And even when they are pre-scribed correctly they might not be the medication the person actually needs. Psychiatrists (they're the ones who can prescribe stuff, psychologists don't, they're more to talk to and try to put your thoughts into a more managable perspective) still have little use of taking bloodsamples, though technology is gladly slowly moving in that direction.
They might for example find that your dopamine levels are low, but, heyho, they probably already guessed after the first three questions, and that doesn't explain why those levels are so low, so they have to rely on interviewing their patients to figure out why their dopamine level is low, classifying when their feelings change and how they change, to figure out what step might be missing in its production or whether the stuff is broken down too fast by the body.
The other option would be taking a brain tissue sample, which isn't exactly a safe procedure for the patient. I really hope there will be some breakthroughs with the blood tests soon, to quicker identify the exact nature of each problem, and better finetune the necessary prescriptions.