The clinical judgment process in the prescribing of psychotropic drugs

The primary objective of this study was to assess the strategies used by general practitioners in prescribing psychotropic drugs. Twenty-four physicians associated with the University of Toronto were presented with 63 vignettes, each describing a patient with a different combination of symptoms for which psychotropics are prescribed, and asked to choose the most appropriate drug for each patient. A mathematical model for each physician's prescribing pattern was developed from the information provided by his/her responses to the case descriptions. The physicians' prescribing strategies were found to be highly idiosyncratic, and were not related to their medical experience nor to the type of private practice in which they engage. The decision-making processes the physicians used to prescribe major tranquilizers and antidepressants were adequately modeled by a multiple linear regression equation, with very few of the predictor variables accounting for a substantial portion of the variance in prescription choice. Policies for prescribing minor tranquilizers were less successfully captured by this technique. Implications for the use of such models for improving prescribing techniques and for their potential utility in clinical training are discussed.