Many authorities recommend a 6-week trial of an antidepressant befire deciding whether the patient will have a favorable response. However, this length of time is fiequently impractical, especially f o r hospitalized patients. Some data suggests that ultimate response to a tricyclic antidepressant may be predicted fiom improvement within the first few weeks. To investigate this issue the authors analyzed data fiom six 6-week double-blind, controlled trials of selective serotonin reuptake inhibitors. Five studies included a placebo control group and all had a n active antidepressant control. Overall the data showed that if there is no improvement early in treatment the chance of meaningfkl response by week 6 is poor. If a patient on active drug does not have at least a 20% improvement in the HAMD score at any point during the first 4 weeks the chance of 6-week response was 3.7%. I f a patient did not have a 20% improvement at 4 weeks but did at one or more previous weeks the chance of 6-week response was 23.1%. The overall response rate f o r the whole sample was 60.7%. Ery similar results were obtained using the Clinical Global Impressions Scale. These data suggest that a f i l l 6-week trial of an antdepressant is usually not warranted i f there is not at least some improvement during the first 4 weeks of the trial. Depression 232-35 (1994). 0 1994 Wiley-Liss, Inc.