Numbers, numbers
It follows, of course, that by the time you read this, you’ll know more than I or anybody else could have hoped to know a couple of weeks ago. As things stand – or rather stood – the spread of the disease in China had stabilised, if not slowed, by mid-February, with 44,000 confirmed cases to date and 1,100 deaths. Although the People’s Republic had effectively quarantined more than 100 million people in their own cities, it was inevitable that some leakage abroad would follow. The first cases of coronavirus were starting to turn up in other countries, including Britain (eight cases in mid-February), and 176 cases (check?) on a cruise liner that was lying in quarantine off the coast of Japan (check).
One of the key problems that health authorities currently face is that nobody knows very much about the epidemiology of this brand new virus. For instance, what is the incubation period? Three days, 14 days? At what point could a recovered patient be considered safe for discharge? And (shudder) could he have been infectious before any of his symptoms emerged?
That, of course, would be a virologist’s nightmare – they’d much rather that people keeled over visibly within 24 hours, because that would make it so much easier to decide who should be on the plane out of China, and who should be in an isolation ward.
The panicky version of the story is that, if every infected person meets 40 other people in the course of a normal day and infects 2% of them every day for fourteen straight days, and if they all then go on to do the same, then that’ll set up a compound growth rate that could stop the country in a matter of months. But can it really be correct to talk about a 23% mortality rate for the coronavirus, as some people do?
Only if you ignore the logic of mathematics……
Consider, for example, the relatively few cases that get to an isolation hospital. Which, in China’s case, was about 10% of all coronavirus diagnoses. Data published on 7th February showed that, of 34,546 confirmed coronavirus cases to date, 6,101 had progressed to “severe”, and that 722 (23% of those) had died while 2,050 (66%) had been declared fit for discharge. And the rest, we presume, were still in care?
Sooooooo, 722 deaths out of 34,546 confirmed cases looks like a 2% danger rate, which is bad but not completely terrifying. The bad news is that those numbers are faulty, because we’re getting our historicity wrong. We really ought to be comparing today’s mortality numbers with the infection rates of two or three weeks ago, when hospital admissions were much lower and when the deceased actually caught their viruses. Time’s winged chariot has a way of skewing the data like that.
The better news is that there are probably twenty undetected cases of coronavirus in China for every case that becomes bad enough to be taken to a doctor. The overwhelming majority of sufferers get nothing more than a mild flu-like infection that clears in a couple of days. Yes, it’s still a worry that my minor sniffle might be enough to kill my elderly neighbour with an existing chest infection, but that’s epidemiology for you.
Equally to the point is the worry that the coronavirus might eventually migrate to less developed countries that wouldn’t be able to isolate all their severely ill in hospitals because they didn’t have the infrastructure to do it. Personally, I’d expect that this might put a question mark over financial markets in some East Asian regions. But for the time being, that’s just guesswork.
Continue reading article…
