[Results of a predictive trial of the response to nomifensine in ambulatory patients].

Nine hundred and fifty eight cases of neurotic-secondary ambulatory depression were treated with nomifensin by 135 psychiatrists. The analysis of visual analogue scales and of the global assessment shows that:--there were 2/3 responders and 1/3 non-responders--the most improved items were: crying, troubles in getting to sleep, abnormal tiredness, anxiety;--associated drugs from the beginning were frequent. They seemed to correspond to different initial symptomatologies but the final state did not differ substantially;--discrimination methods (linear discriminant function, stepwise discrimination) could separate groups of responders and non-responders, but did not allow for a valid individual prediction;--admist possible causes of these misclassifications, several factors are discussed: patient selection, choice of the variables, skewed or bimodal distributions, statistical methods used, acurateness of quotation, spectrum of the stutied drug.