Background: In light of inconsistencies in cut points for identifying non-major depression, this study examined the classification efficiency of the Geriatric Depression Scale (GDS and GDS-15) total scores and individual items, and four additional depressive symptoms for identification of subthreshold, minor, or criterion depression among 166 vulnerable residents of congregate housing.
Methods: Depression (combined categories of major depressive episode, minor, or subthreshold depression) was determined by the Mini-International Neuropsychiatric Interview (MINI) diagnostic interview depression module administered by telephone to 166 older residents of congregate housing facilities who also completed the 30-item GDS and four other yes/no potential indicators of geriatric depression. Classification agreement and ROC curve analysis for the full and 15-item GDS scale scores were calculated. Individual item hit rates for MINI criterion were calculated for GDS items and four new items.
Results: GDS and GDS-15 at standard cut points had 70–75% agreement with MINI. Best sensitivity and specificity were obtained at lower than standard cut points. Some GDS Withdrawal, Apathy, lack of Vigor (WAV) and cognitive items obtained very low hit rates. New items “I just don't feel like myself” and “I feel I am a burden to others” better discriminated MINI depression than most GDS items and had good item-to-total correlations with the GDS.
Conclusions: Diagnostic criteria and GDS screen had partial agreement. Some GDS items did not adequately represent depression among functionally impaired or oldest old older adults. Feeling one is a burden and the sense of feeling “different” from usual may be useful indicators of depression among vulnerable older adults.