The diagnostic accuracy of the Montreal Cognitive Assessment in inpatient stroke rehabilitation.

The Montreal Cognitive Assessment (MoCA) is a commonly used screening measure for cognitive impairment; however, the diagnostic accuracy and optimal cutoff points in inpatients with mild stroke severity is unknown. We examined the diagnostic accuracy of the MoCA in an acute inpatient stroke rehabilitation unit (N = 95). The criterion neuropsychological assessment was the 30-minute National Institute of Neurological Disorders and Stroke-Canadian Stroke Network battery, modified to include the Symbol-Digit Modalities Test and Trail Making Test A & B. The MoCA had moderately strong diagnostic accuracy in receiver operating curve analyses, with areas under the curve ranging from .80 to .89 depending on the threshold for defining cognitive impairment. Sensitivity ranged from .72 to .87, and was generally greater than specificity, which ranged from .60 to .81. The optimal cutoff on the MoCA for detecting mild or greater cognitive impairment was <25/30. The optimal cutoff using more conservative definitions of cognitive impairment ranged from <23-24/30. Exploratory analyses of MoCA subgroups ("normal," "mildly impaired," and "functionally impaired") differed in the frequency and magnitude of impairment on the criterion neuropsychological assessment. These findings inform the clinical use of the MoCA in individuals with mild stroke in an inpatient rehabilitation setting.