A large scale pragmatic validation of the HADS for major depression in an ethnically diverse cancer population
Baker-Glenn, Elena A.
MetadataShow full item record
OBJECTIVES: There is considerable interest in the accuracy of the Hospital Anxiety and Depression Scale in cancer. The HADS is the single most commonly applied tool for depression, tested to date in 22 analyses against major depressive disorder (MDD). However, only one previous study had a sample size over 300 (Walker et al. 2007 J Psychosom Res 63: 83-91). METHOD: We analysed data collected from Leicester Cancer Centre from 2007-2009 involving approximately 1000 people approached by a research nurse, research physician and two therapeutic radiographers. The researcher applied DSMIV criteria of major depressive disorder (MDD). We collated full data on 690 patient assessments of whom 12.9% had MDD, 121 were palliative and 115 (16.7%) were from ethnic minorities (largely British South Asian of India descent). RESULTS: Sensitivity and specificity were as follows HADS-A 87.6%/72.3; HADS-D 86.5%/80.1% and HADS-T 95.5%/76.3%. The AUC was highest for HADS-T > HADS-D > HADS-A. In ethnic minorities sensitivity and specificity were as follows HADS-A 93.1%/66.3%; HADS-D 86.2%/65.1% and HADS-T 96.6%/66.3%. AUC was poor for the HADS-D from the ethnic group and HADS-A/HADS-T were preferred. There was no difference by cancer stage. CONCLUSIONS: This is the largest validation study of the HADS in cancer and suggests that in an ethnically diverse population the HADS-T may be the optimal version based on accuracy alone; regardless of cancer stage. The HADS-D may have weaker validity in patients from ethnic minority populations.