Test-retest reliability of the ultra-brief continuous visual attention test in typically developing children and adolescents

Abstract

Background: Neuropsychological tests require test-retest reliability to ensure score stability and control practice effects. While conventional 2-3 week retest intervals minimize learning effects, there is growing clinical interest in same-day reliability to evaluate acute pharmacological responses, such as stimulant medications in children with Attention-Deficit/Hyperactivity Disorder (ADHD). As attention is foundational to other cognitive domains, its efficient assessment is crucial. The ultra-brief (90-second) Continuous Visual Attention Test (CVAT) offers a standardized solution for pediatric populations. Given that attentional subdomains mature differently, evaluating CVAT’s reliability across developmental stages is essential. Objective: To evaluate the temporal reliability and developmental sensitivity of the ultra-brief CVAT in typically developing children and adolescents over a conventional (2-3 weeks) and same-day interval. Secondary aims included assessing rank-order and test-retest stabilities, measurement agreement, clinical classification stability (≤25th percentile) across development, and internal consistency across age groups. Methods: The cross-sectional study included 282 children and adolescents aged 5-17 years, divided into four age groups. Participants completed the ultra-brief CVAT twice under identical conditions to assess attentional subdomains: 97 participants were retested after 15 minutes, and 185 were retested after 2-3 weeks. Reliability and agreement were analyzed using Spearman’s ρ, Pearson’s r, Intraclass Correlation Coefficients (ICCs), Bland-Altman plots, and Cohen’s κ. Results: Rank-order stability across the whole sample was strong for speed and accuracy (e.g., Mean Reaction Time r=0.859; Correct Answers r=0.762; p<0.001). Most variables exhibited good test-retest reliability (ICC≥0.70, p<0.001) in both intervals, with no significant differences between timeframes (all p > 0.05). Commission Errors demonstrated moderate reliability. Bland-Altman plots revealed minimal systematic bias between the two measurement methods. Furthermore, test-retest agreement for the ≤25th percentile classification demonstrated significant temporal stability (all p <0.001). Reliability generally improved or stabilized for most measures in older groups. Conclusion: The ultra-brief CVAT demonstrates highly stable measurement properties across same-day and 2-3 week retest intervals. Attentional reliability revealed a developmental progression: early processing speed stability matures into highly reliable response consistency and attentional control in adolescence. The CVAT is clinically useful for tracking attention and identifying below-average performance across developmental stages.