1\ A follower asked me how I can be skeptical of Chinese data, yet believe Coronavirus is no big deal.
Here's a thread summarizing why we know a lot more about Coronavirus than you think, and why almost all of it is good news
Here's a thread summarizing why we know a lot more about Coronavirus than you think, and why almost all of it is good news
2\ Before we knew anything about COVID-19, we already had good reason to suspect it's mild. That's because Coronaviruses are well known to humans, which means most of our immune systems will mount an immune response
Look, Coronaviruses made p. 11 of my 2009 virology text:
Look, Coronaviruses made p. 11 of my 2009 virology text:
3\ A disease that infects easily is its own worst enemy, because it likely confers resistance faster than it can kill
"Percolation models" are characterized by nonlinear propagation and dramatic "tipping points" when susceptible patients become too sparse for more transmission
"Percolation models" are characterized by nonlinear propagation and dramatic "tipping points" when susceptible patients become too sparse for more transmission
4\ Flu mortality is bi-modal, killing the very old and very young. As far as I know, CoV is not killing the very young. This may just mean it's a different bug, but it also strengthens the prior probability that we are missing lots of Cov "cases", which inflates CFR
5\ Similarly, known CoV cases in the US are skewing much higher than the average US age of 39
This, too, suggests that we are missing many mild and asymptomatic cases
Which means that CFR is lower than reported, and that conferred resistance is stronger
This, too, suggests that we are missing many mild and asymptomatic cases
Which means that CFR is lower than reported, and that conferred resistance is stronger
6\ Coronavirus has been outside China for at least two months, and probably longer. If it is so lethal and so transmissible, why has it killed so few outside of China?
This, in my opinion, is the strongest evidence that CoV is unlikely to approach the flu's ~600,000 death toll
This, in my opinion, is the strongest evidence that CoV is unlikely to approach the flu's ~600,000 death toll
7\ If CoV test specificity is similar to the 90-95% specificity of PCR flu tests, then with 40,000 CoV recoveries, we should be seeing hundreds if not thousands of remissions
Instead, we're seeing only a handful
This is nearly rock solid proof that CoV confers resistance
Instead, we're seeing only a handful
This is nearly rock solid proof that CoV confers resistance
8\ All CFR estimates are likely over estimated by 5-10x, because mild and asymptomatic cases are unlikely to be caught
Korea appears to have done the most systematic testing, and so far their CF is 0.6%
My guess is that is an upper bound on CFR
Korea appears to have done the most systematic testing, and so far their CF is 0.6%
My guess is that is an upper bound on CFR
9\ First hand reports of non-serious cases sound just like the flu
10\ Like the flu, CoV probably prefers low humidity environments of a certain temperature, with high population density and high smoking rates as a bonus
Many places in the world have a few of these, few have all
So it's unlikely CoV will thrive everywhere
Many places in the world have a few of these, few have all
So it's unlikely CoV will thrive everywhere
11\ Overall, I think the biggest things people overlook are a) the way acquired immunity acts as a non linear fire break to transmission, b) the amount of prior knowledge we have about likelihoods, before even glimpsing non-China data
12\12 To bring it full circle: given everything we know or suspect about CoV, and given two months of international data, I don't find the Chinese data all that suspicious
And if Chinese data is close to right, then Coronavirus lost the battle to immunity nearly a month ago:
And if Chinese data is close to right, then Coronavirus lost the battle to immunity nearly a month ago:
