| Publication Abstract Display | | Type: Published Abstract | | Title: Uncorrected versus Demographically-Adjusted Scores on the NIH Toolbox Cognition Battery: What is the Difference? | | Authors: Casaletto K, Heaton R, Tulsky D | | Year: 2015 | | Publication: 43rd Annual Meeting of the International Neuropsychological Society | | Volume: Issue: Pages: | | Abstract:Objective: Impacts of demographics on neuropsychological test performances are well established, but clinicians and researchers may have difficulty choosing and interpreting uncorrected versus demographically-corrected (DC) scores. Our study compared these two types of scores for clinical and nonclinical groups on the NIH Toolbox Cognition Battery (NIHTB-CB), and examined associations with demographic and clinical variables.
Methods: Our samples were adults with traumatic brain injury (TBI; n=181) or stroke (CVA; n=209), and demographically-matched groups of healthy adults (TBI-matched n=184; CVA-matched n=210). NIHTB-CB Fluid and Crystallized Composites were analyzed, and DC scores controlled for age, education, gender, and race/ethnicity.
Results: Compared to matched controls, both the TBI and CVA cohorts demonstrated larger differences on corrected than uncorrected Fluid Composites (uncorrected to corrected effect sizes: TBI d=0.60 to 0.77; CVA d=0.88 to 1.01). However, on Crystallized scores, differences between clinical groups and their matched controls became slightly smaller with DCs (uncorrected to corrected effect sizes: TBI d=0.23 to 0.21; CVA d=0.40 to 0.17). In the clinical sample, demographic factors accounted for 29-36% of the variance in uncorrected scores (ps<0.001), but only 0-4% of the variance in corrected scores (ps>0.04). Clinical ratings of injury increasing severity were associated with greater impairment on the corrected Fluid scores within the TBI and CVA cohorts (ps<0.01); uncorrected Fluid scores were only significantly associated with injury severity within the CVA group.
Conclusions: DC scores resulted in higher rates of impairment in the clinical groups, and larger differences versus matched controls. Importantly, the DC scores were no longer associated with demographic factors and demonstrated good construct validity with a proxy for brain injury severity. |
return to publications listing
|