Development of an empirically derived neuropsychological screening battery for mild to moderate cognitive impairment

This study sought to develop a screening battery for detecting mild-to-moderate cognitive impairment in adults. Nineteen neuropsychological measures were utilized to develop two proposed screening batteries, and their diagnostic accuracy rates were compared. Demographically corrected norms were utilized to account for the effect that demographic characteristics (i.e., age, gender, education, and race/ethnicity where possible) can have on neuropsychological test performance. The “gold standard” criterion variable utilized was blind clinical ratings derived from test measures and clinical judgments. Proposed Battery 1 was derived from hierarchical multiple regression analysis and included measures of processing speed and visual memory (i.e., WAIS-III Processing Speed Index, WMS-III Visual Delay Index, and Rey-O Complex Figure Test Delay). Battery 2 was derived from stepwise multiple regression, and included the same measures with the addition of a measure of abstraction/cognitive flexibility, Trails B. Both proposed batteries demonstrated excellent accuracy, Battery 1 having 90% sensitivity and 88% specificity and Battery 2 having 93% sensitivity and 79% specificity. Results suggest that demographically corrected T-scores from the commonly used measures in the proposed screening batteries can serve as efficient and cost-effective tools for screening mild-to-moderate neuropsychological impairment in order to identify individuals who could benefit from further assessment