Thread by @AtomsksSanakan

1/C

The infection fatality rate (IFR) states what proportion of people infected with the virus SARS-CoV-2 die of the disease COVID-19.

Many COVID-19 contrarians abuse IFR estimates from the USA's Centers for Disease Control and Prevention.

So let's discuss those estimates.

2/C

For context:

Seasonal flu's IFR is <0.1%:

https://twitter.com/AtomsksSanakan/status/1299429571058696194

If 100 million people (less than 1/3 of the USA's pop.) get infected at an average IFR of 1%, that means 1 million COVID-19 deaths.

medrxiv.org/content/10.110…

https://twitter.com/GidMK/status/1285020775892709377

medrxiv.org/content/10.110…

3/C

Round 1:

The CDC initially gave an IFR of 0.26%.

They gave no evidence for this, and their estimate was below that of published research.

So experts were annoyed:

https://twitter.com/CT_Bergstrom/status/1263608976320507904

https://twitter.com/CT_Bergstrom/status/1279612720489234434

usatoday.com/story/news/fac…

npr.org/sections/healt…

4/C

Round 2:

The CDC updates their IFR from 0.26% to 0.65%.

That makes a bit more sense, but is still an under-estimate, as noted by the source the CDC relied on.
The actual value was closer to 1%:

https://twitter.com/GidMK/status/1285020775892709377

https://twitter.com/AtomsksSanakan/status/1286771217274482689

archive.is/w2xC7#selectio…

5/C

Round 3:

The CDC now gives IFRs for specific age-ranges.

This implies a higher IFR of ~0.72%:

https://twitter.com/zorinaq/status/1309741852661174278

But the CDC again under-estimates IFR, since they leave out those aged 80 and above:

https://twitter.com/AtomsksSanakan/status/1310242459235090434

archive.is/UWGms#selectio…

6/C

Some take-home lessons so far, on the CDC and IFR:

1) They persistently under-estimate IFR.
2) Their IFR estimate increased, undermining COVID-19 contrarians who cited the CDC.
3) COVID-19 is dangerous, even on the CDC's under-estimated IFR.

sciencedirect.com/science/articl…

7/C

And for those who want IFRs for regions of the USA to compare to the CDC's (under-estimated) implied value of ~0.72%:

https://twitter.com/AtomsksSanakan/status/1297654625043132416

USA values circled in red:
medrxiv.org/content/10.110…

8/C

The CDC's age-specific IFR from part 5/C is similar to the age-specific IFR used in March by Neil Ferguson's Imperial College team.

So in citing the CDC, COVID-19 contrarians implicitly affirm Ferguson was right. 🙂

https://twitter.com/AtomsksSanakan/status/1310310480074899456

spiral.imperial.ac.uk:8443/bitstream/1004…