Change in Autism Classification with Early Intervention: Predictors and Outcome

Background: Autism spectrum disorders (ASD) form a heterogeneous set of clinical syndromes. Great variability has been described in the severity of the social-communication symptoms, cognitive and verbal skills at diagnosis and in response to intervention, as some will show limited progress while others make rapid and remarkable gains.
Evidence regarding changes in autism severity and in ASD classification with intervention using standardized tests has been quite limited due to the scarcity of follow-up outcome studies. Previous studies looking at the stability rates of the initial diagnosis when made at two years found mixed findings.  Some reported high rates of change in autism classification including moving to nonspectrum diagnosis, while others reported pronounced stability of autism symptoms and categorical diagnoses.  Studies that looked for predictors of favorable outcome reported that the child's cognitive level and verbal abilities at diagnosis have a significant impact on outcome.

Objectives: The current study had three aims: 1.To characterize the changes and stability of autism diagnostic classification with intervention in very young children. 2. To examine pretreatment factors that might predict changes or stability in autism classification. 3. To compare the intervention outcome in cognitive and adaptive domains between the groups that changed or remained stable in their initial autism classification.
Methods: Sixty-nine children diagnosed with autism (clinical observation and ADI-R), aged 15-35 months (M= 25.2, SD=4.1) participated in the study. Children underwent comprehensive evaluations at pre-intervention time (T1) and after one year of intervention (T2). The evaluation included autism classification based on Autism Diagnosis Observation Schedule (ADOS), cognitive abilities (Mullen) and adaptive skills evaluation (Vineland). The sample was divided into two groups according to their autism diagnostic categories which were derived from the ADOS algorithm scores at T2. The Unchanged Group (N=53) remained in the same autism classification and the Improved Group (N=16) changed classification to Autism Spectrum Disorder (ASD) (N=14) or Off Spectrum (N=2).
Results: After one year of intervention the majority of the examined children (77%) remained with autism diagnostic classification (ADOS algorithm), 20% of the children moved to an ASD classification, and only 3% no longer met the criteria for ASD. The Improved Group had significantly better verbal scores than the Unchanged Group at T1. No significant differences between the two groups were noted in nonverbal ability, Vineland adaptive skills, child's age, parental educational attainment and parental ages.  The Improved Group gained significantly more than the Unchanged Group in cognitive abilities and adaptive skills, and exhibited greater reduction of stereotyped behaviors with intervention.
Conclusions: Two distinct subtypes of autism were identified in very young children with autism. One with stable autism symptomatology, poor verbal abilities and limited cognitive and adaptive gains in response to intervention, and a second type with better baseline verbal abilities and overall better response to intervention in autism symptoms severity and in cognitive and adaptive skills.