Since my first article on “Wuhan Novel Coronavirus 2019-nCoV” there have been a lot of developments; it’s time to catch up.
Science stuff
First of all, what is nCoV? It’s a new strain of a family of viruses that have been known for a long time. Coronaviruses can infect both mammals and birds, and usually, they aren’t serious: in humans, they are one of the viruses that can cause the “common cold”, in cattle and pigs they can cause diarrhea, and they usually cause upper respiratory infections in chickens.
This is kind of sad in chickens since they have no pockets in which to carry handkerchiefs.
Some variations are more dangerous, of course. The SARS and MERS viruses have a relatively high mortality rate, and nCoV is a close cousin.
The nCoV virus was first identified in Wuhan, in Hubei Province in China, and first reported in a paper Clinical features of patients infected with 2019 novel coronavirus in Wuhan, China on 24 January 2020. This paper described the virus based on 41 cases that had shown up when people presented themselves to emergency rooms with fever, diarrhea, and respiratory distress that turned out to be caused by pneumonia, the accumulation of fluid in the lungs.
Of those original 41 patients, 13 were admitted to intensive care, and 6 died. The virus outbreak was originally associated with the Huanan Seafood Wholesale Market (武汉华南海鲜批发市场). I’ve included the Chinese partly because Chinese writing is inherently cool, but it lets me point to a fun fact: “Hua” (华) in scientific Chinese is used for “corona”. So the coronavirus was associated with the Corona Market.
It’s just a coincidence, honest. In any case, further tracking has identified the earliest case as having started about 1 December 2019 in someone who had no connection to the market.
SARS, MERS, and nCoV are all closely related to one another, and all are known to have their primary reservoir in bats in the wild. With the association to the Huanan Market, the notion got started that the disease was first spread by eating bat soup. So, just to head this off:
- bat soup probably isn’t the reason. Coronaviruses don’t survive boiling.
- exposure to live bats in the market might have been the original source, but given that 1 December case it looks less likely
- yes, they do eat bats in China. People eat bats all over the world. People in China think cheese is weird. People in Germany and the US eat fermented cabbage. Try to get out more.
One interesting group of these first cases was a cluster of family members in Hong Kong Shenzhen who had visited another family member in hospital in Wuhan. Six of the family members had visited Wuhan, five of those were sick. A sixth family member who hadn’t visited caught the virus after they returned. None of them have died.
This is interesting for a couple of reasons: first of all, this was one of the first clear examples of human-human transmission; secondly, there’s a very broad range of severity, from very sick to no symptoms at all, asymptomatic.
Since then, there are 12,024 confirmed cases and there have been 259 deaths. (Updated at last look, and probably going to change again.) The link is to the very useful dashboard that’s being maintained at Johns Hopkins University, and those numbers are as of 10 AM Eastern Time on 1 February 2020.
The nCoV outbreak was declared a “public health emergency” by the US Government.
More science stuff
The nCoV virus basically has almost the same symptoms as the usual common cold or influenza: fever, coughing, shortness of breath that appears to be caused by mild to severe pneumonia. People can also apparently develop the infection asymptomatically. This presents an interesting problem because there’s no way to distinguish a mild case from a common cold, so now anyone with a cold has the potential of being quarantined and tested for the virus, and a lot of people with the virus are tempted to dismiss it as just a cold.
This means that the number of people who actually have the virus, or have had the virus, is almost certainly more than that 12,024 identified cases. One model (Gardner, 31 January 2020) suggests that the reported cases are around 10 percent of the real infections, so that means as many as 120,000 actual infected people, of whom 108,000 aren’t sick or aren’t sick enough to have seen a doctor.
Now, people are reacting as “ooh, there are lots more cases than being reported” and even as “They — capital T — are hiding something from us!”
Well, no. It means that most people who get it don’t seem to be getting particularly sick, which is a good thing.
The real effect of this is that estimates of mortality rate are dropping. Right now, based on reported deaths over reported cases (259/12,024), it’s around 2 percent, where the original reports (6/41) it was 15 percent.
If the Gardner et al model is correct, then the real mortality is around 0.2 percent — which would put it squarely in the same range as the normal flu outbreak.
The actual modeling technique is cool and I may write about it separately, but the 25-words version is that they model the spread of infection based on the number of potential contacts using a simulation of how individuals make contact with others and a number R0 that represents the number of successor cases one case causes over its lifetime.
“R naught” — R0 and What It Means
R0 (pronounced “R naught”) is called the basic reproduction number, the basic reproduction ratio, or (incorrectly) the basic reproduction rate. (Incorrectly because a rate is the first derivative of something over time, which this isn’t.)
R0 is quoted a lot more than it’s understood. It’s an important parameter in estimating how quickly an infection will spread, but it is only an estimate and it can change a lot. Quarantines, for example, are meant to reduce R0 — if you have one person with a case of a disease and you limit the number of contacts to 0 until the infection is over, R0 is going to be 0. But R0 sounds real scientific and people can use it to scare others.
It’s also great for frightening news stories because the general population doesn’t have anything with which to compare.
Early reports put R0 for nCoV around 3.8, as the tweet above shows. The predictions of Germpocalypse were a little less than warranted, though — it turns out that even R0 = 3.8 puts nCoV in about the same neighborhood as flu, and lots less than measles, for which R0 is estimated as between 12 and 18. (I suspect it’s no coincidence that that tweet has since been deleted.)
(Measles has a pretty high mortality, too. Get your vaccinations.)
In any case, R0 really is an estimate, a statistical number that is inaccurate early on but gets more accurate as more cases are identified. Right now, one estimate from an NEJM paper puts R0 for nCoV around 2.2, which is more like other common colds.
Where do we stand now?
So this is what the science says: that nCoV is a problem, but the more information we get, the less it looks like Germpocalypse and the more it looks like, say, a bad flu.
Update:
Okay, I tried to link this in comments but it’s not going, so here’s the link to the Johns Hopkins Dashboard.