ABSTRACT
Objective: To find out who is more suitable to be screened by MMSE (mini-mental state examination) Method: A total of 4921 persons aged 55 or above were collected by stratification cluster sampling for urban and rural areas of Xi'an They were screened by MMSE According to different age and education levels, the tenth percentage, median, skewness and kurtosis, and the positive predict value were calculated Results: MMSE scores showed a negative and cliff distributions whether among groups of different education level, among different age groups or in the total sample Moreover, these features was more obvious in relatively younger people (less than 70 years) with higher education level (high school or college), while in those aged more than 70 years with poor education (primary school or illiterate), the distribution of MMSE score was close to normal Consequently, the rates screened out by MMSE was higher in older and poor educated people than in those younger with higher education Multiple regression showed that education level, age, sex, lose of hearing, activity in daily life were factors having influence on score of MMSE Conclusion: In screening dementia, MMSE is more suitable for poor educated older people (older than 70)