Factor structure and diagnostic efficiency of the BDI-II in treatment-seeking substance users

• Published in: Drug and alcohol dependence 2008-03, Vol.93 (3), p.271-278
• Main Authors: Seignourel, Paul J, Green, Charles, Schmitz, Joy M
• Format: Article
• Language: English
• Subjects: Addictive behaviors; Adolescent; Adult; Adult and adolescent clinical studies; Aged; BDI-II; Biological and medical sciences; Depression; Depression - diagnosis; Depression - psychology; Depressive Disorder - diagnosis; Depressive Disorder - psychology; Depressive Disorder, Major - diagnosis; Depressive Disorder, Major - psychology; Diagnostic efficiency; Diagnostic systems; Drug addiction; Factor analysis; Factor Analysis, Statistical; Female; Humans; Male; Medical sciences; Middle Aged; Mood disorders; Patient Acceptance of Health Care - psychology; Psychiatric Status Rating Scales - statistics & numerical data; Psychiatry; Psychology. Psychoanalysis. Psychiatry; Psychopathology. Psychiatry; Reproducibility of Results; ROC Curve; Screening; Sex Characteristics; Socioeconomic Factors; Substance abuse disorders; Substance use disorders; Substance-Related Disorders - diagnosis; Substance-Related Disorders - psychology; Toxicology
• ISSN: 0376-8716; 1879-0046
• DOI: 10.1016/j.drugalcdep.2007.10.016

Abstract Screening instruments are needed to rapidly and efficiently identify depression in drug and alcohol users. The Beck Depression Inventory-Second Edition (BDI-II) has excellent validity in psychiatric, normative and primary care samples, but its diagnostic efficiency has not been examined in substance users. Using a large sample of treatment-seeking substance users and the Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders-Fourth Edition (DSM-IV) to establish diagnosis, we examined the factor structure of the BDI-II and its ability to detect clinical depression, defined as the presence of a Major Depressive Disorder (MDD) or substance-induced depression. A three-factor structure provided the best fit, and the diagnostic efficiency of the BDI-II was moderate, and independent of gender and substance of abuse. Subscores had lower diagnostic efficiency than the BDI-II total score. A range of possible cutoff scores with corresponding sensitivity, specificity, positive predictive power and negative predictive power is provided to aid clinicians and researchers in choosing the optimal parameters for their screening needs.