Abstract
The aim of this study was to clarify the impact of various methodological quality factors on reported outcome of randomised clozapine trials. Trials comparing the atypical antipsychotic clozapine with other antipsychotic drugs were identified in extensive electronic searches. Two independent reviewers extracted data on methodology and primary outcomes, and assessed trial quality by use of three sets of criteria (Cochrane, Delphi, and Jadad). There was no association between trial quality, as measured by any of the criteria sets, and primary measures of outcome. Trials with the best score for randomisation and concealment according to the Delphi scale had a significantly lower relative risk for relapse in clozapine-treated groups, and studies with well reported random order generation according to Jadad criteria tended to have better odds ratios for clinical improvement on clozapine. These findings strengthen the evidence of true clozapine superiority in these aspects. No other quality items correlated to the primary outcomes. Inadequate randomisation techniques may be a source of bias in clozapine studies, but much more research is needed on the connections between trial quality and trial outcome.