Assessing Alzheimer's disease patients with the Cohen-Mansfield Agitation Inventory: scoring and clinical implications.

We explored the applicability of the standard scoring of the Cohen-Mansfield Agitation Inventory (CMAI), a widely used nursing-home derived instrument, to community-dwelling persons with Alzheimer's disease (AD). Item responses to the CMAI were gathered from participants in two large clinical studies, one of which specifically included patients with behavioral disturbances. Confirmatory factor analysis in these two groups of well-characterized AD patients suggested that conventional CMAI subscoring did not adequately describe the responses of these two groups. Exploratory factor analysis indicated that the four CMAI subscores, based on a verbal-physical and aggressive-non-aggressive conceptualization of behavioral disturbance, did not fit community dwelling persons with AD. Based on cross-sectional and longitudinal analyses, there was suggestive evidence for three behavioral clusters, but these clusters did not achieve statistical significance Overall, the CMAI seemed best suited to describe the overall level rather than the specific subtypes of behavioral dyscontrol in community-dwelling persons with AD.