Evaluation of the Strengths and Difficulties Questionnaire for Preschoolers in Urban Primary Care.

The Strengths and Difficulties Questionnaire (SDQ) is a screening tool widely used in both pediatric research and clinical practice. Few studies have examined the psychometric properties of the U.S. version of this measure, and none have examined the structure of the measure with children younger than 4 years of age. The goal of this study was to assess the psychometric properties of the parent informant version of the SDQ administered in a pediatric practice serving a low-income, predominately Black, urban population of children ages 2–5 years. Parents routinely completed the SDQ electronically during annual well-child visits from July 2016 to June 2017. Confirmatory factor analysis was used to test 3-factor, 5-factor, and modified 5-factor (with positive construal response style) models. Responses to age-appropriate SDQ forms were analyzed. There were N = 622 for the 4–12 year old version (only collected on 4–5 year olds) and N = 672 of the 2–3 year old version. In both groups, the modified 5-factor model was the best fitting model. The internal consistency for the modified 5-factor model was lower than the original 5-factor model with the alpha coefficient lower on the Conduct Problems, Emotional Problems and Peer Problems subscales. Measurement invariance testing found non-invariance due to gender among the 4–5 year olds on the emotional problems subscale. Our study found that the SDQ shows satisfactory psychometric properties in preschool-aged children in a low-income urban U.S. population. More research is needed to deepen our understanding of the measure's clinical utility. Highlights: Few studies have examined the psychometric properties of the Strengths and Difficulties Questionnaire in the US. No studies have examined the structure of the measure with children <4 years of age. Participants were recruited from a pediatric practice serving a low-income, predominately Black, urban population of children (2–5 years). A modified 5-factor (with positive construal response style) model fit the data best. Internal consistency was low for some subscales and non-invariance due to gender was found for the emotional problems subscale among 4–5 year olds. [ABSTRACT FROM AUTHOR]