Ann: 83% Survival in COVID-19 ARDS Patients with Remestemcel-L, page-705

Getting very sick of those claiming the n=12 figure determines the result is not sigificant.
"Significance" is a formal criteria decided by the numbers. Lowering the sample size has the effect of reducing likelhood of significance but doesn't exclude it.

A simple Z-test using a base population with proportion 0.12 success against a sample size of 12 with proportion 0.83
The base population was guessed(!) at 500 (probably WAY under the actual pop size).
The result is not very sensitive to this number as base size >> sample size.

RESULT: Z = -7.0764. The value of p is < .00001. The result is SIGNIFICANT at p < .01.

The only two reasons POSSIBLE for this to be wrong are:
1) the sample was not representative of the background population
2) The result was confounded by the use of other experimental agents

From SI's comments & that a legally binding ANN has been produced, I see no sensible reason to believe either of these reasons have arisen.