I don't have a dog in this race since I haven't taken a position on HCQ thus far. Also, I'm a Trump supporter, hate the MSM, and believe that the early use of HCQ was justified given the anecdotal evidence and the great safety profile of HCQ.
But now that we have more evidence, it's time to reevaluate our assessment of HCQ, even if it turns out that in hind sight, we were wrong.
There's two types of studies, 1) observational studies and 2) randomized trials. The former are essentially useless, given the massive omitted variable bias arising from the fact that treatment is not allocated randomly, but correlates with (unobserved) patient characteristics and co-treatments, which both also influence disease outcomes. Consequently, there have been observational studies showing that HCQ does massive harm (hyped by the MSM), and also some showing HCQ to have massive benefits (hyped by right-wing media). The most recent study (
**broken link removed**) mentioned by maldorf above is again an observational study with massive flaws.
So what are some of those issues?
First one is illustrated well by this quote from one of the authors:
Hydroxychloroquine helped Covid-19 patients, a study in Detroit found.
edition.cnn.com
So to oversimplify a bit, you have 2 groups. A healthy group
without cardiovascular risk factors, and an unhealthy group
with cardiovascular risk factors. You only give HCQ to the healthy people and observe that their survival is higher than for the unhealthy group. But this outcome would have been the same if there had been no treatment at all! Sure, there are statistical techniques to try and control for the bias, but they are highly imperfect.
There are a couple of more issues. For example, the use of other medications was not constant across groups:
**broken link removed**
Also, the time from hospitalization until HCQ was given did not significantly differ between this and other studies, so the difference in timing cannot actually explain the differing findings. The exclusion 10% of the sample due to those patients not being discharged yet could also skew results. The finding that azithromycin alone is even more effective than HCQ is also highly dubious, and further indicates that omitted variable bias is driving the results regarding HCQ. There was also no addition of zinc supplements which proponents of HCQ claimed to be crucial when they criticized earlier studies.
So all in all, it is a terrible, useless observational study, just like the ones that came before and showed opposite results. The only type of study one can rely on are randomized trials. And the one randomized trial we have seen did not show statistically significant effects of HCQ treatment. You can criticize that study for being underpowered and argue that it actually does provide some evidence for HCQs effectiveness. But to claim victory (see Rex Feral above) because one of the very flawed observational studies is in line with your priors is disingenuous at best.