Projeto Bambuí: estudo de base populacional comparando a performance do teste do relógio e do miniexame do estado mental (análise de concordância, de correlação e de fatores latentes)

Abstract

Objective: The aim of this study was to compare the performance of the Mini-mental State Examination (MMSE) and clock drawing test (CDT) in a population-based sample of elderly subjects (60 or more years of age) with low schooling level. It is presented in two articles. In the first, correlation and agreement between these tests were evaluated. In the second, MMSE factor structure and the association of MMSE latent factors with CDT scores were investigated. Methods: All members of the population-based cohort of Bambuí that were examined in the fourth follow-up (2001) and had scores on both tests were included in the present study. CDT was administered and scored by the Shulman method. Correlation between the MMSE and CDT for the total sample and according to gender, age and schooling level was estimated bySpearmans correlation coefficient and agreement was evaluated using receiver operating characteristic (ROC) analysis. An exploratory factor analysis based on the tetrachoric correlation matrix of 28 MMSE items was undertaken with unweighted least squares as the estimator. Association between MMSE latent factors and CDT scores was assessed using generalized ordered logistic regression and adjusted by gender, age and schooling level. Results: 1118 and 1113 subjects were evaluated for the first and second articles, respectively. Mean subject age was 72.4 years, 61.6% were women and about 61% of participants had less than 4 years of schooling. Correlation between the MMSE and CDT was moderate (Spearmans correlation coefficient=0.64) and no difference was found according to gender, age or schooling level. For CDT cut-off 3/4, the area under the ROC curve was 0.815 and thehighest agreement (75.1%) was found using MMSE cut-off 27/28. Subjects who performed well on the CDT could be expected to obtain high scores on the MMSE whereas CDTs with low scores had a wide range of corresponding MMSE scores. The MMSE exploratory factoranalysis yielded a four-factor solution: concentration, memory, naming and a factor comprising language, praxis and orientation. CDT scores were associated with the MMSE concentration latent factor for CDTs scored zero, strongly associated with the language, praxis and orientation latent factor for all CDT scores and not significantly associated with thememory and naming factors. Conclusions: A possible explanation for the moderate correlation found between these tests may be the pattern of association between CDT scores and MMSE latent factors. The strongassociation between CDT scores and MMSE language, praxis and orientation latent factor supports the use of the CDT as a brief test that evaluates several cognitive domains. When time for cognitive screening is limited, one could consider the possibility of administering theCDT and extending cognitive evaluation if a low score was obtained.SUMÁRIO