Non-linear demographic correction enhances the diagnostic utility of the Montreal Cognitive Assessment

BackgroundThe Montreal Cognitive Assessment (MoCA) is widely used for cognitive screening. Despite numerous studies showing that MoCA scores are affected by demographic variables including age, education, and race, the instrument is typically evaluated using a raw score or simple one-point education correction in clinical practice.ObjectiveIn this study, we comprehensively evaluate the diagnostic accuracy of the MoCA in a large cohort of individuals being evaluated for mild cognitive impairment or dementia.MethodsWe used data from the National Alzheimer's Coordinating Center (NACC) database to examine diagnostic accuracy of the MoCA, using both raw scores and Z-scores subject to non-linear demographic correction. We present comprehensive accuracy statistics, concordance with <i>APOE</i> ε4 status, and predictive validity of a mild cognitive impairment classification for the subsequent development of dementia.ResultsWe find that non-linear demographic correction modestly enhances diagnostic performance for individuals with mild cognitive impairment and dementia and results in classifying a higher proportion of <i>APOE</i> ε4 homozygous participants as individuals with dementia as compared to MoCA raw scores.ConclusionsNon-linear demographic correction may improve the diagnostic accuracy of the MoCA and increase concordance with a major known risk factor for Alzheimer's disease (i.e., <i>APOE</i> ε4 status).