Validation of 2 Depression Screening Tools in Dialysis Patients

Abstract

Depression is the most common psychiatric disorder in long-term dialysis patients and is a risk factor for morbidity and mortality. An efficient and valid method of diagnosing depression might facilitate recognition and treatment. We sought to validate 2 depression assessment tools, the 21-question Beck Depression Inventory (BDI) and the 9-question Patient Health Questionnaire (PHQ-9), in a dialysis population. We surveyed patients who had received dialysis for at least 90 days in Portland, OR. We excluded patients with dementia, delirium, or a history of major psychiatric disorders other than depression. The Structured Clinical Interview for Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, a gold-standard measure for depression, was administered by trained psychologists within 2 weeks of the BDI and PHQ-9. Of 62 enrolled subjects, 16 were diagnosed with a depressive disorder, including 12 patients (19%) with major depression, 3 patients with dysthymia, and 1 patient with minor depression. Optimal BDI and PHQ-9 cutoff values for depressive disorders combined was 16 or greater and 10 or greater, respectively. Sensitivities were 91% and 92%, specificities were 86% and 92%, positive predictive values were 59% and 71%, and negative predictive values were both 98%, with kappa values of 0.65 and 0.75, respectively. The difference between the 2 receiver operating characteristic curves was not statistically significant (P > 0.9). Our results validate the PHQ-9 and revalidate the BDI against a gold-standard measure for depressive disorders in the dialysis population. Both tools performed equally well. Because depression is prevalent, readily diagnosed, and associated with poor outcomes, screening by means of short and valid measurement tools may lead to better diagnosis and treatment of this modifiable risk factor. This may lead to improved clinical outcomes in dialysis patients.