Interpreting reliable change on the Spanish-language NIH toolbox cognition battery

This study applied a reliable change methodology to the test-retest data from the Spanish-language NIH Toolbox Cognition Battery (NIHTB-CB) normative sample. Participants included Spanish-speaking adults (n = 48; 54.2% women, 100% Latinx) evaluated twice within one to two weeks on the Spanish-language NIHTB-CB, consisting of two crystallized and five fluid cognitive tests. Test-retest means, standard deviations, and intraclass correlations were used to calculate upper and lower bounds of 70, 80, and 90% confidence intervals (CIs) around change scores, with these bounds used as cutoffs for inferring reliable change. Cutoffs were calculated for raw scores, age-adjusted standard scores (SS; M = 100, SD = 15), and demographic-adjusted T-scores (T; M = 50, SD = 10), adjusting for age, gender, and education. Test-retest change scores on the Spanish-language NIHTB-CB exceeding the following cutoffs indicate reliable change based on an 80% CI (i.e., values exceeding these cutoffs indicate greater decline or greater improvement than 90% of the sample): Dimensional Change Card Sort (SS >= 15/T >= 11), Flanker (SS >= 13/T >= 10), List Sorting (SS >= 13/T >= 9), Picture Sequence Memory (SS >= 14/T >= 9), Pattern Comparison (SS >= 14/T >= 10), Picture Vocabulary (SS >= 8/T >= 6), Oral Reading (SS >= 7/T >= 5), Fluid Cognition Composite (SS >= 12/T >= 9), Crystallized Cognition Composite (SS >= 6/T >= 5), and Total Cognition Composite (SS >= 8/T >= 7). These cutoffs are one of few resources to interpret cognitive change at retest among Spanish-speaking patients and participants.