Demographically Corrected Normative Standards for the English Version of the NIH Toolbox Cognition Battery

Abstract

AbstractDemographic factors impact neuropsychological test performances and accounting for them may help to better elucidate current brain functioning. The NIH Toolbox Cognition Battery (NIHTB-CB) is a novel neuropsychological tool, yet the original norms developed for the battery did not adequately account for important demographic/cultural factors known to impact test performances. We developed norms fully adjusting for all demographic variables within each language group (English and Spanish) separately. The current study describes the standards for individuals tested in English. Neurologically healthy adults (n=1038) and children (n=2917) who completed the NIH Toolbox norming project in English were included. We created uncorrected scores weighted to the 2010 Census demographics, and applied polynomial regression models to develop age-corrected and fully demographically adjusted (age, education, sex, race/ethnicity) scores for each NIHTB-CB test and composite (i.e., Fluid, Crystallized, and Total Composites). On uncorrected NIHTB-CB scores, age and education demonstrated significant, medium-to-large associations, while sex showed smaller, but statistically significant effects. In terms of race/ethnicity, a significant stair-step effect on uncorrected NIHTB-CB scores was observed (African American<Hispanic<White). After applying normative corrections, NIHTB-CB no longer demonstrated any significant associations with demographic factors. The previously developed norms still maintained significant associations with demographic factors, and demonstrated more variable impairment rates in segments of the healthy normative sample. Similar to other neuropsychological tests, demographic factors demonstrated significant associations with unadjusted NIHTB-CB scores. Application of fully corrected scores will help account for unwanted variance that is associated with non-clinical factors to more accurately reflect effects of disease-related changes in brain function. (JINS, 2015,21, 378–391)