Time-Lag Between Diagnosis of Autism Spectrum Disorder and Onset of Publicly-Funded Early Intensive Behavioral Intervention: Do Race–Ethnicity and Neighborhood Matter?

J Autism Dev Disord DOI 10.1007/s10803-017-3354-3 ORIGINAL PAPER Time-Lag Between Diagnosis of Autism Spectrum Disorder and Onset of Publicly-Funded Early Intensive Behavioral Intervention: Do Race–Ethnicity and Neighborhood Matter? Marissa E. Yingling1 · Robert M. Hock2 · Bethany A. Bell2 © Springer Science+Business Media, LLC 2017 Abstract Health coverage of early intensive behavioral intervention (EIBI) for children with autism spectrum disorder (ASD) is rapidly expanding across the United States. Yet we know little about the time-lag between diagnosis and treatment onset. We integrated administrative, Medicaid claims, and Census data for children in an EIBI Medicaid waiver (n = 473) to examine the relationship between timelag and (a) child race–ethnicity and (b) neighborhood racial composition, poverty, aluence, and urbanicity. We explored whether the relationship between child race–ethnicity and time-lag varies by neighborhood characteristics. Average time-lag between diagnosis and treatment onset was nearly 3 years. Child race–ethnicity and neighborhood characteristics did not predict time-lag. Reducing time-lag is critical to ensuring that children with ASD receive treatment as early as possible. Keywords Autism spectrum disorder · Early intensive behavioral intervention · Medicaid · Time-lag · Disparities Introduction To address the needs of the 1 in 68 children in the United States who meet criteria for autism spectrum disorder (ASD) (Centers for Disease Control and Prevention 2014), health coverage of early intensive behavioral intervention (EIBI) is * Marissa E. Yingling marissa.yingling@louisville.edu 1 Kent School of Social Work, University of Louisville, 2217 S 3rd St., Julius John Oppenheimer Hall, Louisville, USA 2 College of Social Work, Hamilton College, University of South Carolina, 1512 Pendleton Street, Columbia, USA rapidly expanding. Three critical developments are underway. First, 44 states, the District of Columbia, and the U.S. Virgin Islands each have health insurance mandates requiring coverage of EIBI, and there are ongoing eforts to pass mandates in remaining states (Autism Speaks 2016). Second, in an efort to prevent duplication of services, states who adopted EIBI via 1915(c) Home and Community-Based Services (HCBS) Medicaid waivers (e.g., South Carolina Department of Disabilities and Special Needs 2007) are required by the Centers for Medicare and Medicaid Services to transition the service to Medicaid state plans (Centers for Medicare and Medicaid Services 2014). Finally, 29 states and the District of Columbia require that individual and small business health plans cover behavioral intervention based on the principles of applied behavior analysis (ABA) under their ten “essential health beneits” (Autism Speaks 2014). Despite these changes in healthcare policy and service provision, the Interagency Autism Coordinating Committee (2013) cites disparities in access to early intervention services as among the most pressing yet understudied areas of research on ASD. Similarly, the World Health Organization (World Health Organization 2013) highlights the need for research that focuses on inequitable access to services for children with ASD, as well as the challenges of implementing large-scale, community-based early intervention. Indeed, although states are expanding health coverage of EIBI, research on access to this service is lacking. Evidence suggests that children with ASD experience disparities in access to a range of services (Liptak et al. 2008; Magaña et al. 2012; Murphy and Ruble 2012; Parish et al. 2012; Shattuck et al. 2009; Siller et al. 2014; Tregnago and CheakZamora 2012). Yet states are providing EIBI with minimal research to guide equitable service delivery. The swell in 13 Vol.:(0123456789) J Autism Dev Disord the number of children eligible to receive EIBI is outpacing research that could inform policy and practice. Importance of Early Diagnosis and Treatment Onset Although families use a variety of services to treat ASD (Green et al. 2006; Thomas et al. 2006), EIBI is a wellestablished, evidence-based treatment (Howlin et al. 2009; Lovaas 1987; National Autism Center 2015; Reichow et al. 2014). It involves the application of ABA procedures in oneon-one instruction of adaptive and functional skills (e.g., communication, cognitive skills) in young children (Klintwall and Eikeseth 2014), and it is the most preferred and frequently used treatment approach among parents surveyed in prior research (Stahmer et al. 2005; Thomas et al. 2006). Meta-analyses of EIBI indicate that children make signiicant improvements in intellectual ability, language, social communication, and daily living skills (Peters-Scheffer et al. 2011; Virués-Ortega 2010), and evidence suggests that the younger children are upon treatment entry, the more likely they are to improve on key outcomes (Granpeesheh et al. 2009; Makrygianni and Reed 2010; Perry et al. 2011; Virués-Ortega et al. 2013). For example, unprecedented research on children enrolled in a province-wide program in Ontario, Canada revealed that compared to children who began EIBI at 4 years old or older, children who began EIBI younger than 4 years old completed the program with better outcomes on most measures (Perry et al. 2011). Currently, the dominant focus of research on access to services is the age at which children receive a diagnosis of ASD and the predictors of age of diagnosis. In recent years, evidence of delays and disparities in diagnosis prompted state- and nation-wide initiatives to screen and evaluate children earlier, and there are ongoing developments to improve and expand these eforts (Peacock and Lin 2012). This line of research is critical and well justiied; the sooner children receive a diagnosis, the sooner they can enter treatment. Yet there exists an underlying assumption that after receiving a diagnosis, children immediately access treatment. To the contrary, parents report that post-diagnosis, they are unsure of where to turn or what steps to take. In a study that highlighted parents’ experiences after diagnosis, parents reported feeling “abandoned by the community… like trying to run through a maze blind folded” (Moodie-Dyer et al. 2014, p. 355). In the post-diagnosis period, parents enter an unknown and complex reality. In addition to uncertainty, they face years-long waitlists (L & M Policy Research, LLC 2014), a shortage of providers (Wise et al. 2010), and other factors that act as barriers to treatment utilization, such as competing time demands of school (Yingling et al. 2017). Given the importance of enrolling children in treatment as early as possible, developing an understanding of how children 13 proceed from diagnosis to treatment onset as well as factors associated with this time-lag is imperative. In particular, research suggests that child race–ethnicity and neighborhood characteristics are factors to consider. Race–Ethnicity and Time to Treatment Onset Current literature suggests that black and Latino children with ASD may be less likely to receive timely treatment. In what appears to be the only study in which the impact of race on time-lag to treatment is explored, researchers demonstrated that African American children may be at a disadvantage (Burkett et al. 2015). When asked to explain delays in diagnosis and treatment initiation among African American children, family members cited distrust of healthcare providers as well as a lack of information about ASD and the resources available to them (Burkett et al. 2015). These results add to a larger body of literature on race–ethnicity and children’s mental health treatment, in which African American parents report more negative help-seeking attitudes (i.e., acknowledgement of the existence of psychological problems and propensity to seek assistance from a professional) and greater mental health stigma than European Americans or Hispanic Americans (Turner 2009), and more negative expectations of treatment outcomes than Caucasian parents (Bussing et al. 2003). Among African American mothers living in rural areas, Murry et al. (2011) found that although they were conident that mental health professionals could help their children, the mothers preferred to seek support from family, church, and schools. They also cited community stigma and cultural distrust as barriers. Further, in a study on predictors of EI evaluation and enrollment, children with mothers who identiied as black and reported higher poverty levels were less likely to receive an evaluation and to enroll in services (Clements et al. 2008). To date, Magaña et al. (2013) are the only researchers to examine the role of ethnicity at the individual level on enrollment in a Medicaid-funded EIBI program. In Wisconsin’s program, they documented that despite reporting more unmet needs, parents of Latino children with ASD were less likely to enroll. This aligns with available literature on Part C Early Intervention (EI) services. For example, in a study that investigated diferences in parents’ experiences with the pediatric referral process to EI services, parents with lower health literacy, or “the degree to which individuals can obtain, process, and understand the basic health information and services they need to make appropriate health decisions” (Selden et al. 2000), experienced greater diiculty contacting EI providers and were confused about the referral process, including written materials provided to them by pediatricians (Jimenez et al. 2013). Limited health literacy, particularly in the twenty-irst century when the internet is a primary source of information for ASD services, J Autism Dev Disord disproportionately impacts parents who identify with a minority race and who report limited English proiciency and lower educational attainment (Knapp et al. 2011). In sum, extant literature on race–ethnicity and access to services among children with ASD highlights the importance of examining the role of race–ethnicity in the time-lag between diagnosis and treatment onset. Neighborhood Characteristics and Time to Treatment Onset A child’s neighborhood may also impact timely treatment onset. There is substantial, longstanding evidence that characteristics of an individual’s residence afect healthcare access (e.g., racially/ethnically segregated neighborhoods and neighborhoods with high concentrations of poverty). At the turn of the twenty-irst century, Williams and Collins (2001) argued that racial residential segregation, commonly associated with concentrated poverty, is a fundamental cause of healthcare disparities. Although additional research is needed to target speciic underlying causes of healthcare disparities within racially/ethnically segregated neighborhoods, literature published in the past 15 years underscores their argument (White et al. 2012). Indeed, research points to an association between neighborhoods with socioeconomic disadvantage and poor access to healthcare generally, a relationship that persists after controlling for individual-level characteristics (Kirby and Kaneda 2005). Neighborhoods marked by socioeconomic disadvantage may experience challenges such as recruiting providers to serve the area (Auchincloss et al. 2001; Kim et al. 2009). Notably, racial and ethnic minorities are more likely to live in neighborhoods with high poverty rates. On average, white children live in neighborhoods with a poverty rate of 7%, whereas black and Latino children live in neighborhoods with poverty rates of 21% and 19%, respectively. As Acevedo-Garcia et al. (2008) assert, the “worst-of white children are better of than the majority of black and Hispanic children, and these disparities are not accounted for by diferences in family poverty” (p. 324). White children rarely experience double jeopardy, whereby children live in poor families and in poor neighborhoods. A study by Shattuck et al. (2009) is the only one to date on the relationship between neighborhood characteristics and access to EIBI. Using administrative data of Wisconsin’s Medicaid-funded EIBI program, researchers found that children who lived in census tracts with a higher percentage of families with incomes ≥ 200% of the federal poverty level, a higher percentage of women 25 and older with at least a high school degree, and a higher percentage of people who were white were more likely to enroll. These results are similar to those in related areas of research. In a study on timeliness of provider designation in New York City’s EI program, children who lived in low-income neighborhoods and neighborhoods with large Spanish-speaking populations experienced delays (Kim et al. 2009). Researchers cited the shortage of providers in the state, both monolingual and bilingual, as well as providers who perceive low-income neighborhoods as less desirable work environments, as possible explanations for observed delays. Likewise, children who lived in socially disadvantaged neighborhoods were at increased risk of never receiving EI services (McManus et al. 2013). In addition to racial residential segregation and neighborhood poverty, individuals who reside in rural areas experience numerous barriers to healthcare, including provider shortages, higher healthcare costs, and geographic isolation, which translates to increased travel time and transportation costs (Florence et al. 2012). This appears to be the case among children with ASD. For example, Murphy and Ruble (2012) found that parents who lived in non-metropolitan counties in Kentucky were more likely to report diiculties accessing professionals trained in the treatment of ASD compared to parents who lived in metropolitan counties in Kentucky. Importantly, although rural populations are overwhelmingly white, the number of racial and ethnic minorities who live in rural areas grew 21.3% between 2000 and 2010 (Johnson 2012). This trend suggests that it is more important than ever to examine whether the relationship between race–ethnicity and access to treatment varies by neighborhood urbanicity. Behavioral Model of Health Services Use In response to a lack of research on access to EIBI, it is paramount that researchers attend to the labyrinthine system parents navigate to access treatment for their children. Central to this efort is the recognition that the term “access,” which is commonly plagued by ambiguity and used diferently across studies, is a multidimensional construct. To make progress, it is essential that researchers move beyond the dichotomization of access (e.g., yes/no), and instead identify and more precisely measure multiple indicators of access over time. Andersen’s Behavioral Model of Health Services Use (BMHSU; Andersen et al. 2013), a well-established model commonly used in health services research, ofers a solid foundation for this work. Originally published in 1968, six subsequent revisions relect decades of advancements in health services literature. Broadly, the authors deine access as “actual use of personal health services and everything that facilitates or impedes their use” and “the link” between health services and people who receive the “right services at the right time to promote health outcomes” (2013, pp. 33–34). The goals inherent in the model include the prediction of service use, promotion of social justice, and enhancement of the effectiveness and efficiency of 13 J Autism Dev Disord service provision. In addition to these goals, we selected this model because of its multiple dimensions of access, lexible application to a range of health services, inclusion of both individual and contextual factors, and bidirectional relationship between four domains (i.e., individual characteristics, contextual characteristics, health behaviors, and outcomes). Furthermore, literature on the BMHSU provides insight into explanations for and solutions to disparities in access. The BMHSU includes individual and contextual predisposing, enabling, and need characteristics. In the current study, we focus on child race–ethnicity and neighborhood characteristics (i.e., individual and contextual predisposing social characteristics, respectively), and two of the model’s six dimensions of access, or realized access (i.e., utilization) and inequitable access (i.e., access due to social structure variables, such as race–ethnicity). We identify factors that contribute to the timely onset of EIBI (i.e., utilization of initial therapy session). In particular, we examine the impact of a child’s race–ethnicity and neighborhood racial composition, poverty, aluence, and urbanicity to determine if there is inequitable access to EIBI. For the purpose of this study, we deine realized access as the timely onset of EIBI postdiagnosis. Because all children in the study participated in an initial therapy session, all children realized access. Yet there is greater nuance to realized access than whether or not a child participated in an initial therapy session. More specifically, given the emphasis placed on age of treatment onset, we focus on the time-lag (i.e., number of days) between diagnosis and realized access (i.e., initial therapy session). If results suggest a relationship between child race–ethnicity and time-lag and/or neighborhood characteristics and timelag, then there is evidence of inequitable access. Based on existing literature, the research questions guiding this study include: What is the relationship between time-lag and (a) child race–ethnicity and (b) neighborhood poverty, aluence, racial composition, and urbanicity? Does the relationship between time-lag and child race–ethnicity difer by neighborhood racial composition, poverty, aluence, and/or urbanicity? Figure 1 depicts the relationships examined. Methods In 2007, South Carolina became one of the irst states in the country to ofer a 1915(c) HCBS Medicaid waiver speciically for the statewide provision of EIBI. For 10 years, the South Carolina Pervasive Developmental Disorder (SC PDD) Program provided up to 3 years of EIBI to children between the ages of three and ten who received a diagnosis of ASD made by a professional psychologist by age eight. Based on need, each child received a maximum of $50,000 per year and up to 40 h of direct line therapy per week, of 13 Fig. 1 Empirical model of study aims which at least 50% must have taken place inside the child’s home (South Carolina Department of Disabilities and Special Needs 2013). We partnered with the South Carolina Department of Disabilities and Special Needs (SC DDSN) to create a comprehensive dataset (N = 2338) of all children with ASD who enrolled in the SC PDD Program between the date that the irst child enrolled (February 6, 2007) and the end of the irst quarter of calendar year 2015 (March 31, 2015). To conduct the current study, we irst combined paper case records, electronic spreadsheets, and electronic records from SC DDSN’s Consumer Data Support System (CDSS) to build a base dataset. We then sent the dataset to the South Carolina Revenue and Fiscal Afairs Oice (RFA) to obtain Medicaid claims data and Census data. The RFA returned the base dataset with all identiiers removed along with a dataset that included census-tract data and a dataset that included Medicaid claims data. After data integration, we included children in the study sample if they were diagnosed after SC DDSN established a waitlist to meet the growing demand for the program (August, 2007) and had dates of diagnosis, placement on waitlist, enrollment, assessment, and initial therapy session (n = 473). For families that had two siblings with ASD in the program, we randomly selected one of the children. This reduced the dataset by 50 children. Details on missing data are below. We received institutional review board approval from the University of South Carolina. Measures Time‑Lag A child proceeds through several distinct stages to begin participating in the SC PDD Program. After receiving a J Autism Dev Disord diagnosis, a parent or service coordinator requests that the child’s name be placed on the program waitlist. When a slot in the program is available and the child is at the top of the waitlist, SC DDSN oicially enrolls the child. Parents select a provider from a list of providers that serve their residential area, and then schedule an intake assessment. After an assessment, the parent and provider schedule the irst therapy session, and the child begins therapy on a mutually agreeable date. The intake assessment and the irst therapy session can take place either inside the home or in a treatment center. In the current study, we calculate time‑lag as the number of days between these distinct stages. As illustrated in Fig. 2, we use four measures of time-lag. These include Date of Diagnosis to Date on Waitlist (Time 1), Date of Enrollment to Date of Assessment (Time 2), Date of Assessment to Date of Initial Therapy Session (Time 3), and Date of Diagnosis to Date of Initial Therapy Session (Time 4). RFA provided children’s date of diagnosis, and SC DDSN provided children’s waitlist, enrollment, assessment, and initial therapy dates. The time-lag between placement on waitlist and enrollment is not included because it is administrative in nature and due to funding allocated by the state legislature. Primary Independent Variables We combined data from SC DDSN and RFA to create the primary variable child race–ethnicity (non-Hispanic white, non-Hispanic black, Hispanic, non-Hispanic other, and unknown). Both sources provided the category unknown, and we derived non‑Hispanic other from a range of categories in the original data otherwise too small to analyze (i.e., Asian, Hawaiian/Paciic Islander, and American Indian). We obtained neighborhood variables through RFA, whereby personnel assigned a census tract ID to children based on the residential address recorded in SC DDSN’s CDSS. Racial composition is the percent of white residents in the census tract (grand-mean centered). Neighborhood poverty is a composite variable calculated as a z-score computed from the percent of the following variables commonly used to measure neighborhood poverty (Leventhal and BrooksGunn 2003), or (a) single parent female headed households, (b) percent of people below the federal poverty level, (c) residents who receive cash assistance, (d) residents enrolled in SNAP, (e) residents who receive SSI, and (f) people who are unemployed. Also a composite measure, neighborhood aluence, which is commonly used to measure neighborhood quality, is a z-score computed from the following variables often used to measure neighborhood aluence (Leventhal and Brooks-Gunn 2003) the median household income, percent of residents with professional/managerial employment, and percent of residents with a Bachelor’s degree or higher. We calculated the poverty and aluence composite variables by taking the average of the z-scores for each of the indicators listed above for each composite variable. To measure urbanicity, we used Rural–Urban Commuting Areas (RUCA) to create dummy variables for urban, suburban, and rural census tracts. RUCA codes are at the level of census tracts and are described by the United States Department of Agriculture as codes that use “measures of population density, urbanization, and daily commuting” (United States Department of Agriculture 2016). The most recent RUCA codes are used in this study and are based on the 2010 decennial census and the 2006–2010 American Community Survey. Covariates Covariates collected from SC DDSN include adaptive behavior (Adaptive Behavior Composite [ABC]) standard score on the Vineland-II (grand mean centered), which is a standardized measure of adaptive behavior based on communication, daily living skills, and socialization (Sparrow et al. 2005), sex (1 = female, 0 = male), single parent household (1 = yes, 0 = no), children in household (1 = 3 or more children, 0 = 2 or fewer children), and sibling with ASD (1 = yes, 0 = no). Covariates collected from the RFA included Asper‑ ger’s (1 = yes, 0 = no), Intellectual Disability (1 = yes, 0 = no), and age of diagnosis (measured in months, grand mean centered). Family socioeconomic status is measured by Fig. 2 Process model of timelag in the SC PDD program 13 J Autism Dev Disord the payment category billed by SC PDD program providers, which is most often determined by family income. Sources included payment under the Tax Equity and Fiscal Responsibility Act (TEFRA), which assists families with incomes too high to qualify for Medicaid, Supplemental Security Income (SSI), which assists low-income families, or other payment source (e.g., inpatient psychiatric facility). We used data from both SC DDSN and RFA to create the variable age of enrollment (measured in months, grand mean centered). Missing Data Of the 802 children who met inclusion criteria, 486 children had no missing data. On average, children were missing two items. There was no missing data on child race–ethnicity. Although children with a census tract ID were not missing on any neighborhood variables, 37 children did not have a census tract ID. Variables with strong correlations included Asperger’s, Intellectual Disability, age of diagnosis, age of enrollment, adaptive behavior, single parent, children in household, and sibling with ASD. These correlations indicate that missingness is not missing completely at random, and caution is necessary when interpreting results that involve these variables. However, missing data across all children and all variables was 8.3%. When missing data is less than 10%, listwise deletion does not cause any more bias than imputation (Basilevsky et al. 1985; Roth 1994). Therefore, we chose to listwise delete. The remaining 486 children included children (n = 13) with negative outcome values. Because these values were impossible and most likely attributable to data entry error, we excluded them from the study for a inal sample size of 473. Statistical Analyses We conducted all analyses using SAS® 9.4. Initially, we intended to use two-level organizational models to answer our research questions (i.e., children nested in census tracts). However, due to the number (n = 258) of singletons (i.e., one child in one census tract), the models did not converge. Therefore, we estimated a total of 20 ordinary least squares (OLS) contextual regression models using PROC REG, an appropriate approach when the interest is in the context of the neighborhood rather than diferences between neighborhoods (Diez 2002). Because the outcome did not include zeroes and the means were large numbers, we did not use a Poisson regression. Speciically, we estimated one main efects model and four interaction models for each outcome, for ive total models per outcome. Because we estimated multiple models, we followed guidelines of the Bonferroni correction and used an adjusted alpha (α = 0.0125). To examine model it, we compared changes in R2. This process revealed that for all four outcomes, the main efects 13 models were the best itting models. We examined assumptions associated with OLS. Residuals from the main efects models did not appear to be normally distributed (Shapiro–Wilk p > .05); kurtosis values of the residuals for two models exceeded three. However, given that regression is robust to violations of normality, that skewness values did not violate normality, and that we used an adjusted alpha, we determined that there is little concern for a Type I error (i.e., detecting a statistically signiicant relationship when one does not actually exist). The White test revealed that residuals were homogenous, and the Durbin–Watson test indicated an absence of irst-order autocorrelation and therefore, no violation of independence. Additionally, using studentized residuals and Cook’s D, we identiied no inluential outliers. Although there were four instances of strong zero-order correlations (range of − 0.548 to 0.790), all tolerance values of the best itting models exceeded 0.20, providing no evidence of issues with multicollinearity (Tabachnick and Fidell 2006). Results Table 1 includes descriptive statistics for the 473 children in the sample. There were more males (82.2%) than females (17.8%), most children had a diagnosis of Intellectual Disability (68.9%), and a minority of children had a diagnosis of Asperger’s (17.1%). The average age of diagnosis was approximately 3.5 years, the average age of enrollment was 6 years, and the average ABC score was 65.69. One-third of children identiied as white (34%) and more than two-thirds lived in an urban neighborhood (70%). The average time-lag at Time 1 (Date of Diagnosis to Date on Waitlist) was 333 days, the average time-lag at Time 2 (Date of Enrollment to Date of Assessment) was 68 days, the average time-lag at Time 3 (Date of Assessment to Date of Initial Therapy Session) was 54 days, and the average time-lag at Time 4 (Date of Diagnosis to Date of Initial Therapy Session) was 1041 days. Two of the four main efects models were signiicant. Because there were no signiicant interaction models, we only present results of the main efects models (Table 2). The irst main efects model (Model 1a) accounted for 77% of the variability in the number of days at Time 1 (Date of Diagnosis to Date on Waitlist) [F (20, 452) = 79.94, p < .025, adj R2 = 0.77]. No primary independent variables were associated with the number of days at Time 1. The variables with the strongest relationship with number of days were two covariates, age of diagnosis (β = − 21.57, p < .025,) and age of enrollment (β = 21.09, p < .025), with semi-partial correlation values of 0.46 and 0.61, respectively. The second and third main efects model (Model 2a and Model 3a), or Time 2 (Date of Enrollment to Date of Assessment) [F J Autism Dev Disord Table 1 Univariate descriptive statistics for time-lag (four measures), child race–ethnicity, neighborhood characteristics, and covariates (N = 473) Variable % (M) Dependent variables (in days) Diagnosis to waitlist (332.67) Enrollment to assessment (67.53) Assessment to treatment (53.57) Diagnosis to treatment (1040.89) Child/family variables Race–ethnicity Non-Hispanic white 34.04 Non-Hispanic black 17.55 Hispanic 7.61 Non-Hispanic other 4.02 Unknown 36.79 Payment category TEFRA 40.80 SSI 51.37 Other 7.82 Female 17.76 (41.59) Age diagnoseda (72.96) Age enrolleda ABC score (65.69) Asperger’s 17.12 Intellectual disability 68.92 Single parent 25.37 Sibling with ASD 5.50 > 2 Children in household 33.19 Neighborhood variablesb % White 67.87 Poverty (0) Aluence (0) Urban 69.98 Suburban 25.16 Rural 4.86 a SD Sk strongest relationship with number of days, with magnitudes of 0.54 and 0.74, respectively. Ku Discussion 324.010 59.95 35.60 357.04 1.20 1.29 1.12 0.93 0.43 0.45 1.04 0.08 – – – – – – – – – – – – – – – – 15.99 18.54 13.13 – – – – – – – – – 1.16 0.70 0.45 – – – – – – – – – 1.25 − 0.06 0.20 – – – – – 21.53 0.74 0.94 − 1.04 1.05 0.55 0.57 1.34 − 0.53 – – – – – – Age diagnosed and age enrolled are measured in months b Poverty and aluence are z-scores, so the mean will always be zero. The following values are the minimum and maximum values of poverty and aluence, respectively: − 1.15827, 2.95111 and − 1.63869, 2.94162 (20, 452) = 1.26, p > .0025, adj R2 = 0.01] and Time 3 (Date of Assessment to Date of Initial Therapy Session) [F (20, 452) = 1.56, p > .025, adj R2 = 0.02], were not signiicant. Finally, the fourth main efects model (Model 4a) accounted for 87% of variability [F (20, 452) = 163.86, p < .0251, adj R2 = 0.87] in the number of days at Time 4 (Date of Diagnosis to Date of Initial Therapy Session). Child race–ethnicity and neighborhood variables were not signiicant. As in the irst model, the covariates age of diagnosis (β = − 25.72, p < .025) and age of enrollment (β = 25.84, p < .025) had the The average time-lag between diagnosis and treatment onset was nearly three years, and results suggest that predisposing social characteristics, or child race–ethnicity and neighborhood racial composition, poverty, aluence, and urbanicity, do not predict the time-lag between diagnosis and treatment onset in South Carolina during the time-period examined. Models 1a and 4a explain a substantial amount of variability in Time 1 (Date of Diagnosis to Date on Waitlist) and Time 4 (Date of Diagnosis to Date of Initial Therapy Session), respectively. However, two covariates, or age of diagnosis and age of enrollment, accounted for the majority of variability. It is unlikely that this is attributable to multicollinearity, as residual statistics revealed tolerance values greater than 0.20 (Tabachnick and Fidell 2006). Certainly, these relationships point to the necessity of including these covariates in future research on time-lag. It also suggests that factors other than predisposing social characteristics may play a more important role in timely treatment onset. For example, given that children diagnosed at a later age experienced a shorter time-lag, it is possible that parents of older children are more likely to be aware of available resources and/or to experience a greater sense of urgency to seek treatment compared to parents of children diagnosed earlier. That it takes the average parent nearly one year to request that their child’s name be placed on a waitlist is concerning and underscores the need to explore reasons for this extensive time-lag. The current study conlicts with prior research suggesting that children experience disparities in access to Medicaidfunded EIBI. A detailed review of the diferences between current and prior indings provides a likely explanation for why we did not detect evidence of disparities. Shattuck et al. (2009) compared the census-tract level demographics of children who enrolled in Wisconsin’s EIBI program to the general population of the state, and Magaña et al. (2013) used parent-reported enrollment as an indicator of access in the same program. In both studies, researchers identiied disparities by comparing children who did and did not enroll in the program, whereas in the current sample all children had enrolled in South Carolina’s program. It is possible, therefore, that disparities exist among children who do and do not enroll in Medicaid-funded EIBI, but disparities do not exist in the time-lag between diagnosis and treatment onset between children who do enroll. Clearly, enrollment and time-lag are very diferent indicators of access. In the current study, data for children whose parents never placed them on a waitlist, and children whose parents placed them 13 J Autism Dev Disord Table 2 Estimates from best itting regression models for time-lag (N = 473) Outcome Model Main efects Intercept Child/family variables Race–ethnicity Non-Hispanic black Hispanic Non-Hispanic other Unknown Payment category SSI Other Female Age diagnosed Age enrolled ABC score Asperger’s Intellectual disability Single parent Sibling with ASD > 2 Children in Hsld Neighborhood variables % White Poverty Aluence Rural Suburban Model it R2 Adj R2 Time 1 Time 2 Time 3 Time 4 Model 1a, b (SE) Model 2a, b (SE) Model 3a, b (SE) Model 4a, b (SE) 314.40* (20.50) 49.97* (7.86) 61.97* (4.64) 1058.95* (16.75) 16.60 (24.65) 5.97 (29.31) 53.00* (38.52) − 3.43 (2.50) 15.86 (9.46) 11.08 (11.24) 8.98* (14.78) 1.39 (0.96) 4.06 (5.58) − 5.48 (6.63) − 11.95 (8.72) − 0.53 (0.57) 9.09 (20.14) − 11.60 (23.95) 35.68 (31.48) − 1.43 (2.04) 15.85 (18.20) 16.83 (29.26) − 18.37 (19.09) − 21.57* (0.70) 21.09* (0.60) − 0.71 (0.57) − 0.97 (19.33) 57.06* (16.61) 0.11 (17.84) − 0.12 (32.28) − 6.83 (16.00) 3.67 (6.98) 1.62 (11.22) − 13.11 (7.32) − 0.19 (0.27) 0.25 (0.23) 0.29 (0.22) − 7.65 (7.42) 13.81* (6.37) 1.98 (6.85) − 12.35 (12.38) 6.82 (6.13) − 7.28 (4.19) − 6.33 (6.62) − 4.27 (4.31) − 0.01 (0.15) − 0.05 (0.14) − 0.09 (0.13) − 5.73 (4.38) − 1.22 (3.76) 8.13* (4.04) − 8.94 (7.31) − 0.28 (3.62) − 12.77 (0.39) − 4.97 (23.91) − 17.51 (15.60) − 25.72* (0.57) 25.72* (0.49) − 0.27 (0.47) − 17.53 (15.80) 4.25 (13.57) 33.12 (14.58) − 37.48 (26.37) 9.71 (13.07) − 0.74 (0.48) 3.39 (19.04) 15.92 (13.13) − 11.89 (34.62) − 34.63 (17.03) − 0.03 (0.18) − 12.50 (7.31) − 5.68 (5.04) 0.06 (13.28) 0.41 (6.53) − 0.18 (0.12) 1.16 (4.31) − 0.97 (2.97) − 1.96 (7.84) − 5.66 (3.85) − 0.54 (0.39) − 7.44 (15.56) − 3.69 (10.73) 11.03 (28.29) − 6.91 (13.91) 0.7796 0.7698 0.0527 0.0108 0.0646 0.0232 0.8788 0.8734 Due to extreme values in the outcome most likely due to errors in administrative data entry, we winsorized the dependent variable at the 95th percentile. All time points are measured in the number of days between two dates. Time 1 is date of diagnosis to date on waitlist; Time 2 is date of enrollment and to date of assessment; Time 3 is date of assessment to date of initial therapy session; Time 4 is date of diagnosis to date of initial therapy session *p < .0125 on a waitlist but never enrolled, were unavailable. There is no way to know how many children in South Carolina were eligible to enroll but never did. In other words, it is possible that a replication of the study by Shattuck et al. (2009) in South Carolina’s EIBI program would produce similar indings. Several factors limit the interpretation of study indings. First, the current study is an examination of children who enrolled in South Carolina’s EIBI Medicaid waiver, limiting generalizability to other states. In addition, the sample did not include children whose parents placed them on a waitlist but never enrolled in the program because SC DDSN 13 maintained no records on these children other than their name and social security number. There is no way to determine if the children who never enrolled are qualitatively diferent from children who did, and prior research suggests that this is a possibility (Magaña et al. 2013; Shattuck et al. 2009). Finally, limitations of administrative data are evident, including data entry errors that required the use of winsorizing outcome variables, and although an improvement in data of this type (Shattuck et al. 2009), the high number of children in the unknown racial category (36.8%). Compared to the racial distribution in South Carolina, non-Hispanic black and non-Hispanic white children are underrepresented in the J Autism Dev Disord sample. It is likely that most of the children in the unknown racial category are one or the other, and it is possible that the high percentage of children in this category inluenced indings. Despite limitations, this study lays the groundwork for future research on the time-lag between diagnosis and treatment onset, and it adds to a nascent body of literature on access to Medicaid-funded EIBI. Results demonstrate the importance of conducting research beyond the age of a child’s diagnosis to the time-lag between diagnosis and treatment onset. In this study, an average of 1041 days—or nearly three years—lapsed between a child’s diagnosis and initial therapy session. Notably, nearly one year lapsed before children’s placement on the SC PDD Program waitlist. This time-lag is counterintuitive to the urgency placed on providing EIBI to children as early as possible. Recognizing this, in 2012 SC DDSN collaborated with the South Carolina Act Early Team to spearhead a unique policy. Children at risk for ASD as determined by the Modiied Checklist for Autism in Toddlers and the Screening Tool for Autism in Toddlers and Young Children who were participating in EIBI through the state’s Part C EI services at least 30 days prior to their third birthday bypassed the SC PDD Program waitlist (Rotholz et al. 2017). Presumptive eligibility enables earlier access to services without the need for a formal diagnosis. There is a need to explore the feasibility of similar policies in other states. To build on these results, it is necessary to examine timelag between diagnosis and treatment onset in other states and to investigate factors in addition to predisposing social characteristics that might contribute to this time-lag (e.g., parent perceived need, provider availability), as well as move beyond children’s initial treatment session to investigate treatment utilization over time. It is critical to examine if children receive other therapies while waiting to begin EIBI, and if so, the type and amount of therapy. Although EIBI has the strongest evidence for treating ASD, also recommended are other therapies (e.g., speech-language, occupational, and physical therapy). Exploring whether or not children receive these services as a bridge between diagnosis and EIBI treatment onset is worthwhile, not least of which because an overwhelming majority of states have reported shortages of speech-language pathologists (82%) and occupational therapists (79%; Wise et al. 2010). Finally, there are at least two takeaways regarding data collection for future research and the provision of EIBI. First, there is a need to enhance the collection of demographic data, especially race–ethnicity. This is not a new challenge in health services research. In a study that surveyed hospital patients on this issue, 93% of respondents reported that it is critical to conduct studies that ensure equitable care irrespective of race–ethnicity and 80% of respondents reported that it is important to track race–ethnicity. However, 31% expressed concern that the information could be used to discriminate against patients (Baker et al. 2005). Results of a study by Hasnain-Wynia et al. (2010) indicate that a solution to this perception is to clearly communicate through various mediums the reason for requesting patients’ race–ethnicity. In particular, a majority of study participants were most receptive to the statement, “race/ethnicity information is being collected to ensure that everyone gets highquality care” (2010, p. 378). Adopting this approach when case managers collect information on children and parents could prove useful to future data collection and research. Second, South Carolina is dissolving the SC PDD Program waitlist as the agency transfers the delivery of EIBI from its Medicaid waiver to its Medicaid state plan. As South Carolina and other states begin providing EIBI through Medicaid, it is important to determine how best to capture and record children’s initial contact with the EIBI service system. Acknowledgments We acknowledge the South Carolina Department of Disabilities and Special Needs and the University of South Carolina’s Institute for African American Research for their support of this work. The views and opinions expressed in this article are those of the authors and do not necessarily relect the oicial policy or position of the South Carolina Department of Disabilities and Special Needs. Funding This study was funded in part by the Institute for African American Research at the University of South Carolina. Author Contributions MY conceived of the study, led its design and coordination, and drafted the manuscript; RH advised MY on conceptualization and data collection and provided feedback on subsequent manuscript drafts; BB advised MY on statistical analyses and interpretation and provided feedback on manuscript drafts; All authors read and approved the inal manuscript. Compliance with Ethical Standards Conlict of interest of interest. The authors declared that they have no conlict Ethical Approval This article does not contain any studies with human participants or animals performed by any of the authors. References Acevedo-Garcia, D., Osypuk, T. L., McArdle, N., & Williams, D. R. (2008). Toward a policy-relevant analysis of geographic and racial/ethnic disparities in child health. Health Afairs, 27(2), 321–333. doi:10.1377/hlthaf.27.2.321. Andersen, R. M., Davidson, P. L., & Baumeister. (2013). Improving access to care. In G. F. Kominski (Ed.), Changing the U.S. health care system: Key issues in health services policy and management (pp. 33–64). Hoboken, NJ: Wiley. Auchincloss, A. H., Van Nostrand, J. F., & Ronsaville, D. (2001). Access to health care for older persons in the United States: Personal, structural, and neighborhood characteristics. Journal of Aging and Health, 13(3), 329–354. 13 J Autism Dev Disord Autism Speaks. (2014). The afordable care act and autism. Retrieved from https://www.autismspeaks.org/sites/default/iles/docs/gr/ aca_ehb_11.2014.pdf. Autism Speaks. (2016). Retrieved from http://www.autismspeaks.org/ advocacy/states. Baker, D. W., Cameron, K. A., Feinglass, J., Georgas, P., Foster, S., Pierce, D., … Hasnain-Wynia, R. (2005). Patients’ attitudes toward health care providers collecting information about their race and ethnicity. Journal of General Internal Medicine, 20(10), 895–900. Basilevsky, A., Sabourin, D., Hum, D., & Anderson, A. (1985). Missing data estimators in the general linear model: An evaluation of simulation data as an experimental design. Communications in Statistics, 14, 371–394. Burkett, K., Morris, E., Manning-Courtney, P., Anthony, J., & Shambley-Ebron, D. (2015). African american families on autism diagnosis and treatment: The inluence of culture. Journal of Autism and Developmental Disorders. doi:10.1007/s10803-015-2482-x. Bussing, R., Zima, B., Gary, F., & Garvan, C. (2003). Barriers to detection, help-seeking, and service use for children with ADHD symptoms. Journal of Behavioral Health Services & Research, 30(2), 176–189. Centers for Disease Control and Prevention. (2014). Prevalence of autism spectrum disorder among children aged 8 years-autism and developmental disabilities monitoring network, 11 sites, United States, 2010. Morbidity and Mortality Weekly Report. Surveillance Summaries (Washington, DC: 2002), 63, p. 1. Centers for Medicare and Medicaid Services. (2014, September). Medicaid and CHIP FAQs: Services to address autism. department of health and human services. Retrieved from http://www.medicaid. gov/federal-policy-guidance/downloads/faq-09-24-2014.pdf. Clements, K. M., Barield, W. D., Kotelchuck, M., & Wilber, N. (2008). Maternal socio-economic and race/ethnic characteristics associated with early intervention participation. Maternal and Child Health Journal, 12(6), 708–717. Diez, R. (2002). A glossary for multilevel analysis. Journal of Epide‑ miology and Community Health, 56(8), 588. Florence, J. E., Pack, R., Southerland, J., & Wykof, R. (2012). The depth of rural health disparities in America: The ABCDEs. In R. A. Crosby (Ed.), Rural populations and health (p. 51), New York: Wiley. Granpeesheh, D., Dixon, D. R., Tarbox, J., Kaplan, A. M., & Wilke, A. E. (2009). The efects of age and treatment intensity on behavioral intervention outcomes for children with autism spectrum disorders. Research in Autism Spectrum Disorders, 3(4), 1014–1022. doi:10.1016/j.rasd.2009.06.007. Green, V. A., Pituch, K. A., Itchon, J., Choi, A., O’Reilly, M., & Sigafoos, J. (2006). Internet survey of treatments used by parents of children with autism. Research in Developmental Disabilities, 27(1), 70–84. doi:10.1016/j.ridd.2004.12.002. Hasnain-Wynia, R., Taylor-Clark, K., & Anise, A. (2010). Collecting race, ethnicity, and language data to identify and reduce health disparities: Perceptions of health plan enrollees. Medical Care Research and Review, 68, 367–381. Howlin, P., Magiati, I., & Charman, T. (2009). Systematic review of early intensive behavioral interventions for children with autism. American Journal on Intellectual and Developmental Disabilities, 114(1), 23–41. doi:10.1352/2009.114:23-41. Interagency Autism Coordinating Committee (IACC). (2013). IACC Strategic Plan for Autism Spectrum Disorder (ASD) Research—2013 Update. April 2014. U.S. Department of Health and Human Services Interagency Autism Coordinating Committee. Retrieved from http://iacc.hhs.gov/strategic-plan/2013/index. shtml. Jimenez, M. E., Barg, F. K., Guevara, J. P., Gerdes, M., & Fiks, A. G. (2013). The impact of parental health literacy on the 13 early intervention referral process. Journal of Health Care for the Poor and Underserved, 24(3), 1053–1062. doi:10.1353/ hpu.2013.0141. Johnson, K. M. (2012). Rural demographic change in the new century: Slower growth,increased diversity. Retrieved from http://scholars.unh.edu/carsey/159/?utm_source=scholars. unh.edu%2Fcarsey%2F159&utm_medium=PDF&utm_ campaign=PDFCoverPages. Kim, C., Disare, K., Pfeifer, M., Kerker, B., & McVeigh, K. (2009). Efects of individual and neighborhood characteristics on timeliness of provider designation for early intervention services in New York City. Journal of Developmental & Behavioral Pediatrics, 30(1), 38–49. Kirby, J. B., & Kaneda, T. (2005). Neighborhood socioeconomic disadvantage and access to health care. Journal of Health and Social Behavior, 46(1), 15–31. Klintwall, L., & Eikeseth, S. (2014). Early and Intensive Behavioral Intervention (EIBI) in autism. In Comprehensive Guide to Autism (pp. 117–137). New York: Springer. Knapp, C., Madden, V., Wang, H., Sloyer, P., & Shenkman, E. (2011). Internet use and eHealth literacy of low-income parents whose children have special health care needs. Journal of Medical Inter‑ net Research, 13(3), e75. L & M Policy Research, LLC. (2014). Autism spectrum disorders: State of the states of services and supports for people with ASD (No. HHSM-500-2006-00009I/HHSM-500-T0002). Washington, DC: Centers for Medicare and Medicaid Services. Retrieved from http://www.medicaid.gov/medicaid-chip-program-information/bytopics/long-term-services-and-supports/downloads/asd-state-ofthe-states-report.pdf. Leventhal, T., & Brooks-Gunn, J. (2003). Children and youth in neighborhood contexts. Current Directions in Psychological Service, 12(1), 27–31. Liptak, G. S., Benzoni, L. B., Mruzek, D. W., Nolan, K. W., Thingvoll, M. A., Wade, C. M., & Fryer, G. E. (2008). Disparities in diagnosis and access to health services for children with autism: Data from the National Survey of Children’s Health. Journal of Developmental & Behavioral Pediatrics, 29(3), 152–160. Lovaas, O. I. (1987). Behavioral treatment and normal educational and intellectual functioning in young autistic children. Journal of Consulting and Clinical Psychology, 55(1), 3. Magaña, S., Lopez, K., Aguinaga, A., & Morton, H. (2013). Access to diagnosis and treatment services among Latino children with autism spectrum disorders. Intellectual and Developmental Dis‑ abilities, 51(3), 141–153. doi:10.1352/1934-9556-51.3.141. Magaña, S., Parish, S. L., Rose, R. A., Timberlake, M., & Swaine, J. G. (2012). Racial and ethnic disparities in quality of health care among children with autism and other developmental disabilities. Intellectual and Developmental Disabilities, 50(4), 287–299. doi:10.1352/1934-9556-50.4.287. Makrygianni, M. K., & Reed, P. (2010). A meta-analytic review of the efectiveness of behavioural early intervention programs for children with Autistic Spectrum Disorders. Research in Autism Spectrum Disorders, 4(4), 577–593. 10.1016/j.rasd.2010.01.014. McManus, B. M., Robert, S., Albanese, A., Sadek-Badawi, M., & Palta, M. (2013). Predictors of receiving therapy among very low birth weight 2-year olds eligible for Part C early intervention in Wisconsin. BMC Pediatrics, 13(1), 106. Moodie-Dyer, A., Joyce, H. D., Anderson-Butcher, D., & Hofman, J. (2014). Parent–caregiver experiences with the autism spectrum disorder service delivery system. Journal of Family Social Work, 17(4), 344–362. doi:10.1080/10522158.2014.903581. Murphy, M. A., & Ruble, L. A. (2012). A comparative study of rurality and urbanicity on access to and satisfaction with services for children with autism spectrum disorders. Rural Special Education Quarterly, 31(3), 3–11. J Autism Dev Disord Murry, V. M., Helinger, C. l. A., Suiter, S. V., & Brody, G. H. (2011). Examining perceptions about mental health care and help-seeking among rural african american families of adolescents. Jour‑ nal of Youth and Adolescence, 40(9), 1118–1131. doi:10.1007/ s10964-010-9627-1. National Autism Center. (2015). Findings and conclusions: National standards project, phase 2. Randolph, MA: Author. Parish, S., Magaña, S., Rose, R., Timberlake, M., & Swaine, J. G. (2012). Health care of Latino children with autism and other developmental disabilities: Quality of provider interaction mediates utilization. American Journal on Intellectual and Developmental Disabilities, 117(4), 304–315. doi:10.1352/1944-7558-117.4.304. Peacock, G., & Lin, S. C. (2012). Enhancing early identiication and coordination of intervention services for young children with autism spectrum disorders: Report from the Act Early Regional Summit Project. Disability and Health Journal, 5(1), 55–59. doi:10.1016/j.dhjo.2011.10.001. Perry, A., Cummings, A., Geier, J. D., Freeman, N. L., Hughes, S., Managhan, T., … Williams, J. (2011). Predictors of outcome for children receiving intensive behavioral intervention in a large, community-based program. Research in Autism Spectrum Disor‑ ders, 5(1), 592–603. doi:10.1016/j.rasd.2010.07.003. Peters-Schefer, N., Didden, R., Korzilius, H., & Sturmey, P. (2011). A meta-analytic study on the efectiveness of comprehensive ABA-based early intervention programs for children with autism spectrum disorders. Research in Autism Spectrum Disorders, 5(1), 60–69. doi:10.1016/j.rasd.2010.03.011. Reichow, B., Barton, E., Boyd, B., & Hume, K. (2014). Early intensive behavioral intervention (EIBI) for young children with autism spectrum disorders (ASD): A systematic review. Campbell Sys‑ tematic Reviews. Retrieved from http://www.campbellcollaboration.org/lib/project/338/. Roth, P. L. (1994). Missing data: A conceptual review for applied psychologists. Personnel Psychology, 47, 13–26. Rotholz, D. A., Kinsman, A. K., Lacy, K. K., & Charles, J. (2017). Improving early identiication and intervention for children at risk for autism spectrum disorder. Pediatrics, 139(2), e20161061. Selden, C., Zorn, M., Ratzan, S., & Parker, R. (2000). National Library of Medicine Current Bibliographies in Medicine: Health Literacy. Bethesda, MD: National Institutes of Health, U.S. Department of Health and Humand Services. Shattuck, P., Grosse, S., Parish, S., & Bier, D. (2009). Utilization of a Medicaid-funded intervention for children with autism. Psy‑ chiatric Services, 60(4), 549–552. doi:10.1176/appi.ps.60.4.549. Siller, M., Reyes, N., Hotez, E., Hutman, T., & Sigman, M. (2014). Longitudinal change in the use of services in autism spectrum disorder: Understanding the role of child characteristics, family demographics, and parent cognitions. Autism: The Inter‑ national Journal of Research and Practice, 18(4), 433–446. doi:10.1177/1362361313476766. South Carolina Department of Disabilities and Special Needs. (2007). Enacting legislation: South Carolina Department of Disabilities and Special Needs Pervasive Developmental Disorder Program. Retrieved from http://www.ddsn.sc.gov/Documents/consumersandfamily/pdd/PDDIntro.pdf. South Carolina Department of Disabilities and Special Needs. (2013). Application for a § 1915(c) Home and Community-Based Services Waiver. Retrieved from http://ddsn.sc.gov/providers/ medicaidwaiverservices/pdd/Documents/PDD%20Waiver%20 Doc%201%20Jan%2013.pdf. Sparrow, S. S., Balla, D. A., & Cicchetti, D. V. (2005). Vineland‑II Adaptive Behavior Scales: Survey forms manual. Circle Pines, MN: American Guidance Service. Stahmer, A. C., Collings, N. M., & Palinkas, L. A. (2005). Early intervention practices for children with autism: Descriptions from community providers. Focus on Autism and Other Developmental Disabilities, 20(2), 66–79. Tabachnick, B. G., & Fidell, L. S. (2006). Using multivariate statistics (5th edn.). Boston, MA: Pearson. Thomas, K. C., Morrissey, J. P., & McLaurin, C. (2006). Use of autismrelated services by families and children. Journal of Autism and Developmental Disorders, 37(5), 818–829. Tregnago, M. K., & Cheak-Zamora, N. C. (2012). Systematic review of disparities in health care for individuals with autism spectrum disorders in the United States. Research in Autism Spectrum Dis‑ orders, 6(3), 1023–1031. doi:10.1016/j.rasd.2012.01.005. Turner, E. A. (2009). Parental attitudes toward child mental health services: The inluence of ethnicity and child characteristics on help‑seeking intentions. Texas: A&M University. United States Department of Agriculture. (2016). Rural-Urban Commuting Area Codes. Retrieved from https://www.ers.usda.gov/ data-products/rural-urban-commuting-area-codes/. Virués-Ortega, J. (2010). Applied behavior analytic intervention for autism in early childhood: Meta-analysis, meta-regression and dose–response meta-analysis of multiple outcomes. Clinical Psy‑ chology Review, 30(4), 387–399. doi:10.1016/j.cpr.2010.01.008. Virués-Ortega, J., García, V. R., & Yu, C. T. (2013). Prediction of treatment outcomes and longitudinal analysis in children with autism undergoing intensive behavioral intervention. International Jour‑ nal of Clinical and Health Psychology, 13(2), 91–100. White, K., Haas, J. S., & Williams, D. R. (2012). Elucidating the role of place in health care disparities: The example of racial/ethnic residential segregation. Health Services Research, 47(3pt2), 1278–1299. doi:10.1111/j.1475-6773.2012.01410.x. Williams, D. R., & Collins, C. (2001). Racial residential segregation: A fundamental cause of racial disparities in health. Public Health Reports, 116(5), 404. Wise, M. D., Little, A. A., Holliman, J. B., Wise, P. H., & Wang, C. J. (2010). Can state early intervention programs meet the increased demand of children suspected of having autism spectrum disorders? Journal of Developmental & Behavioral Pediatrics, 31(6), 469–476. World Health Organization. (2013). Comprehensive and coordinated eforts for the management of autism spectrum disorders (133rd session No. EB133/4). World Health Organization. Yingling, M. E., Hock, R. M., Cohen, A., & McCaslin, E. (2017). Parent perceived challenges to treatment utilization in a publicly funded early intensive behavioral intervention program for children with Autism Spectrum Disorder. International Journal of Developmental Disabilities. doi:10.1080/20473869.2017.13243 52. 13