Validity and reliability of a Taiwan Chinese version of the community screening instrument for dementia

Abstract

Objective: The Community Screening Instrument for Dementia (CSI-D) has been reported to be sensitive. The authors examined the reliability and validity of a Chinese (Taiwanese) version of the CSI-D for elderly patients.

Methods: Four groups were tested with the CSI-D: 31 with mild or moderate dementia; 32 non-dementia, depressed subjects; and 34 low-education, and 30 high-education normal-comparison subjects. Patients with dementia or depression were selected from outpatients of two hospitals. Testing was carried out in either the community or outpatient setting.

Results: The internal consistency, interrater reliability, and test-retest reliability of the CSI-D were good. The CSI-D was highly correlated with scores on the Mini-Mental State Examination and 10-word-list-learning task. Correlation with the Montgomery-Asberg Depression Rating Scale was not significant. Receiver Operating Characteristic analysis suggested that the CSI-D was a good instrument in differentiating dementia from depression and normal subjects with low education. It had good sensitivity and specificity. Education was related to the cognitive scores but not related to informant scores. The combination of an informant interview with a cognitive test enhanced the performance of the CSI-D.

Conclusions: The Taiwan Chinese version of the CSI-D is psychometrically sound, brief, easy to complete, and therefore suitable as a screening instrument for dementia in Taiwan. The study reinforces earlier suggestions that the informant interview yields improved validity for detecting dementia.