Determination of a clinically important difference and definition of a responder threshold for the UCSD performance-based skills assessment (UPSA) in patients with major depressive disorder

Background: This article reports an evaluation of the psychometric properties and clinically important difference (CID) threshold of the UCSD Performance-Based Skills Assessment (UPSA) in major depressive disorder (MDD), using data from a large-scale study of the effects of vortioxetine on cognitive functioning and functional capacity in MDD patients. Methods: Adults with moderate-to-severe recurrent MDD and self-reported cognitive dysfunction were randomized to 8 weeks of double-blind treatment with vortioxetine 10/20 mg QD (flexible), duloxetine 60 mg QD, or placebo. Pearson correlation coefficients were calculated between UPSA composite score and demographic/disease characteristics at baseline to examine construct validity. Two methods (distribution-based and anchor-based) were used to establish a CID threshold. Results: A total of 602 patients were randomized; 528 comprised the full analysis set. For the entire sample mean UPSA composite scores were 77.8 at baseline and 83.9 at week 8 (mean change, +6.1). As hypothesized, at baseline, the UPSA composite score correlated with cognitive functioning (Digit Symbol Substitution Test: r=0.36, P < 0.001) and workplace productivity (Work Limitations Questionnaire: r=-0.17, P=0.008), but not depressive symptoms (Montgomery-Asberg Depression Rating Scale: r=0.02, P=0.707) or subjective cognitive dysfunction (Perceived Deficits Questionnaire: r=-0.02, P=0.698). Limitations: Two versions of the UPSA were used and no inclusion/exclusion criteria were based on the UPSA. Conclusions: These results support the construct validity of UPSA for assessing functional capacity independent of mood symptoms. The estimated CID for changes in UPSA scores was quite consistent at +6.4 points and +6.7 based on distribution-based and anchor-based methods, respectively.