Depression Case Finding in Individuals with Dementia: A Systematic Review and Meta‐Analysis

Objectives To compare the diagnostic accuracy of depression case finding tools with a criterion standard in the outpatient setting among adults with dementia. Design Systematic review and meta‐analysis. Setting Studies of older outpatients with dementia. Participants Elderly outpatients (clinic and long‐term care) with dementia (N = 3,035). Measurements Prevalence of major depression and diagnostic accuracy measures including sensitivity, specificity, and likelihood ratios. Results From the 11,539 citations, 20 studies were included for qualitative synthesis and 15 for a meta‐analysis. Tools included were the Montgomery Åsberg Depression Rating Scale, Cornell Scale for Depression in Dementia (CSDD), Geriatric Depression Scale (GDS), Center for Epidemiologic Studies Depression Scale (CES‐D), Hamilton Depression Rating Scale (HDRS), Single Question, Nijmegen Observer‐Rated Depression Scale, and Even Briefer Assessment Scale–Depression. The pooled prevalence of depression in individuals with dementia was 30.3% (95% CI = 22.1–38.5). The average age was 75.2 (95% CI = 71.7–78.7), and mean Mini‐Mental State Examination scores ranged from 11.2 to 24. The diagnostic accuracy of the individual tools was pooled for the best‐reported cutoffs and for each cutoff, if available. The CSDD had a sensitivity of 0.84 (95% CI = 0.73–0.91) and a specificity of 0.80 (95% CI = 0.65–0.90), the 30‐item GDS (GDS‐30) had a sensitivity of 0.62 (95% CI = 0.45–0.76) and a specificity 0.81 (95% CI = 0.75–0.85), and the HDRS had a sensitivity of 0.86 (95% CI = 0.63–0.96) and a specificity of 0.84 (95% CI = 0.76–0.90). Summary statistics for all tools across best‐reported cutoffs had significant heterogeneity. Conclusion There are many validated tools for the detection of depression in individuals with dementia. Tools that incorporate a physician interview with patient and collateral histories, the CSDD and HDRS, have higher sensitivities, which would ensure fewer false‐negatives.