Adapting Mini-Mental State Examination for dementia screening among illiterate or minimally educated elderly Chinese

Background Illiteracy is prevalent among current elderly Chinese. There are few brief cognitive tests in Chinese designed to screen those possibly demented for more detailed evaluation in a clinical setting. Objectives The present study adapted the Mini-Mental State Examination (MMSE) for screening dementia among illiterate or less educated elderly Chinese. Methods Literacy-dependent items of the MMSE were modified or substituted by equivalent items that are not literacy-dependment. Some items were modified to provide socio-cultural compatibility. After developing it, the Chinese adapted MMSE (CAMSE) was administered to 370 elderly outpatients from Xijing hospitals located in Xi'an, China, 93 of whom were found to be demented and 277 non-demented. Sensitivities and specificities for detecting dementia were evaluated by adjusting for different CAMSE cut-off points. The optimal cut-off points of 22 for literates and 20 for illiterates yielded a sensitivity of 83.87% and a specificity of 84.48%. Corresponding positive predictive value (PPV) was 0.65, and negative predictive value (NPV) was 0.94. The impact of literacy on CAMSE and individual test items was also evaluated. Illiterate subjects got a higher CAMSE total score than literate subjects (p < 0.05). Only one out of 12 test items, serial sevens, was negatively influenced by illiteracy (p < 0.01). After an interval of 4-6 weeks, 32 randomly selected subjects were retested with CAMSE. The test-retest reliability for total scores was 0.75 (p 0.01). Conclusions Results suggest that in the socio-cultural context for Chinese, irrespective of their literacy skills, CAMSE proved feasible for use in clinical settings for dementia screening. Copyright (C) 2003 John Wiley Sons, Ltd.