Behaviours associated with agitation are common in people living with dementia. The Cohen-Mansfield Agitation Inventory (CMAI) is a 29-item scale widely used to assess agitation completed by a proxy (family carer or staff member). However, proxy informants introduce possible reporting bias when blinding to the treatment arm is not possible, and potential accuracy issues due to irregular contact between the proxy and the person with dementia over the reporting period. An observational measure completed by a blinded researcher may address these issues, but no agitation measures with comparable items exist.

Development and validation of an observational version of the CMAI (CMAI-O), to assess its validity as an alternative or complementary measure of agitation.

Fifty care homes in England.

Residents (N = 726) with dementia.

Two observational measures (CMAI-O and PAS) were completed by an independent researcher. Measures of agitation, functional status, and neuropsychiatric symptoms were completed with staff proxies.

The CMAI-O showed adequate internal consistency (α = .61), criterion validity with the PAS (r = .79, p = < .001), incremental validity in predicting quality of life beyond the Functional Assessment Staging of Alzheimer’s disease (β = 1.83, p < .001 at baseline) and discriminant validity from the Neuropsychiatric Inventory Apathy subscale (r = .004, p = .902).

The CMAI-O is a promising research tool for independently measuring agitation in people with dementia in care homes. Its use alongside the CMAI could provide a more robust understanding of agitation amongst residents with dementia.