FDA approved remdesivir. I'm in support.

To approve a drug, FDA just needs to know:
• Is there evidence for efficacy (usually at the arbitrary 95% statistical confidence level)?
• Is it safe?

1/n

Some are using the WHO trial to argue remdesivir should not be approved. I invite them to look closely into the WHO trial and use their critical thinking caps.

2/n https://twitter.com/DrEricDing/status/1319425309519859712

If you do an experiment with bad reagents and bad controls, and one with good reagents and good controls, and they have different outcomes, which do you believe?

3/n

I know it's fashionable to bash on the FDA or a pharma company, but the SOLIDARITY trial is unreliable for two reasons.

First, there's no blinding; it's open label. We've been burned time and time again chasing results from non-blinded trials, so we know not to do that.

4/n

You need blinding with an identical-looking placebo to not cause changes in the treatment between the two groups. It could be remdesivir patients were paid less attention because they were expected to do fine on their own for example.

5/n

Now some will point out that in most open-label trials the drug looks effective, then later is found in placebo-controlled trials to not have an effect. They thus conclude remdesivir must really have no effect if it can't beat no treatment WHO's an open-label.

6/n

But as with any experiment, we must consider mechanism. A plausible reason why open-label trials favor the drug is the doctors are motivated for it to work (they won't disappoint he drug company paying them, they will get paid to do the phase 3, they will get a paper)

7/n

In the WHO trial, the doctors were not paid by any of the companies, there isn't future income depending on the outcome, and they will get a paper regardless. These are all good things, yes. Just explaining why there's probably no bias to the treatment group in this case.

8/n

And again there could be bias against remdesivir if the HCWs expect those patients to get better without additional help, and want to help the less fortunate who didn't get remdesivir more.

9/n

The second big problem with the trial is that how long the patients were into their disease course before starting treatment was not clearly assessed. About 1/3 each entered the trial on day 0 of hospitalization, day 1, or days 2+, but 2+ was not defined.

10/n

19% were already ventilated, i.e. on death's door. Also, there were sites in multiple countries, and it's possible that patients may be sicker at hospitalization than in the US ACTT-1 trial due to local conditions.

11/n

But we know from the ACTT-1 and Hubei trials that there's more benefit for patients earlier in the disease, to answer one of @erictopol's questions. (Can see my previous posts on those trials.)

12/n

Overall too many confounding factors in the WHO trial to know what the true effect of remdesivir was. That's why when you run trials you define a tight patient population and you double-blind with a placebo.

13/n

Consumer groups often complain the FDA is too uptight, requiring double-blind placebos when open-label trials already suggest a benefit. Here we have opposite situation, the double-blind with placebo showing good results and the open-label trial not. I know which I believe.

n/n