Objective. Investigation of the efficacy and tolerability of olanzapine (OLZ) as second-line treatment in subjects with a first-episode psychosis (FEP) who had been non-responsive, intolerant or non-compliant to Risperidone (RIS). Methods. The Early Psychosis Prevention and Intervention Centre in Melbourne admitted 786 FEP subjects between 1998 and 2000. Data were collected from subjects' medical records (MR). The objective was to evaluate the efficacy of OLZ as measured by CGI-S, GAF, SOFAS and remission rates as well as tolerability. Results. A total of 104 subjects were switched because of non-response (38%), non-compliance (15%), or intolerance (47%). Independent of reasons for switch, significant symptomatic and functional improvements were detected. Overall, 46 subjects (44%) achieved full remission. Regression analysis indicated that reason for switch did not predict full remission. Significantly more extrapyramidal side effects (P < 0.001) were related to previous RIS and significantly more weight gain to the subsequent OLZ treatment (P < 0.001). Conclusions. OLZ may be an effective alternative for FEP patients who are non-responsive, intolerant or non-compliant to RIS. The decision to switch to OLZ must be weighted against the greater risk of weight gain with its risks for medical complications and its psychosocial consequences