Evaluation of the NIH Toolbox Cognition Battery in Adolescents with HIV Infection in Zimbabwe

Electronic assessments of cognition in low-resource settings have the potential to widen access to cognitive appraisal in at-risk populations. We evaluated the performance of the shorter, tablet-based NIH cognition battery (NIH-CB) against the validated paper-based Kaufman Assessment Battery for Children 2nd Edition (KABC-II) in assessment of neurocognitive function in adolescents living with HIV (ALWH). Adolescents aged 11-19 years living with HIV and an HIV-negative comparison group completed the NIH-CB and KABC-II. Convergent validity between NIH-CB and KABC-II subtest and global scores was measured using Pearson's R and Bland-Altman plots. The ability of the NIH-CB to differentiate between HIV-negative and ALWH participant cognition was evaluated using linear regression. The acceptability of the NIH-CB and KABC-II were evaluated through questionnaires, focus groups and interviews. Forty-two HIV-negative adolescents were enrolled in a pilot acceptability study. Subsequently, 507 (253 HIV-negative, 254 HIV-positive) adolescents (median age 15(IQR:13-17) years) were enrolled into the validation study. Subtest scores on the NIH-CB and KABC-II measuring memory demonstrated low-moderate correlation (R 0.140-0.491). A Bland-Altman plot showed HIV-negative participants had low mean difference between NIH fluid cognition Z-score and KABC-II MPI Z-score (0.02 + 0.21 average); 95.2% of measurements were within limits of agreement. NIH-CB fluid cognition Z-scores were higher in HIV-negative participants than ALWH (mean difference 0.27 [95%CI 0.10, 0.41] p < 0.001). The NIH-CB was highly acceptable to participants and test administrators. The NIH-CB may serve as both a valid and practical cognition assessment tool in adolescents in low-resource settings, including in HIV research.