Background: Little systematic research into the diagnostic performance of instruments used to screen for clinical depression is available for people with diabetes. The objective of this study was to compare performances of the HADS and BDI-SF and their components in association with a standard diagnostic interview.
Methods: In a sample of 298 French outpatients from a diabetes clinic (165 men, aged 59.4 + or - 10.7 years), we assessed diagnoses of clinical depression (CD, n=42) and major depression (MD, n=30) using the MINI and administered the HADS and BDI-SF.
Results: Cognitive symptoms from the BDI-SF (BDIcog) were more closely associated with MD than CD. BDIcog and HADS total scores performed best overall in identifying clinical depression (AUCs under ROC curve 85%). For identification of CD, the sensitivity/specificity of BDI cognitive symptoms was 88/71% (cutoff 3+) and for the HADS 83/65% (cutoff 13+). For identification of MD, BDIcog scored 83/80% (cutoff 4+) and HAD-A 80/76% (cutoff 9+). Logistic regression analyses further suggested that BDIcog and HAD-A discriminated between depressed and non-depressed patients better than the somatic and anhedonia items present in the same scales. The depression subscale of the HADS performed poorly.
Limitations: The consecutive nature of the sample may limit the generalizability of our findings.
Conclusion: Results suggest that, in addition to depressed mood, both negative thoughts and anxiety are core elements for the correct identification of clinical depression in chronic illnesses such as diabetes. It may be more appropriate to use the total score when applying the HADS and distinguish non-somatic symptoms within the BDI.
Copyright 2009 Elsevier B.V. All rights reserved.