"One thing I found a little bit odd was that the three HCQ treatments had odds ratios nearly the same (about 1.3)"
I meant that the 3 mortality ratios were about 1.3, ie the three HCQ groups recorded a 30% increase in mortality over the control group (no HCQ). If you look at the paper (or almost any reporting on it) it indicates that the actual mortality in the HCQ groups was almost double, ie a LOT more than 30% worse. But there was a good explanation for this - the HCQ group had sicker patients; the mathematical model that produced these Odds Ratios(OR) adjusted for those worse risk factors, and arrived at Odds Ratios of about 1.3. I took this as an indicator that the 'correct' adjustment had been made(*) and the authors, at least, had done the stats honestly with the data they had been given (whether they should have taken it on is another matter). I'm not sure history will judge the same of Surgisphere.