Beck Depression Inventory II: determination and comparison of its diagnostic accuracy in cardiac outpatients

Abstract

Objectives: To evaluate the impact of covariates on performance accuracy of the Beck Depression Inventory II (BDI-II) and to determine the optimal cut-off score for the BDI-II in cardiac outpatients. Differences of optimal cut-off scores were also verified across covariate subgroups.

Design and setting: Prospective cross-sectional study at the Department of Nuclear Medicine of the Montreal Heart Institute (Quebec, Canada).

Methods: A total of 750 adult cardiac outpatients (mean ± SD age 58 ± 10 years, 31% women) completed the BDI-II and the Primary Care Evaluation of Mental Disorders (PRIME-MD; a psychiatric interview used as the reference standard for determining diagnosis of major depressive disorder). The receiver operating characteristics (ROC) curve of the BDI-II was adjusted for age, sex, level of education, smoking status, obesity, anxiety disorder, psychotropic medication, and history of coronary artery disease. The ROC analyses were conducted to determine optimal cut-off scores.

Results: Forty-two (6%) patients met criteria for current major depressive disorder according to the PRIME-MD. After adjusted for covariates, the area under the ROC curve was significantly smaller than the unadjusted curve (0.76, 95% CI 0.66 to 0.85 vs. 0.84, 95% CI 0.77 to 0.89; ΔAUC = -0.07, 95% CI -0.13 to -0.02). While the optimal cut-off score was 10 for the total sample (sensitivity 83%, specificity 73%), the analyses indicated different cut-off scores across covariate subgroups: e.g. sex (women 13; men 10), and anxiety disorders (yes 15; no 10).

Conclusions: BDI-II is a good screening instrument for depression in cardiac outpatients. However, the present results suggest that covariates can affect the classification accuracy of the BDI-II's original recommended cut-off score. Scholars and clinicians should be aware of the principle that a screening score established in one population may not be relevant to another.

Keywords: Coronary artery disease; depression; diagnostic accuracy; sensitivity; specificity.