Diagnostic Yield in School-Age Children Referred for Possible ASD

Background: With increased community awareness and media attention, many children are being referred for possible autism spectrum disorder (ASD) for the first time during their school age years. It is not clear how many of these older children are actually diagnosed with ASD and what impact this has on prevalence rates.
Objectives: To characterize school age children undergoing initial evaluation for possible ASD in a diagnostic referral center.
Methods: Medical records from 2008 were examined for all children ages 6 – 18 years evaluated for the first time for possible ASD in the division of developmental and behavioral pediatrics (DDBP) at Cincinnati Children's Hospital Medical Center.  DDBP is the only diagnostic referral center for ASD in a 27 county region. Assessment for ASD consists of a multidisciplinary evaluation including the Autism Diagnostic Observation Schedule (ADOS) administered by a clinician research-reliable with the instrument. Final clinical diagnosis is conferred by a developmental pediatrician following the multidisciplinary evaluation. Descriptive statistics were examined. ADOS classification was compared to final diagnosis (Spectrum vs  Not spectrum) and ANOVA was used to compare group means for cognitive and behavioral test scores.
Results: A total of 333 children ages 6 – 18 years had a first time evaluation for ASD.  Mean age was 9.9 yrs (SD = 3.2yrs). Boys numbered 280 (84%). Final clinical diagnosis was available for 308. Of these, 153 (50%) had a diagnosis within the autism spectrum: 44 (29%) autism; 45 (29%) Asperger’s syndrome and 64 (42%) ASD or PDD-NOS. There were 28 Module 1 ADOSs, 59 Module 2s, 192 Module 3s, and 54 Module 4s. Overall sensitivity of the ADOS compared to final clinical diagnosis was 91%. Specificity was 48% with little difference between original and revised algorithms. Comparison of ADOS classification to diagnosis for 209 children who had a Module 2 or 3 defined four groups in the 2x2 table: (a) 102 classified as spectrum by both ADOS and clinical diagnosis, (b) 75 classified as spectrum by ADOS but not clinical diagnosis, (c) 13 classified as spectrum by clinical diagnosis, not ADOS, and (d) 51 classified as not spectrum by both. Among the 75 children classified as spectrum by ADOS but not clinical diagnosis, 47 (63%) had a diagnosis of anxiety and/or ADHD. Mean VIQ and NVIQ scores for this group were significantly higher than the other groups (p = 0.03). The Communication Domain and composite scores on the Vineland were also significantly higher in the group classified as Spectrum by ADOS but not final clinical diagnosis (p = 0.003 and 0.004 respectively).  Mean core language scores on the CELF did not differ.
Conclusions: Only half of school age children seen in a referral center for the first time for possible ASD are found to have a diagnosis in the spectrum. The most common non-spectrum diagnosis for this group is anxiety and/or ADHD. Children in this age group who meet criteria for ASD on the ADOS, but are not diagnosed in the spectrum have a characteristic profile with higher mean IQ and communication scores.