간편형 단서 회상 검사의 신뢰도와 타당도에 관한 연구 = The Reliability and Validation of The Short Cued Recall Test

다국어 초록 (Multilingual Abstract)

Objectives: We developed two parallel short cued recall tests(each SCRT A and SCRT B) and then carried out a reliability and validation study within a dementia-clinic setting using the tests. Our study aimed to assess its usefulness as a screening tool for memory problems related to dementia, especially the Dementia of Alzheimer`s Type. Methods: All of the subjects were community-dwelling Korean elderly aged 65 and older. The subjects consisted of 337 who were 109 normal controls, 116 patients with mild cognitive impairment, and 112 dementia patients. A psychiatrist with expert in dementia research made the diagnosis of dementia by doing clinical interviews and reviewing neuropsycholgical testing data of the subjects. Severity of dementia was staged using the extended version of Korean Clinical Dementia Rating(CDR) scale. All of the normal elderly were CDR 0 and all of the dementia patients were CDR 1 or 2. A translated Beck depression Inventory was administered to rule out pseudodementia due to depression. Equivalentform reliability coefficients and Test-retest reliability coefficients were calculated. To determine the construct validity of the SCRT, the participants received not only the SCRT but also the standardized Korean-version of Neuropsychological Assessment Battery CERAD(CERAD-K-N). To compare performances on the SCRT among three groups, Multivariate Analysis of Variance(MANOVA) was used. To measure the diagnostic accuracy of the K-MIS for dementia, the receiver operator characteristic(ROC) analysis was used. Results: Three groups did not show significant differences in demographic variables. Significant correlations were found between two alternative forms and between test-retest scores of SCRT. The SCRT scores were significantly correlated with memory measures of the CERAD-K-N(r= .540-.609, p<.01). According to MANOVA results, the performances on the SCRT were significantly differentiated among three groups(Walks Lamda=.134, F(4,316)=510.55, p<.001). The area under the ROC curve(AUC) of each SCRT A and B was .871 and .934, indicating the overall diagnostic accuracy of the SCRT is fairly high. The optimal cut off score for detecting dementia on the SCRT was 5/6. Conclusion: Our results indicate that the SCRT is a very brief and valid memory screening instrument for evaluating memory deficits of dementia patients while showing high sensitivity and specificity.