[MBCook]

Right. What if this were to be conformed and we could find out that A is best treated by Prozac but C responds really badly to it?

Right now we just try lots of drugs, one after another, on each patient until something works or the patient gives up. What if we could confidently predict the drugs that were the most likely to succeed for each group? It may cut treatment times and help people find relief faster.

Even if we don’t find new treatments or gain a better understanding of cause/effect it could still be useful.