Aims/hypothesis  

We compared the screening performance of different measures of depression: the standard clinical assessment (SCA); the Beck Depression Inventory (BDI); the Center of Epidemiological Studies-Depression Scale (CES-D); and the Problem Areas in Diabetes (PAID) questionnaire, which assesses diabetes-specific distress. We also studied the ability of these measures to detect diabetes-related distress.

Materials and methods  

A total of 376 diabetic patients (37.2% type 1; 23.9% type 2 without insulin treatment, 38.8% type 2 with insulin) completed the BDI and CES-D; patients who screened positive participated in a diagnostic interview, the Composite International Diagnostic Interview (CIDI). Also, all patients completed the PAID questionnaire. Results of the SCA that related to depression diagnosis were reviewed to correct for false negative screening results.

Results  

The prevalence of clinical depression was 14.1%, with an additional 18.9% of patients receiving a diagnosis of subclinical depression. Sensitivity for clinical depression in SCA (56%) was moderate, whereas BDI, CES-D and the PAID questionnaire showed satisfactory sensitivity (87, 79 and 81%, respectively). For subclinical depression, the sensitivity of the PAID questionnaire (79%) was sufficient, whereas that of SCA (25%) was poor. All methods showed low sensitivity for the detection of diabetes-specific emotional problems (SCA 19%, CIDI 34%, BDI 60%, CES-D 49%).

Conclusions/interpretation  

The screening performance of SCA for clinical and subclinical depression was modest. Additional screening for depression using the PAID or another depression questionnaire seems reasonable. The ability of depression screening measures to identify diabetes-related distress is modest, suggesting that the PAID questionnaire could be useful when screening diabetic patients for both depression and emotional problems.