19 # 55 (Required Parts Highlighted)

China updated the death tolls in Wuhan, adding over a thousand to the total. Better late than never.

Here's an updated version of those common charts:

Nice that it has a logged axis. Not nice that it allows comparison of one country to another (which is usually good, but which with a logged axis can lead to trouble). Not great that it doesn't have deaths as a percentage of population: that would be more informative, but would obviate the need for logging the axis too. The variety of countries shown is good, as well as the extended time frame on the horizontal axis ... but it also tends to show that the story China is telling is somehow different.

Deaths were way up in the U.S. yesterday. People have been trying to rationalize the reasons, and I haven't seen a good one yet. Oh boy.

Do you know anyone that still thinks this?

1/2 New York City, with population of 8.4 million, has 11,477 confirmed and probable #COVID19 deaths. Even if everyone in the city had covid (8.4M people) which is improbable, then 11,477 deaths would put its case fatality rate higher than what's seen in seasonal flu. #NotTheFlu

— Scott Gottlieb, MD (@ScottGottliebMD) April 17, 2020

If you do know someone like this, they'd better that a virus that somehow has spread all over the globe is going to stop before it gets to their town. BTW: deaths were up 10% in Utah ... in the last day, and cases were up 4.5%. We're up to 42 in Washington County, 17 in Iron County, and 15 in the rest of the Southwest and Central Utah Public Health Districts.

The IMHE model is the primary one everyone is using to forecast COVID-19:

New analysis by @meyerslab finds that the popular IHME model underestimated deaths when applied to Italy and Spain. h/t @nataliexdean pic.twitter.com/ynpHDOOFyc

— Caitlin Rivers, PhD (@cmyeaton) April 17, 2020

Not good. BTW: the model  is doing a pretty good job of forecasting deaths in Utah though. It's still predicting a Utah peak late next week.

Best estimates are the U.S. needs to do about 500K tests per day to be able to model the outbreak effectively. But over the last week we've plateaued at about a third of that. As far as states go, Utah is doing OK:

You have to crunch the numbers a bit, but a 500K/day rate would shift a state from one color to the next every 4 days. Currently we're shifting every ten.

One of the things that is worthwhile hoping for is that a lot of people have already been infected and don't know it. There are serology tests for this. Here's what they found in Wuhan:

Wuhan hot spot Zhongnan Hospital did antibody tests on 3,600 security guards, cleaners, doctors and nurses and 5,000 visitors.

about 2.5% positive for exposure. which surprises on the low side, while still being 5X the rate of confirmed cases in city population. https://t.co/OdAljsf0nE

— Antonio Regalado (@antonioregalado) April 17, 2020

While showing a lot of undiagnosed cases, it's not nearly a high enough proportion to limit the spread of future outbreaks.

There's now video of mass burials coming out of Iran.