The Short Geriatric Depression Scale

This study was undertaken to determine the test-retest reliability and the classification congruence between two popular depression screening instruments, the Short Geriauic Depression Scale (S-GDS) and the Geriatric Depression Scale (GDS). The GDS was administered to 73 independent, community dwelling older adults (mean ± SD: 62 ± 5, range 55-75 years); the S-GDS was extracted from the GDS for comparison. A subset of 49 subjects was reassessed 8 weeks after the initial assessment. There was systematic classification disagreement between the two instruments. Thus, of 15 cases categorized as depressed by the GDS, 9 (60%) were categorized as not depressed by the S-GDS. Of 7 cases categorized as depressed by the SGDS, 1 (14%) was categorized as not depressed by the GDS. Test-retest reliability was high for the GDS (r = .85), but only moderate for the S-GDS (r = .67). The current findings do not support the use of the S-GDS as a surrogate for the GDS. In light of the relatively extensive validation for the GDS and the lack of same for the S-GDS, as well as the demonstrated superior reliability of the former, prudence requires that the S-GDS not be used as a depression screening instrument pending the outcome of criterion-related validation studies.