Aims and method: The study aimed to identify the predictors of drop-out from clozapine treatment by examining the demographic and clinical characteristics of patients registered on clozapine within a 6-month period in one NHS Trust. Results: During the study period, 54 patients were registered and began clozapine treatment and 31% had discontinued within 6 months. Two people died and the remainder discontinued because of non-compliance or side-effects, including neutropenia. Two factors were predictive: the age of the patient (older patients were more likely to discontinue) and the hospital where the initial registration was made. Clinical implications: Neither ethnicity, previous registration nor the individual prescriber are a bar to successful persistence with clozapine. However, one set of hospitals with a history of evidence-based practice and high clozapine prescribing was more successful in retaining patients on maintenance treatment. Although specific data are needed to identify more subtle contributing factors to continuation, it is clear that there is scope for improving the rate of persistence with clozapine treatment.