Evaluation of a Japanese version of the Mini‐Mental State Examination in elderly persons

Abstract

This study aimed to examine the validity and item-response characteristics of the Mini-Mental State Examination (MMSE), which is used for assessing cognitive function, in Japanese older adults. Factor analysis and item response analysis were carried out for MMSE responses (n = 1971) from older adults living in the community (n = 1339) or in a nursing home (n = 632), including Alzheimer-type dementia (n = 330), vascular dementia (n = 36), frontotemporal dementia (n = 7), mixed Alzheimer-type and frontotemporal type dementia (n = 27), and age-related cognitive decline (n = 29). When choosing the cut-off score of 23 points for the MMSE, sensitivity and specificity for each item were calculated. A three-factor solution was found to be most appropriate by factor analysis: complex processing, simple processing and working memory. The item characteristics curves showed unidimensionality with high reproducibility. We identified a simplified scale comprising 10 items in all participants: "naming", "three-step command", "registration", "repeat a sentence", "write a complete sentence", "copies drawing of two polygons", "orientation to place", "delayed recall", "orientation to time" and "serial sevens" tasks. Sensitivity and specificity for both "year" task and "day" task were more than 90% ("year": sensitivity 92.5%, specificity 96.3%; "day": sensitivity 92.4%, specificity 91.7%). For the Alzheimer-type dementia patients, the five-factor solution was suggested by factor analysis and the MMSE also had unidimensionality in terms of level of difficulty. We found that the MMSE had multiple cognitive areas. We showed that the MMSE could be used as an essentially unidimensional measure of cognitive ability and the question about orientation to time might be useful in the simplest assessment to identify cognitive dysfunction.