The Association of Child Characteristics on Outcome in a School-Based Behavioral Intervention

Background: Although interventions based on the principles of applied behavior analysis have been repeatedly cited as evidence-based for individuals with autism, significant discrepancies in outcome are evident within the literature.  That is, in most studies evaluating intervention outcomes for students with ASD, some students make significant progress, while others make minimal to no progress.  The baseline differences in children’s clinical characteristics have been hypothesized as an important predictor of treatment outcome. Few studies, however, have had the sample size, a controlled treatment environment, and the careful characterization of children needed to address this question. 

Objectives: The primary goal of this study was to identify individual characteristics that predicted differences in outcome for elementary-school-age children with autism. Specifically, the study evaluated differences in outcome following one year of a behaviorally-based intervention as a function of children’s cognition, challenging behavior, language ability, autism severity, social skills, adaptive behavior, co-morbidity of psychological diagnoses, and restrictive and repetitive behavior. 

Methods: The sample included 368 students with autism in kindergarten-through-second-grade classrooms in a large urban school district that participated in an intervention study.  As part of the intervention study, students received instruction using the Strategies for Teaching Based on Autism Research (STAR) program as part of their classroom routines.  The STAR program is a behaviorally based program that comprises Discrete Trial Training, Pivotal Response Training, and Teaching within Functional Routines. Intervention fidelity was measured monthly via fidelity checklists and observations by trained recorders. Data were collected at the beginning and end of the school year.  Measures were completed by teachers, parents and direct observation, and included the Differential Ability Scales, 2^nd edition; Aberrant Behavior Checklist; Adaptive Behavior Assessment System, 2^nd edition; Autism Diagnostic Observation Schedule; Child Symptom Inventory; PDD Behavior Inventory; and Social Responsiveness Scale. The outcome of interest was change on the Differential Abilities Scale. Linear regression with random effects for classroom and student were used to evaluate the association between change in outcome and baseline variables of interest following one year of intervention.

Results: Analyses are ongoing. Preliminary adjusted analyses indicate that higher expressive language skills at the start of the year were associated with improved outcome.  Increased anxiety was associated with poorer outcomes.  No other child-level variables were associated with outcomes. 

Conclusions: The finding that children with greater expressive language skills at the start of the intervention year had better outcomes, on average, than other children is in keeping with prior research. These results may imply that an earlier focus on improving language skills in young children with autism could lead to improved outcomes for these students.  The finding that children with greater anxiety were more likely to have poorer outcomes is novel, and suggests the importance of identifying and treating anxiety symptoms early.  The finding that other variables such as baseline cognition, autism severity, challenging behavior, adaptive behavior, and social skills were not associated with outcome is promising, and suggests the broad benefit of behavioral approaches to early elementary education of children with autism.