Web-Based Cognitive Screening Tool Shows Reliable Repeat Scores in Children with Intellectual Disabilities

Researchers have developed an interactive, web-based cognitive screening tool for children with intellectual disabilities and reported that its easiest version produces stable, repeatable scores when the same children are tested twice over a 45-day interval. The instrument, called the Audio-Visual Cognitive Screening Test Kit, or AVCST, was delivered through an HTML-based platform that children used on a touchscreen, answering some tasks verbally and others directly on the screen. In a small study of 22 participants at a special education and rehabilitation centre, the tool showed strong test-retest agreement, with no evidence that scores drifted systematically between the two sessions.

The study, published in the Egyptian Pediatric Association Gazette, was designed around a specific and often overlooked gap. Computerized cognitive assessment in children has expanded rapidly, with established batteries such as the Cambridge Neuropsychological Test Automated Battery and the NIH Toolbox Cognition Battery offering multidomain testing across childhood and adolescence. Yet a 2023 scoping review cited by the authors found that relatively few instruments had been used in web-based settings, and that only 13 percent of the studies it included reported psychometric properties. Evidence becomes thinner still for children with intellectual disabilities, a group for whom conventional paper-based testing can be especially difficult to administer.

The motivation behind the AVCST is practical as much as scientific. Paper-based cognitive tests typically depend on manual scoring, examiner-dependent administration and structured procedures that can be hard to sustain in busy paediatric or special-education settings. Repeated testing for progress monitoring raises the further risk of inconsistent scoring or arithmetic mistakes. The AVCST was built to counter these problems with standardized on-screen instructions and automatic scoring, producing a total score between 0 and 20 in which higher values indicate better performance. Its audiovisual presentation and graded difficulty levels were intended to make a brief screening session manageable for children who often struggle with traditional test formats.

The tool was developed independently, drawing on selected cognitive domains assessed by the Montreal Cognitive Assessment, including memory, attention and executive function, with perception added as an extra domain. The authors are explicit that no MoCA questions, items, wording, images or other test content were reproduced or adapted. The tool was developed in English and a local language, a design choice the researchers suggest may improve its practical relevance in resource-constrained settings.

The study had a narrow and clearly stated aim: to evaluate the test-retest reliability of the AVCST’s Easy level in children with intellectual disabilities. Twenty-two participants aged 10 to 19 were recruited through convenience and purposive sampling at a single institution, and the sample size was set by feasibility rather than a formal statistical calculation. Seventeen of the participants were boys and five were girls. Most were aged 13 to 15 years, with seven aged 16 to 17 and five aged 10 to 12; no participants aged 18 or 19 were included, an uneven distribution the authors flag as a limitation of the small, feasibility-based sample. Severity of intellectual disability was taken from government-issued disability certificates, with 15 children classified as having moderate disability, five mild and two severe. Independent IQ scores were not available, and no separate IQ assessment was performed.

Each child completed the Easy level twice, with the second session 45 days after the first. The retest had originally been planned at roughly one month, a common interval in comparable reliability studies, but institutional constraints, including examination timetables, centre activities, attendance patterns and the limited availability of eligible students, pushed it to 45 days. The researchers argue the interval balanced two competing concerns: long enough to reduce immediate recall and short-term practice effects, short enough to limit genuine change in cognitive functioning between sessions. Only the Easy level was administered, to all participants at both sessions, to keep administration standardized and allow direct score comparison. The Medium and Hard levels were not evaluated and remain untested.

The reliability results were consistent across three statistical approaches. Spearman’s rank correlation between first and second scores was 0.721, with a p value below 0.001, indicating a strong positive association in the rank order of children’s performance across sessions. The intraclass correlation coefficient, computed with a two-way mixed-effects model and an absolute agreement definition, was 0.775, with a 95 percent confidence interval of 0.533 to 0.900 and p below 0.001, a value the authors classify as good reliability. Correlation coefficients of 0.70 or above are conventionally treated as indicating acceptable reliability for health measurement instruments. The Wilcoxon signed-rank test found no statistically significant difference between the two sets of scores, with Z equal to minus 0.476 and p equal to 0.634, meaning there was no evidence of a systematic shift between sessions.

The descriptive statistics reinforced that picture of stability. The median score was 18, with an interquartile range of 5.5, at both the first and second administrations, and the distribution of scores across low, moderate and high bands remained broadly similar between sessions. In the first session, 18 of the 22 children, or 81.8 percent, scored in the high band of 14 to 20; in the second, 17, or 77.2 percent, did so. Movement between the moderate and low bands was slight in both directions. The authors caution, however, that this clustering at the top of the scale may signal a ceiling effect, particularly given that only the Easy level was given to children spanning mild, moderate and severe disability. The high scores, they stress, describe the distribution observed in this sample and should not be read as evidence of uniformly high cognitive functioning, nor do they rule out that the Easy level has limited ability to discriminate between different levels of disability.

The authors place their findings carefully. The observed reliability coefficients fall within the range reported for the original paper-based MoCA’s test-retest reliability across different populations, where moderate to strong correlations have been described, and they suggest that under standardized conditions the digital easy-level adaptation maintains temporal stability comparable to the traditional format. But they are equally clear about what the study does not establish. The sample of 22 is small and affects the precision of the estimates, a point reflected in the relatively wide confidence interval around the ICC, and the authors describe those findings as preliminary, to be confirmed in larger samples with prospectively determined sample sizes. Construct validity, sensitivity to cognitive change and inter-rater reliability have not yet been demonstrated, and the Easy level should not be considered validated for use across all levels of intellectual disability.

Within those limits, the study offers a concrete step for technology-assisted cognitive screening in paediatric disability settings. The AVCST is explicitly a screening instrument, intended to flag potential cognitive difficulties that may warrant further evaluation and to help track broad cognitive performance over time; it is not a substitute for comprehensive developmental, psychological, neuropsychological or diagnostic assessment. The authors argue that automated scoring may improve feasibility and reduce human calculation errors, and that standardized digital administration could lessen examiner-dependent variability in special schools and clinics. The broader significance lies in the validation approach itself: as digital cognitive testing spreads through paediatric care, the authors argue that each individual tool must have its reliability and validity established for its intended population and setting, and this study supplies the first, deliberately cautious evidence that one such tool can yield consistent, repeatable scores in children with intellectual disabilities.

Subject of Research: Test-retest reliability of a web-based cognitive screening tool for children with intellectual disabilities

Article Title: Development and reliability assessment of an interactive web-based cognitive screening tool for children with intellectual disabilities

Article References: Vyas, M., & Solanki, N. (2026). Development and reliability assessment of an interactive web-based cognitive screening tool for children with intellectual disabilities. Egyptian Pediatric Association Gazette, 74(1), Article 126. https://doi.org/10.1186/s43054-026-00626-0

Image Credits: AI Generated

DOI: 10.1186/s43054-026-00626-0

Keywords: cognitive screening, intellectual disabilities, test-retest reliability, digital assessment, pediatrics, AVCST, automated scoring, MoCA, psychometrics, special education, web-based testing, temporal stability