OBJECTIVES: The atypical antipsychotic drugs are rapidly replacing the older typical therapies for the treatment of schizophrenia. The two most widely prescribed atypicals, olanzapine and risperidone, now account for $US4.9 billion in global health care sales per annum (2001) and hence analyses on their respective efficacies are important. METHODS: Several randomised controlled trials have compared these two drugs directly, but there is some inconsistency in the reported relative efficacy. The 8 week study by Conley and Mahmoud (2001) suggested that risperidone is more effective than olanzapine. By contrast, the 28-week trial by Tran et al. (1997) suggested that olanzapine is more effective than risperidone. Two smaller, medium- to long-term trials also showed significant improvements in a variety of efficacy rating scales for patients treated with olanzapine compared with those receiving risperidone (Purdon et al. 2000, Gureje et al. 2002). These trials ran for 52 and 30 weeks, respectively. In order to explore why there was a difference in the findings of the four trials, we examined the mean change in PANSS total score over time with olanzapine and with risperidone in the Tran et al. (1997) study. The original trial dataset was provided by the trial sponsor (Eli Lilly) and the mean change in PANSS total score was plotted for both study arms over the duration of the study. RESULTS: These results suggest that olanzapine and risperidone have similar efficacy in the short-term (up to about eight weeks), after which time olanzapine offers superior effectiveness to risperidone. This is an important finding, which may explain the different outcomes in the four comparative trials. CONCLUSIONS: Since schizophrenia is a chronic condition, understanding the long-term differences between antipsychotic treatments is critical both to the selection of appropriate therapy and to the allocation of antipsychotic drug budgets.