Another poorly designed interpretation of a #HCQ data set for #COVID19. A larger poorly designed "trial" only leads to larger erroneous conclusions. For analysis, see the thread @JamesTodaroMD @niro60487270 @danaparish @marybethpf https://www.thelancet.com/journals/lancet/article/PIIS0140-6736(20)31180-6/fulltext

This study implies that it's an early treatment study because hospitalized patients were treated within 48 hours of diagnosis. It's a late treatment study, and here's why:

About 2/3 of patients were from N. America. Unlike Asia and Europe, we haven't had access to early #C19 testing & quick turn around times. By the time patients get to the hospital, 5-7 days have typically gone by, results have been taking about 4 days, & add 48 hours to that.

Unlike parts of Europe & Asia, #HCQ early #COVID19 treatment isn't embraced in the US. It's given only to the sickest patients, without contrary evidence in this study. To state that the baseline disease severity between treatment & control is equal was incorrect, here's why

The authors provide almost no data to assess disease severity at baseline between groups. They rely on qSOFA which in a study in Annals of Intensive Care, was found to be "...not appropriate to identify Covid-19 patients to have poor outcomes..." https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7167215/

And they don't say who got what other antiviral, grouping 3 different ones together. Big confounders. Except anyone who took remdesivir was excluded from the analysis. Why? Don't want to risk it being associated with a bad outcome in a faulty study?

I agree with one thing the authors said: "Randomized clinical trials will be required before any conclusion can be reached regarding benefit or harm of these agents in COVID-19 patients."

Actually, #HCQ RCT's have been done & they show benefits, but we need more of them.