Diagnostic Testing Practices for Autism Spectrum Disorder in Four US Populations

Background: The use of standardized tools is a recommended professional practice when considering an autism spectrum disorder (ASD), although little research has been done to characterize the use of these tools in community population samples.   

Objectives: To characterize ASD test use in developmental evaluations of children with ASD. 

Methods:  The Autism and Developmental Disabilities Monitoring (ADDM) Network monitors prevalence and describes characteristics of children with ASD in multiple areas of the United States. Descriptive statistics were used to describe the frequency and types of ASD tests documented in developmental evaluation records and type of community professionals who administered the tests. Chi-square analyses were used to assess differences in ASD test use by surveillance site (i.e., AZ, GA, MD, and SC), child sex, child race/ethnicity, and record type (i.e., education or health).

Results: 1,708 eight-year-old children (N=1,438 males and N=270 females) met ADDM DSM-IV-TR criteria for an ASD in the 2008 surveillance year after review of evaluation records through age 8 from multiple health and education sources. Of these children, 76.2% had at least one ASD test documented in evaluation records with an average of 3.7 ASD tests per child. Evaluations with ASD tests were administered between 9 and 107 months of age with a median of 70 months. Of all the ASD tests documented, the Childhood Autism Rating Scale was the most commonly used (CARS; 30.4%), followed by the Gilliam Autism Rating Scale (GARS; 26.6%), Autism Diagnostic Observation Schedule (ADOS; 9.7%), Gilliam Asperger Disorder Scale (GADS; 9.4%), Asperger Syndrome Diagnostic Scale (ASDS; 7.0%), and the Autism Diagnostic Interview-Revised (ADI-R; 2.0%). Other ASD tests (e.g., Autism Behavior Checklist) accounted for 14%. ASD tests were most commonly administered by PhD/PsyD/MA psychologists (45.7%); followed by EDD/EDS psychologists (24.8%); developmental pediatricians (14.2%); educators (3.1%); other types of MDs (e.g., neurologists, psychiatrists) (3.0%); speech language pathologists or occupational therapists (2.4%); social workers or licensed counselors (0.4%); and other or unknown professional types (6.4%). Some surveillance sites documented ASD test use more than others (χ²=55.85, p<0.001), with the greatest proportions of documented testing in SC (86.0%) and the least in MD (68.2%). ASD tests were more frequently administered to boys diagnosed with ASD (77.1%) compared to girls (71.1%) (p<0.05). There were also group differences race and ethnicity of the proportion of children evaluated by an ASD test (p<0.001): Asian or Pacific Islanders (81.6%); black, non-Hispanics (80.0%); white non-Hispanics (75.7%); and Hispanic children (65.0%). ASD tests were more frequently administered in education settings (61.2%) than health settings (38.8%) (p<0.001).

Conclusions: While three-quarters of children identified with ASD were evaluated using an ASD test, very few were administered “gold-standard” tools (the ADOS and ADI-R) in in community health and education settings.  Testing practices varied by site. ASD tests were administered more frequently to boys than girls, and least frequently to Hispanic children. ASD tests were also administered more frequently in education settings, and by psychologists. These results highlight the need for continued education on the importance and use of standardized tools for all children evaluated for an ASD.