The psychometric properties of the Ages and Stages Questionnaires-3 in Arabic: Cross-sectional observational study

Abstract

Background: Screening for developmental delay is recommended for pediatricians, yet validated screening tools in Arabic are scarce. Aims: Assess the reliability, validity, sensitivity and specificity of the Arabic ASQ-3 in detecting developmental delays in children aged 4–33 months. Study design: Cross-sectional observational study. Subjects: A sample of 491 children from all Lebanese governorates from five age groups (4, 10, 18, 27 and 33 months). Outcome measures: Internal consistency using Cronbach's alpha (Cα), convergent construct validity using Pearson Correlation Coefficient (CC) comparing ASQ-3 in Arabic (A-ASQ-3) to Bayley scale for infant development (BSID-III) scores were computed. A subset sample (n = 35) underwent assessment with BSID-III for convergent validity, and sensitivity and specificity. A-ASQ-3 scores were compared to the US mean cutoff scores. Results: Subjects' mean age was 17.75 ± 10.6 months. Cα was 0.85; Pearson CC showed positive moderate correlation between A-ASQ-3 gross and fine motor scores and BSID-III composite motor scores (r = 0.42, p = 0.002; r = 0.39, p = 0.004, respectively); and positive moderate correlation between A-ASQ-3 problem solving and BSID-III cognitive scores (r = 0.43, p = 0.001). A-ASQ-3 had 100% sensitivity and 50% specificity for cognitive scores; 60% sensitivity and 70% specificity for motor scores. Conclusion: A-ASQ-3 has adequate reliability and validity for the tested age groups. It is plausible that this would hold true for the rest of the questionnaires. Further testing is needed to make the five clusters more aligned with the US sample scores and to improve the sensitivity and specificity. © 2019 Elsevier B.V.