This paper shows how decision theory can be used to determine optimal cut-off scores on mental health screeners. The procedure uses (a) the costs and benefits of correct and erroneous decisions, and (b) the rates of correct and erroneous decisions as a function of the cut-off score. Using this information, for each cut-off point expected costs are calculated. The cut-off point with the lowest expected costs is the optimal cut-off score. An illustration is given in which the General Health Questionnaire is employed as a major depression screener. Optimal cut-off points are determined for four different contexts: patients, health service providers, society, and mental health researchers. As in these four situations different costs are encountered, different optimal cut-off points were found.