Abstract
This research aims to describe the characteristics of 1915(c) Home- and Community-Based Services waivers for children with autism spectrum disorder across states and over time. While increasingly popular, little is known about these Medicaid waivers. Understanding the characteristics of these programs is important to clinicians and policymakers in designing programs to meet the needs of this vulnerable population and to set the stage for evaluating changes that occur with the implementation of health-care reform. Home- and Community-Based Services waiver applications that included children with autism spectrum disorder as a target population were collected from the Centers for Medicare and Medicaid Services website, state websites, and state administrators. A data extraction tool was used to document waiver inclusions and restrictions, estimated service provision and institutional costs, and the inclusion of four core autism spectrum disorder services: respite, caregiver support and training, personal care, and evidence-based treatments. Investigators identified 50 current or former waivers across 29 states that explicitly included children with autism spectrum disorder in their target populations. Waivers differed substantially across states in the type and breadth of autism spectrum disorder coverage provided. Specifically, waivers varied in the populations they targeted, estimated cost of services, cost control methods employed, and services offered to children with autism spectrum disorder. Home- and Community-Based Services waivers for children with autism spectrum disorder are very complex and are not consistent across states or over time. Further efforts are needed to examine the characteristics of programs that are associated with improved access to care and clinical outcomes to maximize the benefits to individuals with autism spectrum disorder and their families.
Autism spectrum disorder (ASD) is a complex lifelong neuro-developmental disorder associated with repetitive behaviors, restricted interests, and impairments in social communication that affects 1 in 68 children (Centers for Disease Control and Prevention, 2014). The disorder often co-occurs with significant behavioral and communication challenges (American Psychiatric Association (APA), 2013). While the causes of autism are unclear, early diagnosis and evidence-based treatment, such as therapy delivered by trained behavior analysts for up to 40 h per week, have been shown to enhance cognitive and adaptive functioning (Harris and Handleman, 2000; Lovaas, 1987; Reichow, 2012; Remington et al., 2007; Smith et al., 2000; Warren et al., 2011). Although several states have enacted policies to facilitate access to care, such as insurance mandates or mental health parity laws, many families of children with ASD experience significant barriers to obtaining needed services.
Prior to the passage of the Omnibus Budget Reconciliation Act (OBRA) (1981), children and adults with long-term healthcare needs, such as those that commonly occur with ASD, were cared for primarily in institutional or other restrictive and segregated settings. The OBRA helped change this care model by authorizing state administrators, through Section 1915 of the Social Security Act, to operate Home- and Community-Based Services (HCBS) waivers. These 1915(c) waivers, so named because they allow states to waive certain Medicaid requirements, make it possible for recipients at-risk for being placed in institutions to receive services in their own homes or communities. Waivers became even more appealing to states after the Supreme Court Olmstead ruling, mandating that individuals with disabilities be offered services in integrated settings (Olmstead v. L.C. and E.W., 1999). In 2010 alone, 47 states used 1915(c) HCBS waivers to provide long-term care services to 1.4 million adults and children, 567,117 of whom had intellectual or developmental disabilities (IDD) (Ng et al., 2014).
HCBS waivers allow states the flexibility to define populations by age, medical condition, and geographic location; limit the number of people who receive waiver services at any one time; disregard income and resource rules that are used to determine Medicaid eligibility; and add services not already included in their state plans (Kitchener et al., 2003). In return for this flexibility, states are required to forego the customary state plan amendment process and submit a detailed application. Furthermore, states must demonstrate that the costs of operating the waiver would not exceed the costs that would have been incurred if the waiver had not been available.
New options for delivering Medicaid-reimbursed services, such as the state plan HCBS (1915(i)) introduced by the Deficit Reduction Act (2005) and modified by the Patient Protection and Affordable Care Act (PPACA) (2010), may further influence how states provide HCBS waiver–funded services and the number of people they serve. The HCBS state plan option does not require an institutional level of care determination making it accessible to a broader range of people. However, if states choose to deliver HCBS with this option, they will not be permitted to limit the number of people who receive services and will not be able to put anyone on a waiting list (O’Shaughnessy, 2013). At least one study suggests the potential for waivers to be cost-effective (Cidav et al., 2014). These waivers are growing in importance as a mechanism to provide care to individuals with autism, but to date, there has been little description of these waivers and how they vary across states and over time.
To address this gap, this study details the offerings of states’ 1915(c) HCBS waiver(s) in their provision of services to children with ASD. Prior studies have primarily focused on adults; (Hall-Lande et al., 2011) did not differentiate between target populations of ASD, intellectual disability (ID), or developmental disability (DD) (Rizzolo et al., 2013); or included waivers that did not stipulate that children with ASD were included in their target population (L & M Policy Research LLC, 2014). This is the first review of 1915(c) HCBS waivers looking specifically at children with ASD across time; examining waiver availability, costs, and restrictions; and documenting waiver inclusion of four core ASD services: respite, caregiver support and training, personal care, and evidence-based treatment. This review provides a timely tool to help caregivers, clinicians, and policymakers better understand how HCBS waivers have been used historically to improve access to autism services for children. We anticipate that states’ past experiences will inform current practice and provide a road map for other states that are considering offering a child’s ASD waiver in the future. In addition, our study will help facilitate the evaluation of service and program changes resulting from the PPACA and other policy changes.
Methods
Current HCBS waiver applications are maintained in a database on the Centers for Medicare and Medicaid Services (CMS) website. Applications range from 80 to 300 pages depending upon the target population and the number and breadth of services offered. Between April 2013 and January 2015, investigators examined each 1915(c) waiver application and selected for review all waivers that expressly included children with ASD as a target population. All other waivers were excluded from review. Once we identified eligible waivers, our goal was to review each one from inception (as early as 1981) onward, including all renewals and amendments approved over the years, in order to document changes that occurred over time. Investigators encountered three distinct waiver eras. Between 1981 and 1994, waiver applications were hand-typed, did not follow a uniform format, and contained minimal information. From 1995 to 2005, they were standardized paper-based applications containing basic information. Since 2006, CMS has made available a web-based application system that requires states to provide much more detail about their waivers.
When documents could not be found on the CMS website, because links were either broken or nonexistent for the years in question, investigators next searched state websites, contacted state waiver administrators by email and telephone, and queried CMS representatives. When available, CMS-issued waiver approval letters were used to confirm information and to gather additional details. More than 200 HCBS 1915(c) waiver-related documents were reviewed. Original waiver documents from the 1980s and 1990s were rarely available. CMS does not maintain historical documents and only one state (Mississippi) was able to provide a complete waiver from the 1990s. A few state administrators attempted to find original waivers and while unsuccessful, several were able to provide hand-typed sections from waivers. For the vast majority of states, personnel either did not believe that original paper-based documents had been preserved, did not know where they were stored, or were unable to search for documents or to make photocopies of documents that ran into the hundreds of pages.
A data extraction tool was used to collect data from all 1915(c) waivers identified as serving children with ASD. Among other details, waiver applications were searched for restrictions, service options, and estimated service costs. Relevant restrictions included the following: (a) minimum or maximum age for inclusion; (b) limits on the length of time one could be served by the waiver; (c) individual cost limits (i.e. dollar limits that individuals were not allowed to exceed based on level of service designations or estimated institutional costs); (d) geographic exclusions; (e) limits on the amount, frequency, or duration of specific services; (f) the maximum number of participants who could be served in a year (“slots”); and (g) whether a premium, enrollment fee, co-payment, or other cost-sharing measure was charged to participants. States estimated the costs of services and institutionalization in each waiver application based on past experience.
States are permitted to choose the services they offer through their waivers and have historically been given considerable leeway in assigning service names and developing their definitions. All waiver services were reviewed and categorized. From the multitude of options, four “core” services were selected for inclusion in this review: (a) Respite, which allows caregivers to take a break from their caregiver duties; (b) Caregiver Support and Training, which strengthens a caregiver’s ability to support and meet the needs of the person for whom they are caring; (c) Personal Care, which provides one-on-one care interaction with the participant; and (d) Evidence-based Treatments. Respite was chosen because it is frequently requested by families and at least one study shows that providing respite care results in better outcomes for children (Mandell et al., 2012). Caregiver support and one-on-one care were selected because these methods are most likely to be used in the delivery of evidence-based treatments (National Autism Center, 2015). In addition to these four services, investigators assessed whether waivers included participant-directed services, an option that gives waiver recipients and their families enhanced control over the services they receive and the people who provide them.
In addition to the criteria above, all waiver applications were reviewed to determine the estimated number of days per year that participants were anticipated to receive waiver services, the annual per-person cost of services, and the estimated cost of providing the same group of services to a person in an institutional setting for 1 year. Service cost estimates are required by CMS to help fulfill the federal requirement of cost neutrality. States are expected to base renewal cost estimates on figures derived from experience in prior years.
All documents were independently reviewed by two investigators, and disagreements were settled by the Principal Investigator. Waivers were placed in one of four categories: (a) waivers that exclusively targeted children with ASD (10 waivers), (b) waivers that exclusively targeted children and included children with other IDD as well as ASD (13 waivers), (c) waivers that exclusively targeted ASD but included adults along with children (1 waiver), and (d) waivers that included both adults and children and targeted IDD and ASD (26 waivers). This study was approved by the Penn State University Institutional Review Board.
Results
Investigators identified 50 current or former waivers (43 known to be in effect at time of review) across 29 states that explicitly included children with ASD. A total of 13 states had more than 1 waiver. Within states, waivers differed primarily by the number of individuals served, the targeted population (age and/or condition) and by the breadth of services offered (e.g. number of hours allowed per week). None of the waivers included fees for services or cost-sharing features.
The 10 children’s ASD-specific waivers (top section of Table 1) all began between 2001 and 2012. Eight of the 10 initial applications were available for review. No two states used the same age criteria for eligibility. Some chose younger ages (e.g. 0–5 or 1–4 years) and others covered children through their 18th or 21st birthdays. Two states (Missouri (MO) and South Carolina (SC)) stipulated that children had to be at least 3 years of age to be served through the waiver. Half the states limited the time children could spend on the waiver to 3 years (Arkansas (AR), Colorado (CO), Kansas (KS), Montana (MT), SC) and four imposed individual cost limits (CO and Massachusetts (MA): US$25,000, MO: US$22,000, and SC: US$50,000). None of the children’s ASD waivers limited geographic coverage, but all contained limits on the amount, frequency, or duration of specific services. All children’s ASD waivers included personal care services, caregiver support and training, and evidence-based treatments. All but three (AR, CO, SC) offered respite care and half offered direction of services (MA, MO, MT, SC, and Utah (UT)). The estimated cost of services for a child on an ASD-specific waiver ranged from US$8033 in CO to US$46,111 in North Dakota (ND), averaging US$26,662. The estimated cost of institutionalizing a child and providing her or him with the same services offered by the waiver averaged US$157,593 across the 10 states and ranged from US$38,572 in KS to US$360,287 in ND.
Characteristics of state HCBS waivers with a target population of children with ASD a .
HCBS: Home- and Community-Based Services; ASD: autism spectrum disorder; IDD: intellectual or developmental disabilities; ukn: unknown (information not included in waiver); nl: no limit.
Data were obtained from the most recent waiver available.
First year of waiver available for review.
May only utilize waiver services for 3 years.
Waiver scheduled to end in 2014.
Waiver scheduled to end in 2013.
Waiver terminated in 2013.
Waiver terminated in 2012.
The 13 children’s waivers in eight states that included IDD in addition to ASD were much less likely to restrict age than were the children’s ASD waivers. Only 1 children’s IDD/ASD waiver (1 of the 2 in Idaho) stopped covering children prior to their 18th birthday, and none limited either the time children could remain on the waiver or geographic coverage. The cost of services and cost of institutionalization for children’s IDD/ASD waivers were substantially higher than for the children’s ASD exclusive waivers: the average cost of services per child was US$31,959 and ranged from US$8354 (New York (NY)) to US$161,040 (Maine (ME)), while the average cost of institutionalization was US$241,299 and ranged from US$63,505 (Wisconsin (WI)) to US$493,160 (Michigan (MI)). Three of the four NY waivers were scheduled to end in 2013 or 2014.
Only one ASD exclusive waiver included services for adults in addition to children. This Connecticut waiver was approved in 2013 and allows individuals to receive benefits beginning at the age of 3 years. This waiver provides an average of US$21,232 worth of services to each participant and offers direction of service supports in addition to including all of the core ASD services.
Waivers serving children and adults with IDD/ASD comprised the largest group of waivers, with 26 waivers offered across 18 states. As a whole, these IDD/ASD waivers included very few age restrictions and none limited the time people could spend on a waiver. Unlike any of the other waivers, three of these waivers (MO, North Carolina (NC), NC) restricted waiver access to specified counties within the state. Over half of the waivers in this category were first approved during the 1980s and 1990s and very few of the original documents were available for review. One of the three Indiana waivers ended in 2012 and three of the four Florida waivers were discontinued in 2013.
Across all waivers, the difference between the estimated cost of services and cost of institutionalization within states, which could be interpreted as the savings accruing to the state by providing these waiver services instead of institutionalization, varied substantially. The difference between the estimated cost of providing services under the waiver and the estimated cost of providing the same services in an institutional setting averaged US$200,754 per person and ranged from US$23,228 in Indiana to US$368,587 in NY.
Table 2 displays the maximum number of individuals that could be served on a 1915(c) waiver annually from 2003 to 2014. In 2014, the 10 children’s ASD waivers were limited to an average of 340 children per year and ranged from 47 slots in ND to 1100 in Maryland. The average number of children that could be served on a children’s IDD/ASD waiver was 1338 and ranged from 80 in ME to 5450 in WI. Connecticut’s waiver is the only ASD exclusive waiver serving both adults and children and reserved 115 slots for individuals in 2014. Waivers that targeted adults and children with IDD/ASD averaged 14,574 participants on each waiver in 2014 and ranged from 425 in Louisiana (LA) to 115,000 in California.
Maximum number of individuals served each year from 2003 on HCBS waivers that target children with ASD or IDD/ASD.
HCBS: Home- and Community-Based Services; ASD: autism spectrum disorder; IDD: intellectual or developmental disabilities.
No renewals posted; numbers from previous year.
Numbers obtained from renewal drafts.
ukn—unknown (waiver existed but was not available for review).
nl—no age limit.
“ - - - ” waiver scheduled to end in indicated year.
“–” waiver ended in indicated year.
Discussion
Our review of Medicaid waivers providing services for children with ASD revealed several important findings. Waivers differed in substantively important ways, including (a) the populations targeted, (b) the estimated cost of services, (c) the cost control methods employed, and (d) the services covered.
Waivers differ in the populations they target
Waivers that exclusively target children with ASD provide clear information about children’s coverage. Every element of these waivers—from the number of slots available to the services offered—applies specifically and exclusively to children with ASD. In contrast, waivers that target adults and/or other IDD in addition to ASD do not specify the number of children with ASD served or provide a cost breakdown by age or disability category, and present a challenge to caregivers, clinicians, and stakeholders to understanding the nature of services available to ASD children under such waivers.
Waivers differ in their estimated cost of services per child
Substantial differences in states’ estimated cost of services and cost of institutionalization were evident both within and between target population categories. While the methods used to compute these figures are not known, a possible explanation for higher institutional costs may be that some states rely more on out-of-state facilities or have lower availability of trained personnel to provide services. Differences in the estimated cost of services are also affected by the expected length of time children receive services, as evidenced by the children’s ASD waivers in CO and AR. Although they offer very similar services, CO estimates that they will spend US$25,160 less per child than AR. This is likely related to CO’s services being offered for substantially fewer days each year than AR (178 vs 365). State estimated costs are of particular interest because the difference between the estimated cost of institutionalization and the estimated cost of waiver services could be interpreted as the savings accruing to the state by utilizing waivers instead of institutionalization, particularly important at a time of increased attention to state Medicaid expenditures. Relatedly, no states currently include patient or family cost-sharing or fees as a component of their ASD related waivers, but at a time of rapid evolution of state Medicaid plans and waivers, interested stakeholders should monitor whether this pattern continues.
Waivers differed in the services they covered
Waivers within the children’s ASD waiver category were most likely to offer the four core ASD services, although respite was less available than in other waiver categories. While the majority of children’s ASD waivers provided respite, it is disappointing that coverage was not universal in this category, as respite plays an important role in reducing the risk of psychiatric hospitalization and potentially longer term care among children with ASD (Mandell et al., 2012).
Within the children’s ASD category, Maryland appears to offer the most comprehensive waiver. It is administered by the Maryland Department of Education, which enables better coordination of educational and Medicaid services. They offer all four of the core ASD services; include children of all ages; do not limit the length of time children can stay on the waiver; offer services to the largest number of children; have the longest running waiver; and, until 2014, the estimated annual cost of services per child appeared to be more generous than in any other state. However, the estimated annual cost of services under the Maryland waiver decreased by over US$20,000 per child in 2014, corresponding to decreases in the amount of services offered.
Changes in waiver characteristics occur frequently. Within the past year, three states changed their target ages (CO, ND, Delaware (DE)), two added direction of services (ND, LA), and almost every state had a change in their estimated cost of services and estimated cost of institutionalization. Some of these changes were considerable; for example, the estimated annual cost of services in ND increased by over US$30,000 in 2014. It is unclear whether states made these changes as a direct result of observed waiver effects, changing state environments, or other issues we are unaware of.
There are multiple dimensions upon which one could judge the relative generosity of waivers for children with ASD, and we are unaware of any studies that established an approach to compare generosity of waivers across states or over time. We have sought to be thorough and transparent in our description of the waivers and the various dimensions of generosity, to allow readers to be able to draw their own conclusions. We have also sought to be explicit in situations, such as in waiver applications that included adults and/or individuals with IDD, where we felt that the available information was insufficient to clearly identify the most generous and comprehensive waivers for children with ASD from among the states. Further research is needed to better understand to what extent the generosity of waivers is associated with greater access to services or impact of services for families and children with ASDs.
Our findings must be interpreted in the context of the study’s limitations. First, our examination was limited to 1915(c) waivers that explicitly included children with ASD. States may offer other 1915(c) waivers that do not specify that they serve children with ASD (L & M Policy Research LLC, 2014) or other programs, such as Medicaid Section 1115 Demonstration projects or 1915(i) waivers that provide services to children with ASD who would not otherwise qualify for Medicaid services. However, these programs are less common, and historic information is not collected in a systematic format for these other programs at this time, making comparisons very difficult. As more information becomes available, it will be important to include these programs in future research. In addition, only 19 (39%) of the original 1915(c) waivers were available for review, and all but 1 approved after 2000. This paucity of data prevented us from describing early waivers from the 1980s and 1990s and comparing them with more recent waivers. We are also unable to determine the extent to which services offered under waivers that include both adults and children or that cover individuals with other IDD went to children with ASD. For example, up until 2014, DE’s target population included children as young as 4 years. However, in their recent renewal they increased the minimum age to 12 years, acknowledging that only three children under the age of 12 had ever been served on this waiver. In addition, the recent State of the States report indicates that the waiver is primarily designed for adults (L & M Policy Research LLC, 2014). Hence, we are only able to directly compare states with ASD-specific waivers targeted to children. Finally, we are unable to determine how multiple waivers within the same state interact with each other, since such information is not included in the waiver documents.
Despite these limitations, our study highlights the considerable complexity and heterogeneity in Medicaid waivers for children with ASD. As states and advocacy groups continue to develop these programs to facilitate access to care for individuals with ASD, understanding and evaluating current programs is critically important. Further efforts are needed to examine the characteristics of these programs that are associated with improved access to care and clinical outcomes to maximize the benefits of these programs to individuals with ASD and their families.
Footnotes
Funding
This research was supported by National Institute of Mental Health (NIMH) grant R01MH097298.