The mistake was the inclusion of ventilator dependency. You can...

The mistake was the inclusion of ventilator dependency. You can see why it happened, at the time of this trial ventilation was deemed the best care available and a requirement for moderate/severe Covid-ARDS, it was pretty much the benchmark for how the disease progresses and the world rushed to make ventilators thinking they would run out. Then analysis of results under ventilation showed that it may even be damaging the patients so it quickly became a last resort and well after the time it was being applied earlier in our response to serious cases of Covid-19.

The way to avoid this to to stick to objective pathology criteria rather than the commencement of a particular treatment or the time of admission to hospital etc. CT scans of lungs and blood pathology was enough to provide a clear and objective determination of disease severity and they should have left it at that. Hindsight is great that way. I expect that the delays in ventilation with evolving SOC were a significant reason why enrolment was so slow. It would not surprise me if the opposite of what people are surmising (that SOC reduced death rates) is true, that placebo death rates actually climbed significantly across the trial period - but so did the treated arm - and the gap narrowed dramatically.

I hope the unmasked stats allow for a date based evaluation of the data so that this can be explored - if ventilation delay is causing the placebo arm to also show increased mortality then it would be clear that the timing of first medication is the key to whether Rem-L (or perhaps any treatment) can be effective.