Short Montreal Cognitive Assessment

Abstract

The diagnostic accuracy of the short Montreal Cognitive Assessment (s-MoCA), a cognitive screening instrument recently derived by item response theory and computerized adaptive testing from the original MoCA, for the diagnosis of dementia and mild cognitive impairment (MCI) was assessed in 2 patient cohorts referred to a dedicated memory clinic in order to examine the validity and reproducibility of s-MoCA. Diagnosis used standard clinical diagnostic criteria for dementia and MCI as reference standard (prevalence of cognitive impairment = 0.43 and 0.46 in each cohort, respectively). There were significant differences in s-MoCA test scores for dementia, MCI, and subjective memory impairment (P ≤ .01), and s-MoCA effect sizes (Cohen d) were medium to large (range: 0.65-1.42) for the diagnosis of dementia and MCI. Using the cut-off for s-MoCA specified in the index study, it proved highly sensitive (>0.9) for diagnosis of dementia but with poor specificity (≤0.25), with moderate sensitivity (≥0.75) and specificity (≥0.60) for diagnosis of MCI. In conclusion, in these pragmatic diagnostic test accuracy studies, s-MoCA proved acceptable and sensitive for the diagnosis of cognitive impairment in a memory clinic setting, with a performance similar to that of the original MoCA.