A fundamental question in the medical management of disease is how well treatment works. Although there are many ways of defining improvement, one endpoint of definite interest is the resolution or remission of the disorder. In short-term antidepressant treatment trials, remission is usually defined according to post-treatment scores on symptom severity measures, such as the Hamilton Rating Scale for Depression (HRSD) or the Montgomery-Asberg Depression Rating Scale (MADRS). However, there is uncertainty as to what cut-offs should be used on these measures to define remission. During the past 2 years, as part of the Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project, our laboratory has examined the question of how to define remission on the HRSD and MADRS. In the present report from the MIDAS project, we examined the impact of the cut-off score used to define remission on the percentage of depressed outpatients in ongoing treatment who are considered to be in remission. In addition, we examined the association between remission status and psychosocial impairment for different cut-off scores. Three hundred and three depressed psychiatric outpatients were rated on the MADRS, 17-item HRSD, and an index of DSM-IV remission status. Approximately one-third of patients completed a measure of psychosocial impairment. For both the HRSD and the MADRS, we examined four cut-off scores for remission. For each cut-off, we determined the percentage of patients who met the definition of remission, the percentage of patients who continued to meet DSM-IV criteria for major depressive disorder (MDD), and the percentage of patients without any self-reported impairment from depression. For both scales, the range of cut-off scores was associated with more than a two-fold difference in prevalence of remission. Based on higher thresholds to define remission, a small percentage of patients met criteria for MDD, whereas no patients scoring below the low thresholds had MDD. The threshold to define remission was associated with psychosocial impairment: higher cut-off scores were associated with lower rates of no impairment. The cut-offs used to define remission from depression considerably influence the percentage of patients considered to be in remission. Lower cut-off scores than those most commonly used to define remission appear to be more valid.