Review of cutoff points adjusted for schooling in a large Southern Brazilian sample

The increase in life expectancy can influence the prevalence of dementias in the population. Instruments that evaluate cognitive functions such as the Mini Mental State Examination (MMSE) are necessary for the investigation of dementia. The supposition that patient score on the MMSE can be influenced by academic level points to the need for establishing cut-off values that take into account educational level. The aim of this study was to review MMSE cut-off values adjusted for schooling in a large southern Brazilian sample. Method: Demographic data and MMSE scores of 968 subjects, of which 162 were dementia patients and 806 healthy participants, were analyzed. The sample was grouped according to education. The cut-off values were established by ROC Curve analysis. Results: The total sample mean age was 70.6±7.3 years, and the mean years of education was 7.2±5.3. The cut-off score of 23 points (sensitivity=86%, specificity=83%) was observed as the optimal level to detect dementia on the MMSE instrument for the overall sample. Regarding level of schooling, the cut- off values were: 21 for the illiterate group (sensitivity=93%, specificity=82%), 22 for the low education group (sensitivity=87%, specificity=82%), 23 for the middle education group (sensitivity=86%, specificity=87%) and 24 for the high education group (sensitivity=81%, specificity=87%). Conclusions: The cut-off values revealed by this analysis, and adjusted for level of schooling, can improve the clinical evaluation of cognitive deficits.