In 2014, the Defense Health Agency (DHA), an agency within the Department of Defense (DoD), established the Comprehensive Autism Care Demonstration (ACD) to provide applied behavior analysis (ABA) services to TRICARE-eligible beneficiaries with a diagnosis of autism while evaluating the appropriateness of the ABA services tiered delivery model under TRICARE for military-connected families.1
Under section 737 of Public Law 117-81 as amended, the National Academies of Sciences, Engineering, and Medicine was charged with establishing a committee of experts to conduct an independent analysis of the ACD and submit a report to Congress and the public with its findings and recommendations. The appointed committee includes individuals with expertise in a wide range of areas, including behavior analysis research, psychology, human/child development, public policy, autism spectrum disorder research, lived experience of autistic individuals and parents of autistic children, healthcare systems and public health, clinical care, neurodevelopment, and biostatistics/outcome measurement.
___________________
1 The establishment of the ACD can be found in a 2014 Federal Register Notice (Comprehensive Autism Care Demonstration, 2014). TRICARE is the federal healthcare benefit program for military-connected families administered through DHA in DoD, serving approximately 9.6 million beneficiaries including active-duty personnel, reserve component personnel, military retirees, and their families. Evaluation goals of the ACD included assessing provider qualifications, utilization patterns, beneficiary cost share, effects of different delivery models, and the relationship between ABA and other interventions.
In conducting its independent analysis, the committee was tasked with providing an assessment of methods used under the demonstration, particularly in relation to assessing autism domains, measuring effectiveness of ABA, and adhering to guidelines and industry standards of care. It was also tasked with examining utilization of the demonstration by TRICARE beneficiaries. Further, the committee was asked to consider whether the occurrence, diagnosis, and treatment of autism among children of military families varied from that in the general population.
Autism is a lifelong, complex neurodevelopmental condition, presenting as challenges with social communication and interaction and with restricted or repetitive behaviors, and often accompanied by other morbidities. The latest prevalence estimate from the Centers for Disease Control and Prevention (CDC) of autism in the United States among children aged eight is 1 in 31.2 Autistic individuals can be found among all racial, ethnic, and sociodemographic groups, including within military-connected families.
CDC has monitored the prevalence of autism in select areas of the United States since 2000. This work has been useful for monitoring trends in the prevalence of autism among the general population. Estimates from a national survey of parents and a recent study of autism diagnosis among TRICARE beneficiaries during the pandemic suggest that the prevalence of autism among military families may be slightly higher than the general population (about 3.5–4.5% compared to 3.2%); however, since each study takes a different approach, these estimates are not directly comparable (see further discussion in Chapter 2).
Many individuals with autism have difficulties coping with change and prefer stable routines. For military-connected families, change is inherent in the lifestyle due to frequent relocations and deployments of one or more caregivers. While relocations are difficult and stressful for all families for a variety of reasons, they present additional challenges for families with individuals with autism or other disabilities who have increased care needs and amplified challenges in the face of change. These challenges can affect the well-being and stability of the entire family, especially when there are additional barriers in accessing medical, educational, and therapeutic supports. Family well-being is essential to those serving in the military because the ability of service members to focus on the mission at hand is largely dependent on the stability of the family.
___________________
2 Information about prevalence estimates over time and data sources and methods can be found at https://www.cdc.gov/ncbddd/autism/data/index.html
Broader understanding of autism over the past decades has led to more nuanced diagnostic criteria and ongoing efforts toward better-tailored interventions that address the complex, heterogeneous nature of this neurodevelopmental condition. Recognizing the primary characteristics (often referred to as core symptoms) is just the start of the process for supporting autistic individuals. The medical community now appreciates that clinically identifiable symptoms of autism do not, in and of themselves, offer causal explanations for health outcomes. In the context of developmental conditions such as autism, health outcomes can be affected by skill development in multiple areas such as communication (e.g., able to indicate when feeling ill), activities of daily living (able to practice self-care in areas such as safety, nutrition, and adaptive behavior (able to refrain from maladaptive or injurious behavior). For some individuals with autism, maintaining stability in their daily functioning through interventions is viewed as a positive outcome. Improved understanding has encouraged comprehensive, holistic support systems for autistic individuals that focus more on overall well-being rather than treating specific symptoms. While holistic support can include medical care in addition to educational and community support services, this report is focused on medically necessary care.
Current interventions for autism offer a range of approaches including, but not limited to, behavioral, communicative, developmental, pharmacological, and psychological. These interventions seek to respond to challenges that interfere with individual functioning and to develop skills that improve daily living and quality of life. The support needs of people with autism can vary significantly, and the types of services and accommodations required vary as well. Autistic individuals and their families, in consultation with healthcare providers, can determine appropriate supports and interventions that best support their needs and goals, which can vary throughout an individual’s life.
ABA, a behavioral therapeutic approach, is one such intervention. The ACD is a temporary demonstration that provides coverage for ABA to military beneficiaries with an autism diagnosis. ABA is a set of intervention practices grounded in the science of learning. Behavioral analysts select specific individual practices (e.g., reinforcement, prompting) or combinations of practices (e.g., functional communication training, naturalistic intervention) when designing personalized ABA services that address individual behavioral, health, and learning goals. ABA represents only one autism intervention or support type and is not necessarily appropriate for all individuals with autism nor an exclusive intervention for those with medical support needs. It is, however, the intervention covered by the ACD and therefore the focus of this report.
Since 2001, military beneficiaries with autism have been eligible to receive ABA services through various mechanisms offered by DoD. For most of that time and during this study, ABA has been considered non-medical and has not been covered as a TRICARE Basic benefit.3 Today, roughly 16,000 TRICARE-eligible beneficiaries are enrolled in the ACD to receive ABA services.
The demonstration project authority began on July 25, 2014, and is found in 10 U.S.C. § 1092 as implemented by 32 C.F.R. § Section 199.1(o). Over the years, the ACD and related mechanisms for the provision of ABA services have evolved. Notably, in 2021, DHA revised its ACD policy in a number of ways (discussed more fully in Chapter 3).4 ABA providers and military family advocates have raised concerns in response to certain aspects of this policy revision. These aspects include the mandatory assignment of an Autism Services Navigator (ASN), the required administration of parental stress indices, parent involvement requirements, more frequent and strict administration of assessment tools, and the clarification of coverage in school and certain community settings (see Chapters 3 and 4 for more details).
The authority duration of the ACD has been extended twice and is now set to expire December 31, 2028. Such extensions were made to allow time for DHA to conduct additional evaluation on the objectives of the ACD and to incorporate findings from this study and other independent research. To date, DHA has maintained that ABA services do not meet the hierarchy of reliable evidence standards for proven medical care as defined in the Code of Federal Regulations for the TRICARE program under Title 32 (see Box S-1).
DHA reports that more data collection and analyses are needed to determine whether ABA can be considered medically necessary. While such assertion made sense when the ACD was created, there have been significant developments in the ABA field in the past decade that have made ABA a standard of care for many autistic individuals. In its review of the evidence, the committee finds that DHA’s position deviates from these developments in several ways:
___________________
3 The committee uses the term TRICARE Basic benefit(s) as presented by DHA to reflect covered services and supplies that are determined medically necessary as defined by statute.
4 An announcement of changes is available at https://newsroom.tricare.mil/News/TRICARE-News/Article/2554995/dha-improves-tricare-comprehensive-autism-care-demonstration-program
SOURCE: Definitions, 32 C.F.R. § 199.2.
There has been tremendous growth in both the science and practice of ABA in the last 10 years. During this time, ABA services have been available to TRICARE beneficiaries through the ACD.
ABA is often equated with a single practice or singular application of prompting and/or reinforcement, but there are 20 focused interventions based on principles of behavior analysis that have been designated as
evidence-based by the National Clearinghouse on Autism Research and Practice (see Chapter 5). To determine if ABA meets DoD criteria for reliable evidence (Box S-1), the committee commissioned an updated meta-analysis review of ABA comprehensive programs. The committee observed that findings from this review complemented an existing meta-analysis review of naturalistic ABA strategies.
Basing research evidence for therapeutics and interventions on meta-analyses of the empirical literature is a valid scientific method that follows guidance from the evidence-based practice framework, which is now commonly used to inform treatment decision making in clinical practice. DoD’s criteria for reliable evidence are aligned with the evidence-based practice framework in that controlled trials and meta-analyses are often situated at the top of that framework.
The commissioned meta-analysis review of comprehensive ABA delivery (>10 hours per week for a minimum of six months) found that there are replicated positive outcomes for ABA in the areas of IQ and adaptive behavior for autistic children between ages two and nine. While findings are less robust for the outcome domains of communication, language, and activities of daily living, they are also positive. The review identified 37 controlled trials of comprehensive ABA programs, including four randomized controlled trials. These studies included well-defined clinical endpoints of cognitive ability or adaptive behavior skills that were measured using psychometrically validated and standardized measures commonly used in published autism research. The committee conducted its own review of meta-analyses conducted for focused intervention practices, finding positive effects across a wide range of outcomes and ages.
Conclusion: There is a substantial body of literature, supported by multiple meta-analyses, indicating strong evidence of efficacy and effectiveness of applied behavior analysis (ABA) as an appropriate intervention to support autistic individuals with a range of ABA practices suitable for older individuals as well as younger children. While additional studies will be useful, the totality of evidence for ABA is robust and meets the standards of the evidence-based practice framework and Department of Defense’s own criteria of reliable evidence.
In its review, the committee considered questions around amount of ABA per week and caregiver involvement. In general, DHA’s emphasis on defining a dose response reflects a misunderstanding of ABA services. While there are some studies indicating that higher amounts of ABA services lead to better outcomes, there are other studies that show similar positive effects in IQ and communication behavior from as few as 10 hours of ABA per week. A uniform standard for hours of services is not necessary or
even appropriate for all autistic individuals and their families. Current best practice indicates that amount of ABA be determined by the collaborative care team based on the individual needs of the autistic individuals, family contexts, measurable progress, and roles of other interventions.
While family engagement and co-development of goals and intervention strategies are important and have been shown to improve outcomes, the most effective forms of caregiver involvement will vary, moderated by families’ capacity to engage and the nature of personalized ABA services.
Conclusion: The effects related to the amount of applied behavior analysis (ABA) received (hours per week) as well as caregiver engagement in ABA are moderated by personal needs and families’ capacity to engage. Current standards of care direct individualized amounts of ABA and caregiver involvement.
In 2013, DHA cited the following limitations in its arguments for classifying ABA as non-medical: (a) the absence of American Medical Association (AMA) Current Procedural Terminology (CPT) codes for ABA, (b) limited Medicare coverage, (c) limited coverage in state Medicaid programs, (d) lack of credentialing and state licensing programs and skill standards, and (e) arguments for separate regulatory bodies for behavior analysts that would operate outside of conventional medical or psychological oversight programs.
In the years since this critique, the ABA practice field, which saw demand increase in the mid-1990s, has taken steps to improve quality and professionalize the delivery of ABA services, establishing its position within healthcare systems. Notably, it continues to evolve, as evidenced by recently published guidelines and continued discussions among healthcare, insurance, and policy communities. See Chapter 4 for details on the following developments:
In undertaking an independent, objective analysis of the ACD, the committee took a multi-perspective approach examining the scientific literature on ABA, as discussed above, as well as the experiences of military-connected families, experiences of ABA providers, industry guidelines, and best practices for research and program evaluation. Over the course of the study, the committee heard testimony and received comments from military families, ABA providers, and autistic individuals. It also invited researchers to present on autism diagnosis, frameworks for autism interventions, measurement development and outcome assessment, and appropriate and ethical use of clinical care data. It reviewed surveys, historical documents, and practice guidelines from autism, ABA, and military-relevant organizations, and reviewed public documents on the ACD, including several reports to Congress by DoD. Further, the committee undertook its own analyses with de-identified data from claims activity and ACD assessments of TRICARE beneficiaries diagnosed with autism from 2018 to 2023.
For a six-year period from calendar years 2018 to 2023, there were 109,073 children, ages 1–18, with an autism diagnosis and eligible for TRICARE benefits. During this time, 35,034 (32%) of these children had a claim for ABA services. More than half (58%) were under age six and a
majority (89%) of ABA recipients were under age 12. The committee identified subsets of regular or continuous ABA recipients (75%) and intermittent ABA recipients (25%). The committee observed that continuous recipients tended to be younger children, to have more co-occurring conditions like delayed milestones (more pronounced in older age groups), and to have greater utilization of speech and dietician services in addition to ABA. Among ABA recipients, most hours per week of ABA are very low (less than five hours/week; 60% of ABA recipients), except among the youngest children, and 20+ hours/week is rare (4%). There are a number of possible reasons for why autistic children in the sample did not receive ABA (68%) or used ABA intermittently (8% of total sample): these include, but are not limited to, (a) a larger proportion of older children in non-participants group who may or may not have received ABA prior to 2018; (b) observed hurdles in eligibility requirements to participate in the ACD and to continue access to services; (c) challenges with relocations or finding local ABA providers; (d) no or other interventions selected by families due to individual and family choice or needs; and/or (e) ABA received outside of the ACD. The committee did not have access to data necessary to examine these possibilities. See Chapter 6 for more discussion.
Military-connected families (i.e., TRICARE-eligible beneficiaries) shared their experiences with the committee, expressing that access to ABA services through the ACD is more limited and burdensome for them compared to their civilian counterparts. Families frequently expressed concern that access to ABA services for their children is restricted by extended steps in the process of enrolling in the ACD and requirements that must be completed to continue ABA services (e.g., mandatory assessments at specific time intervals). Parents also shared that the required completion of the Parenting Stress Index–Short Form (PSI-SF) or the Stress Index for Parents of Adolescents (SIPA) every six months is particularly frustrating, overly invasive, and, from their perspective, adds no value to treatment planning. Families also expressed concern about the lack of clarity on the role of the ASN and have asked that access to ABA services should not be contingent upon being assigned to an ASN.
Parents and caregivers expressed to the committee that they appreciate opportunities for parental engagement and education but would prefer not to have mandatory parent training sessions attached to continuation of ABA services. Further, families raised concerns regarding limitations imposed by the ACD on ABA services received through a registered technician in the school and community settings, noting that such services are available to non-military families under other health plans.
These personal experiences mirrored concerns of ABA providers as well as those identified in the committee’s review of assessment practices and industry guidelines.
Conclusion: Military-connected families experience increased burdens to maintain eligibility and enrollment in the Comprehensive Autism Care Demonstration to access applied behavior analysis (ABA) services as well as restrictions on settings in which ABA can be provided.
The committee also heard from ABA providers. ABA providers expressed concerns with the 2021 ACD policy changes and the quality of communication around processes between them, regional contractors, and DHA. The revised auditing process and returned treatment plans for minor corrections have been viewed as overly burdensome, they reported. In addition, ABA providers reported that they are concerned with the lack of flexibility in choosing the timing and the most appropriate assessment tools for their clients. Some ABA providers have reported that they have either limited, dropped, or decided not to serve TRICARE clients due to the 2021 changes to the ACD.
Conclusion: Applied behavior analysis (ABA) providers serving military families experience administrative burdens and limited flexibility to target individual needs through the Comprehensive Autism Care Demonstration (ACD). Overly prescriptive policies and practices are likely limiting the quality of ABA delivery and the ability to adequately assess the implementation of the ACD program. These policies, if continued, may affect the network and availability of ABA providers who are willing to serve TRICARE beneficiaries, consequently limiting access to ABA services for families who need it.
Regular assessment is a critical part of ABA services. In general practice, the assessment approach and appropriate tools are selected by ABA providers and will vary by individual needs and goals. ACD participants, however, must complete a battery of assessment tools prior to the start of ABA services and periodically thereafter in order to process authorization requests for ABA services. DHA selected four standardized assessment tools used in the autism and healthcare communities: the Pervasive Developmental Disorder Behavior Inventory (PDDBI); the Vineland Adaptive Behavior Scales, Third edition (Vineland-3); the Social Responsiveness Scale (SRS);
and the PSI-SF or the SIPA. These four tools have been administered regularly since the 2021 policy changes, with intermittent use of the PDDBI, Vineland-3, and SRS prior to this date. DHA has collected vast amounts of data from the scoring of these assessments on TRICARE beneficiaries and their families.
DHA operationalized the use of these assessment tools by directing the managed care support contractors to help ABA providers use the scores to develop treatment plans and evaluate client progress. Notably, administration of these tools is in addition to the individualized assessments ABA providers normally conduct, and these assessments do not necessarily capture progress for all individuals and treatment goals. Further, DoD has used aggregated scores, primarily from parent-reporting on the PDDBI and often from a small sample of ACD participants, to report health-related observations to Congress. These analyses have been criticized for improperly examining the effectiveness of ABA services. While these reports to Congress acknowledge that findings should be interpreted with caution, they make assertions that many participants are experiencing no change or worsening symptoms after ABA services.
Demonstrations, like the ACD, can be a useful opportunity to test new approaches to delivering services; however, an appropriately designed evaluation plan is essential for guiding the collection of data, weighing the risks and benefits to patients, and designing the methodology appropriate to determine the outcomes of new approaches. The ACD lacks an evaluation plan with clearly stated goals, measurable questions and quantifiable hypotheses, and methodology that justifies the use of these specific assessment tools toward addressing identified questions and goals. As such, the committee determined that the risks to delivery of quality care in collecting data from these tools are significantly higher than any benefits.
The committee’s review of the use of these assessment tools raised several concerns about the data collection practices associated with the ACD, including placing additional burden on families and providers, withholding ABA services if the assessments are not completed, and collecting large amounts of clinical data without clarity about its purpose.
Conclusion: The current approach and use of assessment tools under the Comprehensive Autism Care Demonstration (ACD) is not appropriate for the ACD purposes. Neither an evaluation plan nor research study has been appropriately designed and articulated to inform how data from these tools will address questions of interest for the demonstration or Congress. Further, their required use has placed additional burden on families and providers with limited benefit for treatment planning.
There are important areas where the ACD policy does not adhere to generally accepted standards of care and areas where the ACD policy aligns but has stringent requirements. These differences present administrative burdens and may prevent TRICARE beneficiaries from accessing quality services (see Chapter 4). Key areas for improvement are as follows:
The committee was charged with developing recommendations related to the measurement, effectiveness, and increased understanding of the ACD and its effect on beneficiaries under the TRICARE program. Four recommendations are shown here with additional details and a fifth recommendation provided in Chapter 7.
Scientific evidence indicates that ABA is an appropriate intervention to support health outcomes of autistic individuals. It is the committee’s view that ABA fits DoD’s Criteria for Reliable Evidence of Proven Medical Effectiveness.
Recommendation 1: The Defense Health Agency should discontinue the Comprehensive Autism Care Demonstration and authorize coverage of applied behavior analysis (ABA) as a Basic benefit under the TRICARE program. It should also immediately take steps to identify authorized ABA providers (to include coverage of the tiered service model and behavior technicians) and define their authority, move ABA Current Procedural Terminology codes off the No Government Pay List, establish reimbursement rates consistent with other TRICARE benefits, and ensure its policies align with current generally accepted standards of care.
Since 2017, the ACD has engaged in an onerous collection of data on participants without clear purpose or plan for doing so. Such data derive from required assessment tools administered periodically. Notably, the added requirement in 2021 to administer parenting stress indices seems to have increased stress among military-connected families with no evidence of their value to the delivery of ABA or to additional support services for families.
Recommendation 2: The Defense Health Agency (DHA) should immediately halt the requirement to periodically administer a specific set of assessment tools (PDDBI, Vineland-3, and SRS) purported to monitor health-related outcomes of applied behavior analysis (ABA) and the administration of parenting stress indices. There is no well-designed evaluation plan for the demonstration program that justifies the use of the assessments, and this data collection has placed growing burdens on military-connected families, ABA providers, and DHA itself in the delivery of ABA services.
DHA should establish policies and practices that ensure ABA services are safe and delivered in ways that are in the best interest of the beneficiaries and their families. Particular care should be taken to change administrative practices that place undue burden on access to care.
Recommendation 3: In providing coverage for applied behavior analysis (ABA) to TRICARE beneficiaries, the Defense Health Agency (DHA) should take steps to ensure that administrative processes do not impede access to care. In particular, DHA should eliminate required aspects
of the demonstration that limit flexibility to support individual health needs and are burdensome to military-connected families and ABA providers. Specifically, DHA should do the following:
Recommendation 4: In providing coverage for applied behavior analysis (ABA) to TRICARE beneficiaries, the Defense Health Agency (DHA) should update its health benefit and coverage policies for ABA services to align with generally accepted standards of care and industry guidelines. Specifically, DHA should do the following:
The military medical system was one of the first healthcare systems to embrace ABA services and provide such services to military beneficiaries with autism. Initial programs and pilots and lessons learned helped influence efforts to develop guidelines and standards to improve the quality of services and establish ABA within a broad set of healthcare systems. Despite this initial leading role, the ACD has not kept pace with scientific and medical developments related to autism and ABA. Recent policy changes
have raised significant concerns about restrictions and additional demands around ABA services for military-connected families. The overly prescriptive requirements have reduced flexibility in the delivery of ABA services, which is critical for autistic individuals and military-connected families. The committee recognizes that the ultimate effect of ABA services on the physical and mental health outcomes of autistic individuals will depend on the degree to which the services are implemented in effective and meaningful ways. We encourage DHA to cover ABA services as a Basic benefit under TRICARE, increase flexibility by making some currently required practices optional, and develop appropriate monitoring mechanisms to support quality services.