The Society for Social Work and Research

2013 Annual Conference

January 16-20, 2013 I Sheraton San Diego Hotel and Marina I San Diego, CA

Differentiating Treatment Responsiveness Among Subgroups of Children with Autism Spectrum Disorders

Saturday, January 19, 2013: 4:30 PM
Executive Center 1 (Sheraton San Diego Hotel & Marina)
* noted as presenting author
Robert Hock, PhD, Assistant Professor, University of South Carolina, Columbia, SC
John Kuntz, MS, Doctoral Student, University of South Carolina, Columbia, SC
Brian K. Ahmedani, PhD, Research Associate, Michigan State University, Detroit, MI
Background: The heterogeneity of symptoms and abilities among individuals with autism spectrum disorders (ASD) poses significant challenges to the timely diagnosis and effective treatment of this population. Even the most efficacious treatments, such as Applied Behavioral Analysis (ABA) yield highly variable outcomes. Researchers are now attempting to reduce this heterogeneity by clustering individuals into clinically distinct subgroups. Of particular importance, some studies suggest that specific behavioral, cognitive and language profiles are associated with ABA outcomes. While these studies are yielding promising results, few have examined the utility of subgroups to predict real-world outcomes for individuals with ASD and their families. The objectives of this study are to 1) evaluate the outcomes of a statewide, Medicaid-funded ABA program for children with ASD; and 2) determine whether clinically distinct ASD subgroups predict differential response to ABA.

Methods: This study uses longitudinal data from South Carolina’s Pervasive Developmental Disorder waiver program, which provides three years of ABA intervention to children aged 3-11. A total of 71 children who completed 3 years of EIBI were included in the study.  Each child completed the Vineland Adaptive Behavior Scales (VABS), Peabody Picture Vocabulary Test (PPVT) and the Expressive Vocabulary Test (EVT) at baseline and after the first and second years of treatment. To assess treatment outcomes, paired sample t-tests were used to compare baseline and year 2 scores.  Additionally, a reliable change index was calculated to determine whether each child’s change exceeded statistical variability.  Children were classified into subgroups using statistically derived hierarchical clusters based on baseline scores.  Cubic Clustering Criteria, Pseudo-F statistics and Pseudo-t statistics were examined to determine the appropriate number of clusters.  A 4-cluster solution was specified by all 3 criteria.  A general linear model (GLM) with repeated measures was used to corroborate the observed change in outcomes and to test whether those differences varied by cluster. 

Results: Paired t-test results indicated that children showed significant improvements across all measures of functioning after two years of ABA treatment (p<0.05). Additionally, more than 50% of children achieved reliable change across all domains addressed by ABA. Within the Communication, Social Skills and Behavioral Composite domains, 70% of children achieved reliable change. Similarly, the GLM model yielded a significant difference (p=.001) across all treatment domains.  Despite the small sample size, there is evidence that the cluster of children with consistently high baseline measures showed the lowest improvement, particularly on PPVT and EVT, compared to other clusters.  The interaction between cluster and level of improvement was significant on measures of Social Skills (F=7.29 p=.0002) and PPVT (F=5.01, p=.0035) while other measures were also significantly different by cluster (Adaptive Behavior Composite: F=3.29, p=.026; Expressive Vocabulary: F=4.0, p=.011).

Conclusions: These findings suggest that community-delivered ABA improved the skills and adaptive functioning of children with ASD. Further, this study yields promising results regarding the utility of ASD subgroups for predicting ABA treatment outcomes.  This knowledge may help social workers and other professionals match children with efficacious treatments, leading to improved outcomes and more effective allocation of limited resources.