Mini Exame do Estado Mental e escolarização: propriedades psicométricas para amostras de idosos

Abstract

The Mini Mental State Examination (MMSE) is the screening tool of cognitive status most used in the world. Despite its relevance, its psychometric properties have not been sufficiently analyzed in Brazil. The main purpose of this thesis was to analyze the psychometric properties of a Brazilian version of this measure suggested by Brucki et al (2003) with Classical Theory of Tests and Item Response Theory (IRT), emphasizing education. For this, two studies were developed. First, however, there was a systematic review of empirical research indexed in SciELO database that used the MMSE with elderly (N = 74), which identified 11 versions of the MMSE. More than half of these studies adopted the school to establish cutoffs. It was found that the psychometric properties of this measure have been little investigated in the country and, consequently, lack standardization and evidence of validity. The second study aimed to obtain evidence of validity based on the internal structure for this measure by principal component analysis (PCA) and confirmatory factor analysis (CFA). A sample of 2734 elderly participants of the FIBRA Unicamp study was researched. In the PCA, a solution with five components was the most appropriate. Four internal structure models were tested with CFA. Two models with multidimensional structure of five factors and a factor of second order showed satisfactory fit indices, which did not happen with the unidimensional model. Therefore, there is evidence that it is a multidimensional measurement. The third study presents an analysis with IRT to evaluate the differential item functioning (DIF), difficulty and discrimination of the items. It was found that four items of serial seven were the most difficult and the items 9, 13, 22 and 23 were the easiest ones. The skill level of the people was higher than the level of difficulty of the items. There was DIF for education (27 items), sex (18 items) and age (16 items). We conclude that the use of MMSE with community dwelling elderly should be cautious due to the large number of biased items, especially for education. In conclusion, the version of the MMSE analyzed the version of MMSE analyzed obtained validity evidences regarding the internal structure, and is strongly influenced by education.