Screening for depression in older people in the acute general hospital setting : the performance of the Edinburgh Depression Scale (EDS)

• Main Author: Esiwe, Collins Iheanyichukwu
• Published: University of Leicester 2011
• Subjects: 616.85
• Online Access: https://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.548431

Depression affects approximately 25% of older people in general hospitals, yet it is often undiagnosed and under treated leading to extended lengths of stay and increased morbidity and mortality. Recent UK depression guidelines recommend screening at risk populations such as the physically ill. Several depression screening tools exist, but they include somatic symptoms that are common to both physically ill and depression. Many practitioners also find them cumbersome and time consuming for routine use. Aim The aim of this study is to assess the performance of the EDS as a brief screening tool for depression in older adults in the acute general hospital setting, and compare its performance with the Geriatric Depression Scale (GDS,4-item and 15-item) and the two questions suggested in the NICE depression guidelines (NICE-2; NICE, 2009). Methods 118 eligible older adults under the care of 4 acute medical teams at the Leicester General Hospital between October 2006 and June 2010 were assessed for depression using the Present State Examination – Schedules for Clinical Assessment in Neuropsychiatry (PSE-SCAN; WHO, 1982).Participants were also screened for depression with the EDS, GDS and the two NICE questions .The PSE-SCAN determined ICD-10 diagnosis of depressive episode acted as „gold standard‟. The test characteristics of the screening instruments were calculated at various cut-offs. Receiver Operating Characteristic (ROC) curves enabled comparisons of the various scales and determined the most appropriate cut-off threshold. Results The estimated prevalence of ICD-10 depressive episode was 22%.All scales had comparable sensitivity (80 – 100%), specificity (70 – 86%), and negative predictive values (93-100%).Areas under the ROC curves for the scales were comparable(range 0.85 -0.92) with no statistically significant difference between them. Conclusion The EDS and the two NICE questions performed just as well as other established screening tools and may be utilized in screening for depression in this population.