Second-generation atypical antipsychotics improve the outcome of patients with schizophrenia, although studies of their cost efficacy in comparison to first-generation conventional antipsychotics have yielded mixed results. This study examines the cost effectiveness outcome of olanzapine treatment in veterans with schizophrenia (n 5 22) or schizoaffective disorder (n 5 4). Health-care utilization and costs associated with prospective olanzapine treatment were compared with those of retrospective first-generation neuroleptic treatment in a mirror-image design. The analysis of variance with repeated measures for the Positive and Negative Syndrome Scale (PANSS; n 5 22) showed a significant main effect of olanzapine treatment (p , .025), and the effect was of medium-to-large size (h2 5 .13). The PANSS-positive subscale (p , .005) and the PANSS general subscale (p , .005) significantly decreased, but the PANSS negative subscale did not change. The quality of life survey (n 5 21) significantly increased (p , .025), and the effect size was large (h2 5 .14). For VA outpatient and inpatient care, study patients incurred an average cost difference of 2$1,289 (NS) and 2$6,682 (NS), respectively. Combining inpatient and outpatient VA care, patients incurred an annual difference of 2$7,971 per patient (NS). These numerically lower costs were due, in part, to a slower growth rate in outpatient encounters (p 5 .013), lower overall cost per outpatient encounter (p 5 .008), and a lower overall inpatient encounter rate (p 5 .005). Olanzapine treatment resulted in improvements in positive and general psychiatric symptoms, as well as quality of life. Negative symptoms did not change significantly. Though not statistically significant, the postbaseline health-care costs and utilization declined.