Randomized, double-blind, placebo-controlled trials suggest that medic
ations can be efficacious in several psychiatric disorders. The effect
iveness of medications as they are used in practice remains to be seen
. Automated pharmacy data give evaluation researchers tools to examine
the quality of care for people using psychotropic medications. Reason
ably convincing data indicates that automated pharmacy records are yal
id measures of medication consumption. Evaluation research is needed i
n areas such as standard setting, risk adjustment, and the linking of
pharmacy data to outcome measures.