Reliability and validity of the Beck depression inventory in patients with Parkinson's disease

Abstract

We evaluated the validity, reliability, and potential responsiveness of the Beck Depression Inventory (BDI) in patients with Parkinson's disease (PD). In part 1 of the study, 92 patients with PD underwent a structured clinical interview for DSM major depression and based on this patients were considered depressed (PD-D) or nondepressed (PD-ND). Subsequently, patients filled in the BDI. In part 2, a postal survey consisting the BDI was performed in 185 PD patients and 112 controls. Test-retest reliability was assessed in 60 PD patients. The factor analysis revealed a cognitive-affective and a somatic factor. Cronbachs alpha for the BDI was 0.88. Mean BDI indicated significant differences (P<0.001) between the PD and control group, between the PD-ND and PD-D group, and between PD-ND and control group. In part 1, the receiver operating characteristic curves showed that the area under the curve for the total BDI was 0.88. A cutoff was calculated for the BDI (14/15) that had the highest sum of sensitivity (0.71) and specificity (0.90). In part 2, the test-retest reliability for the BDI total score was 0.89 (intraclass correlation coefficient). The smallest real difference was 3.3 for the total BDI. The BDI is a valid, reliable, and potential responsive instrument to assess the severity of depression in PD. However, an adjusted cutoff is recommended.