They call it disinfection? Indonesia… no words. pic.twitter.com/FUS85aJ5n1
— Russian Market (@russian_market) February 2, 2020
My first thought was this was an Indonesian wet hijab contest.
But, no. They are spraying people so that the spray seeps into skins and into bodies to kill the home of the virus. Maybe. I hope. Or perhaps they are just panicking.
On Friday the Drudge headline was School District Closes For Two Days After 600 Students Call In Sick.
An entire school district in Ohio was shut down for two days this week after nearly 600 students called in sick, the district’s superintendent said…
Many of the students have the same symptoms, such as high fever, vomiting and body aches…
There’s nothing in the story about the new coronavirus, but the story was embedded in a series of such stories. Like how the USA is going to quarantine foreigners for two weeks at some (at the time) undisclosed location if they had traveled to the States from China’s Hubei’s province. Is that happening? Or was this only rumor?
Italy had barred all flights from China by Thursday night. They defined China as China, Macau, Hong Kong, and Taiwan. Taiwan? Ah, politics. The WHO lists Taiwan as part of China, and WHO’s on first. Singapore followed suit, but was more careful about their country definitions. Vietnam wasn’t as careful.
Lots of airlines have cut off new flights to China.
China themselves have quarantined several cities, leading to all sorts of difficulties. Like how to feed the quarantined.
Is this level of hyper-concern warranted?
Here are the Numbers!
There were, in the United States alone, in the week ending 25 January, over 19 million cases of disease. There were 180,000 hospitalizations. And—the big one—about 10,000 deaths!
That’s a lot, isn’t it? Worse, 68 of these deaths were in infants. The rate of infection for 65+ was 71.3 per 100,000, which is huge.
Quoting directly: “6.7% of the deaths occurring during the week ending January 18, 2020 (week 3) were due to P&I. This percentage is below the epidemic threshold of 7.2% for week 3.”
Quoting who? The CDC. A known authority.
About this year’s flu.
Which has not yet reached its peak, or is maybe at the peak, but in any case with several more weeks of flu season to go. That “P&I” is pneumonia and influenza.
This year’s flu, which sounds awful, and is as far as body count goes, is not as bad as it was two years ago. Indeed, this is a mediocre year, as this plot shows:
Two years ago was twice as bad as this year’s horrible numbers. But I don’t recall a panic then, and there isn’t one now, either. About the flu. Importantly, there weren’t any travel restrictions or extraordinary measures taken to contain the flu then or now, yet the flu each year, as it does, burned itself out for all the reasons you already know (it gets warmer, people go outside, etc.).
What about the coronavirus? Here are the numbers, as of 7 PM Sunday night EST (Monday morning in Asia).
According to one source, cases were 17,386 worldwide with deaths 362. Are these the true numbers? We hear all kinds of rumors out of China, and China being China some of these are probably true, or they are in the right direction. Perhaps the death counts are higher. I do not know.
However, conditional on these numbers, as all probability is conditional on the assumptions we make, I fit a naive R nls/SSlogis logistic curve to both data sets. You can see the predictions. Unfortunately, the nls method doesn’t have prediction intervals implemented yet, so I can’t show you the +/-. There are many other ways to fit this curve which will give them, and maybe we can do them together in a class.
That’s not the point of the picture though. I do not swear to the prediction (fit is lousy in the beginning), but I do caution against plotting raw exponential-only curves as I’m seeing elsewhere (e.g. this). It cannot be that the number of new cases and deaths will increase without limit exponentially. When the panic/concern started a week or so back, I saw some projections of 1,178,144 deaths by 21 February using these kinds of exponential models.
Most infections follow something like a logistic curve, like SARS, which might also with MERS (via a camel intermediary) have originated in bats. The naive curve above might not be catching the recent increase, but the idea is likely not far wrong. Yet there could be another strong upward kick, though its a good bet it settles down globally like a logistic curve. Maybe the middle prediction at this link is better: 60,000 total cases by the end, 2.2 times this naive fit, which would give about 1,355 deaths.
Here’s more numbers:
When epidemics start only the sickest go to hospital to get treated, hence the early high mortality rate. As publicity increases, more people who really didn’t need special treatment go to get checked, and the mortality rate decreases. A good indication we’re settling down is the flattening of the mortality rate. If these are the right numbers; again, I don’t know.
Publicity thus also drives cases. On the other hand, since the coronavirus is a cold virus which in robust people will seem like a cold, so many cases will go unreported. How many I don’t know. Obviously these unreported cases are not dying, though.
Some of the panic seems to be settling down. Yet breathless is one word to describe how people react to each new report. Here, for example, is one Dr. Eric Feigl-Ding, who tweeted “Whoa- the rate of increase ***outside of China*** is steeper than inside of China or Wuhan! Figure 1A. From: @TheLancet ‘Nowcasting and forecasting the potential domestic and international spread of 2019-nCoV http://bit.ly/2GF6gZP’)”. I replied (including my typo), “Well of course it is. Being a novel virus to a new, uninfected area, and the rate of increase necessarily is higher than in an area where is it extant. The rate of increase on the price a new widget, previously non-existent, is infinite.” Make that bring not being.
I never did hear back from Fiegl-Ding (what a fun name!).
Am I wrong in sensing that some want this to be worse than it seems to be?
It is obvious that all the control measures governments and people are taking will reduce the infection rate. Would the spread be worse had these measures not been taken? Well, to some extent, yes. But to what extent? That’s a counterfactual question you can’t answer by appealing to observations.
For example, it must be that a strict quarantine (true isolation) of a infected man keeps that man from spreading his disease. The quarantines in place appear to be rather leaky, though. And on the other hand, the far worse flu of two years ago burned itself out without extraordinary precautions, as it does every year. So maybe we’re fooling ourselves with our “solutions.” But then flu deaths are huge!
The global instant news cycle is driving some of the panic/concern. Officials, or some of them, must be reasoning that it’s better to do something just in case. If the something they do wasn’t needed, they’ll be able to point to all the other officials and say “Everybody’s doing it.” On the other hand, whatever deaths there are outside China will be blamed on those same officials.
As of this writing, and using the same data source, all the deaths save one were in China. The other was in the PI. The USA had 9 cases, no deaths.
Sure, the new coronavirus might bust out everywhere. Or again it might not. Taken together, this is a tautology that does not given any information about what to do.
To support this site and its wholly independent host using credit card or PayPal (in any amount) click here