Grip - it's a body of work: 84 m/m + 195 s/c + 61 EIND = 340 COVID pats already, plus
(1) Safety - we have >1k pats w/ safety data - others have proven worse record or are newbies lacking data, and
(2) Efficacy: square root rule - if you have half the effect, you need four times the trial size to detect it.
All longs here believe in a minimum 50% reduction in mortality in S/C and 50% reduction in progression of m/m to s/c (data: m/m NEWS2 80% LL improved vs. 50% control)
So, considering the RCTs together, compared to a drug targeting half that, in terms of stat power for proving efficacy, we already have data matching a 1,000 person trial.
Of course, each trial was/will be on its own statistically significant on its relevant endpoint. The body of work will be "big enough" and have unmatched consistency.