Autism Definition 'Too Broad,' Expert Warns of Clinical Misdiagnosis Risk

Uta Frith, a professor at University College London and a pioneering figure in autism research, has raised a concern that sits at the heart of modern diagnostic medicine: the definition of autism may have expanded so far that it has lost its clinical meaning. The warning comes as the field grapples with a paradox—efforts to cast a wider net and identify more people on the autism spectrum have, in her view, created a classification so broad that it risks becoming useless.

The expansion of autism diagnostic criteria over the past two decades was well-intentioned. Clinicians and researchers wanted to capture the full diversity of how autism presents, to move away from narrow stereotypes and ensure that people who struggled but didn't fit the old profile could still access diagnosis and support. The criteria widened. The spectrum stretched. More people received diagnoses. But Frith's concern cuts deeper than simple numbers: if the definition becomes a catch-all that applies to nearly anyone with certain traits, then the diagnosis itself stops doing the work it's supposed to do.

A diagnosis, at its core, is supposed to tell you something useful. It should point toward a specific set of challenges, suggest which treatments or interventions might help, and allow clinicians to distinguish between different conditions that might look similar on the surface. When a definition becomes too expansive, it blurs these distinctions. Someone with significant social communication difficulties and repetitive behaviors might receive the same diagnosis as someone with milder traits, even though their clinical needs, prognosis, and treatment pathways could be entirely different. The label stops being a tool and becomes merely a category.

Frith's warning also touches on a more immediate human problem: misdiagnosis. If the criteria are so broad that they capture people who don't actually have autism, those individuals might be directed toward autism-specific interventions and support systems that don't address their actual condition. A person whose difficulties stem from anxiety, trauma, or another neurodevelopmental condition might receive an autism diagnosis instead, delaying proper treatment and potentially wasting resources meant for people who genuinely need them. Conversely, the resources available for autism support could become stretched thinner across a larger population, reducing the intensity and specificity of help anyone receives.

The tension Frith identifies is not new in medicine, but it has become acute in psychiatry and neurodevelopmental assessment. The field has learned, rightly, that autism looks different in different people—that it's not a single presentation but a spectrum of presentations. Yet somewhere in the process of becoming more inclusive, the definition may have lost its boundaries altogether. The question now facing clinicians and researchers is whether those boundaries can be restored without losing the genuine progress made in recognizing autism's diversity.

What comes next will likely involve difficult conversations about where to draw lines. Healthcare systems may need to develop more nuanced diagnostic frameworks that preserve the breadth of recognition while restoring clinical precision. This could mean distinguishing between different profiles within the autism spectrum, or it could mean being more rigorous about which traits actually constitute autism versus which are simply human variation or symptoms of other conditions. The stakes are real: get it wrong, and people either go undiagnosed when they need help, or receive diagnoses that don't serve them.