Psychometrics of the Montreal Cognitive Assessment (MoCA) and its subscales: validation of the Taiwanese version of the MoCA and an item response theory analysis

Background: The Montreal Cognitive Assessment (MoCA) is an instrument for screening mild cognitive impairment (MCI). This study examined the psychometric properties and the validity of the Taiwan version of the MoCA (MoCA-T) in an elderly outpatient population. Methods: Participants completed the MoCA-T, Mini-Mental State Examination (MMSE), and the Chinese Version Verbal Learning Test. The diagnosis of Alzheimer's disease (AD) was made based on the NINCDS-ADRDA criteria, and MCI was diagnosed through the criteria proposed by Petersen et al. (2001). Results: Data were collected from 207 participants (115 males/92 females, mean age: 77.3 +/- 7.5 years). Ninety-eight participants were diagnosed with AD, 71 with MCI, and 38 were normal controls. The area under the receiver operator curves (AUC) for predicting AD was 0.98 (95% confidence interval [CI] = 0.97-1.00) for the MMSE, and 0.99 (95% CI = 0.98-1.00) for the MoCA-T. The AUC for predicting MCI was 0.81 (95% CI = 0.72-0.89) using the MMSE and 0.91 (95% CI = 0.86-1.00) using the MoCA-T. Using an optimal cut-off score of 23/24, the MoCA-T had a sensitivity of 92% and specificity of 78% for MCI. Item response theory analysis indicated that the level of information provided by each subtest of the MoCA-T was consistent. The frontal and language subscales provided higher discriminating power than the other subscales in the detection of MCI. Conclusion: Compared to the MMSE, the MoCA-T provides better psychometric properties in the detection of MCI. The utility of the MoCA-T is optimal in mild to moderate cognitive dysfunction.