RESUMO
In order to examine elderly depression in the elderly that too weakened cognitively, and physically to check depressive symptoms reliably, the authors deviced collateral source version of elderly depression scale and tested it's validity. Sixty elderly depression patients and forty-eight old person in normal group completed the KGDS (Korean Form of Geriatric Depression Scale), and their collateral source (CS) completed a CS version of the KGDS (CS-KGDS). The differences of CS-KGDS means was signified at alpha=.001 level (patients group mean=18.33 (4.71), control group mean=7.60 (5.25), t=11.18). The values of Cronbach's alpha and Split-half reliability were .88 and .81 respectively, and correlational coefficent with KGDS was .68. On factor analysis, 5 factors were extracted. They were labeled 'Physical weakening and decreased vitality' (factor 1), 'Emotional discomfort' (factor 2), 'Negative thinking and Unhappiness feeling' (factor 3), 'Cognitive dysfunction' (factor 4), 'Decreased social interest and activity' (factor 5), which congruous with internal structure of KGDS. The result of discrimination analysis showed 87.09% of hit ratio, and suggested the score of 13 as optimal cut-off score. This cut-off score was identical to that of KGDS. Conclusively, CS-KGDS could be a useful tool for evaluaion of elderly depression instead of KGDS in such case that KGDS can not be administered reliably.