Latent classes of nonresponders, rapid responders, and gradual responders in depressed outpatients receiving antidepressant medication and psychotherapy.

BACKGROUND We used growth mixture modeling (GMM) to identify subsets of patients with qualitatively distinct symptom trajectories resulting from treatment. Existing studies have focused on 12-week antidepressant trials. We used data from a concurrent antidepressant and psychotherapy trial over a 6-month period. METHOD Eight hundred twenty-one patients were randomized to receive either fluoxetine or tianepine and received cognitive-behavioral therapy, supportive therapy, or psychodynamic therapy. Patients completed the Montgomery-Åsberg depression rating scale (MADRS) at the 0, 1, 3, and 6-month periods. Patients also completed measures of dysfunctional attitudes, functioning, and personality. GMM was conducted using MADRS scores and the number of growth classes to be retained was based on the Bayesian information criterion. RESULTS Criteria supported the presence of four distinct latent growth classes representing gradual responders of high severity (42% of sample), gradual responders of moderate severity (31%), nonresponders (15%), and rapid responders (11%). Initial severity, greater use of emotional coping strategies, less use of avoidance coping strategies, introversion, and less emotional stability predicted nonresponder status. Growth classes were not associated with different treatments or with proportion of dropouts. CONCLUSIONS The longer time period used in this study highlights potential overestimates of nonresponders in previous research and the need for continued assessments. Our findings demonstrate distinct growth trajectories that are independent of treatment modality and generalizable to most psychotherapy patients. The correlates of class membership provide directions for future studies, which can refine methods to predict likely nonresponders as a means to facilitate personalized treatments.