The Montreal Cognitive Assessment and the Mini-Mental State Examination as Screening Instruments for Cognitive Impairment: Item Analyses and Threshold Scores

Abstract

Objective: This study was performed to provide a detailed analysis of the MoCA versus the MMSE, including an item analysis and an examination of threshold scores appropriate for use in different clinical settings. Methods: 135 subjects enrolled in a longitudinal clinicopathologic study were administered the MoCA and MMSE. Subjects were classified as cognitively impaired or cognitively normal based on neuropsychological testing and consensus conference diagnosis. Results: 89 subjects were cognitively normal, 46 cognitively impaired (20 dementia, 26 MCI). ROC analysis showed that, for any threshold value selected for the MMSE to identify cognitive impairment, a MoCA value with better sensitivity and specificity could be identified. Recall performed best among individual items on the MMSE, and Orientation performed best on the MoCA. Overall, the best discrimination was obtained using a weighted combination of four items (2*MoCA-Orientation + MMSE-Recall + MoCA-Language + 0.5* MoCA-Visuospatial/Executive; AUC 0.94). A MoCA threshold score of 26 had a sensitivity of 98% and specificity of 52% for identifying cognitive 5 impairment. A MoCA threshold score of 21 had a sensitivity of 57% and specificity of 96%. Conclusions: The MoCA was superior to the MMSE in detecting cognitive impairment. Individual domains on the MoCA and MMSE made substantially different contributions to each instrument’s sensitivity, and a weighted subset of items from both instruments performed best in detecting cognitive impairment. A lower MoCA threshold score may be appropriate in a population with a higher prevalence of cognitive impairment such as a memory clinic.