RESUMO
BACKGROUND: Cutoff values of cognitive screen tests vary according to age and educational levels. OBJECTIVE: The objective of this study was to compare the accuracy and determine cutoffs for 3 short cognitive screening instruments: the Mini-Mental State Examination, Montreal Cognitive Assessment (MoCA), and Quick Mild Cognitive Impairment Screen-Turkish version (Qmci-TR), in older adults with low literacy in Turkey. METHODS: In all 321 patients, 133 with subjective cognitive complaints (SCC), 88 amnestic-type mild cognitive impairment (aMCI), and 100 with probable Alzheimer disease (AD) with a median of 5 years education were included. Education and age-specific cutoffs were determined. RESULTS: For the overall population, the Qmci-TR was more accurate than the MoCA in distinguishing between aMCI and AD (area under the curve=0.83 vs. 0.76, P=0.004) and the Qmci-TR and Mini-Mental State Examination were superior to the MoCA in discriminating SCC from aMCI and AD. All instruments had similar accuracy among those with low literacy (primary school and lower educational level or illiterate). CONCLUSIONS: To distinguish between SCC, aMCI, and AD in a sample of older Turkish adults, the Qmci-TR may be preferable. In very low literacy, the choice of the instrument appears less important.