Limitations for Interpreting Failure on Individual Subtests of the Montreal Cognitive Assessment

Background: The Montreal Cognitive Assessment (MoCA) is sensitive to mild forms of cognitive impairment in geriatric populations and asks questions under the subheadings visuospatial/executive, naming, attention, language, abstraction, delayed recall, and orientation. This study examined the extent to which these subsets of MoCA items evaluate their intended cognitive domains. Methods: Clinical data from 185 geriatric memory clinic outpatients who underwent cognitive screening and subsequent neuropsychological assessment were analyzed. Factor analysis of their neuropsychological test scores identified 5 cognitive domains memory, language, visuospatial ability, attention/processing speed, and cognitive control. Scores on MoCA subtests were examined for their correlations with individual factor scores and for their sensitivity and specificity in predicting impairment within each domain. Results: The MoCA subtest scores correlated significantly but modestly with neuropsychological test factor scores in their corresponding domains, for example, the correlation between 5-word recall and the memory factor was 0.46. However, subtest scores were poor predictors of impaired performance on the tests contributing to each cognitive domain. The best predictive accuracy was seen for the visuospatial/executive subtest that showed fair accuracy at predicting impairment on tests in the visuospatial domain. Other subtests showed unacceptably poor levels of accuracy when predicting impaired scores in their respective domains (60%-67%). Conclusions: In a sample of geriatric outpatients referred for cognitive assessment, performance on individual items and subtests of the MoCA yields insufficient information to draw conclusions about impairment in specific cognitive domains as determined by neuropsychological testing.