Thread by @iowahawkblog

For all the people in my TL who spent a week screaming HYDROXYCHLOROQUINE WILL MAKE YOU DIE, icymi
sciencemag.org/news/2020/06/m…

The Cliff Notes: there is a little known medical data analytics company named Surgisphere in Chicago that purportedly gathered 100k cases of COVID patients treated with HQC.

/1

Surgisphere is the sole source of the HQC study that appeared in the blue ribbon, peer reviewed medical journal The Lancet. They also provided data another COVID study that appeared in New England Journal of Medicine.

/2

Surgisphere claimed to have data sharing agreements with hundreds of hospitals around the US and the world, despite having, it appears, only 5 employees, only one of whom has any medical background.

/3

Two of the other employees, according to the Guardian, include a science fiction writer and a male model for hire.

/4

How this ragtag Fab Five let's-put-on-an-international-medical-data-gathering-show was able to sign up data sharing contracts with hundreds of hospitals around the world is a bit of a mystery, but I'd be happy to hear the explanation.

/5

Apparently, the peer reviewers at The Lancet and NEJM didn't share my curiosity about this, before putting their imprimatur on studies that relied solely on Surgisphere data. You know, that one that proved HQC will make you DIE

/6

But don't worry, the Lancet and NEJM are out there busily issuing concerns and checking things and whatnot, now that their peer reviewers have been peer reviewed.

/7

Does HQC kill COVID patients? Hell if I know. Does it make COVID patients better? Hell if I know. Will I trust The Lancet or NEJM when they finally issue conclusions on these questions? Hell if I know.

/8

I'm not claiming HQC is good for COVID-19 patients, I have no idea if that's true. And I'm not necessarily criticizing news outlets for running with the Lancet study, because, I mean, it's the freaking Lancet fer crissakes.

/9

It's just remarkable how a fly-by-night sketchy data shop that would never pass muster on the simplest due diligence test was accepted as a valid data sources all the way through the trust chain from researchers, to top scientific journals, to mainstream media.

/10

The end information product of it all was so tasty and delicious, though, why check the ingredients?

And it comes on the heels of Theranos and the irreproducibility crisis in psychology.

Get your shit together, science.

/11

Please note that this was not *a* COVID/HCQ study, it was *THE* COVID/HCQ study, the Big Kahuna, with 96k patients in 600+ hospitals on 6 continents, with shit-tons of control variables, that would determine treatment guidelines in the midst of a global pandemic.

/12

And as far as I can tell, nobody in the entire public health information supply chain once thought to ask, "is this data, you know, real?"

/13

Everybody is focused on the HQC thing in The Lancet, due to Trump, but the same fucking questionable data set was used to study the impact of ACE inhibitors (blood pressure meds) on COVID patients in a published NEJM article. Previous research suggested they were harmful.

/14

The study published in NEJM concluded ACE inhibitors had no impact on death rates. Whew! Don't worry Doc, keep those COVID patients on their BP meds.

What if that conclusion, in contradiction to previous research, was based on a potentially fabricated data set?

/15

There are no heroes in this, save for some skeptical statisticians and amateur sleuths, and the Guardian who to their credit actually digged into Surgisphere.

In a sane world this would lead the network news and be a Page 1 banner at the NYT. I'm not holding my breath.

/fin