OBJECTIVES: To examine the impact of having a newly prescribed second generation anti-psychotic (SGA) claim rejected due to formulary restrictions on subsequent medication persistence to the SGAs. METHODS: A retrospective cohort study was conducted using data from a national pharmacy benefit management company. Newly initiated antipsychotic users, aged 18–64 years who had a rejected SGA (aripiprazole, olanzapine, quetiapine, risperidone, ziprasidone) claim due to benefit reasons (step therapy, prior authorization, not in formulary) between January 1, 2005 to December 31, 2006 but who subsequently filled an SGA or conventional antipsychotic within six months of the rejected claim, formed the case group (n=328). Newly initiated anti-psychotic users who were in health plans with an open formulary and thus did not experience rejection of an SGA claim formed the control group (n=1097). All patients were followed up for 13 months. Cox regression models were used to estimate the effect of having rejected claims on all-cause discontinuation of the index drug, defined as discontinuation, add-on or switch. The model controlled for age, sex, co-morbidities, geographic locations, index drug, prescription and co-payment. RESULTS: Reasons for rejected claims were distributed as follows: 1) drug not on formulary (72.9%); 2) required prior authorization (19.5%); and 3) required step therapy (7.6%). Median time to discontinuation was 120 days for the case group and 127 days for the control group. The adjusted hazard for discontinuation of the index drug (HR = 1.29, 95% CI: 1.08–1.53) was significantly higher for patients with rejected initial SGA claims compared to controls. Co-payments ranging from $20 to $39 were associated with lower discontinuation compared with copayment ranging from $0 to $4 (HR=0.75, 95% CI: 0.60–0.93). CONCLUSIONS: New antipsychotic users with rejected initial SGA claims due to formulary restrictions were more likely to discontinue their anti-psychotic drugs compared to users who did not face such restrictions.