Abstract
OBJECTIVES: Site-independent ratings surveillance assessed ratings reliability in a clinical trial. METHODS: Inter-rater reliability was assessed at the screen visit in a 6-week, double-blind, placebo-controlled study of lurasidone for the treatment of major depressive disorder (MDD) with subthreshold hypomanic ("mixed") symptoms (clinicaltrials.gov NCT01421134). Site-based Montgomery-Asberg Depression Rating Scale (MADRS) and Young Mania Rating Scale (YMRS) interviews were paired with 184 site-independent ratings derived from audio-digital recordings. RESULTS: The paired MADRS and YMRS scores were highly correlated (r = 0.708 and 0.885, respectively) and generated minimal scoring discordance. The surveillance program identified 14 MADRS scores (7.6% of this sample) that were below the study entry criterion (MADRS ≥26) and resulted in screen failure. When present, paired scoring discordance was associated with symptom severity, interview length, interview quality, and the level of patient cooperation. Higher severity scores (MADRS ≥40 and YMRS ≥15) were associated with greater paired scoring discordance. Further, MADRS scores <30 and short MADRS interviews conducted in ≤12 min revealed significantly more pairs of discordant outliers (p = 0.04 and 0.009, respectively). CONCLUSIONS: The findings suggest that MDD patients with mixed features can be reliably assessed, that paired site-based and site-independent assessments were generally concordant, and that ratings surveillance may reinforce ratings precision.