Background Clozapine is the only drug licensed for treatment-resistant schizophrenia (TRS) but the real-world clinical and cost-effectiveness of community initiation of clozapine is unclear. Aims The aim was to assess the feasibility and cost-effectiveness of community initiation of clozapine. Method This was a naturalistic study of community patients recommended for clozapine treatment. Results Of 158 patients recommended for clozapine treatment, 88 (56%) patients agreed to clozapine initiation and, of these, 58 (66%) were successfully established on clozapine. The success rate for community initiation was 65.4%; which was not significantly different from that for in-patient initiation (58.82%, χ2(1,88) = 0.47, P = 0.49). Following clozapine initiation, there was a significant reduction in median out-patient visits over 1 year (from 24.00 (interquartile range (IQR) = 14.00-41.00) to 13.00 visits (IQR = 5.00-24.00), P < 0.001), and 2 years (from 47.50 visits (IQR = 24.75-71.00) to 22.00 (IQR = 11.00-42.00), P < 0.001), and a 74.71% decrease in psychiatric hospital bed days (z = -2.50, P = 0.01). Service-use costs decreased (1 year: -£963/patient (P < 0.001); 2 years: -£1598.10/patient (P < 0.001). Subanalyses for community-only initiation also showed significant cost reductions (1 year: -£827.40/patient (P < 0.001); 2 year: -£1668.50/patient (P < 0.001) relative to costs prior to starting clozapine. Relative to before initiation, symptom severity was improved in patients taking clozapine at discharge (median Positive and Negative Syndrome Scale total score: initial visit: 80 (IQR = 71.00-104.00); discharge visit 50.5 (IQR = 44.75-75.00), P < 0.001) and at 2 year follow-up (Health of Nation Outcome Scales total score median initial visit: 13.00 (IQR = 9.00-15.00); 2 year follow-up: 8.00 (IQR = 3.00-13.00), P = 0.023). Conclusions These findings indicate that community initiation of clozapine is feasible and is associated with significant reductions in costs, service use and symptom severity.