5
Early Identification and Interventions for Developmental Disabilities Emerging in Childhood
Speakers on the next panel considered a variety of factors affecting a particular set of conditions: developmental disabilities emerging in childhood, with a particular focus on autism. Settings, approaches, funding, families, research, practice, and policy are all important dimensions of interventions for these conditions, observed moderator Craig Newschaffer, and all of these dimensions need to be considered to get a full picture of the potential of those interventions.
ENSURING QUALITY AND ACCESSIBLE CARE
Heather Johnson, acting director and chair of the Family and Women’s Health Nurse Practitioner Programs at the Uniformed Services University of the Health Sciences (USUHS) Daniel K. Inouye Graduate School of Nursing, talked not only about her experiences as a family nurse practitioner but about her experiences as the mother of a son on the autism spectrum. Her son was born a few weeks early, and other than wanting to be swaddled in his swing rather than lying flat in his crib, he was a typically developing baby. But between his 15- and 18-month pediatric visits, he did not seem to be developing language in the same way that other children were doing. The family’s pediatrician thought that he was just developing slowly and that Johnson might be overly concerned because she was a nurse practitioner trained to detect problems. But a month after his 18-month visit, Johnson called the Babies Can’t Wait system in Georgia, where the family was living at the time, and, after an assessment, was told that her son did need services.
Speech services were not available through Babies Can’t Wait, so the family needed to work through their insurance company to secure services and had to take at least 1 hour off from work 3 days a week to drive their son to his appointments. Similar arrangements were made for a special education teacher to come to his daycare program for a half-hour twice a week and for appointments with a developmental pediatrician, though the closest such specialist covered by their insurance was more than a 2-hour drive.
When the family moved to the Washington, D.C., area, Johnson lined up a new set of services ahead of relocating, but because of administrative insurance complications, the family had a significant lapse in care while they waited for a new referral. Meanwhile, their son’s new school system wanted to do a parallel set of evaluations while they were trying to access care. “It was very challenging, [but] we did learn a lot from the process,” Johnson said.
One complication was that many services in the Washington, D.C., area are cash-only practices. “You have to be able to keep your practice afloat, [but] there’s no cloud where money comes from,” she said. When getting a referral to see a speech therapist or occupational therapist in the area, they had to decide whether to get on waiting lists for a therapist who would take their insurance or pay out of pocket and hope to get reimbursed. Parent and family support systems also were often cash practices, or insurance would not cover benefits for families. Driving distances did not correspond to driving times in their new home. “It might be a 20-mile drive, but it’s an hour to get there,” she remarked.
Johnson drew several lessons from her family’s experiences. First, she warned against a cookie-cutter approach to health care. Even though autism is now much more widely recognized and treated, treatment still often consists of 30 minutes a week of speech therapy and 30 minutes a week of occupational therapy despite the distinct needs of the patient. “We’re forgetting that everybody is an individual,” she said. Evidence-based practice does not just mean using practices that are based on good evidence but using evidence to inform clinical practice. Incorporating the individual characteristics of patients and family members is fundamental to evidence-based practice.
The preparation of clinicians, and especially their training to be culturally competent when treating people with disabilities, is another major consideration, she said. For example, the nurse practitioner students at USUHS receive training on typical and atypical development, 504 plans, individualized education programs, individual family service plans, and advocacy for children.
Another lesson is that parents have a major role as advocates and not just bystanders, Johnson said. In particular, primary care physicians can help parents become advocates for their children.
Schools and education systems provide health care services to help children access a free and appropriate education, but a paradigm shift is needed so that the education and health systems can work seamlessly together, Johnson said. Parents could be partners in this collaboration to reduce fragmented and redundant care, for example, decreasing the need for parallel evaluations and focusing on the same problem. “That’s something we really need to address,” she said.
BRIDGING THE GAP BETWEEN RESEARCH AND PRACTICE
The Los Angeles Unified School District is the second largest school district in the country. Three-quarters of its students are Hispanic, many of whom are English learners, and 80 percent are on free and/or reduced-price lunch. The school district has 14,000 children with autism spectrum disorder (ASD).
Connie Kasari, professor of human development and psychiatry at the University of California, Los Angeles (UCLA), does research in the school district on novel interventions for students with ASD. She began her presentation by emphasizing that public schools should be places without disparities in services. Yet white students, who make up only 9 percent of the student body, represent 30 percent of the ASD population in Los Angeles public schools. “Disparities are real,” she said. “That’s very concerning to those of us who are trying to do research in schools and to bring evidence-based practices to those settings.”
She moved on to three issues involving research. First, a huge gap exists between research and practice, she observed. Researchers might perfect an intervention in a laboratory and roll it out to a community. But most children with a disability have never been in a research study, partly because many children with comorbidities are excluded from research studies, and most interventions developed for children with disabilities have never been tested in any controlled way. “The evidence base doesn’t really reflect most children in the community,” she said.
Deployment-focused models may help, where an intervention is tested with participants within the context, such as homes or schools, for which the intervention is intended and by people, such as parents or teachers, who are intended to deliver the intervention. In addition, interventions need to be tested against usual care, she said, “because if you don’t find differences, we don’t need to change usual care.” However, as was pointed out during the discussion period, usual care as the comparison to evidence-based practice can be uninformative. A complication for research is that access to services is variable across the United States. If research is being conducted over multiple sites, some sites may have fairly good access to services, while others have poor access. In addition, different places may be more or less
willing to adopt practices backed up by evidence. Usual care must be carefully considered as a comparison to evidence-based practice.
The second issue Kasari emphasized is that a single treatment is not effective for all. Many children need a combination or sequence of interventions. Methodologies need to be personalized or tailored for individuals and for sites through such approaches as sequential multiple-assignment randomized trials to see what works for which children, she said. Children who are not responding then can be provided with an increased dose, switched onto another intervention, or given a combination of interventions. “Our goal is to have all children make progress. . . . These kind of models could help us think this through,” she said.
The third issue she mentioned is that sustainability is often poor. Partly, training models tend to be complicated, and implementation fidelity is often low. One way to address this issue is to think about the “active ingredients” of effective treatments, she suggested. These active ingredients may be reflected in modules or targeted treatment strategies that can be applied more flexibly than multicomponent, comprehensive treatments. Knowing what elements of an intervention make it effective can inform implementation strategies and likely lead to greater sustainability of an intervention.
Another way to increase sustainability is to foster academic-community partnerships, she said. Equitable relationships between community partners and researchers can increase the feasibility and acceptability of sustainable models. As Kasari said, “It’s not easy, it takes a lot of time, yes, but it’s worth it.”
TREATING PROBLEM BEHAVIORS
Although estimates of prevalence vary (Emerson et al., 2011), problem behaviors are more common in individuals with ASD than same-age peers, observed Nathan Call, director of the Severe Behavior Program at the Marcus Autism Center in Atlanta. Common problem behaviors include aggression, tantrums, self-injury, disruptive behavior, property destruction, pica (the persistent eating of nonnutritive items), elopement (wandering or running from supervision), and encopresis (incontinent bowel movements). Call showed several examples at the workshop, including a heartbreaking example of a boy who engaged in self-injury in the form of head hitting, averaging about 8,000 blows to his head per 6-hour therapy day. Though his hands were padded, he had blinded himself in both eyes through repeated blunt force trauma that disconnected his retinas. “These are behaviors that can have a significant effect,” said Call. They “are the number one causes for seclusion, restraint, overmedication, and isolation, in many cases both academically and residentially.”
Problem behaviors also negatively affect almost every facet of a care-
giver’s daily life. These effects include poorer family functioning, lower marital satisfaction, higher rates of problem behavior in siblings, parental unemployment, and a significant increase in total lifetime costs of care (Caicedo, 2014; Davis and Gavidia-Payne, 2009; Hall et al., 2012; Hartley et al., 2012; Hastings, 2007; Sikora et al., 2013). Families, he related, “report to us that they don’t leave their home, and that caregivers cannot work because they’re taking care of their child.”
The current state of treatment for problem behaviors is inadequate, Call said. Treatment is often limited to crisis stabilization, with a short-term and usually pharmacologic focus. Many families do not have access to services that can address the problems in a meaningful way. Families may wait for a crisis to occur and then call first responders who are not well equipped to deal with these behaviors. “We have patients who get tazed and have had all kinds of run-ins with juvenile justice,” he said. “They end up being dropped off in an emergency room, or if they are lucky, they end up in an inpatient or residential psychiatric center.”
The Severe Behavior Program at the Marcus Autism Center focuses on producing long-term effects, which requires several treatment components, said Call. The first is function-based treatments. Behavioral treatments for problem behavior that are based on the type of problem behavior, such as aggression or self-injury, are largely hit-and-miss. “They work for some kids but not all, or they work some of the time but not all of the time,” he explained. Treatment components based on the purpose the behavior serves for the individual—that is, its “function”—are consistently more effective (Campbell, 2003; Heyvaert et al., 2014). He gave as an example two children who engage in self-injurious behavior, like head hitting. One may be hitting himself because he has learned that it will make people back away, whereas another might engage in the behavior because it is stimulatory and part of the repetitive nature of ASD. “Why would we expect the same treatment to work with behaviors that have completely different etiologies?” asked Call. In fact, Call and his colleagues are attempting to move pharmacologic treatment toward the same approach.
A second factor behind successful long-term outcomes is a level of service that fits where the individual falls on the continuum of severity. Tertiary care is for the most severe behaviors, which tend to be less prevalent. It requires a high investment of resources, and relatively few providers are able to deliver this care. However, many individuals are at risk for these behaviors, and if the need for intensive services can be avoided, it should be, Call observed. A group of his colleagues have developed a parent training program that is universal for children who are at risk for more significant problem behaviors, such as children who are engaging in tantrums.
Treatment of problem behavior exhibited by children with autism and other developmental disorders has generally taken one of two perspectives
that have tended to be distinct in the past. One is applied behavior analysis, which tends to focus on the most severe cases, uses highly individualized treatment components, and emphasizes the basic mechanisms behind specific treatments and internal validity. The other is rooted in clinical psychology and focuses on less severe cases, group designs, manualized interventions, exportability, and external validity. There is a significant opportunity to build upon the strengths of both approaches by bridging the divide between these perspectives and answering questions on each side using whatever methods are most appropriate.
Between universal and tertiary care are individuals who have moderate intensity conditions with prevalence between the highest severity and lowest severity cases. A treatment model for these cases uses master’s-level clinicians who can provide in-home community-based services to address problem behaviors. These individualized treatments focus on a handful of specific caregiver priorities and have been shown to be successful, according to Call.
All of these treatments are mediated by caregivers, Call noted, which means that caregivers have to be able to implement the treatment. This requires both buy-in and training. Although parents tend to say they will do anything to make a problem behavior get better, they can have second thoughts when they learn what they have to do. Call and his colleagues have done research into factors that affect whether caregivers follow through with treatments, such as how long it takes for a treatment to have the desired outcome and how hard a treatment is to implement. For example, the longer a treatment takes to have an effect, the less value it has for caregivers. The goal, he said, is to customize treatments not just for the child but for the family to take into account how sensitive they are to delays in outcomes or the effort the treatment requires.
Call also observed that dealing with a problem behavior is only part of the puzzle for these families. Many have a host of other issues with which they must deal, such as poverty, underemployment, lack of social supports, or barriers to treatment success. The Severe Behavior Program at the Marcus Autism Center does preadmission assessment of the home environment and identification of barriers to treatment success, with care coordination to provide wrap-around services. A problem, he acknowledged, is that care coordination is generally unreimbursed. “It isn’t something that we can continue to do forever. This is something that we cover through philanthropy, but not a lot of programs are going to be able to do the same,” Call said.
Finally, he observed that a major barrier to effective treatment for problem behaviors is the limited number of providers. Not many people have the level of expertise to address severe problem behaviors. This is another reason to emphasize prevention, said Call: “to do more up front, when kids
are younger, smaller, when these behaviors can be managed much more easily, and to invest more resources into those types of interventions to prevent the long-term need.” Intensive services will always be needed, but caregivers and health care providers both would prefer to treat children before they have lengthy histories of problem behaviors.
The prevention-based model has been conducted in clinics, but the Marcus Autism Center also has been developing a telehealth-based model in which families drive to a satellite clinic and receive the same training via video link. In the secondary care model, families download a free smart-phone app so that care can be maintained from the clinic to the home. Families receive coaching in their homes from a therapist located in the center. However, this, too, is not a reimbursed service, and the regulations surrounding such care are not clear. For example, it is not clear how this service would be provided over state lines. “This is just one more area that needs to be clarified,” Call said.
IMPROVING THE QUALITY OF POLICIES
A variety of policies, from the Community Mental Health Act and Medicaid in the 1960s to the Mental Health Parity Act and the Affordable Care Act in the past decade, have affected people with ASD, noted David Mandell, associate professor of psychiatry and pediatrics at the University of Pennsylvania’s Perelman School of Medicine. “What these things do is open doors,” said Mandell. “In a lot of ways they have been very successful.”
These policies and interventions have greatly increased access to care. The number of children with a diagnosis of autism served through the special education system has risen dramatically, to 8.2 percent of all special education students. The number of children with autism who are served through the Medicaid system has similarly expanded. Laws or guidelines that require pediatricians to screen for autism and other developmental delays have led to more screening (“although not as much as we would like,” said Mandell), and more children have been identified. While autism insurance mandates have not had the effect that was hoped in increasing access within the private insurance market, they have increased the number of children in private insurance who are identified with autism.
At the same time, a wealth of new interventions has improved communication, adaptive behavior, and social behavior (Weitlauf et al., 2014), Mandell said. At least in a laboratory setting, these interventions have demonstrated their potential to create profound changes for children with autism.
However, this increase in access and the creation and testing of these interventions have not resulted in a commensurate increase in quality or
important outcomes, Mandell continued. For example, according to data from a Pennsylvania survey in which parents were asked if their child had ever been psychiatrically hospitalized, children in later birth cohorts had a higher risk of hospitalization than children in earlier birth cohorts (Mandell, 2008). “We are psychiatrically hospitalizing more young children now than we were a decade ago,” said Mandell.
Data from the U.S. Department of Education reveal that more children are in more segregated environments in public schools than in the 1990s, after a trend toward greater inclusion in the prior years. Furthermore, the effects of community intervention on children with autism are, according to Mandell, “woefully small.” The effect size of usual care on improving cognition is just 0.26, and the effect sizes on adaptive behavior, improvements in social ability, and improvements in communication are also small. “Whatever we’re providing in community care is not meeting the benchmark that we’ve established in efficacy trials. There is something that is lost in translation about what we know works in efficacy trials versus what we are doing in communities,” he said.
In a research project in the Philadelphia school district, Mandell and his colleagues compared two interventions to each other. When teachers implemented the interventions the way they were designed, one intervention showed much better outcomes than the other. “But the most startling statistic to me was the one about fidelity,” he continued. Teachers were implementing about half of the components of these interventions, despite extensive training and support.
This experience led to the use of findings from social psychology to think about practitioner behavior change. In particular, Mandell’s group borrowed from the Theory of Planned Behavior, which states that intentions—the degree to which an individual plans to execute a particular behavior—are the strongest predictors of that behavior. Intentions are in turn affected by attitudes, norms, and self-efficacy. The most common strategy to get practitioners to change their behavior (namely, increase their use of evidence-based practice) is to train them. “If the use of the desired practice does not increase, we train them some more,” said Mandell. “We think of it as a skill deficit that we are going to improve, but the reality is that it often isn’t.” Policies usually offer little support or incentive in this area, given that they dwell on access without mentioning the need for practitioners to use evidence-based practices.
The comprehensive interventions that make up the bulk of tested interventions generally involve complex repertoires of behaviors. Especially for school-age children, the use of visual schedules, positive reinforcement, one-to-one instruction, and data collection form the backbone of interventions that can help children with ASD, Mandell said. His team has found that to affect behavior change among practitioners, these intervention
components must be introduced and supported individually. He and his colleagues initially focus on positive reinforcement and schedules, getting buy-in from participants before moving on to one-to-one instruction and data collection.
Mandell said his take-home message was that it is practitioners’ behaviors that result in a change in child outcome, so policies need to be focused on what practitioners should do. He said questions that should guide these policies are: Are the desired behavior changes clear? Who are the people whose behavior needs to change? Are the resources available to change? Is the behavior intrinsically reinforcing and, if not, what incentives should be put in place to change behavior? What is the short-term and long-term implementation policy? How can a policy and the behavior changes it seeks be made sustainable?
APPROACHES AND COMPARISONS
A major topic of conversation during the discussion session was how to identify practices that are effective with ASD and how to implement those practices more widely. Comparisons of interventions to each other or to usual care can be difficult to interpret for several reasons. One issue is that a particular intervention can be effective with some children but not others. Also, as Mandell said, in randomized trials in which a university-based intervention is compared to treatment as usual, it can be hard to distinguish between the effects of the treatment per se and the effects of having skilled clinicians deliver the intervention.
In response to a question about the National Database for Autism Research (NDAR) developed by the National Institutes of Health, Johnson pointed out that one of the difficulties with the database is the lack of people who can translate information into practices that can be implemented in schools, homes, and communities. Mandell added that NDAR has “beautiful clinical characterization” for understanding clinical features of people with autism, but “if you want to look at trajectory and treatment effects, it’s not a good database.” One problem he identified is that a standardized set of instruments does not exist for measuring treatment components or fidelity of implementation. A better approach might be to look at moderators and mediators of treatment effects across interventions, he suggested.
Different approaches to autism treatment tend to emphasize different treatment components and outcomes, such as observable behaviors of children versus their level of functioning and that of their families. For example, in the case of the boy hitting himself in the head, a 60 percent reduction in self-injurious behavior would be less meaningful than a measure of the family’s functioning, said Call. The boy’s mother “told us that the only way that she could keep him from hitting himself in the head [was] to
sit by his side and restrain him all day long. She didn’t take showers very often, she didn’t cook meals, she didn’t have very much quality of life at all. So for that mom a treatment goal would be, I need to be able to cook a meal, I need to be able to go to the store. Those are meaningful treatment goals, and when you approach it in that way, when that’s what you are working toward, that changes everything else,” he said.
The implementation of a particular approach is also critical, said Mandell. Two different approaches to autism treatment may be similar in essence, but the implementation of the approaches differs, which in turn has an influence on outcomes. For that reason, he said, different strategies “all should be at the table.”
Outcomes are in turn related to the evidence base supporting an approach. An intervention may produce an effect but not a meaningful outcome, said Kasari. Yet reimbursement may hinge upon this effect. “I don’t get reimbursed,” she pointed out, “but I do care that something has a meaningful outcome and that can be implemented by somebody who can get the outcome to happen.” To this, Johnson observed that “I do care about reimbursement because otherwise I can’t live” while agreeing that “this is a really great point, because what is evidence, and is all evidence the same?” For example, Johnson continued, specific reading interventions might improve phonemic awareness or fluency, “but what we don’t have is meaningful data showing that this intervention or this combination of interventions improves the ability to read, including understanding and being able to articulate back to somebody what was read.”
Call pointed out that caregivers are highly constrained by the options that are available to them. They tend to put much of their effort into well-known interventions that they consider effective while also devoting some time to less substantiated approaches. “Caregivers are making rational choices within the sphere of information that they have. When we give them good information about treatments, they make rational choices,” he said. However, when they do not have good information, they tend to make much worse choices.
A fundamental question, observed Kasari, is whether to incentivize particular practices or particular outcomes. Incentivizing outcomes assumes that people know what they should do or that there is not enough evidence to choose a particular practice. But with practices that do have evidence to support them, teachers, for example, might be expected to justify why they are not using those practices. From that perspective, case management might include doing quality assurance and improvement with care providers, with the collection of data to determine what to incentivize.
On the issue of outcomes, Mandell pointed to the examples set by the pediatric oncology and cystic fibrosis communities. Within each research and practice community, different treatment sites agreed on a common set
of outcome measures and committed to sharing data so that everyone can learn from each other’s practices. The challenge for the autism community is agreeing on the important outcomes not only within the medical setting, but also within the other settings in which children live, including their schools and homes. “We need to think about how to pull those groups together to have a common voice to create the same kind of successful networks that we see for condition-specific networks,” Mandell said.
This page intentionally left blank.
