A comparison of alternative approaches to the scoring of clock drawing

Although a number of scoring procedures for clock drawing have emerged in the literature, no systematic comparison of the psychometric properties of various approaches has yet been conducted on a large sample of persons over 64 years of age diagnosed with dementia or deemed cognitively intact. The present study examined the reliability and validity of five scoring approaches (Doyon, Bouchard, Morin, Bourgeois, & Cote, 1991; Shulman, Shedletsky, & Silver, 1986; Tuokko, Hadjistavropoulos, Miller, & Beattie, 1992; Watson, Arfken, & Birge, 1993; Wolf-Klein, Silverstone, Levy, Brod, & Breuer, 1989) among the 493 participants of the Canadian Study of Health and Aging who completed clock drawing and who had a final diagnosis assigned at the conclusion of a comprehensive clinical examination. Inter- and intra-rater reliabilities were highest for the Tuokko et al. method. The Tuokko and Shulman scoring procedures had the highest sensitivities and relatively low specificities. The Wolf-Klein procedure had relatively low sensitivities and high specificities. Estimated areas under receiver operating curves were relatively high for all scoring methods. However, the area under the curve for the Watson procedure was significantly lower than the other procedures. All claims to the utility of clock drawing for differentiating between normal persons over 64 years of age and those with dementia appear validated.