OBJECTIVES To measure healthcare utilization and costs among new initiators of LAI and oral second generation antipsychotics with schizophrenia or bipolar disorder. METHODS A large database of a commercially insured US population was used to index patients on their first treatment between 1/1/2011 and 12/31/2011. Patients were required to have ≥ 12 months pre-index and ≥ 12 months post-index, were new users of a second generation antipsychotic, and diagnosed with schizophrenia or bipolar disorder during their pre-index period. LAI and oral patients were matched 1:3 using propensity scores. Mean differences in annual resource use and costs were compared across groups in an unadjusted difference-in-difference analysis: [(LAI post - LAI pre) - (Oral post - Oral pre)]. RESULTS Initial selection identified 250 LAI and 8,356 oral treatment patients. Matching resulted in balanced cohorts of 204 LAI and 612 oral initiators. Annual hospitalizations and ER visits from pre-index to post-index was significantly lower in LAI initiators compared to oral initiators. Mean annual hospitalizations per LAI patient reduced from 1.09 to 0.51 (p < 0.0001) while that of the oral cohort reduced from 0.53 to 0.39 (p = 0.0011). This resulted in a net reduction of 0.45 annual hospitalizations per patient in the LAI cohort, using the oral cohort as a reference (p < 0.0001)