The Minilinguistic State Examination for Primary Progressive Aphasia: Validity evidences for the Argentinean version

Abstract

Background: This work presents a new brief language assessment tool for Primary Progressive aphasia: the Minilinguistic State Examination (MLSE). It has been developed with the intention to have equivalent version in different countries. We translated and adapted the test to Argentinean Spanish. Here we present the psychometric properties of the Argentinean Spanish version of the MLSE. Method: The sample comprised 29 aphasic patients from different etiologies and 94 healthy community-dwelling adults. All of them were native Argentinean Spanish speaking. The Argentinean MLSE was administered as well as the following equivalent subtests with validated Argentinean versions in order to analyze the concurrent validity through a Spearman´s Rho correlation coefficient: the Boston Aphasia Battery, the BEA (Batería de Evaluación de la Afasia), Pyramids and Palm tree test and Repeat and Point. Score´s reliability was analyzed according to K-R 20 coefficient for each kind of possible error: motor, phonologic, semantic and working memory. Three independent clinicians scored a sub-group of 16 patients in order to determine interraters reliability. Krippendorff´s alpha was calculated. Result: All correlations with the equivalent tasks were statistically significant except for semantic association task that was near statistical significance (p = .061). Score´s reliability for motor errors was .825, for phonological errors .889, for semantic .841 and for working memory .874. Krippendorff´s alpha obtained for the total MLSE score was .877, while the following values were obtained for each type of error identified in the MLSE: motor = .398; phonological = .707; semantic = .729; syntactic = .831; working memory = .539; omissions = .672. Conclusión: All these analyses provide evidence of validity of the scores of the test and suggest that it can be used with confidence by clinicians. They also showed in general acceptable interrater reliability. However, some clarifications should be added to the motor and working memory scoring guide in order to avoid interrater´s divergencies, as the obtained values were low.