Ann: Mesoblast Update on COVID-19 ARDS Trial, page-490

Ok people for non trialists or scientists let me try to give you an important point, others back me up here:

a primary endpoint of reducing mortality by 46% is huge right? Why would they set themselves such a high bar? The answer is statistical power. The smaller the effect size you are trying to detect, the larger the number of test subjects you need to recruit is. In theory you are supposed to let the previous research tell you or use a previously defined minimally worthwhile effect size. The reality is often that trialists take a pragmatic approach along the lines of “how many people can we recruit how quickly that will give us a decent effect size? If we get enough and it works it will detect a smaller effect with sufficient precision. This is not theoretically what you are supposed to do but it happens a lot. To detect a more reasonable effect for ARDs of like 20% they may have needed 700 people which would have taken too long? We need to see the full results before we can judge the science and potential.