Fake news kills during a pandemic but wild 5G conspiracies may be less dangerous than the lowering of standards in mainstream science. Friday saw the most egregious example of this so far -
"Gilead drug shows positive signs in early testing" from @FT
ft.com/content/c59a38…
"Gilead drug shows positive signs in early testing" from @FT
ft.com/content/c59a38…
The Washington post used the word "hopeful" in their headline.
washingtonpost.com/business/2020/…
washingtonpost.com/business/2020/…
The headlines are about this paper on Remdesivir, an antiviral developed by pharma company Gilead for Ebola and similar infections published in the prestigious @NEJM nejm.org/doi/full/10.10…
The paper (NOT a trial or even a study) has many flaws - If you’re a parent trying to homeschool science then get your kids to critique the paper. It’s very empowering for them to discover they can spot flawed methods in top journals. Ask them whether it’s an “experiment”.
The article describes what happened to 53 COVID patients who were given Remdesivir for “compassionate use”.
36 out of 53 patients needed less oxygen
25 out of 53 were discharged from hospital.
7 out of 53 patients died.
So is this good or bad?
36 out of 53 patients needed less oxygen
25 out of 53 were discharged from hospital.
7 out of 53 patients died.
So is this good or bad?
One of several pithy critiques from @HTropDis colleague @dr_michaelmarks is that while there are only 53 patients in this paper there are 56 authors (although the word "author" may also be a stretch - more on that later). He's written more here
More authors than patients is never a good sign but many other expert clinicians (inc @dr_michaelmarks & @DrNeilStone) have pointed out deeper issues. This excellent blog by @PulmCrit summarises the most important points.
emcrit.org/pulmcrit/pulmc…
emcrit.org/pulmcrit/pulmc…
First, there’s no placebo group. So all that has happened is that a drug (new, possibly toxic) has been given to patients: some got better some died. We know that this happens in COVID infection anyway - this study doesn't help us to know if the drug changed outcomes.
There’s a reason that when we trial new drugs we have to randomly assign people to get the drug or not. It means you know that if the group taking the drug does better (or worse) it’s due to the drug. This is the basis of all scientific experiments - we learn this as children.
Sometimes we do studies without controls but only when necessary. It would have been very possible to do a good trial here. Loads of patients, masses of resources. Being in a rush to get information makes it MORE not less important to do good research.
The drug here is given "unblinded". Ideally trials are double blind so that neither clinician nor patient knows who is getting what. This is really important - doctors and patients behave differently if they think that they’re giving or getting a placebo.
It's not clear how patients were selected here - it's obviously important not to cherry pick patients and give the drug to those who seem more likely to recover. In this case the company who makes the drug picked the patients.
And there are a lot of patients included who weren’t very ill. To be considered for the drug you needed to have an oxygen level of 94% or less on air. We send some patients home with oxygen levels of 94% so this is hardly being given to the sickest patients.
Patients were recruited from 9 countries. There are only 50 patients. This is a bizarre way to do science - imagine all the things that vary from hospital to hospital let alone country to country.
Around a quarter of patients had really serious problems - organ failure, sepsis - because this is such a poor quality study it’s impossibly to know if these are due to COVID (possible) or Remdesivir (also possible).
The hand of Gilead is heavy here - we know that when studies (and remember this isn’t a study it’s just words and numbers) are funded by industry (even when they are well designed) find in favour of the drugs more often than independent studies. cochrane.org/MR000033/METHO…
Gilead picked the patients and did the stats
“The sponsor collected the data, monitored conduct of the program, and performed the statistical analyses.”
“The sponsor collected the data, monitored conduct of the program, and performed the statistical analyses.”
They also wrote the paper.
“The initial draft of the manuscript was prepared by a writer employed by Gilead Sciences along with one of the authors, with input from all the authors.”
“The initial draft of the manuscript was prepared by a writer employed by Gilead Sciences along with one of the authors, with input from all the authors.”
As @pulmcrit points out - between 56 authors there was surely enough expertise to write the manuscript. And physicians these days should be seeking distance from this much pharma involvement.
When you’re reading an article about a new drug for COVID consider that all drugs do harm. Then ask what type of study was done. A double blinded placebo controlled trial independent of the manufacturers is good news but still needs careful critique.
This isn’t research - it’s simply a few observations about an experimental therapy written down by people selling that therapy. It might best be described as marketing.
If this data was submitted for another condition in another situation it wouldn’t be published anywhere. It is misleading, potentially dangerous and bordering on unethical to publish it.
A final pithy observation from friend and colleague
@neilstone - “Interesting side effect of COVID - the big journals abandoning their standards”.
@neilstone - “Interesting side effect of COVID - the big journals abandoning their standards”.
This conclusion from @geraldineaohara sums it all up nicely.
