Objective: Most screening instruments for externalizing disorders have been developed and validated in Western children. We developed and validated a brief screening instrument for predicting externalizing disorders in native Dutch children as well as in non-Dutch immigrant children, using predictors that can be easily obtained from teachers.
Method: Teachers completed the Strengths and Difficulties Questionnaire for an ethnic diverse sample of 2,185 children ages 6 to 10 years. In a stratified subsample, 254 children and their parents were additionally interviewed regarding psychiatric disorders and sociodemographic data. In this group, stepwise logistic regression was used to derive a score from sex and all items of the Hyperactivity and Conduct Problems Scale of the Strengths and Difficulties Questionnaire, for predicting a best-estimate diagnosis of any externalizing disorder. The accuracy of the score was compared between native Dutch and non-Dutch immigrant children.
Results: Ninety-one cases of externalizing disorders were identified. An externalizing disorder could be predicted by the items restless, obeys, lies, and concentrates. Sex and ethnicity did not contribute to a prediction of an externalizing disorder. The area under the receiver operating characteristic was 0.84 (95% confidence interval 0.79-0.89), indicating good discriminatory power with no substantial differences between native Dutch and non-Dutch immigrant children.
Conclusions: Externalizing disorders in both native Dutch and non-Dutch immigrant children can be predicted with a scoring rule, based on only four items that can be easily assessed by teachers. Before this internally validated prediction tool can be implemented, external validation in another sample is necessary.