Autism prevalence has risen dramatically over the past two decades, from 1 in 150 children in 2000 to 1 in 31 in 2022, according to the CDC’s most recent surveillance data. That’s roughly a fivefold increase in just over 20 years. But most researchers agree the rise is largely driven by changes in how autism is defined, detected, and diagnosed, not by a sudden surge in a single biological cause.
The real answer is that several forces are working at once. Some are purely about detection. Others may reflect genuine shifts in risk. Here’s what the evidence actually shows.
Broader Diagnostic Criteria
The biggest structural change came in 2013, when the American Psychiatric Association replaced several separate diagnoses with a single umbrella category: Autism Spectrum Disorder. Before that, clinicians chose between Autistic Disorder, Asperger Disorder, and a catch-all category called PDD-NOS (Pervasive Developmental Disorder Not Otherwise Specified). The new system collapsed all of these into one diagnosis with varying levels of support needs.
This merger didn’t expand the net as much as people sometimes assume. A study comparing the old and new criteria in the same population found that 99% of children previously diagnosed with Autistic Disorder and 92% of those with Asperger Disorder still met the new criteria. The combined prevalence under the new system was virtually identical to the old one. What did change, though, was visibility. A single, well-known label replaced a confusing patchwork of subtypes, making it easier for parents, teachers, and doctors to recognize and discuss the condition.
The criteria also shifted in subtler ways. The age-of-onset requirement loosened from “before age three” to “early childhood,” and sensory sensitivities (like strong reactions to textures, sounds, or lights) were formally added as a diagnostic feature for the first time. Both changes meant some people who would have been overlooked before now had a clearer path to diagnosis.
More Children With Fewer Challenges Are Being Identified
One of the most telling patterns in the data is where the growth is concentrated. Between 2000 and 2016, the prevalence of autism among children with no significant day-to-day functional challenges increased by 464%. Among those with borderline challenges, it rose 382%. Meanwhile, autism prevalence among children with moderate to profound difficulties actually decreased slightly, from 1.5 per 1,000 to 1.2 per 1,000.
This pattern strongly suggests that the core population of autistic people with high support needs has remained relatively stable. The increase is coming almost entirely from identifying children who, in earlier decades, would have gone undiagnosed or received a different label entirely. Many of these children function well enough in early childhood that their traits only become apparent when social and academic demands increase.
Diagnostic Substitution
Part of the rise in autism numbers reflects a shift away from other diagnostic categories. As autism prevalence tripled between 2000 and 2016, the proportion of autistic children who also had an intellectual disability dropped. Children who might once have been labeled with an intellectual disability, a speech impairment, or a general developmental delay are now more likely to receive an autism diagnosis, either instead of or in addition to those other labels.
This doesn’t mean these children don’t have real needs. It means the same child is being categorized differently than they would have been 20 years ago. In the U.S. school system, autism now accounts for 13% of all students receiving special education services, making it the fourth most common category. For Asian students, it’s the single most common one, at 31%.
Girls and Women Are Finally Being Recognized
Autism has historically been seen as a condition that primarily affects boys, and diagnostic tools were largely developed and validated using male subjects. That’s changing. Over a 20-year period, the male-to-female ratio among those diagnosed at statewide autism clinics in North Carolina dropped from 5.64 to 1 down to 3.07 to 1. The CDC’s national estimate shows a similar trend, narrowing from 4.2 to 1 in 2016 to 3.8 to 1 in more recent data.
Females are still diagnosed an average of 18 months later than males and are more likely to receive a first diagnosis at age 13 or older. Many autistic girls and women develop social coping strategies that mask their traits in clinical settings, which has historically led to missed or delayed diagnoses. As awareness of this pattern grows, more females are being identified, and that alone pushes overall prevalence numbers upward. Between 2011 and 2022, diagnosis rates increased particularly fast among female children and adults.
Adults Getting Diagnosed for the First Time
The rise isn’t limited to children. Growing numbers of adults are receiving first-time autism diagnoses, many after decades of being misdiagnosed with anxiety, depression, or personality disorders. Diagnosis rates among adults increased substantially between 2011 and 2022, with notable growth among women. In 2022, roughly 20,000 females and nearly 58,000 males received a new autism diagnosis. While these adult diagnoses don’t directly affect the childhood prevalence numbers the CDC reports, they contribute to the broader cultural sense that autism is “everywhere now” and they reflect just how many people were missed by earlier, narrower criteria.
Better Screening and Awareness
Pediatricians in the U.S. now routinely screen for autism at the 18-month and 24-month well-child visits, something that simply didn’t happen a generation ago. Parents, teachers, and childcare providers are also far more familiar with early signs than they were in the 1990s. This combination of formal screening and informal awareness means children are being flagged earlier and referred for evaluation in much greater numbers.
Access matters too. Communities that historically had lower diagnosis rates, including Black, Hispanic, and lower-income families, have seen their numbers climb as screening becomes more widespread and culturally responsive. Much of the recent increase in overall prevalence reflects these communities catching up to diagnosis rates that wealthier, white families reached years earlier.
Biological Factors That May Play a Role
While most of the increase appears to be driven by detection, some biological trends could be contributing a real, if smaller, rise in actual prevalence.
Parental age is the most studied factor. A meta-analysis covering nearly 30 studies found that mothers over 35 and fathers over 40 have roughly 40% to 50% higher odds of having an autistic child compared to younger parents. The average age of first-time parents in the U.S. has been climbing steadily for decades, which means this risk factor is affecting a growing share of births. The leading explanation is that older fathers accumulate more spontaneous genetic mutations in their sperm over time, and some of those mutations are linked to autism.
Other environmental exposures have shown associations with increased autism risk, though none as consistently as parental age. Birth complications involving oxygen deprivation carry one of the stronger links. Prenatal exposure to air pollution, particularly fine particulate matter and nitrogen dioxide, has been associated with higher rates in retrospective studies. Maternal obesity and gestational diabetes show a modest but significant association. Prenatal exposure to certain pesticides, especially organophosphates, has also been flagged.
These factors are difficult to study in isolation because they often overlap with each other and with socioeconomic variables. None of them, individually or collectively, come close to explaining a fivefold increase. They may contribute a genuine uptick in prevalence layered on top of the much larger effect of improved detection.
So Is Autism Actually More Common?
The honest answer is: probably somewhat, but nowhere near as much as the raw numbers suggest. The jump from 1 in 150 to 1 in 31 is mostly a story about a society getting better at recognizing a condition that was always there, especially in girls, in communities of color, in adults, and in people without intellectual disabilities. Broader criteria, routine screening, reduced stigma, and greater public awareness have all pulled back a curtain on a population that was previously invisible to official counts.
At the same time, trends in parental age, prenatal exposures, and birth complications leave open the possibility that the true underlying rate has increased modestly. Disentangling detection from biology remains one of the central challenges in autism research, precisely because both are moving in the same direction at the same time.