LAUSR

OBJECTIVE The primary aim of the study was to determine accuracy, sensitivity and specificity of the Clock Drawing Test (CDT) in detecting probable dementia as compared to the multi-domain dementia screening test DemTect. METHODS The sample was derived from the general practitioner (GP)-based, cluster-randomized controlled intervention trial DelpHi-MV (Dementia: life- and person-centered help in Mecklenburg-Western Pomerania). Selected from 6.440 patients systematically screened for dementia in primary care, we examined three groups (a,b,c) where the CDT (as index test) as well as the DemTect (as reference standard) were available. After excluding cases with missing values, we included a sample of n=462 with "probable dementia", n=586 with "mild cognitive impairment" and n=553 with "no cognitive impairment" matched for age and gender. We analyzed the accuracy of the CDT in identifying people with probable dementia by the DemTect and report sensitivity, and specificity for the CDT. We further analyzed age and gender differences associated with the groups. RESULTS In comparison to the DemTect the CDT identified more than twice as many of the screened patients as cognitively impaired (63.1% in the CDT vs. 28.9% in the DemTect). The sensitivity and specificity for the CDT were 84.4% and 45.6% respectively. We found considerable age and gender differences for the performance of the CDT. Higher age (p < 0.001) and female sex (p < 0.001) were associated with incorrect clock drawings. CONCLUSION The CDT shows a considerably high rate of false positive screening outcomes compared to the DemTect and disadvantages older people and women. Thus, in contrary to previous findings our results indicate that the CDT should not be used as exclusive instrument to screen for probable dementia in primary care.