Reliability and validity of four scoring methods of clock drawing test for screening dementia and mild cognitive impairment

Abstract

Backgrounds: In order to evaluate the suitability of the clock drawing test (CDT) for a detection of dementia and mild cognitive impairment (MCI), we assessed interrater reliabilities, correlations, and validity of four common scoring methods of the CDT.

Methods: The study subjects consisted of 524 community elderly aged between 60 and 85 years. Clock drawings were scored independently and blindly by three raters according to four different methods (Freedman, Rouleau, Todd, and CERAD-CDT). Subjects were assessed of their cognitive function and activity of daily life. Dementia was assessed by a psychiatrist using DSM-IV or NINCDS-ADRDA criteria. Sensitivity and specificity of the each scoring methods was estimated. By comparing area under the curve, the diagnostic efficiency of the CDT methods for dementia and MCI were evaluated.

Results: Inter?rater and intra-rater reliabilities were high for all four scoring methods. The area under the curve was 82.0%-89.7% for the dementia and 65.3%-66.9% for the mild cognitive impairment. Sensitivity of clock drawing test was between 0.75 and 0.85 for dementia and 0.41 and 0.85 for mild cognitive impairment, and specificity was between 0.80 and 0.84 for dementia and 0.43 and 0.83 for mild cognitive impairment.

Conclusion: The reliability and validity of CDT for dementia and MCI in the elderly living in the community was not high. Therefore, CDT is not recommended to be used alone to screen for possible prodromal stages of dementing illnesses and dementia.