Nearly half of children who receive an autism diagnosis from community clinicians do not meet the clinical criteria when assessed by specialist research teams, according to a stark new warning from leading researchers who argue the condition is being “wildly overdiagnosed”.
A Warning of Harmful Implications
The findings, published in the journal JAMA Pediatrics, come from Dr Lester Liao, a paediatrician at Montreal Children’s Hospital and McGill University, and Professor Eric Fombonne, director of autism research at Oregon Health & Science University. Both experts warn that current diagnostic trends carry “harmful implications” for children and risk diverting finite resources away from those with the most profound needs.
They argue that the expansion of autism as a diagnostic category means children facing the severest challenges are increasingly overlooked. “Overdiagnosis diverts resources from children with the most significant challenges,” they stated. Their research suggests it is often those with greater “cultural capital” or functional capabilities, such as the ability to speak and self-advocate, who successfully navigate the system, leaving the most vulnerable families struggling without adequate assistance.
How Everyday Behaviours Are Misread
At the heart of the issue, the researchers contend, is the frequent misinterpretation of common childhood behaviours as definitive indicators of autism. They highlight concerns about tools like the Autism Diagnostic Observation Schedule (ADOS), a 40-60 minute assessment involving play or conversation, which clinicians may apply incorrectly.
“For example, poor eye contact may be due to inattention or social anxiety rather than to limited social reciprocity,” the researchers noted. Other behaviours, such as toe walking or a sensory aversion to certain clothing, are also often misread as clear signs of autism. This diagnostic blurring, they argue, is a key driver of overdiagnosis, alongside greater public awareness and lowered diagnostic thresholds.
The study also challenges the popular concept of “camouflaging”, where individuals are thought to mask autistic traits in social settings. Professor Fombonne and Dr Liao argue that “camouflaging involves a much milder form of autism” and that children with more profound needs may not understand social norms in the first place, let alone strategically adjust to them.
The Real-World Consequences: From Self-Fulfilling Prophecies to Crumbling Services
Labelling a child as autistic when they do not meet the strict clinical criteria can create a damaging “self-fulfilling prophecy”, the researchers warn. They explain that a child perceived as socially withdrawn may be permitted to isolate, “minimising habituation to social circumstances and decreasing social opportunities, thereby reducing social practice.” There is a danger, they state, of attributing every difficulty a child experiences to autism rather than to circumstance, reinforcing a fixed and limiting perception.
This clinical warning arrives against a backdrop of skyrocketing diagnosis rates across the UK, which are placing immense strain on support services. In England, more than 166,000 autistic pupils currently attend schools, an 8% rise since 2020. More recent NHS data from June 2024 shows 193,203 patients in England had an open referral for suspected autism. By September 2025, that waiting list had ballooned to over 227,000, with average waiting times exceeding 16 months; some NHS trusts report waits of over eight years for a child’s assessment.
The prevalence varies significantly across the UK. As of December 2023, autism affects 4.3% of children in Scottish schools. In Northern Ireland, the 2024/25 figures show a prevalence of 5.9% among school-aged children. While global studies suggest the underlying prevalence of autism may have remained stable over the last 15 years—with a 2021 study estimating 1 in 127 people globally are autistic—the recorded rates in the UK have surged due to improved awareness and changing diagnostic practices.
These rising numbers hide significant diagnostic disparities. Research indicates that in England, Black children have a higher recorded prevalence (around 2.1%) compared to other groups, while children who speak a language other than English at home are less likely to be diagnosed. Autism is also more commonly diagnosed in males; in England in 2021, boys had a prevalence rate of 2.8% compared to 0.65% in girls, though some research suggests girls may be underdiagnosed or diagnosed later.
Notably, the children in the JAMA study who were misdiagnosed with autism showed higher rates of other psychiatric disorders, suggesting clinical complexity contributes to errors. The researchers’ central caution is that the drive to diagnose risks obscuring individual needs. “There’s a significant difference between a child who has difficulty doing something and a child who utterly cannot,” they wrote. “A spectrum does not negate this.”