Are ADHD and autism overdiagnosed? What England's independent review and the wider research found

Few questions in medicine generate as much heat as whether too many people are being told they have attention-deficit/hyperactivity disorder (ADHD) or autism. On October 9, an independent review commissioned by the government in England delivered its answer, in a report of about 600 pages. Headlines summed it up as a warning about overdiagnosis.

The review’s own wording is more careful than that, and it matches what researchers have found when they have looked at the question directly. Recorded diagnoses have risen many times over. The underlying conditions do not seem to have become much more common. That gap makes overdiagnosis possible, and for ADHD there is some direct evidence of it. The same evidence shows people still being missed.

What overdiagnosis means for ADHD and autism

The word is often used loosely, to mean that a condition is not real or that people are exaggerating. In medicine it has a narrower sense. The review’s executive summary defines it as giving a person a label for a disease or condition that either they do not have, or that would not have caused them harm or real difficulty if it had never been found.

An overdiagnosed person may fit the written criteria. The question is whether the diagnosis leaves them better off. Australian researchers who reviewed the ADHD evidence in the journal JAMA Network Open put the test this way: “for increased detection to represent current overdiagnosis rather than previous underdiagnosis of ADHD, we also need evidence that these additional cases do not derive a net benefit from diagnosis.”

Overdiagnosis is one of three separate errors. Misdiagnosis is the wrong label, or an incomplete one. Underdiagnosis is a missed condition. Underdiagnosis, misdiagnosis and overdiagnosis are different problems, and all 3 can occur within the same system at the same time.

Both conditions are hard to pin down for the same reason. There is no biological test, the review notes. ADHD and autism are neurodevelopmental conditions, meaning they arise as the brain develops, and their traits are spread across the whole population in varying degrees. Diagnosis depends not simply on the early developmental presence of traits, but on their severity, persistence and impact on functioning.

What England’s review concluded about overdiagnosis

The review was chaired by renowned psychoanalyst and clinical psychologist, Professor Peter Fonagy, in the words of ITV News. Its central sentence on this question is a negative one: “On the basis of evidence available to us we cannot conclude that overdiagnosis is occurring at scale, particularly given the gaps in population prevalence data and data reporting.”

Recorded diagnoses in some young age groups now exceed the best estimates of how many young people have the conditions. Of that, the review says, this does not, by itself, establish that overdiagnosis is currently occurring at scale. Population prevalence estimates are dated, use different methods and should not be treated as fixed biological ceilings. But the direction worries the authors. If present trends continue, they write, the balance of risks is changing, away from the earlier problem of missed cases and in the direction of mistaken and unnecessary ones.

The review also rejected the idea that the people being assessed are less affected than before. The rising numbers coming forward for assessment have not been accompanied by any fall in severity or impairment, it says. And it addressed them directly. People are seeking answers and support for very real problems, and they are not the problem.

Where it places the fault is in how help is rationed. The problem is a system that leans too heavily on diagnosis as the route to help and cannot keep pace. Fonagy made the same point to ITV News. “We made a mistake in placing support behind the door of a diagnosis,” Fonagy said.

How fast ADHD and autism diagnoses have risen

The review separates two kinds of number. Population prevalence is how many people have a condition, whether or not anyone has recognized it, and it is measured by surveying a sample of the public against fixed criteria. Administrative prevalence is how many people have a diagnosis on their records.

On the first, the review finds little movement. The available evidence suggests that the population prevalence of ADHD has remained broadly stable or increased only modestly over recent decades, although uncertainty remains because England lacks up-to-date population prevalence data. The most recent survey estimated ADHD prevalence at 3% among 5 to 15 year olds in 2017. For autism, the best available estimate remains around 1.2% among young people.

On the second, the change is dramatic. Administrative prevalence of ADHD has increased among children and young people aged 6 to 11 years by more than 6-fold (from 0.45% in 2000 to 2.70% in 2025) and among those aged 12 to 17 years by more than 15-fold (from 0.26% in 2000 to 3.88% in 2025). For autism the rise is steeper. Recorded autism at ages 6 to 11 went from 0.24% in 2000 to 4.58% in 2025, and it rose among those aged 12 to 17 by more than 40-fold (0.10% in 2000 to 4.06% in 2025).

Recorded autism in school-age children is now more than three times the survey estimate. Recorded ADHD in teenagers is above the 2017 survey figure and close to the review’s own rough projection of 3.8% for today.

Independent research had already traced the earlier part of this curve. A study of UK family doctors’ records found that, except among 3- to 5-year-olds, the incidence and prevalence of ADHD diagnoses and prescriptions have increased from 2000 to 2018 in all age groups. The largest proportional rise was in adults: among men aged 18-29 years, approximately 20-fold and nearly 50-fold increases in diagnoses and prescriptions, respectively.

Studies that measure the conditions directly suggest the rise is in recognition, not in the conditions. An international team pooled prevalence surveys of childhood ADHD. The team identified 154 original studies, analyzed 135 of them together and concluded that over three decades there has been no evidence to suggest an increase in the number of children in the community who meet criteria for ADHD when standardized diagnostic procedures are followed.

For autism, Swedish researchers compared two things in children born over a ten-year period: traits reported by parents in a standard interview, and diagnoses in the national register. The prevalence of the autism symptom phenotype has remained stable in children in Sweden while the official prevalence for registered, clinically diagnosed, autism spectrum disorder has increased substantially. Phenotype means the observable traits themselves.

The youngest-in-class effect in ADHD

Rising numbers alone cannot show overdiagnosis. They could equally reflect children who were missed in the past and are now found. To show overdiagnosis, researchers need cases where something other than the child’s condition is driving the label.

The clearest example is age within a school year. A child born just before the enrollment cutoff can be almost a year younger than a classmate born just after it. A younger child is, on average, more restless and less able to concentrate. If teachers and doctors read that immaturity as a disorder, the youngest should be diagnosed more often.

They are. A study in the New England Journal of Medicine used insurance records to test it. The study population included 407,846 children in all U.S. states who were born in the period from 2007 through 2009. In states where children had to be five by September 1 to start kindergarten, August-born children were the youngest in class and September-born children the oldest. Among the August-born, 309 of 36,319 had an ADHD diagnosis, or 85.1 per 10,000. Among the September-born it was 225 of 35,353, or 63.6 per 10,000. That is about a third higher for the youngest.

The authors tested whether the calendar explained the gap. These differences were not observed for other month-to-month comparisons, and they did not appear in states with a different cutoff date.

The Australian review counted similar studies. Of the 12 included studies that focused on relative age, 11 studies showed that the youngest children in class were more likely to be diagnosed with ADHD than the oldest children.

The same review found a second mechanism, in the written criteria. Eighteen of 20 studies that compared diagnostic prevalence between 2 or more diagnostic criteria described a concurrent increase in potential cases with the broadening of criteria. England’s review makes a similar observation about both conditions. Diagnostic concepts have also broadened, which has contributed to some increase in administrative prevalence.

Does a diagnosis help milder cases of ADHD?

The Australian review, led by Luise Kazda, is the most thorough attempt to answer the overdiagnosis question for ADHD in young people. It sorted 334 studies against five questions. Evidence that actual ADHD diagnosis had increased was found in 45 studies. Twenty-five studies showed that these additional cases may be on the milder end of the ADHD spectrum.

On whether those milder cases benefit, the evidence ran thin. Only 5 studies evaluated the critical issue of benefits and harms among the additional, milder cases. What they showed pointed one way. These studies supported a hypothesis of diminishing returns in which the harms may outweigh the benefits for youths with milder symptoms.

The reviewers’ overall verdict was firmer than England’s. This review found evidence of ADHD overdiagnosis and overtreatment in children and adolescents. They also acknowledged weak foundations: many of the studies behind three of their five questions were at high risk of bias. The two reviews differ on severity as well. The Australian one found signs that newer cases are milder, and England’s found no fall in severity. They cover different countries, periods and measures, and neither settles the point.

The same group then tested the question in a single cohort of Australian children. They found 393 adolescents with an ADHD diagnosis and matched each to another child of the same age and sex with the same level of hyperactive and inattentive behavior but no diagnosis. At 14 to 15, the diagnosed group reported no better quality of life (QOL). ADHD diagnosis was not associated with any self-reported improvements in adolescents’ QOL, the authors wrote. The diagnosed adolescents felt less sense of belonging at school and were more likely to harm themselves, with about two and a half times the odds.

That study cannot prove the diagnosis caused the difference. Children who end up diagnosed may have had difficulties the matching did not capture. It does show that, at similar levels of symptoms, a label was not followed by the benefits families might expect.

No comparable body of work exists for autism. The Australian studies concern ADHD in children and adolescents, and nothing here measures whether milder autism diagnoses help or harm.

Underdiagnosis of ADHD and autism has not gone away

The evidence that points to overdiagnosis also points the other way. The Australian review found that equally symptomatic girls were less likely to be diagnosed than boys in 2 studies. England’s review reached the same view of the past. Research suggests that at least part of the historical sex difference reflected lower rates of recognition rather than true differences in prevalence.

A study of autism diagnoses in UK family doctors’ records from 1998 to 2018 read the newer diagnoses as catching up. Its authors wrote: “Rising diagnosis among adults, females and higher functioning individuals suggest augmented recognition underpins these changes.”

England’s review is explicit that this is unfinished. Under-recognition nevertheless continues for some groups.

What England’s review recommends

The main recommendation is to stop making a diagnosis the price of help. Diagnosis has increasingly become a requirement to access help at school, at work and in the benefits system, the review says, and that drives people to seek one. It argues that the public argument itself has become part of the problem. The heated debate about whether there is too much diagnosis or too little risks reinforcing the centrality of diagnosis rather than need and timely access to support and treatment.

It also wants tighter control of who assesses. Much of the work is now contracted out by the National Health Service (NHS). Independent providers now carry out more than half of all NHS-funded ADHD assessments and more than a third of autism assessments. The review recommends making all diagnostic assessments for ADHD and autism Care Quality Commission (CQC) regulated activity and banning direct-to-public advertising of NHS-funded ADHD and autism assessments.

What the overdiagnosis research cannot settle

The true rate. England has no recent survey of how common ADHD or autism is in children, and none at all in adults. Without one, nobody can say how many of the recorded diagnoses are in excess. The review says as much: “Where evidence is incomplete, outdated or uncertain, we say so.”

Individuals. Evidence about populations says nothing about any one person. A child born in August who has an ADHD diagnosis may well have ADHD. The birth-month studies show only that some such children would not have been diagnosed had they been born a month later. Nor can they say whether the youngest are diagnosed too often or the oldest too rarely.

Autism. Almost all the direct evidence on overdiagnosis concerns ADHD in young people. For autism the case rests on the gap between recorded and surveyed rates.

Other countries. The figures in the review are for England. Diagnostic habits, school systems and health funding differ elsewhere, and so will the balance of over- and underdiagnosis.

Whether a particular child or adult meets the criteria for ADHD or autism is assessed by their doctor.

The fairest reading of the evidence is that all three errors can be happening at once, and that the review’s refusal to pick one is a conclusion in its own right.