MODIFICATION OF THE MINI-MENTAL-STATE-EXAMINATION FOR USE WITH THE ELDERLY IN A NONWESTERN SOCIETY .2. CUTOFF POINTS AND THEIR DIAGNOSTIC VALIDITIES

Diagnostic power of the Korean version of the Mini-Mental State Examination (MMSE-K) (Park and Kwon, 1990) for DSM-III-R dementia was studied in 406 persons aged 60 years or more from three sources. Psychiatric clinic patients (N = 177) were diagnosed using standard clinical procedures, laboratory tests and psychological tests; in residential home elderly (N = 128) consensus diagnoses were reached by two psychiatrists after joint interviews using the CAMDEX (Roth et al., 1988) including physical examinations but without laboratory tests; and members of psychiatric patients' families (N = 101) were diagnosed after a brief psychiatric interview and assessment of activities of daily living by one psychiatrist. At the cutoff point of 24/23 of MMSE-K score, sensitivity was 92.0% and specificity 91.5%. 23.9% of demented people and 39.3% of non-demented obtained similar scores between 20 and 26. With the introduction of the arbitrary criteria of 'questionable dementia' (MMSE-K score 21-24) the false positive rate was reduced to 1% and the false negative rate to 3%. 74% of males with 'questionable dementia' scores were clinically demented while only 12% of females with the same scores were demented.