A normative calculator for MoCA domain scores: proxy for Z-scores of conventional neuropsychological tests.

Abstract

Background: Test items in MoCA (Montreal Cognitive Assessment) can be used to generate 5 domain scores (i.e. Memory, Language, Attention, Executive and Visuospatial) which have been shown to approximate well-established neuropsychological tests. As neuropsychological tests are known to be affected by age, sex, education, and language of administration, this study derived a regression-based Z-score calculator for MoCA Domain Scores (MDS) that adjusts individual performance for these key confounders; with the intention of improving the clinical utility of MDS as a proxy for conventional neuropsychological tests.

Methods: Participants ≥ 50 years were recruited from Alzheimer's Disease Centers across USA (n = 25,330), and completed MoCA and conventional neuropsychological tests. A subset with normal cognition and global Clinical Dementia Rating of 0 (n = 11,371) was used to derive the Z-score calculator for MDS; while the full sample (n = 25,330) verified the performance of MDS Z-scores in detecting domain-specific impairments (as defined by conventional neuropsychological tests), using areas under the receiver operating characteristic curve (AUC).

Results: MDS varied significantly by age, sex, education, and language of administration even among participants with normal cognition. Based on age-, sex-, education-, and language-adjusted Z-scores, the respective AUCs were 91.2% for MoCA-Memory (95%CI 90.7-91.6), 83.6% for MoCA-Language (95%CI 83.0-84.3), 88.7% for MoCA-Attention (95%CI 88.0-89.4), 85.5% for MoCA-Executive (95%CI 84.8-86.1), and 81.0% for MoCA-Visuospatial (95%CI 80.2-81.8). At the commonly-used cut-off of Z-scores ≤ -1.50, all the MDS had specificities of ≥ 80%.

Conclusions: MDS Z-scores can be easily computed using the newly-developed Excel-based calculator, and provide a viable alternative when conventional neuropsychological tests are needed but cannot be feasibly administered, such as in non-specialty clinics with large volume of patients at high-risk of cognitive impairment (e.g. primary-care, geriatric, and stroke-prevention clinics), and, with further validation and calibration, plausibly also in other resource-limited healthcare settings (e.g. in lower- and middle-income countries). They can complement neuropsychological tests as part of the systematic evaluation of cognitive impairment, and help reserve neuropsychological tests for patients most likely to benefit from further evaluation.