My Authors
Read all threads
A new thread on another #hydroxychloroquine study in a major medical journal from the same study group as posted in @TheLancet and for which the data has been shown to be impossible.

This time the @NEJM

Hold on to your hats
You need to have a bit of statistics background to understand this. It's all about the variance. When you have different groups they vary. The more numbers you have in each group, the less the variance (proportionally).

Different groups would never have the same variance
Except in this study!

Not only do all these age groups have the same variance in this chart, but what *should* happen is that very small groups would have big variance (confidence intervals on this chart).

Well, it appears there are a lot of hypertensive teens with #covid19 Image
Yeah, that's right. Either the 11-20 group had roughly equal numbers as all the other groups or they were a group of clones (because similarity would reduce the confidence intervals in a small group).

#Covid19 really doesn't affect this age group.…
(more to come later today, I'm just getting started)....
OK just to clarify, this paper isn't about #hydroxychloroquine *but* it is from the same group that published a paper in the Lancet, from the same data, that they showed (using debunked data) that #HCQ increased the death rate in #COVID19
Here's the link to the paper.

Despite being not that useful, on a smaller data set than the debunked #HCQ paper in the Lancet, it is available in full on the @NEJM website.…
So, let's get back to the paper itself. Yes, the paper that has equally distributed age groups including a whole bunch of hospitalised 11-20 year olds (who just don't get covid)

Apart from the magical @Surgisphere database, where is the data from?
Oh look!

Can you imagine? Study subjects from Chinese hospitals contemporaneously uploading huge datasets to a US company.

Really? Image
This is from the country that vets all data, images, conversations etc etc.

So @Surgisphere are trying to tell us that the #CCP just allow all this data to flow to their servers in the US, in real time from 7 hospitals...…
Yeah, OK. Let's ignore the huge #China link for a second and get back to the data.

I'll just say it's weird that 121 US hospitals provide 1499 patients (12 pt/hosp) but 7 Chinese hospitals provide 1507 patients (215 pt/hosp) - for a supposedly US company....

Hmmmm... Anyway.. Image
OK let's get back to propensity score matching (PSM). What is that? It's where you have a prior dataset and want to try to compare similar patients so you can see the effect of the intervention or study factor.
In this case they would want to take patients with similar age, BMI, race, gender, comorbities, medications and compare those that took the drugs of interest (ACE inihibitors and statins) against similar patients that didn't.

It's a reasonable statistical tool...
...but PSM requires (usually) a big reduction in your data set because you have to match similar patients on many factors. The more factors, the smaller the suitable patient pool.

If you don't (PSM) match, the groups will be very different (by nature)
This study did *not* perform any matching. The groups were just the people they found in the database from different countries.

Except, they look like perfectly matched groups.

I mean *perfectly* matched. Image
So, without propensity matching the authors want you to believe that the variance of age, diabetes, hypertension etc in their groups across three different continents was the same, with a mean of 50 (when most covid patients and hypertensive patients are much older).
Nah, that's BS. It's also BS when you use the same (unmatched) data set and sort by a different factor (high income vs low income countries) and still maintain perfect matching...

Nope. This is not possible. Image
So, what on earth is going on?

This is my (educated) guess. #covid19 was a potential death knell for #bigpharma. #HCQ is a treatment and is out of patent, and costs pennies.

It's so important that the Shermans were killed to stop them producing it cheaply
Oh, you mean you didn't know the chairman of @Apotex was murdered?…

Hey, that's fine nothing to see here, anyway moving on...
The point is there is a *huge* industry in drugs that you are told to take to prevent future health issues.

The big ones are antihypertensitves (blood pressure meds) and anti-lipid agents (statins).

Neither has been shown to prevent deaths significantly, but keep taking them
But when #covid19 came out there was a lot of concern about these drugs - especially ACE inhibitors (because the virus uses the ACE2 receptor to gain entry).

So, how convenient that a large study shows.... (guess what)...
That the risk of dying from #covid19 is *lower* if you're taking ACE inhibitors (not those other cheap blood pressure meds of course) and statins.

And the risk reduction is more pronounced than any other factor (even age) - imagine that! Image
Bear in mind that the same data showed no difference in the risk of mortality for hypertension - exactly the indication for ACE inhibitors.

Strange eh?

Of course there are more strange things about this data. Like the 7 UK hospitals involved. Image
You see, just like the inclusion of Australian data in the debunked @TheLancet study, including UK hospitals is a *real* problem.

If you think data sharing in the US or Australia is a problem, try it in the UK.


The #NHS has their e-data locked down tight.
Anyway, that's enough for now. (I have to keep some in reserve for when @Surgisphere find a "mistake")

Hope you enjoyed the thread.

Hope that @NEJM didn't, because they need to retract the paper.
However you spin this #Hydroxychloroquine story one thing that @Surgisphere cannot deny.

If you are treated in a hospital with Surgisphere's wonderful IT system your chance of dying is 5-10x higher than in @raoult_didier's hospital.

Moral - don't go to a Surgisphere hospital ImageImageImageImage
Missing some Tweet in this thread? You can try to force a refresh.

Enjoying this thread?

Keep Current with arkancide_is_real

Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

Twitter may remove this content at anytime, convert it as a PDF, save and print for later use!

Try unrolling a thread yourself!

how to unroll video

1) Follow Thread Reader App on Twitter so you can easily mention us!

2) Go to a Twitter thread (series of Tweets by the same owner) and mention us with a keyword "unroll" @threadreaderapp unroll

You can practice here first or read more on our help page!

Follow Us on Twitter!

Did Thread Reader help you today?

Support us! We are indie developers!

This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3.00/month or $30.00/year) and get exclusive features!

Become Premium

Too expensive? Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal Become our Patreon

Thank you for your support!