The clinical utility, reliability and validity of the Rivermead Behavioural Memory Test—Third Edition (RBMT–3) in Hong Kong older adults with or without cognitive impairments

Hundred older adults residing in community dwellings were recruited from three non-government organisations and completed the RBMT-3: 29 patients with mild to moderate dementia, 34 persons at risk for MCI, and 37 matched older adults with everyday functional cognition for a healthy control group (NC). The test has excellent inter-rater (ICC [2, 1] = 0.997), intra-rater (ICC [3, 1] = 0), and parallel version (ICC [3, 1] = 0.990) reliabilities, as well as satisfactory internal consistency (Cronbach’s alpha: 0.643–0.832). The scores of the MCI group were significantly lower than those of NC group in four subtests. The optimal cut-off scaled scores of ≤ 41.5, ≤ 102.5, and ≤ 131.5 are suggested for the RBMT-3 to discriminate between patients with mild and moderate dementia, mild dementia and MCI, and MCI and NC, with sensitivities 73%, 100% and 94.1%, respectively. This version is useful to differentiate those with or without risk of cognitive impairments.