Fonagy tells MPs:
“no evidence of
overdiagnosis at
scale” — but a
“warning sign”
in the figures
The Fonagy review’s chair and a vice-chair told MPs of no evidence of overdiagnosis at scale. Fonagy saw a “warning sign”, and a system that gatekeeps “in order to be able not to treat”. Fisher and Frith’s BMJ piece wants more than an advertising ban.
Overdiagnosis “at scale” and “a warning sign”: what the Fonagy review’s authors told MPs
On 14 October 2026 Joe Robertson put it to Professor Peter Fonagy that his review’s report “identifies a population prevalence of autism in young people of about 1.2%,” “but a diagnosis that has increased to over 4%,” and asked: “How is that anything other than evidence of overdiagnosis, based just on your figures?” (Q260). The figures are Robertson’s reading of the report. Fonagy’s reply was exact: “What we said was that we have no evidence of overdiagnosis at scale.”
The setting was joint evidence to the Health and Social Care and Education Committees on children and young people’s mental health (HC 239). Fonagy is Chair of the Independent Review into Mental Health Conditions, ADHD and Autism; Professor Gillian Baird, one of its two Vice-Chairs, gave evidence with him. The Directory has covered the final report and where it stops and the government’s response on CQC regulation, advertising and a tariff, both published on 9 October 2026.
Fonagy then defined the word: “Overdiagnosis implies that there is nothing wrong with that person”. “We find absolutely no evidence of that.” But: “What we do point out is that there is a warning sign in the figures that we present.” “Unless the system is changed, it is almost inevitable that the problem of overdiagnosis will arise” (Q261).
The report’s executive summary agrees: “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.” Neither the report nor its authors said overdiagnosis is impossible. They said it has not been shown at scale, and that unless the system changes it is heading towards it.
Baird was blunter. “The first thing to say is that there is no evidence of overdiagnosis.” “To some extent, the studies have not been done” (Q226). Nobody, she said, has taken a very large group with one diagnosis and had another group of people examine them “to see whether they would come to the same conclusion.” What the review did find was “evidence of misdiagnosis”. A separate data gap sits under the same argument: Sima Rafiei and colleagues, in one of three BJPsych Bulletin opinion papers on autism and ADHD the Directory reported on 4 November 2026, write that national reporting “cannot yet reliably report how many people receive an ADHD diagnosis following assessment”, though they make no claim about overdiagnosis either way.
“There simply has not been a recent population prevalence study,” Baird said (Q262); the review’s eighth recommendation asks for regular ones. The last was in 2017. “We do not know what the population prevalence is” (Q263). The Directory’s prevalence statistics collect the estimates that do exist. The baseline has also moved: with DSM-5 in 2013, “There were specific changes for ADHD that raised the possible population prevalence by 2% in adults and 3% in children.” So no evidence at scale is partly a finding about the evidence, not a clean bill of health. The gap between diagnosis and prevalence is not only a British question: in Japan, a whole-nation claims study published in October 2026 projected that, at 2023 rates, 15.7% of people would receive an autism diagnosis by 49 — data its authors say reflect treated rather than community prevalence.
“Gatekeep in order to be able not to treat”: the gate, and the treatment behind it
Fonagy’s first change is the order of care: from “Wait, assess, diagnose, support” to “Recognise need, then understand need, then begin support, then establish diagnosis where it changes, and then review and adapt as needed.” Then: “At the moment, we have a system where you gatekeep in order to be able not to treat” (Q236). Outside the NHS, one gate has already moved from diagnosis to need: since 30 September 2026, neurodivergence can qualify for the Disabled Persons Railcard on its impact on rail travel, not on a diagnosis alone, though a professional still has to put that impact in writing.
He did not pretend it is free: “To have a needs-based system without adequate support creates a problem rather than solving one.” Baird warned against the opposite mistake: “If all you did was to increase capacity, you would not do a very good job at all” (Q235).
Baird described what sits behind the gate, answering Jodie Gosling: “The opportunity to move from a diagnosis of ADHD to implementing a very specific treatment,” “first of all a psychosocial intervention, which is the number-one treatment, and, secondly, a medication intervention,” “has halved since covid.” “There is no follow-through at the moment” (Q269). She gave the figure orally and without a source: it is her evidence, not a statistic the Directory has checked.
For anyone seeking an ADHD diagnosis in the UK, what sits behind the gate matters as much as the gate.
Right to Choose: “prioritisation”, not abolition — and the families who “fall into a hole”
Josh Fenton-Glynn said of Right to Choose: “In West Yorkshire, there is a two-year minimum wait for that” (Q245) — a two-year minimum wait the Directory has reported on.
Fonagy replied: “It is important to say that we did not recommend abolishing Right to Choose, but we did recommend prioritisation.” “If we change the system so that what it offers is support for the parents, Right to Choose becomes an irrelevance except for in the cases where it is not.” A British Journal of Psychiatry Letter published on 15 October 2026 cautioned against deprioritising ‘mild’ ADHD, reporting that in the 1958 birth cohort childhood ADHD symptom scores as low as 2 were associated with higher odds of death by 58.
On the share of ICB resource going to Right to Choose, Baird said “the costs of that are far higher than perhaps the NHS cost would be.” “It then means that there is nothing available to support the families afterwards.” Of such families: “They fall into a hole” (Q245). The Directory’s national Right to Choose figures show the growth: across 29 ICBs with comparable figures, Right to Choose spending on ADHD and autism rose 141% in a year, from £57.6m (2023/24) to £138.7m (2024/25).
Source: FOI requests by Ronnie Cane to all 42 ICBs in England, January 2026.
On standards for private providers, Fonagy said: “It needs to be something that is probably developed by NICE and then implemented with CQC oversight” (Q246). Two days earlier, on 12 October 2026, the Minister of State for Health, Karin Smyth, said in a written statement: “In future, we will ensure that NHS funding is only used for assessments delivered by CQC-registered providers”. That decides who may hold the key. It does not change what the gate is for.
“Does not go far enough”: Fisher and Frith’s case against the market — and Fonagy’s own account of the incentive
On 20 October 2026 the BMJ published an Opinion piece by Naomi Fisher and Uta Frith, “Private providers are profiting from the expansion of ADHD and autism diagnosis”. It is argument, not research. Its opening examples — a billboard reading “Skip the waiting lists”, a discount code, a bus advert — are the authors’ own anecdotes, with no source, date or named location. Fisher declares that she is author of the forthcoming book This Explains Everything: How Autism Expanded and Why That Matters to Everyone; no other competing interests are declared.
They write that the review “rightly recommended” banning direct-to-public advertising of NHS-funded assessments, and that “this does not go far enough.” More stringent regulation of private providers, they write, “would be a start, it is not enough”: the “conceptual and cultural problems” behind the rise in demand have to be confronted too. Private providers, they argue, “generally have an incentive to increase demand through marketing and to provide assessments as cheaply as possible”. On scale, the review’s executive summary says independent providers “now carry out more than half of all NHS-funded ADHD (attention deficit hyperactivity disorder) assessments and more than a third of autism assessments.” Fisher and Frith write that where assessors specialise only in autism and ADHD, “no other diagnostic outcomes are possible”, and that “high confidence in autism and ADHD diagnoses derived in this way is misplaced”. Without a biomarker, they argue, definitions “have broadened over time”, and people with severe or profound autism and ADHD “have become a minority” — a pattern the Directory traced in profound autism’s falling share of US autism cases.
A declaration, because it bears on the reading. The Directory lists private and independent assessment providers and hosts reviews of them. Fisher and Frith do not name directories. But they note that a customer denied the diagnosis they hoped for “may leave poor reviews or complain”, a pressure on providers that runs through any site that hosts reviews, this one included. Read this piece knowing the Directory is part of the market it describes.
Fonagy’s own account of the incentive blames no one. Asked by Gregory Stafford for the evidence behind the advertising ban, he said: “You have a system where people are incentivised by the number of diagnoses they do,” “and you enable them to drive up demand through advertising, which the NHS has to pay for.” “That is an avoidable problem” (Q237). Stafford replied: “Demand is driven up, but previously you said that levels of diagnosis had not gone up.” The Chair said, “We are going to come back to this,” and the session moved on. One answer is implicit in his Q261: the claim is not that diagnoses have not risen, but that there is no evidence they are wrong at scale.
A gradient, not a conspiracy: where the incentive sits — and what neither remedy reaches
The two accounts agree more than their framing suggests. Fisher and Frith write that people seeking assessment “are incentivised to do so by a system that positions diagnosis as the key to services and self understanding.” Fonagy told MPs the system gatekeeps in order not to treat. Both put the incentive in the system, not in the people seeking help — the point the Directory made against the “incentivised” framing of the interim report. In The Neurodiversity Book (chapter X, “The Economics of Inevitability”) I put it this way: “That is not a conspiracy. It is a gradient. And things roll downhill whether or not anyone pushes.”
They part on the remedy. Fonagy’s is the order of care, plus a regulator over those who assess. Fisher and Frith want the categories’ “unclear and elastic boundaries” confronted, and call diagnoses “descriptions, not explanations.” The Directory has argued that part of the demand starts upstream of the clinic, in what a diagnosis has come to mean as an identity. If so, neither tighter criteria nor CQC registration reaches it.
The hearing added a plainer account of what happens if nothing changes: a “warning sign” in the figures; treatment after diagnosis “halved since covid”, on Baird’s unsourced account; and families who “fall into a hole”. By 14 October 2026 the government had promised to change who may assess and what the NHS will fund. The test for whatever follows is whether it changes what sits behind the gate. If not, a system that gatekeeps in order not to treat will simply get better at gatekeeping — and the overdiagnosis Fonagy warned of may arrive anyway.
Citations
House of Commons Health and Social Care Committee and Education Committee (2026) — Oral evidence: Children and Young People’s Mental Health, HC 239, Wednesday 14 October 2026, Questions 221–294 — UK Parliament
Fisher, N. & Frith, U. (2026) — Private providers are profiting from the expansion of ADHD and autism diagnosis — BMJ 2026;395:e101196 (Opinion)
Department of Health and Social Care (2026) — Independent review into mental health conditions, ADHD and autism: final report — GOV.UK, 9 October 2026
Smyth, K. (2026) — Health and care reviews, improvement and regulation (HCWS364) — UK Parliament written statement, 12 October 2026
News and updates.
Straight to your inbox.