Thread by @jeremyfaust

The time has come to defend science against its own practitioners. Thread!!!!!!!!!! Spoiler: Remdesivir may work, but we won't find that out from a "study" with no fricking control group and with patients who statistically are *not expected to die anyway*.

STAT News reports that TWO out of 113 patients with “severe” covid19 receiving Remdesivir died so far. Here's the headline

"Early peek at data on Gilead coronavirus drug suggests patients are responding to treatment"

statnews.com/2020/04/16/ear…

hmm...

Let's look at some assumptions, shall we? Assumption 1: These are SEVERE patients so they're the most sick patients possible. Right? Wrong. Look at how they define severe...

The study *excludes* those with multi-organ failure, abnormal liver tests, moderate kidney disease, needing mechanical ventilation at time of first evaluation, needing mech ventilation for 5+ days. Reality: *SEVERE cases not the sickest patients. *CRITICAL is the worst category.

Reminder: "critical" covid19 patients are the ones with all the bad things in the previous tweet...patients who had those things...this "study" won't accept. I suppose the rationale for that is that the critical patients are "too sick" for the drug to work. Noteworthy, eh?

Ok. Let's move on to....Assumption 2: SEVERE cases are...SEVERE. Right? They are very likely to die, right? So we should definitely be studying these ones. Right? Well... How many patients with SEVERE covid19 usually die. Wait for it. Waaaait for iiiiit...

Well, in China, ZERO patients with “SEVERE” covid19 died (Wu and McDoogan, JAMA)

That's zero with a capital Z

EVERY patient who died had “CRITICAL” covid19, not "SEVERE" (which starts to look not-so-severe, by comparison.) Suddenly "severe" aint so severe.

(Not saying I want severe covid19, thank very much. But I'll take it a million times over "critical." Cuz that's when a ton of patients start dying.

"Severe" is not too well defined. But it appears to mean "low oxygen" under 94% on a finger oxygen probe. Signs of fast breathing. In other words: genuine symptoms beyond fever and a cough (though now a fever isn't even required to meet "severe" definition).

Now...I think I hear you saying....wait a sec! In China (where they weren't all getting Remdesivir), ZERO severe cases died? And now in the US out of 113 patients with "severe" covid19 getting the drug, and TWO patients died. Uh oh... Does that mean....

Does that ACTUALLY mean that this Remdesivir drug is killing people? Should you sell your Gilead stock? No. Not necessarily. Which is my point. We don't know. This is not a controlled trial.

But the two ways you can spin the data proves my point. It shows why you have to do randomized controlled trials.

You know...like grownups!

What are we going to find when we randomize patients with severe (or critical) covid19 to get either Remdesivir or placebo? NO ONE KNOWS. I don't. You don't. The doctors "researching this" don't. Gilead doesn't. The president doesn't.

To reiterate...until we get REAL randomized controlled data (double blinded, ideally), you could spin information like this "leaked video" any way you'd like....

You could spin it as GREAT NEWS ("Only two people died so far out of 113!"). Or you can spin it to be TERRIBLE NEWS ("Holy shit, two people died, and probably ZERO should have!").

In conclusion, this study is not science. I beg of you, though: Let's insist upon science. Let's get GOOD data. REAL data. Controlled trial data. Then let's give the medications that are proven to work to the patients in whom they have been proven to work on, only.

It is not complicated. But it does take a modicum of discipline and a hint of scientific literacy. (Which is apparently under-taught in med school and under-emphasized on board recertification exams.

Doing things the scientific way...is the only thing that separates us from the Medieval alchemists selling potions to the gullible and illiterate.

Let me put it to you this way. This is Kansas, not Oz. Fantasy doesn't cut it. Lives are on the line, now and in the future.

So demand better from science from researchers and better reporting from the journalists covering them. For more analysis like this, check out @brief_19. We don't fall for b.s. And we have no paywall. Ever.