BACKGROUND: The availability of atypical antipsychotics varies substantially in the UK. Several drug protocols are available for treatment of people with first episode schizophrenia. Current evidence doesn t compare the relative efficiency of these protocols for first episode schizophrenia. OBJECTIVES: To compare the relative costs and quality adjusted life years (QALY s) at three years of 8 antipsychotic protocols for first episode schizophrenia. METHODS: Data were estimated from systematic reviews, UK specific statistics and published literature. A decision analytic model was used with monte-carlo simulation analysis to estimate the expected costs and QALY s (95% confidence intervals). RESULTS: The 8 protocols gave 32 combinations of drug sequences. The mean (95%CI) expected costs and QALY s per person for 1st line therapy were: chlorpromazine, 19600- 21100 ( 19400- 21300) and 2.20-2.35 (2.19-2.35), haloperidol 22800- 24300 ( 22600- 24600) and 2.18-2.27 (2.17-2.27), risperidone, 22911- 22961 ( 22700- 23200) and 2.35 (2.34-2.36), olanzapine 26800- 26994 ( 26500- 27300) and 2.28-2.31 (2.28-2.31). The incremental cost/QALY of risperidone versus chlorpromazine was 54755- 663170. The ranking of expected costs and QALY s were similar for 2nd, 3rd and 4th line therapy. The incremental cost/QALY of risperidone versus chlorpromazine was 14300- 26000. The incremental cost/QALY of clozapine versus chlorpromazine for 3rd and 4th line only was 13700- 18000. CONCLUSIONS: Chlorpromazine may be more efficient as 1st line therapy. Risperidone and clozapine may be more efficient for patients failing 1st and subsequent therapy. The results are uncertain: (i) the analyses were exploratory; (ii) differences in expected costs and QALY s may not be economically important (iii) the quality of clinical and utility data was poor; (iv) some side effects were excluded due to lack of data.