Objective: Developmental delay is relatively common and produces serious impairment. Efforts to screen for delay often include parent-completed instruments. We evaluated the agreement between the most popular such instrument, the Ages and Stages Questionnaires (ASQ) and the third edition of the Bayley Scales of Infant Development (BSID-III).
Methods: We analyzed a community sample of 587 children aged 1 month to 36 months who received both the ASQ and the BSID-III. We calculate sensitivity, specificity, and positive and negative predictive values. Because published BSID-III norms produced unexpectedly low prevalences, we also derived a set of distribution-based thresholds using quantile regression, and we repeated the validation analysis using these results.
Results: BSID-III prevalence was 2.9% (95% confidence interval [CI] 1.7-4.6) with published norms and 7.7% (95% CI 5.6-10.1) with distribution-based thresholds, while 18.2% (95% CI 15.2-21.6) of children were positive on the ASQ. For published BSID-III norms, sensitivity was 41% (95% CI 18-67) and specificity 82% (95% CI 79-85). Results with distribution-based thresholds were essentially identical. Performance was somewhat better among children over 1 year (sensitivity 50%, specificity 87%). For subscales, sensitivities were generally lower (range 0-50%) and specificities higher (range 92-96%).
Conclusions: Agreement between the ASQ and BSID-III was relatively poor. Previous studies have reported somewhat better agreement. There are numerous possible explanations for differences, including the age ranges used, the risk profile of children, and differences in the ASQ administration. Results raise concerns about the performance of this instrument in primary care and community settings.
Keywords: Ages and Stages Questionnaire; Bayley Scales of Infant Development; development screening; sensitivity and specificity; validation.
Copyright © 2015 Academic Pediatric Association. Published by Elsevier Inc. All rights reserved.