Abstract
Disparities exist in the early identification of underserved children with ASD. Our study used a large university sample to examine subtypes of children with an eventual ASD diagnosis based on parent concerns and sociodemographics. Results suggest that children with speech parent concerns are most likely identified earlier, regardless of race, ethnicity, or SES. However, our findings point to the difficulty in identifying girls with ASD and children with social and behavior parent concerns.
Primary Author and Speaker: Anna Wallisch
Contributing Authors: Lauren Little, Evan Dean, Winnie Dunn
The American Academy of Pediatrics (AAP) highlights the importance of listening closely to parent concerns to identify children with autism spectrum disorder (ASD) earlier, given that parent concerns often precede an ASD diagnosis (AAP, 2014). However, disparities exist in the early identification of underserved (i.e., low socioeconomic status [SES], racial/ethnic minorities) families and girls with ASD (Daniels & Mandell, 2014). Occupational therapists (OTs) often serve families and children prior to a formal diagnosis; if we understand how parent concerns and socio-demographics group by subtypes we may better identify children and tailor services across diverse groups. Research questions include: 1) Among children with ASD, how do early parent concerns, child (i.e., age, gender), family (i.e., race, ethnicity, SES) and community (i.e., provider access) characteristics group by subtypes; 2) To what extent do subtypes of children with ASD differ by chronological age (CA) at the diagnostic evaluation and age of child when a parent was first concerned (AFC)?
We performed a secondary analysis with data drawn from a large university diagnostic center. The sample included 712 children, 12 months to 12 years-11 months old who received a diagnostic evaluation and were later diagnosed with ASD. Data was drawn from intake paperwork parents completed prior to a diagnostic evaluation. Parents reported their top 3 concerns, which were coded into 6 categories including: 1) behavior, 2) medical, 3) development, 4) speech, 5) social, and 6) stereotyped behaviors. Parents also reported AFC, race and ethnicity. SES was determined by use of Medicaid or non-Medicaid insurance at the evaluation. Community factors were determined by the number of service providers in the county a family lived. For question 1, we used latent class analysis (LCA) to examine subtypes based on parent concerns, child (i.e., age, gender), family (i.e., race, ethnicity, SES) and community characteristics (i.e., provider access). For question 2, we used the nonparametric Kruskal Wallis H test.
Parent concerns and socio-demographics distinguished 5 latent classes. Two subtypes were identified around 3.5 years of age and were differentiated by 2 parent concerns: speech and medical concerns. One of the younger subtypes included non-white, Hispanic children utilizing Medicaid. One subtype was identified around kindergarten and was differentiated by stereotyped and developmental parent concerns. Two subtypes were identified around 9 years of age with either developmental concerns, or social and behavior concerns. One of the oldest subtypes was characterized by girls with ASD. The Kruskal Wallis H tests indicated a significant main effect for CA x 2(4)= 419.363; p<0.001 and for AFC x 2(4)= 40.647; p<0.001.
Results suggest our systems are capturing subtypes of ASD with speech parent concerns earlier, regardless of race, ethnicity and SES; however, we are missing girls and children with social and behavior parent concerns. A substantial time gap exists between CA at the evaluation and AFC for all subtypes, and especially when subtypes consisted of girls or children with social and behavior concerns. This is especially concerning as it takes years for our systems to capture these children. We need to become more vigilant when listening to various parent concerns.
Many socio-demographics are linked to later diagnoses of ASD, and we have yet to determine methods to identify these children earlier. These gaps are especially pressing as a goal of Healthy People 2020 is to lower the age of identification. We used a novel LCA approach to parse the variability in ASD to inform screening methods. By listening to parent concerns OTs may identify children earlier and tailor services to diverse needs.
American Academy of Pediatrics. (2014). AAP Publications Reaffirmed or Retired. Pediatrics, 134(5). doi:10.1542/peds.2014-2679
Daniels, A. M., & Mandell, D. S. (2014). Explaining differences in age at autism spectrum disorder diagnosis: a critical review. Autism : The International Journal of Research and Practice, 18(5), 583–97. doi:10.1177/1362361313480277