Psychiatric Diagnoses and Concordance with Clinician Diagnosis of Children with Autism Spectrum Disorders Served in Community Mental Health Settings

Background: Research suggests that discrepancies exist in child psychiatric diagnoses between those assigned by community mental health (CMH) therapists and those determined by standardized assessment tools (Haine et al., 2007; Lewczyk et al., 2003). Accurate classification of psychiatric co-morbidity is critical to determine diagnostic prevalence and inform development, implementation, and dissemination of evidence-based practices in CMH settings. Children with an autism spectrum disorder (ASD) are often served in CMH settings due to their high rates of co-occurring psychiatric disorders. Information on the types and diagnostic accuracy of co-occurring mental health problems is critical to bolster efforts to improve care for youth with ASD served in CMH settings.

 Objectives: This study reports preliminary psychiatric diagnostic data from an ongoing randomized community effectiveness trial of AIM HI (“An Individualized Mental Health Intervention for ASD”) conducted within publicly-funded community and school-based MH services.  AIM HI is a clinical intervention targeting challenging behaviors in children with ASD and designed to be delivered by MH therapists. The prevalence of child diagnoses for children receiving community and school-based MH services are presented. Further, concordance between MH clinician-assigned diagnoses and those derived from an adapted MINI-KID (Sheehan et al., 1998), a structured diagnostic interview that corresponds with the DSM-IV, are reported. 

Methods: The current sample includes 69 children (86% male, 52% Hispanic) ages 5-14 (M = 8.81; SD = 2.63) with an existing ASD diagnosis that was validated with the ADOS-II, and their primary caregivers. Children were drawn from 12 participating publicly-funded community and school MH programs, receiving care from 60 MH therapists. Four trained study personnel administered the MINI-KID to all caregivers. Prevalence data for diagnoses from the MINI-KID and clinician-report were analyzed using descriptive statistics, Chi-square analyses were used to compare differences in prevalence between clinician-assigned and MINI-KID diagnoses. Data from the MINI-KID were compared to clinician-report diagnoses by calculating Cohen’s kappa statistic. The latter analyses were examined by four diagnostic categories: ADHD, disruptive behavior disorder, mood, and anxiety.

Results: Approximately 90% of children met diagnostic criteria for at least one non-ASD diagnosis based on the MINI-KID. The most prevalent MINI-KID diagnoses were: ADHD (88%), a disruptive behavior disorder (57%), an anxiety disorder (49%), and a mood disorder (26%). There were no statistically significant differences in prevalence rates between clinician-assigned and MINI-KID diagnoses. Overall, diagnostic agreement between MINI-KID and clinician assigned diagnoses was poor and below the standard for moderate agreement beyond chance (Cohen, 1960). The κ values for ADHD was 0.07 (poor), 0.08 for disruptive behavior disorders (poor), 0.12 for anxiety (poor), and 0.15 for mood (poor).

Conclusions: Results suggest that diagnostic co-morbidity is high among children with ASD receiving MH services. Further, diagnostic concordance between clinician report and the MINI-KID was poor for this sample of school-aged children with ASD served in CMH settings. Additional analyses will be conducted to identify factors that may explain this poor diagnostic agreement. Preliminary results underscore the need for improved diagnostic evaluation to ultimately inform efforts to improve care for youth with ASD served in CMH settings.