Validation of the Telephone
MoCA
in
iRBD
and Parkinson's Disease with Conversion to the Standard
MoCA
Ludo van Hillegondsberg, Tanja Zerenner, Michael Lawton, Karolien Groenewald, Jamil Razzaque, Jessica Welch, Yoav Ben‐Shlomo, Michele Hu Abstract
Background
Telephone‐based cognitive assessment may improve follow‐up when face‐to‐face testing is impractical. The Telephone Montreal Cognitive Assessment (T‐MoCA) omits visually dependent items from the standard MoCA, but its validity in isolated REM sleep behavior disorder (iRBD) and Parkinson's disease (PD), remains uncertain.
Objectives
To assess agreement between the T‐MoCA and face‐to‐face MoCA‐22 in iRBD and PD, and derive and externally validate a MoCA‐22 to MoCA‐30 conversion.
Methods
In the Oxford Parkinson's Disease Centre (OPDC) Discovery Cohort, 137 participants (38 controls, 47 iRBD, 52 PD) completed T‐MoCA and MoCA‐22 assessments. Agreement was evaluated using Lin's concordance correlation coefficient, Bland–Altman analysis, and equivalence testing with ±1‐point bounds. Equipercentile equating with log‐linear smoothing was used to derive a MoCA‐22 to MoCA‐30 conversion using Discovery data (3561 tests), with validation in Tracking Parkinson's (5767 tests) and Parkinson's Progression Markers Initiative (5047 tests).
Results
T‐MoCA and face‐to‐face MoCA‐22 scores were statistically equivalent, with minimal bias; 83.3% of scores differed by ≤1 point. However, concordance between test formats was moderate (Lin's CCC = 0.58). Across 14,375 MoCA assessments, equipercentile equating outperformed proportional scaling in development and validation cohorts. Propagating T‐MoCA–MoCA‐22 variability resulted in wider prediction intervals for T‐MoCA‐derived MoCA‐30 estimates, indicating individual‐level uncertainty, with intervals estimable only for T‐MoCA scores ≥17 due to sparse data.
Conclusions
At group level, T‐MoCA is a valid remote measure of non‐visual MoCA domains in iRBD and PD. The equating function supports cohort‐level harmonization of MoCA‐22 and MoCA‐30 scores, while individual T‐MoCA‐derived MoCA‐30 estimates should be interpreted with caution.