Re:Coronavirus: A third of hospital patients develop dangero...
by kaoD on Saturday May 16, 2020 @03:05PM
(#23206229)
> They’re in the bar graph, expressed as rates per 100k.
AFAICT that's per 100k total population, not per 100k infected (which amounts to 0.2% total FR citywide, not even IFR).
Or am I missing the graph you mean?
> The values cited by the OP (>>1.5%) are far too high, and based on severity-biased samples were gathered early in the pandemic.
I am OP. They're based on an antibody study currently ongoing in my country so... no, they're not far too high nor severity-biased. (I'm trying to find links in English, but no luck so far).
The link I provided has IFR, and is where I got the NYC numbers I quoted. They’re in the bar graph, expressed as rates per 100k.
The best reference I know of for a global, age-adjusted IFR as of early April is here (their estimate is 0.66%):
https://www.thelancet.com/pdfs/journals/laninf/PIIS1473-3099...
However, more recent papers have come out, suggesting much lower values. For example, this paper estimates a a value of 0.08% for patients under 70, using antibody data in blood donors in Denmark:
https://www.medrxiv.org/content/10.1101/2020.04.24.20075291v...
My stated estimate is conservative. If anything we are continually reducing the number as we gather more data on prevalence. Regardless, a range of 0.5-1% is a safe bound for the highest quality estimates of IFR. The values cited by the OP (>>1.5%) are far too high, and based on severity-biased samples were gathered early in the pandemic.