Predictors of response and remission in a naturalistic inpatient sample undergoing multimodal treatment for depression

Background: Many depressed patients do not achieve response or remission despite adequate treatment. Identifying predictors of outcome can contribute to developing therapeutic algorithms for difficult-to-treat depression. Therefore, we examined clinical predictors of response and remission in a naturalistic inpatient sample undergoing multimodal treatment for depression. Methods: Three hundred and fifty-one consecutive inpatients admitted to a tertiary care university hospital (specialized psychiatry unit for treatment of unipolar and bipolar depression) between January 2014 and December 2016 were characterized by a set of sociodemographic and clinical variables. The predictive value of these variables for response >= 50% decrease from baseline Montgomery-Asberg Depression Rating Scale (MADRS) score) and remission (MADRS score at discharge < 10) were explored using bivariate analysis and logistic regression. Results: Greater symptom severity and fewer psychotropic medications at the time of admission predicted response. Remission rates were higher for patients with non-chronic depression, higher number of previous depressive episodes, fewer psychotropic medications and less severe depression at admission. Limitations: This was a retrospective study without a control group. The sample was drawn from a single inpatient ward specialized for difficult-to-treat depression. Conclusions: Greater baseline depression severity might be a proxy for a less chronic course of depression thereby explaining its association with greater response rates. Fewer episodes in the past and polypharmacy could indicate treatment-resistance and chronicity, contributing to lower remission rates. Therefore, preventing chronicity should be a central aim of depression treatment.