To settle the best cutoffs for inventories to diagnose depression in chronic hepatitis C (CHC) patients.

Seventy-five CHC patients were assessed using a standard psychiatric interview (Mini International Neuropsychiatric Interview) to establish Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) major depression diagnosis and the following inventories: the Beck Depression Inventory (BDI), its short form for Primary Care (BDI-PC) and the Hospital Anxiety and Depression Subscale for Depression (HAD-D). According to the receiver operative characteristic (ROC) curve, the best cutoff to screen for depression was settled, and sensitivity, specificity, and positive and negative predictive values were calculated. Agreement between each inventory and the diagnosis of depression was calculated through Cohen's Kappa. Internal consistency was measured through Cronbach's alpha.

Twenty-one (28%) individuals met criteria for a depressive disorder. BDI, BDI-PC and HAD-D best scores were settled at 18, 5 and 8 points, respectively. They showed the following discriminative properties: sensitivity 81%, 76.2%, 85.7%; specificity 92.6%, 88.9%, 77.8%; positive predictive value 81%, 72.7%, 60%; negative predictive value 92.6%, 90.6%, 93.3%, respectively. ROC curve areas were similar between BDI and BDI-PC, but slightly lower for HAD-D. The agreement between inventories and DSM-IV depression diagnosis was substantial for BDI (0.91) and BDI-PC (0.91) and moderate for HAD-D (0.86). Internal consistency was 0.90, 0.86 and 0.75, respectively.

BDI, BDI-PC and HAD-D showed good discriminative properties to screen for depression in CHC patients and should be considered in clinical practice.