[jacobion]

Yes, for headaches no one wants to read a meta-analysis, they just want to buy something quickly and feel slightly better.

On the other hand, for depression medication, they don't want their doctor to look at an online tool and choose the most effective antidepressant. They want their doctor to look at THEM and say "hmm, we'll try you on X but if it doesn't help with the intrusive thoughts we'll maybe switch it out to Y and up the dosage of Z". Or they want to tell the doctor what they think they want to be prescribed. They are paying big bucks to see a psychiatrist. Most of them are not on a self-optimization trip, they just want to feel better, and also feel like someone takes an interest in how they are getting on. Using a snazzy tool would probably lead to the patient being less satisfied with the doctor's service, even if they have slightly better outcomes by whatever questionable metric is in the study.