SUMMARYThe clock-drawing test was used in a longitudinal study of 183 dementing individuals and their caregivers. Clockdrawing performance was measured by a simple standardized score which correlated significantly with other measures of cognitive function. Clock performance showed high individual consistency of performance as well as a significant deterioration from the initial level of performance to that obtained at 1-year follow-up. Dementing individuals who had experienced a significantly greater decline in clock-drawing performance at 1 -year follow-up were more likely to have caregivers who had already decided to institutionalize them. This suggests that the caregiver's decision to institutionalize was based in part on the perception of a rapid decline in their dependant's cognitive function. Thus, rate of change in cognitive function may prove to be as important a variable as the level of deterioration. The clock-drawing test appears to be a useful adjunct in the assessment and monitoring of the progressive dementias in the community.KEY woms-Clock-drawing, dementia, community, longitudinal.The utility of the clock-drawing test in the assessment of dementia has been investigated in a number of recent studies (Shulman et al., 1986;Sunderland et al., 1989;Wolf-Klein et al., 1989;Dastoor et al., 1991;Huntzinger et al., 1992;Tuokko et al., 1992). Clock-drawing appears to subsume a diverse range of higher-level cognitive abilities including visuospatial organization, integrative functions and abstract thinking. Because of the wide range of intellectual functions tapped by the clock-drawing test it is particularly suited for the purpose of assessing global cognitive function and reflecting subtle changes in overall brain functioning. As the test is still in its early stages of development, a variety of scoring systems have been used. While there is still no system that is widely accepted, Dastoor (1992) recently confirmed the test's usefulness in differentiating Alzheimer's disease from normal ageing changes. They used a more detailed scoring system dividing performance into three categories including clockdrawing, setting and reading.Clock-drawing as a measure of cognitive function is appealing for a number of practical reasons. It is easy to administer and well tolerated by patients and does not appear to be noticeably influenced by language, cultural, ethnic or educational factors. Since it provides a graphic measure of cognitive function it is useful in chart review and clinical follow-up. For clinicians and epidemiologists alike, a short easily administered test that correlates well with other more detailed and timeconsuming cognitive screens is attractive (Huntz-

Clinicians should inquire about the positive aspects of caregiving if they are to fully comprehend the caregiver experience and identify risk factors for negative caregiver outcomes.

This longitudinal study of 196 caregiver/care receiver dyads was undertaken to determine the variables predictive of caregiver decision to institutionalize a dependent with dementia. Seven variables (use of services, enjoyment of caregiving, caregiver burden and health, caregiver rating and reaction to care receiver behavior and memory problems, and presence of troublesome behaviors) predicted the decision to institutionalize. Six variables (caregiver health and burden, use of services, care receiver cognitive function and troublesome behaviors, and caregiver reaction to behaviors) predicted actual institutionalization at 18 months.