This is not a rigorous study and doesn't pretend to be but those are sufficiently impressive results that it is still a cause for hope.

You need meticulous study design when you're looking for small effect sizes, with a large enough effect you can extract useful information even from flawed data.