If you work backwards from the number of deaths, you get some troublesome numbers
Testing and confirmed numbers all over the world are basically only indicative of how much you're testing. Even among nations there are testing at high rates, there are numbers that don't add up.
New York State and the UK have done about the same amount of tests and yet the UK has found one-third as many cases. So there must be fewer overall cases in the UK, right?
Yet the UK has 759 deaths compared to 519 in the US. The UK's median age is 40.5 compared to 38.1 in the US but it's healthier on some other metrics so I don't think that's a significant factor. Is the healthcare system that much more effective in the US? Possibly but by that much?
It's the same thing with a dozen other countries. There are quirks everywhere with the testing population and the quality of the tests themselves.
Even deaths themselves aren't iron clad. The first US confirmed death was recorded more than a week after the person died when they went back and tested. Surely many people who have died of 'pneumonia' in many countries had the virus but I still think that's the best number we've got.
So if you work back from deaths and you make a few assumptions here's what you get.
Assumptions:
- 0.8%-1.0% mortality
- 5 days from time of exposure to first symptoms
- 25 days on average from time of exposure to death
- Cases double every 3 days
Now take the US where there are at 1,358 as of Friday. If that's 1% of the people who had the virus 25 days ago (or longer) then 25 days ago then 135,800 people had it on March 2, including those who had contracted it but didn't have symptoms.
If you assume a doubling every three days, that gives you 43.8 million cases right now with about 33 million not yet showing any symptoms. Given how many celebrities, executives and politicians have tested positive already, that's entirely believable.
That's too high because the US started some social distancing measures on March 16 and got more aggressive on the 23. At best that gives you aggressive growth for another 4 days but 10 days is more realistic given the Italian experience.
As for places like Italy and Spain with unusually high mortality rates, you have to add adjustments for age and things like smoking but smoking levels are also high in China and the Wuhan mortality rate (at least officially) was 1.4%.
Take Spain, it has 4858 deaths as of Friday, which would mean 485,800 cases a month ago with three-quarters not showing any symptoms. That would have been March 2. There was no national lockdown until March 14. Some simple math from there and you have to assume everyone in the country is exposed to it, eventually.
