This study investigates the extent to which the diagnostic performance of the Center for Epidemiologic Studies Depression Scale (CES-D; both 20- and 10-item versions) varies with cognitive status, and whether the same threshold can be applied regardless of cognitive status. Three hundred and ninety-six persons aged 60+ referred for psychiatric assessment were broken down into four groups depending on age (< 70 and ≥ 70) and dementia status (mild vs. none). All were independently interviewed using the CES-D before their first evaluation by a psychiatrist. Receiver operating characteristic curves showed that both versions of the CES-D produced essentially identical results, regardless of age and dementia status. Both versions were more or less robust to the effect of mild dementia but were vulnerable to the effect of age itself. Furthermore, the optimal threshold for the 20-item version varied somewhat across the different age-dementia groups, and no clear-cut threshold existed in old-old persons with dementia. On the contrary, the same threshold of 12 can be adopted for the 10-item version, regardless of age and dementia status. Compared with the full 20-item scale, the 10-item version has the added advantage of an identical threshold across age and cognitive status. Copyright © 2007 Elsevier Ireland Ltd. All rights reserved.
CitationCheng, S.-T., & Chan, A. C. M. (2008). Detecting depression in Chinese adults with mild dementia: Findings with two versions of the Center for Epidemiologic Studies Depression Scale. Psychiatry Research, 159(1-2), 44-49. doi: 10.1016/j.psychres.2007.06.023