The clock drawing test and questionable dementia: reliability and validity
Objectives In order to evaluate the suitability of the clock drawing test (CDT) for the detection of questionable dementia (QD), we assessed the inter–rater reliabilities and intercorrelations of four common scoring methods of the CDT in a sample of non‐demented subjects and determined the concurren...
Saved in:
Published in: | International journal of geriatric psychiatry 2002-11, Vol.17 (11), p.1048-1054 |
---|---|
Main Authors: | , , , , |
Format: | Article |
Language: | English |
Subjects: | |
Citations: | Items that this one cites Items that cite this one |
Online Access: | Get full text |
Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Summary: | Objectives
In order to evaluate the suitability of the clock drawing test (CDT) for the detection of questionable dementia (QD), we assessed the inter–rater reliabilities and intercorrelations of four common scoring methods of the CDT in a sample of non‐demented subjects and determined the concurrent validity.
Methods
The study sample consisted of 253 discharged general hospital patients, aged between 65 and 85 years. Subjects were screened for cognitive impairment during their hospital stay. Four to six weeks after discharge each non‐demented patient was visited at home and interviewed by a trained psychologist. The interview procedure included a cognitive test battery incorporating the CDT, the Mini‐Mental State Examination (MMSE), the Syndrome Short Test (SKT), and a verbal fluency test (VF). The criteria of the Clinical Dementia Rating (CDR) were used to differentiate between cognitively normal subjects and those with QD. Clock drawings were scored independently and blindly by two raters according to four different methods. The agreement between raters was assessed, as well as the agreement between the different scoring methods. The association of gender, education, age, test performance and CDR‐rating with CDT scores was examined. Accuracy of the CDT for the detection of QD was calculated.
Results
Inter–rater reliabilities were high for all four scoring methods. However, substantial differences among the scoring methods were observed, the proportion of abnormal test results varying between 9% and 50%. The CDT correlated significantly with MMSE, SKT and VF, but correlation coefficients were low (r = 0.13 to r = 0.32). Furthermore, CDT scores were influenced by age, gender, and education. Sensitivity of the CDT for QD was 66%, specificity was 65%; the negative predictive value was 73%, the positive predictive value 58%.
Conclusion
In a sample of non‐demented elderly, the reliability of the CDT was sufficiently high, but the different scoring methods were not equivalent. When established cut‐off scores were used, the proportion of abnormal CDTs were significantly different. Concurrent validity with other common cognitive tests was unsatisfactory. The CDT lacks sufficient sensitivity and specificity for the identification of QD and should not be used alone to screen for possible prodromal stages of dementing illnesses. The association of age, gender and level of education with CDT scores should be taken into account by clinicians using the CDT. Copyright |
---|---|
ISSN: | 0885-6230 1099-1166 |
DOI: | 10.1002/gps.747 |