The use of therapeutic drug monitoring (TDM) in the authors' hospital was studied for various antidepressants (maprotoline, clomipramine, imipramine, desipramine, nortriptyline, amitriptyline) in a two part-study (retrospective and prospective). Criteria for appropriate requests for TDM from the resident psychiatrists and for appropriate interpretation of the results were defined according to a clinical effectiveness/toxicity model. For this purpose, a "therapeutic window" (between 150 and 250 ng/ml) was chosen. Whereas the quality of the TDM requests varied in the two studies (41% and 26% inappropriate requests in the retrospective and prospective study respectively), the interpretation of the results was the same in the two studies (19% inappropriate therapeutic adjustments). Considering both the inappropriate requests and the inappropriate therapeutic adjustments, if appears that 52% of the assays in the retrospective study and 40% in the prospective study are inappropriate. In common with previous studies with other drugs, the present study concludes that the use of the antidepressant TDM could be improved. To this end, closer attention must be given to the nature of TDM requests.