OBJECTIVE The clock-drawing test has achieved widespread clinical use in recent years as a cognitive screening instrument and a significant amount of literature relates to its psychometric properties and clinical utility. This review aims to synthesize the available evidence and assess the value of this screening test according to well-defined criteria. DESIGN A Medline and Psycho-info literature search of all languages was done from 1983 to 1998 including manual cross-referencing of bibliographies. A brief summary of all original scoring systems is provided as well as a review of replication studies. Psychometric data including correlations with other cognitive tests were recorded. Qualitative aspects of the test are also described. RESULTS Among published studies, the mean sensitivity (85%) and specificity (85%) of the clock-drawing test are impressive. Correlations with the Mini-Mental State Examination and other cognitive tests was high, generally greater than r = 0.5. High levels of inter-rater and test-re-test reliability and positive predictive value are recorded and despite significant variability in the scoring systems, all report similar psychometric properties. The clock test also shows a sensitivity to cognitive change with good predictive validity. CONCLUSIONS The clock-drawing test meets defined criteria for a cognitive screening instrument. It taps into a wide range of cognitive abilities including executive functions, is quick and easy to administer and score with excellent acceptability by subjects. Together with informant reports, the clock-drawing test is complementary to the widely used and validated Mini-Mental State Examination and should provide a significant advance in the early detection of dementia and in monitoring cognitive change. A simple scoring system with emphasis on the qualitative aspects of clock-drawing should maximize its utility.