When Matthew Maenner, PhD, was in college, he started working with autistic children as a behavioral therapist. In those days—the early 2000s—autism was still a relatively new diagnosis, and there were few supports available for autistic people. “No one knew what to do to help,” Maenner recalled in a recent interview with JAMA. Parents were desperate to find ways to support their children, sometimes turning to unproven diets or pricey speech-generation devices.
Their challenges stuck with him. After graduating in 2003 with a degree in psychology, Maenner joined the Wisconsin branch of the Autism and Developmental Disabilities Monitoring (ADDM) Network, a multistate autism prevalence-tracking project run by the US Centers for Disease Control and Prevention (CDC).
The ADDM Network—the largest ongoing autism tracking system in the US—is the source of frequently cited statistics on the prevalence of autism in children. The sites aren’t nationally representative, but the data they produce can be used to monitor trends and determine whether screening and other autism services are reaching different demographic groups.
A key question when Congress sponsored the surveillance effort as part of the Children’s Health Act of 2000 was whether increased rates of autism observed in prior studies represented an “epidemic, or an epidemic of diagnosis,” said Maenner, who is now chief of the Child Development and Disability Branch of the CDC’s National Center on Birth Defects and Developmental Disabilities.
More than 2 decades later, the CDC’s biennial reports have revealed a steady uptick in autism prevalence among 8-year-old children, from an estimated 1 in 150 children in the year 2000 to 1 in 44 in 2018. By 2020, an estimated 1 in 36 children were identified with autism spectrum disorder, according to the most recent report, published this March.
To get at those numbers, Maenner led a team that combed through medical and educational records of children at 11 surveillance sites, looking for an autism diagnosis, an autism special education classification, or an autism spectrum disorder International Classification of Diseases code.
Why has autism prevalence increased so dramatically over time? Maenner said the evidence points to more awareness, improved identification, and a better understanding of the ways autism manifests in people. What looks like an autism surge is attributable largely to external advances in the ability to understand and characterize the condition. Although researchers can’t rule out that more children today than in the past develop signs of autism, there’s currently no good evidence to support this, Maenner and other experts say. Evidence for an “epidemic of diagnosis,” on the other hand, is plentiful.
But the progress in recognizing autism hasn’t been uniform. And although disparities in who gets diagnosed have narrowed over the years, these improvements haven’t necessarily translated to better access to services for all autistic kids.
Closing the Gaps
In 2001, before Sarabeth Broder-Fingert, MD, MPH, started medical school, she worked at a preschool program for children with autism and other developmental disabilities. It was about a year after the revised fourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM) was published, and many people hadn’t even heard of autism, she remembers. By the time she started her pediatrics residency in 2011, most of the public was aware of the condition, although there was substantial stigma around the diagnosis.
Today, with the efforts of groups like the Autistic Self Advocacy Network, “we’re moving to a place where [autistic] people are really proud of who they are,” said Broder-Fingert, who is now a pediatrician at the UMass Chan Medical School and a member of the American Academy of Pediatrics (AAP) autism subcommittee.
The shift in awareness—and acceptance—Broder-Fingert has seen during her career is part of what underlies the ongoing increase in autism prevalence, she says.
Since psychiatrist Leo Kanner first described autism as a condition in 1943, health professionals’ ability to identify autistic children—particularly those who are part of historically underserved groups—has also steadily improved, thanks in part to initiatives like the CDC’s “Learn the Signs. Act Early” campaign that launched in 2004 and the AAP’s 2006 recommendation that children receive screening specifically for autism starting at 18 months.
“[W]e are doing a better job identifying kids early,” Kristin Sohl, MD, a professor of clinical child health at the University of Missouri, said in an interview. Sohl founded and leads a project that uses telementoring to increase access to early diagnosis—ideally by 24 months—in local communities. With the help of programs like these, autistic children are now more likely to be identified by 48 months, according to CDC data.
In 2020, an estimated 4% of 8-year-old boys in the ADDM Network were identified as autistic. Autistic girls have been chronically underdiagnosed or misdiagnosed, and tend to be diagnosed later than boys, but the data show that this is improving. Identification of autism in girls reached an all-time high in 2020, and for the first time, more than 1% of 8-year-old girls in the Network were estimated to be autistic, a threshold recognized in boys 2 decades ago.
Some of the biggest strides in identifying autism have been seen among children from racial and ethnic backgrounds that historically have had less access to diagnoses and services. In the ADDM Network’s early years, a greater percentage of White children were identified as autistic than were those of other racial and ethnic groups. That gap has narrowed as clinicians have identified more autistic Asian or Pacific Islander, Black, and Hispanic children. The 2018 report found no difference in the percentage of 8-year-old autistic children from these groups compared with White children.
In 2020, the autism prevalence increased for all of these groups and, for the first time, the prevalence among Asian or Pacific Islander, Black, and Hispanic children was higher than among White children. In addition, higher autism prevalence was associated with lower socioeconomic status—the reverse of what researchers found in earlier reports.
To Sohl, who chairs the AAP’s autism subcommittee, the increasing identification of autistic children from historically underserved groups is evidence that the delivery of screening and other services is improving. Researchers will be looking to understand the factors, including the social determinants of health, that may lead to a higher prevalence of autism in different populations of kids.
Inequities in identifying autism persist. As in previous reports, Black autistic children were more likely to have a co-occurring intellectual disability than autistic children from other ethnic or racial groups in 2020. This suggests that Black children with less pronounced traits may be slipping through the cracks, Sohl said. And even after an autism diagnosis, research has shown that there are still significant racial and ethnic disparities in access to resources.
Diagnostic Dynamics
Altogether, the reason researchers are finding more autism is “simple,” 2 autistic adults, one of whom is a neurodiversity expert, wrote in a recent editorial: “Clinicians are getting better at spotting what was always there.”
Plus, the value of having an autism diagnosis has changed over time. “It used to be that if you had an autistic child, you would try and get a different diagnosis for that child because there were no services for autistic children,” David Mandell, ScD, a professor of psychiatry at the University of Pennsylvania, said in an interview. It wasn’t until 1990 that Congress began including autism under what is now called the Individuals with Disabilities Education Improvement Act, for example, and not until the early 2000s that states started requiring private health insurers to cover autism services, such as behavioral interventions.
Because a diagnosis can unlock a number of supports for children and their families, “parents might go for specific autism testing now in a way that they never did before,” said Mandell, whose research focuses on the delivery of services to those with autism.
The criteria for an autism diagnosis are also different today. The diagnosis evolved as researchers and clinicians better understood the similarities and differences across other conditions, Sohl explained. In 1980, the DSM-III defined autism as its own condition, infantile autism, for the first time instead of as a type of childhood schizophrenia. At the time, an autism diagnosis required that 3 traits be present before the child reached 30 months old: a lack of social relationships, impaired communication, and peculiar responses to the environment, such as resistance to change.
The DSM-III-R, published in 1987, eliminated the age-at-onset requirement and added the diagnosis “pervasive developmental disorder—not otherwise specified (PDD-NOS).” And in 1994, the DSM-IV restructured the category to 5 different conditions—autism, PDD-NOS, Asperger disorder, childhood disintegrative disorder, and Rett syndrome—under the “pervasive developmental disorders” diagnosis.
In 2013, the DSM-5 changed the condition’s name to “autism spectrum disorder.” The criteria—which continue to guide clinicians today—require a child to have differences in 3 areas of social communication and interaction, as well as at least 2 of 4 types of restricted, repetitive behaviors.
Today, “we have a much better understanding of what is and what isn’t autism,” Sohl said. “Therefore, we diagnose it more clearly.”
Incredible Need
Despite these advances, the latest CDC report found considerable variation between states in the prevalence of probable or diagnosed autism, ranging from a low of 1 in 43 children in Maryland to a high of 1 in 22 children in California.
There must be something about local characteristics—either diagnostic practices or the availability of services for autistic children—that explains the different prevalence, according to Mandell. “I’m pretty confident that autism doesn’t respect geopolitical boundaries,” he said.
But Mandell, who worked on the CDC’s surveillance studies as a postdoctoral fellow, draws a distinction between autism’s true prevalence and what’s measured by surveillance efforts. “In a ‘true’ prevalence study, the information a child has in their clinical or educational record is irrelevant,” he and a coauthor wrote in a 2014 editorial in the journal Autism, and is instead determined by in-person clinical assessments by researchers. The CDC’s numbers, on the other hand, “measure the extent to which clinicians and educators test for and document the symptoms of autism,” they argued.
The distinction matters, Mandell says. He’s concerned that people who don’t understand the methodology—and the external forces that can affect who receives autism services—will wrongly assume that the increasing prevalence results from an increase in pollutants, or vaccinations, or prenatal factors.
A fundamental shift in how autism has been understood and identified over the past few decades is the far better explanation.
“There’s really not good evidence that there’s something else that would be changing over time in a way that would amount to an appreciable increase in prevalence,” the CDC’s Maenner said. A mathematical modeling study he coauthored in 2011, for example, found that changes in several perinatal risk factors for autism, either alone or in different combinations, could account for less than 1% of the 57% increase in US autism prevalence from 2002 to 2006.
The bottom line, Maenner says, is that many more children are being identified as autistic today than in the past. The CDC data “underscore an incredible need and likely demand for services across their lifetimes,” he said.
Autism doesn’t end when children turn 18 years old. “We know that autistic children become autistic adults,” Mandell said. This places autism in the realm of clinicians ranging from pediatricians to adult specialists.
Identifying autism is just the beginning. An autism diagnosis on its own can help children and their families better understand and relate to each other. And although there are long wait times for some behavioral interventions, other supports to manage co-occurring traits and conditions, such as speech and occupational therapies, are easier to access with a diagnosis. But there’s much more work to be done to support all autistic people. As Broder-Fingert puts it, “it’s not that helpful to get someone a diagnosis if you can’t get them all the treatment services that they need.”
Published Online: May 3, 2023. doi:10.1001/jama.2023.6078
Conflict of Interest Disclosures: Dr Broder-Fingert reported being a principal investigator or coinvestigator on multiple National Institutes of Health (NIH) awards to study early autism diagnosis and treatment and serving on the Scientific Advisory Board for EarliTec Inc. Dr Mandell reported receiving funding from the NIH; being a coinvestigator on grants funded by the Patient-Centered Outcomes Research Institute (PCORI), the Institute of Education Sciences, and the Eagles Foundation; and receiving honoraria for speaking engagements from a variety of universities and stipends from Sage Publishing and PCORI for editorial responsibilities. Dr Sohl reported serving on the advisory board of Quadrant Biosciences (parent company to As You Are), and being a board member of the Missouri American Academy of Pediatrics (immediate past president) and chair of the American Academy of Pediatrics’ autism subcommittee within the Council on Children with Disabilities; receiving institutional research support from Cognoa and grants from the NIH, Department of Defense, PCORI, and Autism Speaks; consulting for Medscape and Autism Navigator; and providing expert testimony to the Missouri Board of Healing Arts. No other disclosures were reported.
This content was originally published here.