Comparing the cost effectiveness of risperidone and olanzapine in the treatment of schizophrenia using the net-benefit regression approach.

To estimate the cost effectiveness of olanzapine and risperidone for the treatment of schizophrenia in Belgium. Data were retrieved from a prospective, observational, non-randomized, follow-up survey. Clinical investigators assigned 265 patients with schizophrenia to either olanzapine (n = 136) or risperidone (n = 129). Patients were followed up for 2 years. Total healthcare costs were determined from the public payer perspective and calculated by multiplying resource use with official tariffs; effectiveness of the drugs was measured with the EQ-5D. This study uses a net-benefit regression approach to accommodate for baseline differences between treatment groups and uncertainty. Total 2-year costs were very similar for patients receiving risperidone and olanzapine (euro20 915.33 and euro20 569.69, respectively; p = 0.925) [year 2002 values]. The health condition of the patients receiving risperidone was better than that of patients receiving olanzapine but not significantly so (1.46 and 1.41, respectively; p = 0.191). Simple ordinary least squares (OLS) regressions indicated that, for lambda = euro40 000, we could not reject the null hypothesis that the drugs provide similar net monetary benefits to the patient (risperidone vs olanzapine euro2046.95; p = 0.656). When we controlled for several patient characteristics, risperidone moved further away from olanzapine but the difference did not reach statistical significance (risperidone vs olanzapine euro3198.07; p = 0.595). Numerous sensitivity analyses confirmed the robustness of the results. Results of this study suggest that it is important to control for baseline patient characteristics when performing a cost-effectiveness analysis. No significant difference in net monetary benefit was found between risperidone and olanzapine.