Development and Analysis of Item Response Theory-based Short-form Depression Severity Scales Based on the HDRS and MADRS

Abstract

Objectives The Hamilton Depression Rating Scale (HDRS) is the most frequently used primary endpoint for antidepressant clinical trials. This study developed and evaluated the psychometric characteristics of 3 item response theory (IRT)-based short-form depression severity scales based on combinations of the HDRS and Montgomery-Asberg Depression Rating Scale (MADRS) items. Study Design A secondary analysis was completed using data from 1027 subjects with major depressive disorder participating in 2 antidepressant clinical trials. Data were collected using the HDRS and MADRS throughout the 6-week clinical trials. Maier, Bech, and Gibbons brief depression scales were calculated based on the HDRS. Results Three short-form depression severity (DS) scales were developed based on clinician recommendations and IRT analyses, (DS-1, 7 items; DS-2, 8 items; DS-3, 10 items). Internal consistency reliability of the short forms was 0.87 to 0.93. DS were more reliable across the range of the depression than the HDRS or MADRS. The DS scales were correlated 0.27 to 0.29 with HDRS, 0.55 to 0.85 with MADRS, and −0.25 to −0.34 with Quality of Enjoyment and Satisfaction Questionnaire scores at baseline. In 1 clinical trial, none of the depression outcome measures demonstrated statistically significant differences between the paroxetine and placebo groups. In the second clinical trial, there were significant between-group differences in DS-1 ( P = .004; ES = 0.46), DS-2 ( P <.001; ES = 0.59), DS-3 ( P <.001; ES = 0.63), Bech ( P = .007; ES = 0.43), Maier ( P = .009; ES = 0.41), Gibbon ( P = .003; ES = 0.47), HDRS ( P = .007; ES = 0.43), and MADRS ( P = .001; ES = 0.54) scores. Conclusions The IRT-based short-form depression measures were reliable, valid, and responsive in patients with major depressive disorder. Effect sizes were comparable or better to other depression severity scales.