Background: We present information regarding the standardization, reliability and clinical validity of two versions of the Behavioural Neurology Assessment (BNA). The BNA-Long Form consists of 24 subtests within separate domains: Attention, Memory, Language, Visuospatial Function, Executive Function, and Praxis. The BNA-Short Form consists of 13 subtests within the domains of Attention, Memory, Naming, Visuospatial Function and Executive Function. In addition to individual domain indices, a Grand Total score was calculated for both BNA versions.
Objective: To standardize the administration and scoring and validate the BNA for detection of dementia.
Methods: Standardized normative data were obtained on 115 healthy subjects ranging in age from 50 to 95. Test-retest stability was obtained on 19 subjects and clinical validity was investigated by administering the BNA and Mini-Mental Status Examination (MMSE) to 29 patients with dementia and 29 age-matched healthy subjects (controls).
Results: Age had a significant effect on all but the Visuospatial and Praxis indices of the BNA-Long Form and an effect on Naming and Grand Total score of the Short-Form. Internal consistency (Cronbach's coefficient alpha) was .87 and .67 for the Long and Short Forms (.95 and .96 for dementia and control groups combined). Test-retest stability was acceptable. Grand Total indices of both BNA versions showed significant, positive correlations with the MMSE. Both BNA versions had superior sensitivity to dementia relative to the MMSE (.93 versus .79). Specificity was equivalent to the MMSE (.93 versus .97).
Conclusions: Positive predictive values of the BNA and MMSE are equivalent but the BNA provides superior negative predictive value.