Poster #20 - “You're simply the best”: IQ and baseline Vineland better than all the rest at predicting adaptive functioning among youth with Autism Spectrum Disorder

• In Event: Poster Session 02

Abstract

Background: Standard cognitive assessments have been identified as the most robust predictor of later adaptive functioning in youth with autism spectrum disorder (ASD), but IQ may be less accessible or limited in its predictive ability for children with greater autism severity (e.g., Liss et al., 2001). Common ASD comorbidities, like psychiatric issues, gastrointestinal problems, and sleep disturbance, are more accessible for families and clinicians to assess and are mostly unrelated to ASD symptom severity (e.g., Simonoff et al., 2008; Neumeyer et al., 2018). Previous literature found associations between adaptive functioning and psychological problems and physical health problems (Baghdadli et al., 2012; Hus Bal et al., 2015; Szatmari et al., 2009; Taylor et al., 2012). However, past studies have not assessed behavioral and physical health issues in comparison to previously identified predictors of adaptive functioning (e.g., IQ).

Objectives: The objective of this study is to compare the predictive strength of multiple baseline predictors (i.e., demographic information, cognitive ability, adaptive skills, behavioral health issues, and physical health issues) of adaptive skill development four years later.

Methods: A total of 565 children with ASD (ages two to 17; 80% male; average IQ of 77), who were previously enrolled in the Autism Treatment Network (ATN) Registry and follow-up, were selected for the current study. The follow-up assessment took place roughly four years after their baseline assessment. Prediction of follow-up VABS was estimated in three models using multivariable regressions. Each model groups the variables into different groups to assess how these variables function as scale scores or as larger sets of predictors. Model 1 evaluated three sets of predictors: demographics (age, race, sex, income, education, Vineland Aberrant Behavior Scale (VABS) scores, and IQ), behavioral health (ADOS scores and Child Behavior Checklist internalizing and externalizing problems scores), and physical health (gastrointestinal and constipation problems, seizures, and Child Sleep Health Questionnaire score). Model 2 treated each measure’s scale scores as predictors. Lastly, model 3 treated each individual subscale score as a variable. Semi-partial coefficients of determination (R-squared) were used to provide estimates of the amount of variance in follow-up VABS that could be independently explained in each model.

Results: A little less than half of the variance in follow-up VABS was explained by the model with all focal predictors (42.4%). Almost all the variance was accounted for by the demographic variables. Baseline VABS significantly and independently explained the largest portion of variance (13.4%, p < .001) followed by age (1.6%, p < .001) and IQ (1.5%, p < .001). Behavioral measures as a group collectively independently explained 0.5% of the variance and physical health measures independently explained 0.7% but were not significant. Our findings replicate the pre-existing literature among a diverse and representative sample of children in the autism population and suggest clinicians and caregivers should seek out IQ and Vineland assessments to best assess later social and functional outcomes for children with ASD. And although our model did not find health measures to be significant predictors, these measures may still indirectly contribute to adaptive functioning outcomes.

Group Authors

Autism Care Network (formerly Autism Speaks - Autism Treatment Network)

Author

• Siobhan S Glynn, University of California - Los Angeles
  Presenting Author