#chloroquine https://thelancet.com/journals/lancet/article/PIIS0140-6736(20)31180-6/fulltext Maybe there is a surprise in this observational study.

If we treat intubation as an independent event, there is a weak survival benefit from the CQ+macrolide group: This treatment lowers the mortality by ~5% if patient will be on ventilator.

Decision is tricky b/c Dr doesn't know a priori if a patient needs intubation. If a patient will be 100% on ventilator, CQ+marcolide might save a few lives. However, if patient won't be on ventilator, current data suggested a higher mortality by HCQ/CQ(I think it's biased,though)

Similar observation is also in the previous NEJM study, in which the sample size is much smaller & unadjusted. Please DM/reply if you think there is major confounder in this analysis.

Disclaimer: as one of the first anti-Raoulters, the purpose of this analysis is not to escalate partisan fight. We already had enough!

Hypothesis can be derived from a different angle. But ultimately, all hypotheses should be validated by randomized controlled trial.

I will give 24hr before I explode this thread. Possible bias has been found which can completely ruin the conclusion here. https://statmodeling.stat.columbia.edu/2020/05/25/hydroxychloroquine-update/#comment-1345717