Pharmaceutical decisionmaking: pharmacoepidemiology or pharmacoeconomics--who's in the driver's seat?

Although it is well known that clinical trial findings and actual clinical experience can differ substantially in pharmaceutical decisionmaking, it is our working hypothesis that this divergence is critically important in the area of psychopharmacology. We support this contention with a discussion of recent findings from post-marketing pharmacoepidemiologic and pharmacoeconomic investigations of clozapine. The pharmacoeconomic evaluations purport to show cost savings of clozapine versus standard neuroleptic therapy but these conclusions are flawed, in large part because the epidemiologic investigations on which they are based are inadequate. To correct this situation, long-term, randomized field trials (usual practice settings) are needed to compare costs and outcomes of clozapine versus standard therapy. The design of these studies should incorporate multidimensional outcomes, including social function, employment, and rehospitalization, as well as measures of symptoms and self-reported quality of life. Pharmacoeconomic evaluations that adopt the designs of typical clinical trials with limited outcome measures, such as symptoms or self-reported quality of life measures, will not be sufficient to determine cost-effectiveness for psychopharmacologic therapies of severe mental disorders.