There are clearly 2 "camps" now. One thinks the IFR is very low and a lot more than we know are infected.
Michael Mina on Twitter
Could the US have >2 million
#COVID19 cases by now?
I'll be astounded if not! We simply don't test.
The ratio of who gets a test vs who should, likely worse than 1 in 10.
So given 220K cases reported ->
>2.2 million COVID cases in US already?
>4 million? Perhaps...
Last note - and *speculative* – given very limited testing and relatively short windows of time to capture virus in a nasal swab, I really won't be surprised if even 50x more people have acquired the virus than cases confirmed If so ~11 million in US could have acquired
#COVID19
And those who think IFR is actually > 1% and not that many are infected. I'm using the following reply to the above tweet by a Harvard epidemiologist because it was retweeted by Marc Lipsitch. I think this is the mainstream epidemiologist view.
Pierre Andurand on Twitter
We have way enough data to know that it is not 0.1%, but more than 1%... already 0.024% of both the Spanish population and Italian population died from it officially, with reports that death rates are actually 3-4 times higher in many places.
Regions in Italy and Spain have more than 1% of their population that died from it. SKorea with extensive testing and contact tracing have a 1.7% CFR. It obviously cannot be a 0.1% IFR. Way enough evidence now
That's a good articulation of the two camps of thought on this issue, but I think it may be in the end a seductive but perhaps ultimately specious achievement to chase down a single number as though that number means a great deal. And I'll bet you dollars to doughnuts that that final case fatality rate number varies enormously across societies even when you remove the possibility of cooked data from motivated underreporting, along with all the well-documented uncertainties in determining case fatality numbers at the early end of a pandemic. And I'll tell you why any single number for CFR or IFR is misleading – it may be that there is simply
enormous differential vulnerability and what your final case fatality number ends up being in a given population depends on the many and disparate risk variables expressed in that population.
Even at the very incomplete state of our science on this question, I can put many highly plausible and several proven relevant risk variables on the table right now:
1) Demographics of age In your population, and more specifically, the state of the adaptive immune system. See previous posts
here and
here on this and the relevant concept of 'inflammaging' as a potent mortality risk variable. In aging the innate immune system is disinhibited to partially compensate for declining adaptive immunity.
2) So-called age-related comorbid illnesses which as I have already argued are proxies for or at least express some of the degree of #1 penetrance. Classically of course type II diabetes, coronary artery disease, COPD, recent cancers, and I suspect Alzheimer's disease will prove to be huge although that's not been proven, and there is no data on that.
3) Degree of moderate or worse smoking (or perhaps vaping?) in your demographic.
4) Degree of antibiotic resistance in your demographic – and this, #1 and #3 may explain why Italy has been devastated, As Italians reportedly have relatively high levels of antibiotic resistance presumably from antibiotic overprescription. This may of course index also possible problems with the Demographic microbiome which we now appreciate as a front-line in the immune system. One would think that the Italian diet would mitigate this but perhaps not.
5) Other protective lifestyle variables particularly degree of regular aerobic exercise and presence or absence of sleep disorders as both of these affect immunocompetence directly and potently.
6) Degree of social support versus social stress – people are surprised to hear that immunocompetence is modulated by this but it is. Chronic upregulation of the stress axis subjects us to a daily bath of corticosteroids and upregulation of pro-inflammatory cytokines. See previous discussion on dangers of upregulated innate immunity when facing novel pathogens.
7) The role of gender with significant increased vulnerability in males to bad outcome in respiratory infection.
8) Likely to be dozens of unmapped polymorphisms affecting vulnerability, including ACE2 Polymorphisms (Receptor by which Covid 19 Virus gains entry In the cells).
9) degree to which there is partial recognition by adaptive immune system of this coronavirus by virtue of prior exposure to one of its relatives. This is totally unmapped but it perhaps in combination with other risk reducing issues may explain why some people are completely asymptomatic.
What is clear is that a subset of the population is extremely vulnerable. And as I indicated before, whether that ultimate case lethality is under 1%, if we bust our healthcare capacity, our case lethality becomes the percentage of folks that need ICU level care that can't be cared for in ICUs on respiratory support. And that's going to be way more than 1%. Possibly 5%. And that's a terrifying number.
