Service Deserts and Service Oases: Utilizing GIS to Evaluate Service Availability for Individuals with ASD

Background: ASD is a pervasive neurobehavioral disorder characterized by deficits in social communication and interactions, and restricted/repetitive behaviors, interests, or activities.^1,2 ASD’s core symptoms and common co-occurring psychological disorders require specialized behavioral and psychological services throughout the life of individuals with ASD.^3,4 However, access to high quality specialty services can be hindered by socio-environmental determinants, such as a lack of available ASD providers within communities where individuals with ASD and their families live.^5,6

Objectives: As a response to the paucity of literature related to the availability of community-based ASD services, this study utilized GIS methods to map the locations of ASD agencies across the state of Michigan. We hypothesized: 1) Population will be correlated with the availability of service such that fewer agencies will be located in less populated areas as compared with more populated areas; 2) SES will be correlated with the availability of ASD services such that areas with low SES will have fewer ASD services as compared with areas with medium and high SES; and 3) An interaction will be found between population and SES such that areas with low SES and high population densities will have fewer available ASD services as compared to areas with medium or high SES and high population densities.

Methods: We compiled a list of ASD providers within Michigan by conducting a systematic, iterative web-based search by county. We then geocoded the location of verified ASD specialty service providers using ArcGIS 10.4.1, and used network analysis to assess available ASD services in relation to population distribution, socioeconomic disadvantage, urbanacity and immobility.

Results: Individuals living in rural areas had significantly fewer ASD agencies available within 10 miles than individuals in suburban and urban areas—average distance to nearest ASD provider was 12.43 miles for rural areas versus 2.82 miles in urban and 3.36 miles in suburban areas. Spatial autocorrelation (Moran’s I) indicated that the spatial distribution of ASD agencies was not random (p<0.01, Moran’s Index<0.01); areas with greater socioeconomic disadvantage had fewer nearby ASD agencies available. Finally, Getis-Ord GI* clustering analysis was used to determine statistically significant spatial disparities. While most urban areas had good availability of ASD services, the region of Metro Detroit exhibited a significant disparity in availability, such that wealthier suburbs had good availability of ASD agencies while few agencies were available in the poorer urban area. Thus, high population and low SES alone do not seem to determine the distribution of ASD agencies.

Conclusions: Knowing the landscape of ASD availability (and the specific areas that are particularly poorly serviced) gives service providers and healthcare planners the opportunity to advocate for increases in ASD services in these underserved areas. Additional services research is needed to identify system-level barriers contributing to disparate service availability. While addressing gaps in urban and rural regions may be difficult, use of evidence-based dissemination and implementation strategies may facilitate adoption and uptake of more nimble treatment options such as integration of behavioral health and primary care, school-based health centers, mobile autism clinics or telehealth services.^7-10