I think we need to listen to those types of anecdotes. Anecdotes are a form of evidence and if someone has a dramatic improvement where they were completely non-functioning and with the addition of one particular treatment they then resumed employment, became functional and their symptoms regressed dramatically, that's a very important piece of information to work with.
As a researcher, if I was going to test that on an individual, I would embark on something called an n-of-1 trial, which is really a randomized controlled trial done at the level of the individual. I would randomize the intervention between the active cannabis and a placebo in a random order and I would see if the resolution of symptoms tracked according to what they were getting. There is a way to do studies at the level of the individual. I do take Dr. Zhang's point that a lot of the research we have talks about average effects. Even if something isn't working for the average person, there may be benefit for select individuals. What we then want to learn is what features distinguish the individuals who are more likely to receive the benefit, and what features distinguish an individual who is more likely to experience harm, or even to become addicted, so that we can make clinical decisions that are more responsible to those characteristics.