Access to Therapy for Children with Autism: A Population-Based Analysis

Background:  Improving early access to effective care for children with autism spectrum disorder (ASD), while minimizing costs, is an important health policy objective.  Occupational, physical, and speech-language therapy are frequently utilized non-behavioral services for children with an ASD (Thomas, Ellis, McLaurin & Morissey, 2007), and these services are utilized significantly more by children with ASD than by children with other special health care needs (SHCN) (Chiri & Warfield, 2012). Historically, disparities have been found in access to other services for children with ASD, particularly for underserved populations. However, existing research on therapy access is dated and studies were conducted prior to major policy and social changes that may impact access. 

Objectives:  The aim of this research was to examine population-based trends in access to therapy services for children with ASD under the age of 18 years, to compare differences in access to therapy services between children with ASD and children with SHCN, and to identify individual, family enabling, and need characteristics that contribute to reduced access to therapy services.

Methods:  A retrospective, cross-sectional analysis using the 2005-06 and 2009-10 National Survey for Children with Special Health Care Needs (NS-CSHCN) was conducted. Weighted logistic regression models were used to examine access to therapy services at two time points for children with ASD between 3 and 18 years of age, and between children with other SHCN. NS-CSHCN datasets from the two time periods were concatenated after data cleaning to ensure appropriate merging of similar variables at both time points, and to ensure appropriate structure for use of imputed datasets (Blumberg et al., 2008). A series of nested logistic regression models were developed for analysis in Stata to control for the complex survey design and multiple imputation structure.  The outcome variable was unmet need for therapy services, defined as children identified as needing therapy services in the past year, but who were reported to not receive all services in the past year (access variable). Predictor variables included diagnosis (ASD, SHCN), survey year, and a diagnosis by year interaction term, as well as models that included stepwise individual, enabling, and need characteristics that may be associated with unmet need for services.

Results:  Results suggest that access rates were significantly poorer for children with ASD sampled in 2009 compared to 2005, with 24.7% (21.7-28.0, 95%CI) of children with ASD not receiving services in 2009 compared to 17.9% (15.0-21.3, 95%CI) of children with ASD not receiving services in 2005. Additionally, these rates of therapy access problems are significantly greater for children with ASD compared to children with other SHCN, with 12.8% (11.4-14.2, 95%CI) and 16.0 (14.6-17.5, 95%CI) of children with other SHCN not receiving needed therapy in 2005 and 2009 respectively.  Additional analyses on groups with similar physical (e.g. fine/gross motor) and mental/behavioral need factors (e.g. socialization, behavior problems) will be described. Predisposing, family enabling, and need characteristics contributing to reduced access for children with ASD will be presented.

Conclusions: Results will be discussed in the context of potential policy implications and areas for future systemic interventions.