We need really good science and statistics, far superior to p-values, to allow this potentially life saving treatment without the inevitable deaths from people at the limits returning a "not significant" finding for something that actually works.
Maybe something like allowing people who are too far gone to qualify for treatment as part of a study to instead qualify for the experimental treatment with their results only being included as a special "humanitarian" cohort.
I'd also like studies to be able to compare a treatment group against prior control groups, so that everyone today can get the experimental treatment instead of some of them being given a placebo (if there's a lack of another decent treatment to use as a control). If there's nothing effective that can be done as a control, every one should get the treatment.