Clock-drawing test in vascular mild cognitive impairment: Validity of quantitative and qualitative analyses.

Introduction: The clock-drawing test (CDT) has been used as a screening tool to identify cognitive deficit in patients with dementia. However, it has not been extensively evaluated for categorizing patients with vascular mild cognitive impairment (vMCI). This study aimed to examine the discrimination of vMCI using various CDT scoring methods.

Method: A total of 120 vMCI patients and 119 normal control (NC) subjects were tested using three CDT quantitative scoring systems: the one from the Montreal Cognitive Assessment (MoCA) (CDT3) and the systems of Rouleau (CDT10) and Babins (CDT18). We used a revised scoring method to evaluate the effectiveness in differentiating vMCI patients from NC subjects, which combined the CDT10 quantitative score and three qualitative errors with a significantly higher prevalence in vMCI group (called hereinafter CDTcomb, including CDTcomb13 and CDTcomb16 based on different weights of the three error types). The sensitivity and specificity of the CDT methods were determined by the receiver operating characteristic (ROC) curve. The results of the scoring systems were compared with those of the Mini-Mental State Examination (MMSE).

Results: The results of the ROC analyses with the CDT3, CDT10, and CDT18 systems produced a sensitivity of 71.1%, 81.8%, and 60.3%, and a specificity of 66.12%, 58.68%, and 73.55%, respectively, for the diagnosis of vMCI. Compared with the separate MMSE score, the combination of MMSE with the CDT3, CDT10 and CDT18 scores did not increase the sensitivity and specificity. When three qualitative errors were incorporated into the CDT10 quantitative score, CDTcomb13 and CDTcomb16 provided a sensitivity of 87.6% and 86.78%, and a specificity of 74.79% and 80.67%, respectively, in differentiating vMCI patients from the NC group.

Conclusion: Our findings suggest that the combination of CDT quantitative score with qualitative observations of the clock-drawing errors can provide a better discrimination between vMCI patients and cognitively normal subjects.