This paper analyzes local and regional geographic variability in the use of antidepressant, antipsychotic and stimulant medications in the United States. Using a data set that covers 60% of prescriptions written in the United States, we find that use of antidepressants in three digit postal codes ranged from less than 1% of residents to more than 40% of residents. Stimulant and antipsychotic use exhibited similar levels of local geographic variability. A Kulldorf Spatial Scan identified clusters of elevated use of antidepressants (RR 1.46; p<0.001), antipsychotics (RR 1.42; p<0.001), and stimulants (RR 1.77; p<0.001). Using a multilevel model, we find that access to health care, insurance coverage and pharmaceutical marketing efforts explain much of the geographic variation in use.