Adults awaiting an autism diagnosis face a postcode lottery in England, with some NHS centres more than twice as likely as others to give a positive assessment of the condition.
Landmark research from University College London (UCL) suggests people have an 85% chance of being diagnosed as autistic in some centres compared with a 35% chance in others.
The findings, from the Improving Adult Autism Assessment study, have not yet been peer-reviewed but have already been picked up by NHS England for further investigation.
William Mandy, a professor of neurodevelopmental conditions at UCL, who led the research across several NHS foundations in southern England, said the wide variation suggested a lack of consistency between centres. He said it raised fears that lives could be badly affected by inaccurate diagnoses and called for “radical changes” to the system.
“Our results worryingly suggest that people are being overdiagnosed in some areas and underdiagnosed in others,” said Mandy.
The study considered a range of possible reasons behind the variation. But, he said, his conclusion was that “someone could go to one well-established clinic and get one answer as to whether they’re autistic, then go to another well-established clinic and get a completely different answer despite displaying exactly the same behavioural traits”.
“The fact is that the diagnostic manuals are open to a lot of interpretation and so there is fundamentally a lack of consensus as to where the right boundaries of autism are,” he added.
Dr James Cusack, the chief executive of Autistica, the UK’s leading autism research charity, said Mandy’s results did not surprise him. “Autism diagnosis can be a wild west in terms of inconsistencies in approach,” he said. “We know of many good quality assessment centres where people aren’t getting access to proper autism diagnoses because that centre follows untested diagnostic practices, sometimes even ones they’ve developed themselves.”
Cusack suggested the research indicated people were being overdiagnosed and said that “unless something is done to regulate this, it will affect public trust in diagnoses per se”.
He also said there were large regional variations in diagnoses. He pointed to research that found up to 4% of children aged 10 to 14 years old are being diagnosed in some areas, compared with 1.5% in others.
“That’s really dangerous because it means there’s no clarity in what a diagnosis is for, who it is for and how systems support the wide range of people being given a positive diagnosis,” he said.
Unlike Cusack, Mandy guesses that there is more underdiagnosis of adults than overdiagnosis. “But the key question is, ‘What level of diagnosis is correct?’ and the honest answer is nobody knows,” he said. “But the fact that there isn’t a consistent conversion rate between referrals and diagnoses is very troubling.”
Mandy said his results “show we need radical change in the way autism diagnoses are given”. He said: “Our findings worry me because this is a high-stakes assessment that is really important in affecting somebody’s identity, their access to services and funds, and their life decisions.”
The government’s 2010 autism strategy recommended the establishment of specialist, community-based, multidisciplinary teams to provide, coordinate and oversee services. However, there is no guidance around the staffing of these teams or how they should operate. As a result, a number of different models have emerged.
Bryony Beresford, a professor at the University of York, conducted the first national evaluation of specialist autism team service models in 2020. She also found a wide variation in reported rates of diagnosis, ranging from less than 50% to more than 80%.
“Each team differed in their diagnostic assessment protocols and each was unique,” she said. “In addition, unlike some diagnoses, clinical judgment is central to the autism diagnostic process and this will contribute to variation. Does this mean we can’t trust diagnoses? We just don’t yet have the evidence to say yes or no to that question.”
Dr Peter Carpenter, the chair of the Neurodevelopmental Psychiatry Special Interest Group, agreed that it was “very much in the eye of the clinician-beholder as to whether someone is diagnosed as autistic or not”.
He said there needed to be a national agreement of common practice and thresholds but that a level of variation was inevitable. “How do you get a definition of autism that experts can work to precisely and consistently which covers everybody at all ages and includes both members of Mensa and the most profoundly intellectually disabled?” he asked. “All our criteria involve a large amount of room for interpretation by individual clinicians.”
Many are now turning to independent diagnostic assessments but the National Autistic Society (NAS) said their decisions could not necessarily be relied on either. “There are many providers out there now, some for profit and others not for profit,” said Dr Sarah Lister Brook, NAS’s clinical director. “Some are regulated – many are commissioned by the NHS to deal with NHS waiting lists – but some operate with minimal regulation or oversight so the quality is variable.”
An NHS spokesperson said: “The NHS is fully committed to supporting and improving the lives of autistic people, and published new national guidance for autism assessment services to ensure local areas can manage the 50% increase in referrals they have seen over last year while ensuring people have the support they need as they wait to be assessed.”
