A retrospective cohort study, with a mirror-image design, was used to measure inpatient service utilization in 63 consecutive patients started on clozapine from a geographical catchment area compared to a control group matched for previous inpatient service use. An intent-to-treat analysis, including those patients (n = 28) who discontinued clozapine during the study period, showed a significant reduction in number of admissions and total time spent in hospital in the 2 years following clozapine initiation compared to the previous 2 years and to the follow-up period in the control group. This translated into a reduction of £7300 in hospitalization costs per patient started on clozapine, over the 2-year period. In those patients who continued clozapine treatment for the whole of the 2-year period, there was a two-thirds reduction in number of admissions and total time spent in hospital compared to no change in the clozapine discontinuers. These findings suggest that clozapine is a clinically and cost-effective intervention for severe schizophrenia in routine clinical settings.