Thread by @AtomsksSanakan

1/U

Sometimes John Ioannidis just makes me laugh. 😀

In the slide below, Ioannidis discusses age-specific IFR (infection fatality rate), i.e. what proportion of SARS-CoV-2-infected people die of the disease COVID-19 at various ages.

22:23 - 23:18:

2/U

Here's the work Ioannidis is likely referring to:

Axfors (Ioannidis is the lead author)
ncbi.nlm.nih.gov/pmc/articles/P…

Levin (from @GidMK + @BillHanage)
link.springer.com/article/10.100…

O'Driscoll
nature.com/articles/s4158…

CDC
web.archive.org/web/2021032419…

https://twitter.com/AtomsksSanakan/status/1381074021580685315

3/U

Ioannidis says his Axfors estimates mostly agree with O'Driscoll:

from 23:04


Yet experts noted for around year that his Axfors estimate is a low outlier.

So what's going on here?

publichealthontario.ca/-/media/docume…

4/U

O'Driscoll et al. leave out nursing home deaths, while Levin + CDC include them.

Including nursing home deaths increases IFR in older groups, which is one of the main reasons why CDC + Levin are higher than O'Driscoll.

https://twitter.com/AtomsksSanakan/status/1367303395976097795

https://twitter.com/GidMK/status/1323379424784392192

5/U

So Ioannidis should compare O'Driscoll (purple) to his *blue* Axfors results that *exclude* nursing home deaths.

But he compares purple to his *red* Axfors results that *include* nursing home deaths.

Thus, he conceals his work being an outlier. 🤦‍♂️

6/U

John Ioannidis either:
- misremembered
- didn't read O'Driscoll et. closely
- willfully misled his audience

In any event, @GidMK was again more forthcoming than Ioannidis. Makes it even more ridiculous that Ioannidis smeared @GidMK multiple times.

https://twitter.com/GidMK/status/1376304539897237508

7/U

And, of course, later research supported @GidMK's and @hanage's Levin / CDC results that Ioannidis objected to.

That isn't just my interpretation; that's what the authors of the research state:

https://twitter.com/AtomsksSanakan/status/1405623744773640193

mdpi.com/1999-4915/13/6…

8/U

This is not the first time Ioannidis has criticized that paper by distorting other research. It's just the first time I've seen him be so blatant + brazen in his distortions.

https://twitter.com/AtomsksSanakan/status/1376924754494566408

https://twitter.com/AtomsksSanakan/status/1375935382139834373

onlinelibrary.wiley.com/doi/10.1111/ec…

9/U

Ioannidis also says that the papers agree IFR was "very low" until about age 50.

Well, "low" is relative. Compared to other causes of death, it wasn't low. The Levin paper was pretty explicit on that, as were others.

sciencedirect.com/science/articl…

Levin:
link.springer.com/article/10.100…

10/U

The CDC previously cited another age-specific IFR estimate from @C_Althaus, that largely matched Levin + was larger than Ioannidis' estimate.

Ioannidis left it out, of course

web.archive.org/web/2020091122…

https://twitter.com/AtomsksSanakan/status/1304640560385359874

https://twitter.com/C_Althaus/status/1288190016288686083

journals.plos.org/plosmedicine/a…

11/U

Ioannidis' image also left out the results from Neil Ferguson's team at Imperial College, which again matched Levin.

When he *conveniently* leaves stuff out, it helps conceal the fact that his results are a low outlier.

imperial.ac.uk/mrc-global-inf…
[imperial.ac.uk/media/imperial…]

12/U

Looks like criticism of the video's content is being blocked from the video's comments section.
Oh well. 🤷‍♂️

https://twitter.com/AtomsksSanakan/status/1374617361194565634

- Ghisolfi: gh.bmj.com/content/5/9/e0…
- WHO documents:
apps.who.int/iris/handle/10…
who.int/publications/i…

13/U

Re: "3/U
Ioannidis says his Axfors estimates mostly agree with O'Driscoll"

Axfors update is out.

Ioannidis knew O'Driscoll showed a higher IFR than Axfors results. So he was just being misleading when he said otherwise in the video, as usual.

medrxiv.org/content/10.110…

14/U

The Axfors update even adds Imperial College's analysis (ICCRT; see part 11/U), again illustrating how much Ioannidis' Axfors results are a low outlier.

Remember, he did a misleading comparison of orange vs. red, when it should be orange vs. blue.

medrxiv.org/content/10.110…

15/U

19:48 - 20:16 :

"In Africa, [COVID-19 deaths were] tremendously undercounted. In India, not as much as it is feared. If you put the numbers in the model, it suggests that there's a little bit of an undercounting. But not tremendous undercounting"

16/U

Several studies present observational evidence of tremendous under-counting of COVID-19 deaths in India, contradicting Ioannidis' model-based claim.

https://twitter.com/AtomsksSanakan/status/1418039216580304896

medrxiv.org/content/10.110…
archive.is/JXUtp#selectio…
github.com/akarlinsky/wor…

medrxiv.org/content/10.110…

17/U

Researchers keep getting age-specific infection fatality rates larger than Ioannidis' work. Amazing how he can still claim his work isn't a low outlier.

https://twitter.com/dherrera1911/status/1421792190456553472

medrxiv.org/content/10.110…

18/U

Ioannidis continues claiming he was right all along + that COVID-19 is less dangerous than it actually is.

Of course, the usual COVID-19 minimizers like it.

@JHowardBrainMD
sciencebasedmedicine.org/what-exactly-d…

https://twitter.com/Potholer54T/status/1322460853409644544

https://twitter.com/VPrasadMDMPH/status/1422305032381231136

https://twitter.com/AtomsksSanakan/status/1418977661402161161

19/U

Re: "Ioannidis continues claiming he was right all along + that COVID-19 is less dangerous than it actually is.
Of course, the usual COVID-19 minimizers like it."

Thread on it:

https://twitter.com/AtomsksSanakan/status/1422403327019692060