I think the problem here is that the predictive value of the "pretty good evidential" support and success in a future (acceptable to the FDA) clinical trial isn

I think the problem here is that the predictive value of the "pretty good evidential" support and success in a future (acceptable to the FDA) clinical trial isn't as high as you might think. Admittedly it's for a different indication, but the Medivation experience with dimebon, which is a drug used in Russia as an antihistamine for decades, which had "pretty good evidential support" that the drug might work for Alzheimer's disease is, unfortunately, more likely to be par for the course.