Apart from its methodological originality, which may serve for subsequent studies, the main interest of this study lies in the pioneering nature of its massive use of population micro-data to evaluate prescription quality. These data are taken from the computerised clinical records in primary care.