bout one-third of the annual $51 billion cost of mental illnesses is related to productivity losses. However, few studies have examined the association of treatment and productivity. The purpose of our research is to examine the association of depression and its treatment and work productivity.The special extract of St. John's wort, LI 160, exhibited a superior antidepressant efficacy compared to placebo in several controlled trials. Two further trials demonstrated a similar reduction of depressive symptomatology under LI 160 compared to tricyclics. All these trials were performed in mildly to moderately depressed patients. The present investigation was a randomized, controlled, multicentre, 6-week trial comparing 1800 mg LI 160/die to 150 mg imipramine/die in severely depressed patients according to ICD-10. The main efficacy parameter, a reduction of the total score of the Hamilton Depression Scale, proved both treatment regimens very effective at the end of the 6 week treatmend period (mean values 25.3 to 14.5 in the LI 160 group and 26.1 to 13.6 in the imipramine group), but not statistically equivalent within a a-priori defined 25 % interval of deviation. The analysis of subgroups with more than a 33 % and 50 % reduction of the HAMD total score justified the assumption of equivalence within a 25 % deviation interval. This view was also supported by the global efficacy ratings from patients and investigators. Regarding adverse events, the nonrejection of the nonequivalence hypothesis denotes a superiority of the herbal antidepressant. These main results indicate that LI 160 might be a treatment alternative to the synthetic tricyclic antidepressant imipramine in the majority of severe forms of depressions. However, more studies of this type must be performed before a stronger recommendation can be made.