Screening for late life depression: cut-off scores for the Geriatric Depression Scale and the Cornell Scale for Depression in Dementia among Japanese subjects

Background Proper screening of depression among older adults depends on accurate cut‐off scores. Recent articles have recommended the Geriatric Depression Scale (GDS) and the Cornell Scale for Depression in Dementia (CSDD) for this screening. However, there has been no investigation of the sensitivity and specificity of either scale using Japanese subjects. The purpose of the present study was to identify appropriate GDS and CSDD cut‐offs for Japanese older adults. Methods The GDS and the CSDD were interview‐administered to nondepressed Japanese older adults (n = 74) and to Japanese older adults with a SCID‐IV diagnosis of major or minor depression (n = 37). Depressed subjects were also administered the Hamilton Depression Rating Scale (HDRS). Data were also collected on demographic variables, mental status, health status, and medication use. Results ROC curve analysis identified a cut‐off score of 6 for the GDS which had a sensitivity of 0.973, a specificity of 0.959, a False Positive Rate (FPR) of 0.894, and a False Negative Rate (FNR) of 0. A cutoff score of 5 for the CSDD yielded a sensitivity of 1, a specificity of 0.919, a FPR of 0.942, and a FNR of 0. Comparisons indicate current HDRS cut‐offs may overlook subthreshold depression. The GDS cut‐off score identified among Japanese subjects was the same as that reported for Western subjects. Conclusions Due to the substantial prevalence of psychiatric disorders found in false‐negative subjects, the above cut‐off scores were chosen to optimize the potential for true positives. These scores are recommended for alerting physicians and other caregivers as to when more intensive depression evaluation is needed. Copyright © 2003 John Wiley & Sons, Ltd.