Tablet-Based Neurocognitive Tests for Ugandan Adolescents Living With HIV: Cross-Sectional Validation Study.

Abstract

Background: Adolescents living with perinatally acquired HIV (PHIV) remain at risk of neurocognitive difficulties, yet access to valid neurocognitive tests is limited in sub-Saharan Africa (SSA), where most of this population resides. Tests that can be administered by nonspecialist personnel are culturally and linguistically appropriate for SSA's diverse populations and are brief enough for routine use are needed to expand access, but few such tools exist.

Objective: We adapted NeuroScreen, a brief tablet-based neurocognitive test battery, into Luganda and evaluated its validity and acceptability for use among Ugandan adolescents with PHIV.

Methods: In this cross-sectional validation study, 100 Ugandan adolescents aged 11 to 20 years (50 with PHIV; 50 HIV-negative controls) completed the Luganda- or English-language NeuroScreen battery and a commonly used traditional test battery. Participants were recruited from an existing longitudinal cardiovascular cohort at the Joint Clinical Research Centre in Kampala, Uganda. We examined convergent validity with correlations between the 2 batteries' individual and global z-scores, known-groups validity with unadjusted and age-adjusted group comparisons, criterion validity by examining classification agreement between the 2 batteries quantified with Cohen κ, and acceptability through self-reported technology use and ease of use.

Results: NeuroScreen and the traditional battery global z-scores were strongly correlated (r=0.75; P<.001). Correlations between tests assessing similar domains were generally small to moderate, and were strongest for processing speed, attention, and working memory. Adolescents with PHIV scored significantly lower than controls on global scores for both NeuroScreen (P=.008) and the traditional battery (P=.003), even after adjusting for age. Agreement between batteries in classifying participants ≤1 SD (κ=0.56) and ≤2 SD (κ=0.66) below the global z-score mean was moderate, with both batteries classifying more adolescents with PHIV than controls below these thresholds. Although fewer than half of participants (42/97, 43%) had previously used a tablet, most rated NeuroScreen as easy to use, and prior tablet experience was not independently associated with NeuroScreen performance after age adjustment.

Conclusions: The Luganda-adapted NeuroScreen demonstrated convergent, known-groups, and criterion validity in Ugandan adolescents with PHIV. By delivering automated administration, automated scoring, and standardized multimedia instructions in the local language on a portable tablet, NeuroScreen addresses several persistent barriers to neurocognitive assessment in low-resource settings, including shortages of specialist personnel, the time burden of traditional batteries, and the scarcity of culturally and linguistically appropriate tools. The NeuroScreen battery was found to be highly acceptable among adolescents with little or no prior tablet experience. These findings add to a growing body of evidence supporting tablet-based assessment as a practical and scalable mobile health approach to expanding access to neurocognitive testing. Future research should evaluate test-retest reliability, sensitivity to change over time, normative benchmarks, and implementation within real-world clinical workflows.