OBJECTIVE: To compare the treatment costs of schizophrenia patients who relapsed in the European SOHO study with those who never relapsed over three years. METHODS: The SOHO study is a 3-year perspective, outpatient observational study of 10,972 patients across 10 European countries. The Clinical Global Impression (CGI) Scale was used to measure clinical effectiveness on a scale of 1(not ill) to7 (most severely ill). The definition of relapse was based on increases in the CGI Scale and/or inpatient admission. Kaplan-Meier analysis was used to estimate the relapse rate over three years. The resource use (inpatient, day care, psychiatrist visits and medications) for those who ever relapsed during the three years of study was compared to those who never relapsed. UK costs were applied to the resource use. RESULT: A totla of 6397 patients with complete data on CGI and resources were used in the analysis. A totla of 2489 (39%) patients relapsed over the 3 years. The total cost for patients who relapsed (16724.27 pounds) was higher than those who did not relapse (7724.93 pounds). 6368.88 pounds (71%) of cost difference was due to inpatient stays, 2020.13 pounds was due to day-care, 416.11 pounds due to medications and 194.22 pounds due to psychiatrist visits. The costs will also be modelled adjusting for patient characteristics and taking account of the fact that the cost data is not normally distributed. CONCLUSION: Relapsed patients used more health care resources in the treatment for schizophrenia. The cost of antipsychotic treatments which prevent relapse could be offset by the savings associated with preventing relapse.