Clinical trials of antidepressant medications have shown that, overall, these drugs are effective, as measured by a > or = 50% decrease in Hamilton Rating Scale for Depression (HAM-D) total scores in about two thirds of patients. However, the results of long-term trials under rigorously controlled conditions show that, even with close follow-up and provision of interpersonal psychotherapy, a third or more of the patients will not achieve or maintain a response to medication for depression. Nevertheless, the improved efficacy of some antidepressants for certain features or types of depression has been shown. Factors associated with a better response to a specific agent or class of drugs include severity of symptoms, patient age, and the symptom profile of the depressive episode, as revealed by assessment scales or subscale scores for selected symptoms. Moreover, a number of studies indicate that a patient's early response to a given medication may assist in predicting long-term outcome. However, outcome measures in traditional trials of antidepressant drug efficacy, such as a 50% reduction in scores on one or more depression rating scales, do not necessarily reflect an improvement in the patient's ability to function in the workplace; they only show that a particular patient at a particular time has responded to treatment in a significant manner by measurement of a depression scale.