The leading U.S. therapy for autism is no panacea and has shown mixed results. Critics say it’s time to look for alternatives—and consult people with autism.
The dramatic rise of autism diagnoses in recent decades has played out in a barrage of media coverage, from scare stories over an autism “epidemic” to claims of miracle cures. Today, this increase is understood to be largely the result of better screening to identify individuals with autism symptoms—in addition to evolving (and widening) diagnostic criteria. We now have a better understanding of the varied ways people experience autism spectrum disorder, which research has shown is significantly related to underlying genetics.
The public conversation about autism has long centered on one question—why?—with feverish scrutiny lavished on the potential “causes,” from debunked vaccine fears to genetics to dietary factors. For decades, autism research has reflected this focus, overwhelmingly focusing on the basic science of causation. This has included not just a search for the underlying “genes of autism” but also efforts to uncover environmental factors like atmospheric toxins, microbiome signatures, and telltale vitamin deficiencies. Other environmental factors being investigated include maternal immune response and complications during delivery.
Often lost in the search for genetic and environmental risk factors are countless other vital questions, says Eric Garcia, a Washington, D.C. journalist and author of the book We’re Not Broken: Changing the Autism Conversation. The United States continues to be more interested in finding out why people are autistic and less interested in “understanding what autistic people need right now,” says Garcia. “I want to be clear about this: I don’t think studying things like biological causes of autism is a bad thing,” he adds. “But when we spend that much of a chunk of research dollars on that, compared to services, that’s a problem.”
Services for people with autism have lagged far behind investments aimed at uncovering its cause, consistently accounting for less than 10 percent of federal funding, despite the fact that investment in services ranks as a top priority among people with autism. “We’re only in the womb for nine months!” jokes Garcia. “I think that it’s really crucially important that we spend research dollars for other things.”
Life expectancy among people with autism has been estimated between 36 and 54, with the broad range explained by injury, suicide, as well as conditions that are more common among people with autism, such as epilepsy. Meeting the service needs of people with autism and researching lifespan issues received just 6 percent and 2 percent of research funds in 2018. A 2021 review paper found that nearly a decade of advocacy by stakeholders in autism care had done little to drive up levels of investment in autism services by the National Institutes of Health, despite the 2012 Combating Autism Act calling for a more coordinated and comprehensive approach to federal investment.
What do these services look like today?
They are wide ranging. In different ways, they aim to tackle symptoms that impede day-to-day functioning and reduce a person’s quality of life, from relationship-based play time to occupational therapy that helps autistic people negotiate schools or workplaces. Given the diverse ways that autism spectrum disorder impacts the lives of children and adults, such a range of options is not unsurprising. But recent years have seen the growth of a single school of therapy, called applied behavior analysis or ABA, as the dominant, one-size-fits-all treatment. Its emergence is one of the most surprising—and controversial—realities of autism treatment in the United States today.
Although it was developed by psychologists, applied behavior analysis is distinct practice, which acts at the level of behavior. Rather than delving into the brain’s workings., it teaches skills and pushes modified behavior through systematic reinforcement. Many applied behavior analysis clinics today target a “developmental window” in which they seek to enroll children from around 18 months to six years old and work with them intensively for up to 40 hours a week, often for years. The most common form of applied behavior analysis is “discrete trial training,” which breaks down daily situations and social interactions, such as playtime with other children or brushing one’s teeth, into organized cue-and-response tasks, in which the child receives praise or reward for responding appropriately.
Jon Bailey, a board certified behavior analyst and professor emeritus of psychology at Florida State University, who has worked in applied behavior analysis for four decades, says that sessions must be tailored to each child’s needs and should be continually adapted to their progress: “It’s a diagnostic method of determining which behaviors need to be changed, in what order, and to what amount. And going through those in a very systematic fashion.”
At the core of the practice is individualized data. When applied behavior analysis is done correctly, Bailey says, it cannot be a cookie-cutter system. Records are gathered from each session, then analyzed to fine-tune an evolving treatment plan. “The nature of the treatment is such that it’s constantly being evaluated,” he says. “If it’s done right.”
Unlike many other therapies, which emphasize qualitative, often self-reported improvements in a person’s wellbeing, applied behavior analysis produces piles of quantifiable data, which its proponents claim makes this approach “empirically proven to be the most effective method for treating individuals with autism.” When states began to be aware of a need to provide care for autism, this data brought applied behavior analysis to the front of the line, says John McLaughlin, who has led autism and special needs education businesses in Minnesota, Tennessee, and South Dakota for 45 years. Advocates call it the “gold standard” for autism treatment, and it is recognized as a best practice therapy by the U.S. Surgeon General and by the American Psychological Association.
Still, applied behavior analysis is one of the most divisive and emotive subjects in autism today, with debate reaching the American Medical Association (AMA) in the summer of 2022. In response to a petition to withdraw its previously explicit support for applied behavior analysis, the AMA instead removed a specific reference to the therapy and declared its endorsement for all “evidence-based” care, opening the door to non-behavioral approaches. Applied behavior analysis’ position as the sole therapy for autism that is covered by insurers in most states is a barrier to the development of alternative therapies, at the very least.
Scores of studies demonstrate that applied behavior analysis can help reduce so-called problem behaviors, but at the same time, the therapy has become “a quasi-monopoly” that critics say goes against the best interests of autistic Americans. “Treating a spectrum disorder with a uniform model is unique as well as paradoxical,” says John Summers, a Massachusetts-based writer whose son has autism, writing in The Nation. “In no other area of child development does government prescribe and mandate access to one—and only one—packaged therapy.”
By 2019, some two-thirds of Americans with autism had received applied behavior analysis therapies, which are overwhelmingly designed for young children. This makes the United States an outlier, with the uptake of this therapy much lower in comparable countries worldwide, and sporadic across Europe. Apart from the United States, only Canada and Australia certify behavior analysts, with the United Kingdom set to join in 2025. But a recent report that made recommendations for the British National Health Service found intensive applied behavior analysis was unlikely to provide sufficient health improvements to justify its high cost.
The use of intensive, regimented 40-hours-a-week schedules for young children is particularly disputed. Some studies have found the 40-hour weekly regimen effective at supporting long-term change, but the overall evidence base is low and the United States remains a global outlier in its wholehearted adoption of intensive applied behavior analysis. “As far as the actual empirical evidence for the 40-hours-a-week claim, it’s paper-thin,” says Zack Williams, chair of the International Society for Autism Research’s committee for autistic researchers.
A hugely controversial 2019 report for the Department of Defense came to similar conclusions when analyzing the experience of veterans’ families insured under Tricare, a health care program for U.S. service members and their families. It found that 76 percent show little to no change in symptom presentation over the course of 12 months of applied behavior analysis services, with an additional 9 percent demonstrating worsening symptoms. Some advocates for applied behavior analysis criticized the findings as “flawed,” but in 2021, Tricare reformed its autism program to stop funding for some behavior services and expanded coverage for other services. Another DOD report in 2021 found that 57 percent of participants in the program made “statistically significant improvements” with applied behavior analysis compared to 43 percent showing no improvement or worsening symptoms.
“The data is at best inconclusive,” Garcia says.
Rarest of all are studies that compare applied behavior analysis to alternative therapies, which could provide better information to help people make informed decisions about what’s best for people with autism, says Williams. A 2021 study by Sally Rogers at University of California, Davis compared discrete trial training with a play-based therapy, each in doses of 15 and 25 hours a week. Surprisingly, neither the therapy nor the dose had overall effects on child outcomes.
“These are the kinds of trials that we need,” says Williams, “and are very rarely done.”
But the AMA’s debate also reflects a more fundamental disagreement. The growth in applied behavior analysis therapies has coincided with the rising prominence of autistic self-advocacy organizations that typically embrace neurodiversity—the belief that differences in brain function are natural and should be accepted, not corrected. Many autistic self-advocates say applied behavior analysis is harmful compliance training, designed to make neurodivergent people act “normal.” These advocates have referred to behavior therapies as torture, trauma-inducing, or akin to gay conversion therapy—the highly controversial and discredited practice of seeking to change the sexual or gender identities of LGBTQ+ youth.
Critics tell parents to do their research and look into the field’s origins. Go back to behaviorism’s “father” B. F. Skinner or further and this becomes a mind-boggling journey through controversial human experiments, Cold War espionage, and basketball playing racoons, not to mention dehumanizing language. Discrete trial training, the earliest form of applied behavior analysis for autism, was first known as the Lovaas model, developed at UCLA by Skinner’s student Ivar Lovaas, who pioneered behaviorism as an early intervention for autism in the 1960s.
In a 1974 interview with Psychology Today, Lovaas explained that his approach had been built on the shocking-to-our-modern-eyes view that people with autism are not yet “people” and must be constructed by therapy: “You have a person in the physical sense—they have hair, a nose, and a mouth—but they are not people in the psychological sense,” said Lovaas in the 1974 interview.
In its earliest form, it was a system of drills and routines designed to modify undesirable behavior that included both positive feedback and systematic punishments called “aversives.” Lovaas became known to the public through a photo essay published in Life magazine in 1965 titled “Screams, Slaps, and Love,” showing autistic children receiving slaps and shocks, which Lovaas’ supporters argued made up a minute fraction of the program.
By the 1990s, such aversives had largely been phased out in favor of withholding rewards or ignoring undesired behavior. Yet controversies have dogged the practice of ABA, whose critics question its academic separation from mainstream autism research—research results are often found only in behavioralist journals and conferences but not typically in their neuroscience and medical journal counterparts. Some even question its basis of evidence. “I think it expanded without a lot of evidence to support it,” says Connie Kasari, a professor at UCLA who focuses on non-ABA autism therapy. “There’s evidence, but it works for just some people at some ages. So we’re not training providers in the community to be, you know, a better therapist or better interventionist.”
Significantly, states mostly adopted laws requiring insurers to cover applied behavior analysis before the rise of the self-advocacy movement, says Garcia. “I think that is important because it was done without the input of autistic people. And therefore, autistic people weren’t able to say, either: ‘If we’re going to do this, here’s how we would do it. What are the standards we need to put in place to make sure that it isn’t too torturous?’ or ‘No, we don’t want this at all.’”
“You want to remember this idea that autistic people can advocate for themselves, that’s a fairly new idea,” Garcia says.