Need a New DSM-5 ASD Assessment? Just AASC?

Background:  

The under-identification of high-functioning autism (HFA) in school age children is an ongoing problem, especially more able students with high-functioning autism (Wilkerson, 2010).  Lost time due to under-identification or failing to provide needed services will diminish the developmental potential of children with ASD (Pool & Hourcade, 2011).  More than half the children with autistic impairments, at the same levels as those with an ASD label, are not identified even though they have the same needs for support in educational settings (Russell et al., 2010). 

Objectives:  

The need to assess individuals with high-functioning autism is well demonstrated (Barnhill, 2007).  A simple pre-screening assessment, with input from both a parent and an educator, would provide a method for identifying individuals in need of evaluation.   The objective of this research is to design an assessment that is both DSM-5 compliant and easy to use.

Methods:  

This study looks at existing data consisting of participant demographic data and item responses to the 272 question Ellis Functional Assessment (EFA) for high-functioning autism.  The EFA is a measurement assessment that examines areas of functional difficulty for people with high-functioning autism.  There were 538 participants in the original study.  The data set was later expanded to include 740 participants.

This data was analyzed using exploratory factor analysis to assess how many latent variables are included in the assessment. This information was used to design a short 15 to 25 question assessment pre-screen students for referral to the child study team for evaluation to receive special educational services for autism spectrum disorders.  This assessment targets identifying students at the high-functioning end of the spectrum.

Confirmatory factor analysis (CFA) was used to test the model which was specified in advance to run the analysis (Thompson, 2010). The goal of CFA is to test a specific model or hypothesis (the shortened assessment) (Osborne, 2008).

Results:  :  The results for the final model of this confirmatory factor analysis produced a model with χ² = 8730.610, df = 655 (p < .001) and resulted in CFI = 1.00, NFI= 1.00 and RMEA = .243.  After examining the Modification indices, the parameters with indices over 80 were freed and the resulting model was then analyzed. This produced a model with χ2 =3294.790, df = 627 (p < .001) and resulted in CFI = 1.00, NFI= 1.00 and RMEA = .104.   This represents a significantly better fit (χ2 = 5435.82, df = 28, p < 0.00001).  The assessment was also evaluated for internal reliability.  This resulted in a Cronbach’s Alpha = .941, indicating internal reliability.

Conclusions:  

These five identified factors, covered the majority of the variance in the model and covered the most important aspects of ASD identified in literature.  These factors cover all of the elements used for diagnosis in DSM-5.  Because of the data in the EFA included behavior from early childhood the third requirement of DSM-V, symptoms must be present in the early developmental period is satisfied.  The AASC is a valid, reliable screening instrument for Autism Spectrum Disorders.