Project Description:
We propose to study the long term outcome of Early Intensive Behavioral Intervention (EIBI) for children with
autism spectrum disorder (ASD and to test promising predictors of outcome. We will follow up at age 14-15 71
children treated in the preschool years with EIBI in publicly funded community agencies and 65 children
contemporaneously diagnosed with autism but who did not receive EIBI (Treatment as Usual, TAU). ABA
intervention models such as EIBI have empirical support from many studies, but it remains important to assess
their long term impact, particularly when delivered in community agencies, where most intervention is
administered. Previous studies have evaluated outcome primarily with respect to IQ and academic functioning.
Therefore, we will compare the outcome of EIBI and TAU with a comprehensive battery that includes
measures of ASD symptoms, intellectual function, adaptive skills, social competence, language, academic
achievement, and psychopathology. These variables not only have empirical support as being core and
associated features of ASD, but are related to the skills acquired in EIBI. We expect children with EIBI to have
better outcome than the TAU sample overall. We will use marginal structural models (MSMs) to examine
group differences; MSM uses the probability of treatment selection as a weight to adjust for the bias in
treatment selection, a technique known as the inverse probability weighting. Finally, noting the labor
intensiveness and expense of EIBI, we will test pre-treatment predictor variables with some empirical support
from prior studies by our research group and other investigators: social engagement and stereotypy. We will
use multiple linear regression to test the hypothesized predictors. To improve robustness, we will use the
generalized estimating equations (GEE) for inference. Identifying predictors of outcome of EIBI will permit
therapists to avoid focusing on children unlikely to benefit from this treatment and facilitate future efforts
to.develop effective interventions for these children. The proposed study represents the longest follow-up of
large EIBI and TAU samples, such that we would have a uniue opportunity to examine long-term outcomes
and the prediction of outcomes for EIBI.
Project
Unserved or Under-served Populations:
Racial or Ethnic Minorities, Disadvantaged Circumstances, Limited English, Geographic Areas, Urban
