Evidence for Undertreatment of ADHD in Girls with ASD in the National Survey of Children's Health

Background: Numerous studies have established that pharmacological treatment of ADHD is more effective than behavioral intervention alone for ADHD symptom improvement, although effect sizes are typically lower in children with autism spectrum disorders. Untreated ADHD has been associated with worse long-term outcomes in a wide range of functional domains. Objectives: To elucidate national patterns in ADHD medication use among girls and boys with and without autism in the US. Methods: The 2016 National Survey of Children’s Health (NSCH), conducted by the US Census Bureau from June 2016 to February 2017, provides state-level and nationally representative data of parent report on the health and wellbeing of noninstitutionalized children aged 0-17 with and without special healthcare needs. The 2016 NSCH surveyed the parents/caregivers of 50,212 children in all 50 states and the District of Columbia. Data were adjusted and weighted by the survey team to create state and nationally representative samples of noninstitutionalized children aged 0-17. The overall weighted response rate was 40.7%. Weights were adjusted to account for nonresponse, and a nonresponse bias analysis is available on the survey website. In the current study, chi-square and logistic regression analyses were performed examining rates of ADHD diagnosis and medication use among children with and without ASD. SPSS Version 25 Complex Samples Module with subpopulation routine was use to estimate variance per the study documentation. Results: For boys, unweighted N=917 for current ASD and 430 for current ASD+ADHD; for girls, unweighted N=214 for current ASD and 91 for current ASD+ADHD. ASD children were age 3-17 (mean age 10.04). Race/ethnicity distribution was 53.3% white, 24.1% Hispanic, 14.4% black, and 8.2% “other.” There was not a significant difference by gender of ADHD diagnosis among children with ASD (OR 0.575, 95%CI 0.327-1.010, 51.2% of ASD boys and 37.6% of ASD girls); girls who currently had both ASD and ADHD diagnoses had lower odds of currently using medication for ADHD (OR 0.318, 95% CI 0.149-0.677, 69.1% of ASD+ADHD boys and 41.5% of ASD+ADHD girls) compared to boys. Non-ASD girls had lower odds of ADHD diagnosis (OR 0.439, 95% CI .381-0.507) versus boys; there was not a significant difference by gender of medication use among children with ADHD but not ASD (OR 1.156, 95% CI 0.868-1.539). In logistic regression, the odds ratio for ASD girls’ ADHD medication use remained at 0.311 (95% CI 0.128-0.757, p<0.01) even after controlling for race/ethnicity, household income, service use (current mental health therapy [p<0.05] and special education services), and parent rating of ADHD severity and three ADHD behaviors (finishing tasks, doing all homework, and staying in control when challenged [p<0.05]). Conclusions: This nationally representative dataset indicates that girls with both ASD and ADHD are being undertreated for ADHD. The one-third of ASD girls with ADHD diagnoses therefore might have improved long term outcomes in academic and occupational achievement, antisocial behavior, unsafe driving, drug/alcohol use, obesity, services use, and self-esteem, social functioning if their ADHD were treated in accordance with practice guidelines.