An attempt to integrate information from young children into parent/teacher-based mental health evaluation of child internalizing disorders in epidemiological surveys

Background: To date, most large surveys on the mental health of children under 11 years old have relied exclusively on data collected from parents and teachers, with no information from children themselves. However, these children are valuable informants despite their young age. Objective: To evaluate the effects of adding young children’s self-evaluation of internalizing disorders to a combined teacher/parent-based algorithm using data from a European survey carried out in Italy, Bulgaria, and in the Netherlands. Method: Comparing concordances with the Development and Well-Being Assessment (DAWBA; considered as the gold standard) of a Strengths and Difficulties Questionnaire’s (SDQ) parent/teacher-based predictive algorithm, with and without integrating a child self-evaluation. The child self-evaluation was using the Dominic Interactive (DI), a computerized, pictorial, self-administered questionnaire for young children, which followed DSM 5 diagnoses definitions for internalized disorders. The concordance comparisons concerned 238 children, 5 to 12 years old from out-patient clinics, double evaluated, using kappa concordance test . Subsequent variations in disorder prevalence were evaluated in cross-sectional surveys in the general populations of the three study countries ( n = 3,442). Results: Adding self-evaluated child DI-based ‘probable’ internalizing disorder diagnoses to the SDQ parent/teacher algorithm increased concordance with gold standard: kappa increased from .25 (.14–.37) to .36 (.23–.49). In terms of population-based data, this addition increased the prevalence of ‘probable’ emotional disorders from 2.00 % (1.56–2.53) to 3.63 % (3.03–4.31) and revealed an expected gender difference. Conclusion: The inclusion of young children’s self-reported data on internalizing disorders is a valuable addition to the generally used combined parent/teacher-based SDQ algorithm, and therefore better reflects the true prevalence of these disorders in these children. Further evaluation on larger validation samples is required.