Prevalence of Common Child Mental Health Disorders Using Administrative Health Data and Parent Report in a Prospective Community-Based Cohort from Alberta, Canada: Prévalence des troubles communs de santé mentale de l’enfant à l’aide des données de santé administratives et des rapports des parents dans une cohorte prospective communautaire d’Alberta, Canada

Objective. Knowing the prevalence of mental health difficulties in young children is critical for early identification and intervention. In the current study, we examine the agreement among three different data sources estimating the prevalence of diagnoses for attention deficit hyperactivity disorder (ADHD) and emotional disorders (i.e., anxiety or mood disorder) for children between birth and 9 years of age. Methods. Data from a prospective pregnancy cohort was linked with provincial administrative health data for children in Alberta, Canada. We report the positive agreement, negative agreement, and Cohen's Kappa of parent-reported child diagnoses provided by a health professional (“parent report”), exceeding a clinical cut-off on a standardized questionnaire completed by parents (the Behavior Assessment System for Children, 3rd edition [“BASC-3”]), and cumulative inpatient, outpatient, or physician claims diagnoses (“administrative data”). Results. Positive and negative agreement for administrative data and parent-reported ADHD diagnoses were 70.8% and 95.6%, respectively, and 30.5% and 94.9% for administrative data and the BASC-3, respectively. For emotional disorders, administrative data and parent-reported diagnoses had a positive agreement of 35.7% and negative agreement of 96.30%. Positive and negative agreement for emotional disorders using administrative data and the BASC-3 were 20.0% and 87.4%, respectively. Kappa coefficients were generally low, indicating poor chance-corrected agreement between these data sources. Conclusions. The data sources highlighted in this study provide disparate agreement for the prevalence of ADHD and emotional disorder diagnoses in young children. Low Kappa coefficients suggest that parent-reported diagnoses, clinically elevated symptoms using a standardized questionnaire, and diagnoses from administrative data serve different purposes and provide discrete estimates of mental health difficulties in early childhood.