Thread by @alexandrosM

I've criticized Kyle's group for not sharing the data they use to make sweeping statements. I'll demonstrate why this matters:🧵

Take this quote:"Every randomised trial that found a statistically significant benefit for survival was fake or did not actually happen as described."

https://twitter.com/K_Sheldrick/status/1445914957942444033

First of all, it's important to say that while the tweet is far more carefully worded, the quote offered to the BBC, and not corrected when the article is shared, is far less careful, and basically false:

That aside, what's actually being said, is that if the bar is set to:

- Randomized Clinical Trial
- Endpoint is survival/death
- Result reaches p<0.05

All the trials with the above characteristics are "fake or did not actually happen as described".

Who has made such a claim? That would be:

Niaee (Iran)
Hashim (Iraq)
Elgazzar (Egypt)
Okumuş (Turkey)

Notice a pattern? Alongside the other retracted study from Lebanon (Samaha), all the studies are from the middle east. Another I've heard whispers on is also in that set.

Which ivermectin RCTs have been conducted in the middle east? That would be:

Aref (Egypt)
Abd-Elsalam (Egypt)
Shouman (Egypt)
Elgazzar (Egypt)

Niaee (Iran)
Shahbaznejad (Iran)

Samaha (Lebanon)
Okumuş (Turkey)
Hashim (Iraq)

So, what if the claim was instead: "6 of 9 RCTs conducted in the Middle East were fake or did not happen as described"?

This is why sharing the dataset of studies investigated is so important. Science is about generating hypotheses and testing them to see their validity.

So, on the basis of the above, I hypothesized that there may be an issue with falsification of randomized controlled trials from the middle east. And a trivial googling does appear to lend credence to the notion.

If we look at the editorial from John Ioannidis "Hundreds of thousands of zombie randomised trials
circulate among us", a very familiar set of countries tend to appear (Turkey, Iran, Egypt). …-publications.onlinelibrary.wiley.com/doi/epdf/10.11…

In the BMJ opinion piece that set the 20% baseline for trials being fraudulent, the same three names come up over and over again, linking to the work of John Carlisle as well as that of Ioannidis.
blogs.bmj.com/bmj/2021/07/05…

The research on the matter is broad and deep, and having lived half my life in the East Mediterranean (though things in Greece are not nearly this bad) I can intuit some of the relevant cultural underpinnings. [1,2]

[1] scholar.google.com/scholar?hl=en&…

[2]:researchgate.net/profile/Meirav…

So when looking at a pattern of dishonesty, is it the medicine or the region? Correlation does not equal causation, so I don't think the group have a basis to make either claim, but they seem to be picking the weaker claim to make over the stronger one. This requires explanation.

What kind of claims are being made? I'll post a few here, but this is far from a complete set.

James Heathers:

https://twitter.com/jamesheathers/status/1445961513106173955?s=20

Kyle Sheldrick:

https://twitter.com/K_Sheldrick/status/1445961035886706689?s=20

Gideon Meyerowitz-Katz:

https://twitter.com/GidMK/status/1440797094852239363?s=20

Nick Brown (with a political twist!):

https://twitter.com/sTeamTraen/status/1371907698158080001?s=20

It really should not have to get to this for me to demand that the researchers publish their full list of investigated papers and their findings on each, before making pronouncements that implicate an entire field of research. To keep doing so is unethical and unjustified.

As a final thought, what we have here is a case of the what @slatestarcodex calls an "isolated demand for rigor". By demanding far more rigor of one subset of research than others, and interpreting findings in specific ways, falsehood can be built on truth.slatestarcodex.com/2014/08/14/bew…

If you got here without having seen my analysis of the claims being made in the original BBC article, please see this thread for more:

https://twitter.com/alexandrosM/status/1446271495454429215?s=20

More resources roll in (thanks @JoomiKim1) al-fanarmedia.org/2019/07/false-…

Addendum:

Kyle questions my numbers on a few of the studies. First, thanks for the quick and factual response. I'll work through the implications of his response by extending this thread:

https://twitter.com/K_Sheldrick/status/1446829525954793480?t=iikAaZG1V5CXgA5h91HtqA&s=19

1. His response is to say that two of the studies are not implicated. OK. That's good news.

I sourced my data from ivmmeta.com, and I would not have to, if they had been public about their top-secret list of 26 studies.

2. This still means "all implicated RCTs are from the middle east" so my point in this thread stands firm.

3. But Kyle didn't stop there. He had imply a racist motive. And to say that to someone whose mother's maiden name ends in "-oglu" is to show, once again, that people without adequate data should pause before making accusations.

https://twitter.com/K_Sheldrick/status/1446829944550526987?t=iikAaZG1V5CXgA5h91HtqA&s=19

I'll end this addendum with the end of my response to Kyle:

https://twitter.com/alexandrosM/status/1446832811852242955?t=iikAaZG1V5CXgA5h91HtqA&s=19

More extreme claims by @erictopol this time. @gidmk @k_sheldrick @steamtraen @JackMLawrence @jamesheathers why do you sit idly by when your work is being used to support unscientific claims? Call out bad science or be shown to endorse it.

I'll note that Eric Topol has me blocked ever since I wrote the thread on his podcast with @SamHarrisOrg showing him spouting falsehood, defamation, and disinformation galore.

(warning, it's *long*)

https://twitter.com/alexandrosM/status/1418820817161363456?s=20

Kyle attempts another response that addresses approximately zero of the concerns raised here. It does accuse me of racism 3 times, though, which is great.

https://twitter.com/alexandrosM/status/1447179226180771849?t=ZTotWkB3LKiXuU6fi8jDcQ&s=19

The pressure this group applies is highly asymmetric: (1) insist on the highest standards on trials, (2) recommend making meta-analyses far harder to do, while (3) making sweeping public statements implicating most scientists studying ivm, with no published analysis of their own.

As a result, they abandon the scientific method in favor of drip-feeding results and unsupported statements to a press that is hungry for material that confirms their biases.

They fight against actual grassroots science and replace it with narrative crafting and media savvy.

Ultimately, their actions boil down to deferring to a bias for narratives backed by raw power. The very thing science is supposed to save us from.

John Campbell's response makes very similar points to mine, and seems to be getting on the nerves of the fraudbusters crew. "who watches the watchers?"

https://twitter.com/nashville_brook/status/1447207745715834898?t=2HAeQsnFJsuSHBq5OLzSEA&s=19