[javagram]

> For example, COVID's current "death rate" estimate is ~3x worse than flu. Versus Trump's "~5x" in your quote from Feb 2020¹.

COVID-19 Infection Fatality Rate is estimated at around 0.6% whereas H1N1/09 Flu infection fatality rate was estimated to be less than 0.03%. Not sure where you get “3x”, I think the correct number would be 20x based on these numbers.

* Covid-19 IFR source: https://www.cdc.gov/library/covid19/112420_covidupdate.html where a chart indicates an IFR between 0.5 and 0.75% for USA and England

* H1N1/09 IFR source: https://www.bmj.com/content/339/bmj.b5213.full “We estimate that 0.026% (range 0.011-0.066%) of these individuals died from causes related to this infection”

[Radim]

I compared Covid's estimated IFR of < 0.23% to flu's < 0.07% [1].

But that's exactly the discussion I called out as not terribly relevant: The variance within Covid (due to age, location, comorbidities…) is much greater than variance between Covid and flu.

In other words, even a Covid-vs-flu rate difference of 20x (as you say) is dwarfed by the 1000x difference between Covid age groups (which you didn't contest). Can we agree on that?

My point is that a broadly aggregated statistics like "global IFR" is too crude to be actionable. Easy to put in a headline, sure, but more potential for confusion than good.

[1] https://www.who.int/bulletin/online_first/BLT.20.265892.pdf

[javagram]

Disagree strongly. An IFR of 0.6% when considered across the entire population is indeed cause for much more alarm than one of 0.02%. It is very relevant.

0.6% of the population in the USA, for instance, would mean about 2 million people would be killed by uncontrolled and massive COVID-19 spread.

Focusing down to small subgroups would only be relevant if you have a magic wand and could, for instance, seal off all over-60 year olds from human society for a year.

P.S. the paper you cited is by John Ioannidis. He has become notorious in 2020 attempting to prove Covid-19 isn’t very dangerous. Worth consideration as this version has managed to pass peer review, but keep in mind it’s outside the mainstream of opinion. IMO, the 0.2% estimate is pretty clearly low and I’ve read a good debunking of that specific paper in the past. https://threadreaderapp.com/thread/1316511734115385344.html

[Radim]

I'll put it bluntly: fixating on an average ("average IFR") of a wildly heavy-tailed distribution (e.g. exponential for IFR-vs-age) is idiotic.

Technically yes, such average exists – the population is finite. But taking a population-wide decision based on such estimate is suboptimal. We already know a population parameter (age, comorbidities) that gets us an actionable segregation!

I personally see such "hiding behind an average of an exponential" as scientific fraud. Misinformed at best; disingenuous and murderous at worst (such as with Covid).

> He has become notorious in 2020

Interesting, thanks. I wasn't aware of John Ioannidis' pedigree. For those curious – this article does a good job summarizing the controversy (April 2020):

https://undark.org/2020/04/24/john-ioannidis-covid-19-death-...

> Focusing down to small subgroups would only be relevant if you have a magic wand and could, for instance, seal off all over-60 year olds from human society for a year.

A magic wand to seal off over-60 olds? How do you feel about sealing off everyone?

[javagram]

> A magic wand to seal off over-60 olds? How do you feel about sealing off everyone?

Worked great in New Zealand and Taiwan, and pretty dang good in Singapore, Australia, and even China. A pity we weren’t fast enough to do it here.

I haven’t seen any society successfully execute a strategy to exclude the aged from human society or viral transmission yet, and we’ve had 9 months to experiment. If such a strategy existed and was proven to work, your theory that it was idiotic to ignore it would make sense. Unfortunately, it doesn’t seem to exist.