Objectives: To present a new version of the Modified Mini-Mental State Examination (3MS-R), provide normative information extending to individuals in the 10th decade, and examine the effects of demographic variables on test performance.
Background: The Modified Mini-Mental State Examination, based originally on the Mini-Mental State Examination, has been used to screen populations for dementia. Providing normative information and an analysis of demographic variables on test performance for this version would support broader use in clinical and other settings.
Methods: Two thousand, nine hundred thirteen elderly individuals determined to be free of dementia and other neurologic and psychiatric conditions served as subjects. An analysis of variance was conducted to examine the effects of age, gender, and education on test performance. Descriptive statistics (means, standard deviations, and percentile ranks) were calculated to summarize the range of normal performance. To examine the sensitivity/specificity of the suggested cut-off points at the 7th and 10th percentiles, two subsamples of elderly individuals, on whom clinical dementia assessments were available, were used to classify individuals with regard to dementia status.
Results: Lower age, higher education, and female gender were associated with higher 3MS-R scores. Gender effects were among the weakest, but most important at lower levels of education. Education effects were most prominent in the youngest age groups. Selection of a cut-off point at the 7th percentile revealed 69%-70% sensitivity for detecting dementia, and higher sensitivity for individuals in the youngest age groups. Specificity at this cut-off point was 89%. Raising the cut-off point to the 10th percentile improved sensitivity to 73%-76%, but reduced specificity to 85%-86%.
Conclusion: We present a version of the Modified Mini-Mental State Examination that has demonstrated utility in screening a population for dementia. An analysis of normative information and the effects of demographic influences suggest that the 7th percentile cut-off point performs very well in detecting dementia in 65-79-year-old individuals but less well for individuals in their 80s and 90s. To increase the sensitivity of the 3MS-R to detect dementia or other forms of cognitive impairment, particularly among the "old-old," the test user may wish to raise the cut-off point for impairment in some demographic groups or to supplement the test with additional cognitive measures.