Assessing social-emotional development in infants and toddlers using parent-reports: Comparing the ASQ-SE-NL to the Social-Emotional Scale of the Bayley-III-NL. (October 2021)

• Record Type: Journal Article
• Authors: Krijnen, Lisa J.G.; Verhoeven, Marjolein; van Baar, Anneloes L.
• Is Part Of: Early human development. Volume 161(2021)
• Publication Date: 2021-10
• Subjects: Social-emotional development -- Ages and Stages Questionnaire Social-Emotional -- Bayley-III -- Infants -- Toddlers -- Community sample; ASQ-SE-NL the Dutch translation of the Ages and Stages Questionnaire - Social-Emotional -- AUC Area Under Curve -- Bayley-III-NL-SE the Dutch translation of the Social-Emotional Scale of the Bayley-III -- PPV positive predictive value -- NPV negative predictive value -- CUI Clinical Utility Index -- UI+ screen-in accuracy -- UI− screen-out accuracy -- ROC Receiver Operating Characteristic; Fetus -- Periodicals; Neonatology -- Periodicals; Prenatal influences -- Periodicals; 612.65
• DOI: 10.1016/j.earlhumdev.2021.105439
• Languages: English
• ISSNs: 0378-3782

Abstract

Abstract: Background: The Ages and Stages Questionnaire Social-Emotional (ASQ-SE) is a parent-report screening instrument designed to assess children's social-emotional development and detect those at risk for delay or problems. Psychometric properties of this questionnaire have been studied before, but the ASQ-SE has never been compared to the Social-Emotional Scale of the Bayley-III (Bayley-III-SE). Aim: To compare the Dutch ASQ-SE (ASQ-SE-NL) to the Dutch Bayley-III-SE (Bayley-III-NL-SE; criterion measure). Method: A Dutch community sample of mothers with children aged 3–41 months ( n = 1014) filled out both questionnaires. Cut-off scores for the ASQ-SE-NL were determined using: 1) >1 SD above the mean and 2) ROC curves. For the Bayley-III-NL, Dutch norm scores were used. Results: Specificity (70.8% and 88.5%) and screen-out accuracy (0.65 and 0.77) of the ASQ-SE-NL were good. Sensitivity was only sufficient (70.6%) when using ROC curves and only for the ASQ-SE-NL ≥18 months age versions. Screen-in accuracy was insufficient (<0.49). Positive predictive value was 34.7% and 32.7%, and negative predictive value was 87.5% and 92.3%. False positive cases on the ASQ-SE-NL scored significantly lower on the Bayley-III-NL-SE than true negative cases. Conclusion: Using the Bayley-III-NL-SE as the criterion, the ASQ-SE-NL performed well in identifying children not at risk for delay or problems. The ASQ-SE-NL sufficiently detected children at risk for delay or problems in the ≥18 months ASQ-SE-NL age versions when cut-off scores were determined by ROC curves. The ASQ-SE-NL can be used in a monitoring routine, but early rescreening is advised after a positive test result, given the number of false positive results. Highlights: This is the first study comparing the ASQ-SE to the Bayley-III-SE. Specificity and screen-out accuracy were good. Screen-in accuracy was insufficient. Sensitivity was only good in the ≥18 months ASQ-SE-NL age versions using ROC curves. The PPV was low and the NPV was high due to the low prevalence of delay in our sample. False positive cases scored lower than true negative cases on the Bayley-III-NL-SE.