Factor structure and diagnostic validity of the Beck Depression Inventory-II with adult clinical inpatients: comparison to a gold-standard diagnostic interview.

Little is known about the psychometric properties and clinical utility of the Beck Depression Inventory-II (BDI-II) among adult clinical inpatients, a group at high risk for major depressive disorder (MDD). Data from 1,904 adult inpatients were analyzed using confirmatory factor analysis (CFA), Cronbach's alpha, and Pearson's correlations. Receiver operating characteristic (ROC) analyses evaluating MDD diagnostic performance were conducted with a subsample (n = 467) using a structured diagnostic interview for reference. CFA of 3 previous 2-factor oblique solutions, observed in adolescent and older adult inpatient clinical samples, and 3 corresponding bifactor solutions indicated that BDI-II common item variance was overwhelmingly accounted for by 1 general factor specified to all items, with minor additional variance contributed by 2 specific factors. Analyses revealed high internal consistency (Cronbach's α = .93) and significant (p < .01) intercorrelations between the BDI-II total scale and Behavior and Symptom Identification Scale-24's Depression/Functioning (r = .79) and Overall (r = .82) subscales. ROC analyses generated low area under the curve (.695; 95% confidence interval [.637, .752]) and cutoff scores with poor sensitivity/specificity balance. BDI-II use as a screening instrument for overall depressive symptomology was supported, but MDD diagnostic performance was suboptimal. Clinicians are advised to use the BDI-II to gauge severity of depression and measure clinical changes to depressive symptomology over time but to be mindful of the limitations of the BDI-II as a diagnostic tool for adult inpatients. (c) 2014 APA, all rights reserved.