OBJECTIVES: This meta-analysis compares effects directly of olanzapine and haloperidol and indirectly olanzapine and quetiapine via haloperidol in the treatment of schizophrenia. The indirect comparison provides a method to compare safety and efficacy of different medications in the absence of head to head studies. As many of the trials had fixed dosing regimes this analysis examines effect of varying the dose of all three antipsychotics on the results. METHODS: Trials of olanzapine or quetiapine versus haloperidol were identified. Each was compared using all doses of olanzapine, quetiapine and haloperidol. A second comparison analysed clinically relevant doses of these antipsychotics based upon their approved Product Information. RESULTS: The analyses with all doses of haloperidol and olanzapine and the clinically relevant dose analysis showed that for the efficacy and tolerability outcomes, olanzapine was statistically superior to haloperidol. When comparing olanzapine and quetiapine via haloperidol, the all dose analysis highlighting superior efficacy and safety of olanzapine include more dropouts due to lack of efficacy for quetiapine, greater change in PANSS total for olanzapine and better CGI severity score. The superior efficacy and safety of olanzapine when analysing trials using clinically relevant doses of olanzapine (10-15mg), quetiapine (300-450mg) and haloperidol (5-15mg), is shown by statistically more dropouts for any reason and lack of efficacy for quetiapine and a higher likelihood of response and a better improvement in PANSS total score for olanzapine. CONCLUSION: The analysis of clinically relevant doses is considered to be most informative as it looks at efficacy and safety in patients taking doses of drugs most likely to be used in clinical practice. In fixed dose trials sub or supra-therapeutic levels of the study drug can be used which could affect side effects and efficacy of therapy.