There is a long discussion about this study with the virologist Christian Drosten on a German podcast (transcript here: https://www.ndr.de/nachrichten/info/coronaskript134.pdf), though in german only.

He was quite skeptical about the study, in particular as it doesn't have randomized groups, and e.g. the age difference between control and treatment group was very large. He also didn't consider the virus concentration in the throat to be a good endpoint to measure as from their experiments they know that it goes down over time anyway, and the virus tends to move to the lungs.

It's an antibiotic. Severe patients will be dealing with secondary (bacterial) infections, so it makes sense to use both. But on its own, I expect it would have little effect.

It’s already commonly used as both treatment and profylactic (in healthy risk group patients) for viral pneumonia. Why it works for I think is still not entirely clear.

Being an antibiotic is also convenient because many times you would also get antibiotics to prevent a secondary infection from bacteria.

It’s not exactly a miracle cure (like an antibiotic is against a bacterial pneumonia), but doctors have few tools when it comes to viral pneumonia, so even something that helps a little is welcome.

As I understand it - it interferes with mRNA transcription. Given that COVID-19 is RNA that may be useful.

The study wasn't randomized. It tells you basically nothing.

I really don't understand why anyone approves such studies. They could've done something useful and we'd know a little bit now.

to the best of my knowledge, there has been no randomized study against placebo of several important things - like parachutes.

Next time I am in a burning plane, I'll make sure to refuse a parachute, as the lack of studies means anecdotal evidence tells us basically nothing.

You may want to read this: http://cmajopen.ca/content/6/1/E31.full

I don't know if Chloroquine works, but I'm pretty sure Chloroquine is no parachute (i.e. it's not a drug where the benefits are so obvious that it doesn't need a proper trial).

Nice article, thank you. Still it doesn't adresse the question: statistically, the benefits can sometimes be so obvious that no randomized trial is needed. RCT and large series are needed when trying to prove some minimal effects. The article is more like a sociological study to say invoking this parachute argument is no definite proof based on cases where the parachute argument was invoked and in retrospective was misguided.

That is very true. But unless we are to argue that the virus is not what causes the disease, using the viral load as a proxy for clinical status and contagiousness is acceptable. And unless the effect is small enough that sample selection can affect the conclusion, it doesn't matter much in practice.

I meean, for chloroquine, we have studies showing non inferiority and strong effects. We will not know if it works according to modern standards of statistics using direct survival statistics for another two weeks at best. What to do in the meantime?

In this article, the comparison to flossing in the intro is not good: lack of flossing does not entail mortality rate >.5%.

And all the other approaches selected for study do not feature infectious agents creating a risk of death, so the conclusions of the article are not very transposable to the problem here. For chemotherapy, surgery, etc. yes, they have a point. For microbiology, no.

Everybody seems to be doing armchair critique of the methodology used. I respect the desire to understand, but I believe the critique stands: we do not have RCT for parachutes. And at the moment, we do not have RCT for chloroquine either. We have a long experience of using it for other diseases. It's an old drug with well known side effects and toxicity.

If you are infected, you are free to wait until there is a trial to accept the drug. Yet, based on the information available, I still think it is a wrong decision, akin to asking for a RCT on parachutes aboard a burning plane.

> but it's something!

And that something may turn out to be noise.