Effect of education on the clock-drawing dementia screen in non-demented elderly persons

To examine the effect of education on clock-drawing ability in non-demented elderly persons.

Descriptive study of 187 elderly persons, 77 demented, 110 non-demented, 54 with 9+ years of education, 133 with 8 or fewer years of education, from three university medical center geriatric divisions.

Subjects took the Folstein Mini-Mental State Exam and were asked to draw a clock showing a time of 3 o'clock. Clocks were scored using three previously described scoring scales (Shulman, Sunderland, and Wolf-Klein). Mean scores and proportions of normal and abnormal clocks were compared for well and poorly educated non-demented subjects. Sensitivities and specificities for detecting dementia were calculated.

Mean scores of the well educated non-demented subjects were significantly better than mean scores of the poorly educated non-demented subjects on all three scales. However, proportions of abnormal clocks were not significantly different between well and poorly educated on the Wolf-Klein scale. For the poorly educated subgroup, sensitivity and specificity for detecting dementia by clock drawing were 90% and 42% by the Shulman scale, 74% and 44% by the Sunderland scale, and 48% and 90% by the Wolf-Klein scale.

Clock-drawing ability is affected by education in non-demented elderly persons. The scoring method of Wolf-Klein is least educationally affected and maximizes specificity for detecting dementia but has low sensitivity. Educational effects make clock drawing a poor single screening test for dementia in a poorly educated population.