The autism diagnosis has become a victim of its own success. Once recognised as a rare and profound developmental disorder, it has gradually expanded into a diagnosis encompassing an extraordinarily wide range of presentations. In a new essay in the journal Psychological Medicine, Uta Frith, one of the world’s foremost autism researchers, argues that the spectrum has become so broad as to risk losing its clinical meaning. She is right.

Frith began studying autism in 1966, when it was estimated to affect around one in 2,500 children (0.04%). Today’s studies suggest a prevalence of around one in 57 children (1.76%), a 44-fold increase. Across the UK, between 1.8% and 2% of the population is now estimated to be autistic. But why are so many more people now being diagnosed with the condition?

Since autism first became a formal diagnosis in 1980, successive revisions to the Diagnostic and Statistical Manual of Mental Disorders (DSM) have broadened the criteria and lowered the threshold for diagnosis. What began as a narrowly defined condition affecting children with serious social, communication and behavioural difficulties now encompasses people of any age and intelligence level, including those with no language impairment.

Few researchers doubt that autism exists or that it is highly heritable. High heritability, however, does not mean the condition is determined by genes alone; nor does it explain precisely how genetic and environmental factors interact in its development. Recent genetic studies have nevertheless raised important questions about the coherence of today’s autism spectrum. People diagnosed with the disorder in early childhood appear to have a different genetic profile from those diagnosed later in life. Rather than representing milder versions of the same condition, this may reflect fundamentally different conditions grouped under a single diagnostic label.

Greater awareness and reduced stigma have undoubtedly contributed to rising autism diagnoses, but they cannot fully explain the pattern. The increase has been concentrated among adolescents and adults without intellectual disability, particularly females, suggesting that changes in diagnostic practice have played a major role.

In her essay, Frith identifies several forces that have steadily widened the autism spectrum. The first is, of course, inclusivity. What began as an attempt to recognise overlooked cases may, she argues, have ended in over-inclusion. She also draws on what the philosopher Ian Hacking called “looping”. As autism has become part of popular culture, autistic people, clinicians and diagnostic manuals have begun to influence one another, creating a feedback loop that continually reshapes the diagnosis itself.

Social media has further amplified this process. According to Frith, autism has become “a popular cultural meme”. Online narratives spread rapidly, encouraging people to interpret their own experiences through an autistic lens, because human beings naturally adopt the language and identities available to them. If subjective experience begins to outweigh careful clinical assessment, ordinary differences in personality and temperament risk becoming self-diagnosed medical disorders. Online questionnaires, TikTok videos and identity-based communities have made autism a powerful framework through which many people, particularly adolescents, now understand themselves.

If autism can be diagnosed even when no observable symptoms are present because they are presumed to be hidden, then the diagnostic threshold risks becoming impossible to define. An unfalsifiable diagnosis is certainly an unusual way for any diagnosis to work. Modern medicine increasingly prioritises avoiding missed diagnoses over avoiding false positives. The result is a steady expansion of diagnostic categories, with autism becoming one of the clearest examples.

The explosion in the number of people diagnosed with the condition highlights the growing tension between the medical and social models of disability. According to the social model, disability is not located within the individual but instead arises from a lack of accommodation in society. Within this framework, an autism diagnosis is seen not as a tragedy but a potential “superpower”. Meanwhile, the medical model views autism as a neurodevelopmental disorder, and disability may result from that disorder.

Perhaps the clue lies in the comorbidities. More than half of autistic people have four or more co-occurring conditions, while almost three-quarters have at least one. The crucial point is that they are not distributed evenly. Early-diagnosed autism clusters with intellectual disability and language delay, while later-diagnosed autism clusters with anxiety, depression and other psychiatric disorders. That pattern may help explain why the two groups increasingly look like different conditions rather than different expressions of the same one. If so, our diagnostic system needs to reflect that reality.

Greater diagnostic precision would reduce misdiagnosis, improve care and ensure that those with the most profound autism are no longer lost within an ever-expanding spectrum. Only by describing people’s difficulties more accurately can we hope to understand them properly and offer the support they actually need.

---

Stella O’Malley is a psychotherapist and bestselling author. She is Founder-Director of Genspect, an international organisation that advocates for a healthy approach to sex and gender.

stellaomalley3

---