Objective: To develop a simple, readily administered and scored screening test for dementia utilizing the clock-drawing task.
Design: Retrospective analysis of clock-drawing errors and prospective validations.
Setting: Hospital-based outpatient geriatric assessment clinic, rehabilitation service, apartment building for older adults, and long-term care facility.
Participants: Convenience sample of patients attending the geriatric assessment clinic, patients on the rehabilitation service, or residents of the above sites.
Measurements: Sensitivity and specificity of a clock-scoring system in identifying patients with dementia and the comparison of this system with the Short Blessed Test (SBT) in the diagnosis of dementia and in the prospective validation of the test.
Results: Of the 10 clock-drawing errors evaluated, placement of digits in a pre-drawn circle had the greatest sensitivity and specificity in distinguishing patients with irreversible dementia from patients with other disorders who did not meet NINCDS-ADRDA criteria for probable dementia. The derived scoring system had a sensitivity of 87% and a specificity of 82%, compared with a sensitivity of 82% and a specificity of 88% for the SBT in identifying dementia. Test-retest reliability for the distinction between demented and non-demented was 82%, with a Kappa of 0.63 for the clock completion, and 82%, with a Kappa of 0.62 for the SBT. Inter-rater reliability for clock completion was 0.90 to 0.93.
Conclusion: A simple, completely objective scoring system for a clock completion test has been developed which involves only the number of digits placed in the fourth quadrant of a pre-drawn circle. This readily administered test is as effective in screening for dementia as the longer six-item SBT.