Use of the Geriatric Depression Scale in dementia of the Alzheimer type.

Abstract

The Geriatric Depression Scale (GDS) has been shown to be an effective screening test for depression in selected geriatric populations. However, it has not been evaluated as a screening test for depression among elderly adults with dementia of the Alzheimer type. Over a two-year period 283 patients were seen in a geriatric assessment center and were screened for depression using the Geriatric Depression Scale. They also received a clinical psychiatric diagnosis by one of two geropsychiatrists. Patients with a Clinical Dementia Rating (CDR) of 0 (cognitively intact) (n = 70) and those with mild Alzheimer's disease (CDR of 1) (n = 72) were selected for comparison. The data were analyzed using Receiver Operating Characteristic Curves (ROCs) in order to compare the utility of the Geriatric Depression Scale in these two groups. ROC curves, which plot sensitivity against false positives, have come into increasing use as a method of examining the clinical performance of tests. The area lying beneath the curve (AUC) can be estimated and used as a quantitative measure of test performance (equivalent to the Wilcoxon rank sum). In the intact group, the Geriatric Depression Scale produced a ROC curve with an AUC of 0.85 (percent score = 1), which is significant (z = 7.28, P less than .0001). In the group composed of those with Alzheimer's disease, the Geriatric Depression Scale yielded a ROC curve with an AUC of 0.66, which was not significantly different from chance (z = 1.92, P = NS). This study provides empirical evidence that while the Geriatric Depression Scale is an accurate screening test for depression in cognitively intact geriatric populations, it does not maintain its validity in populations that contain large numbers of patients with dementia of the Alzheimer type.