Reliability and validity of a self-report scale for daily assessments of the severity of depressive symptoms

Abstract

To evaluate the efficacy of rapidly effective treatments for depression it is necessary to use measures that are designed to assess symptom severity over short intervals. In the present report from the Rhode Island Methods to Improve Diagnostic Assessment and Services (MIDAS) project, we modified our previously published depression scale and examined the reliability and validity of a daily version of the Clinically Useful Depression Outcome Scale (CUDOS-D). One thousand one hundred and fifteen patients presenting for treatment of DSM-IV/DSM-5 major depressive disorder (MDD) to a partial hospital program completed the CUDOS-D as part of their initial paperwork and on a daily basis thereafter. Test-retest reliability was examined in 50 patients who completed the CUDOS-D twice on the same day. A subset of 69 patients were interviewed by a trained diagnostic rater who administered the 17-item Hamilton Depression Rating Scale (HAMD) at baseline and on the day of discharge. The CUDOS-D had high internal consistency (Cronbach's alpha at intake = 0.82; Cronbach's alpha at follow-up = 0.93) and test-retest reliability (r = 0.91 and 0.98 at intake and follow-up, respectively), and was more highly correlated with another measure of depressive than nondepressive symptoms. CUDOS-D scores progressively declined during the course of treatment, and scores on each successive day were significantly lower than the preceding day. The change in CUDOS-D scores was significantly correlated with a change in HAMD scores (r = 0.65, p < 0.001). A large effect size was found for both measures (CUDOS-D: d = 1.63; HAMD: d = 1.56). In a large sample of partial hospital patients, the CUDOS-D was a reliable and valid measure of the DSM-5 symptoms of MDD assessed on a daily basis.