Autism Diagnosis Today (1): Overdiagnosis

Autism has undergone a remarkable transformation in recent decades. Once thought of as a rare and narrowly defined childhood condition, it is now recognised as a broad and heterogeneous Autism Spectrum Disorder[1] (autismS) that includes people with very different causes, experiences and needs. This expansion has helped many children and adults finally receive recognition, understanding, and support. Yet alongside these benefits, experts are increasingly voicing concerns about diagnostic inflation and overdiagnosis—the risk that autism is sometimes identified when another explanation may fit better.

Several diagnostic tools have been developed.[2] However, diagnosing autism is not as straightforward as ticking off items on a checklist. Discrepancies between test results are common; there is no simple algorithmic solution to resolve them, and expert clinical judgment is necessary. It is the work of experienced clinicians trained to distinguish autism from overlapping conditions. This is particularly critical when evaluating impairment. Social difficulties may stem from anxiety, disruptive behaviour, or lack of social knowledge—but not all social impairment is “autistic” in nature. Careful differential diagnosis is essential (Bishop and Lord 2023).

A child who talks at length about dinosaurs after a museum trip may simply be sharing an age-appropriate interest. For that interest to count as “circumscribed” in an autistic sense, it would need to be unusually intense, inflexible, or disruptive. Similarly, unusual eye contact is not unique to autism – it can appear in anxiety disorders, ADHD, and some other psychiatric conditions. What matters is not simply whether a child looks or doesn’t look, but whether their ability to modulate gaze in the give-and-take of social interaction is impaired in a specifically autistic way.

To assess the psychopathology of children and adolescents, it is necessary to rely on data from multiple informants and sources, to evaluate the situational specificity or pervasiveness of relevant behaviours, and to assess functional impairment beyond merely taking symptom inventories (Fombonne 2023). Evaluation of an individual for possible ASD requires a process (not just test results) with flexible administration and scoring of tools, followed by integration of all findings by a clinician with autism knowledge and experience.  Inadequate diagnostic processes, inexperienced testers or shortcomings of the diagnostic instrument can result in overdiagnosis (Bishop and Lord 2023).

For example, in one study, Duvall et al. (2022) re-evaluated over two hundred children who already had an autism diagnosis. After a comprehensive assessment, fewer than half of them—just 47% – met research criteria for ASD Many of the others had indeed been judged to “meet DSM criteria” in their medical records, but a more thorough process revealed different explanations. This finding illustrates how relying too heavily on tools, or on the letter of diagnostic criteria without considering context, can lead to false positives and inflate the number of individuals labelled with autism.

Diagnostic expansion and its cultural dimensions

Overdiagnosis is not only a matter of instruments but of the diagnostic culture itself. Autism, as Poletti, Preti, and Raballo (2025) argue, has become stretched to encompass a growing range of presentations to the point of blurring its neurobiological distinctiveness. Two recent trends are particularly striking: the rise of adult diagnoses in people without documented childhood traits, and the tendency to interpret transdiagnostic social difficulties as signs of subthreshold autistic traits. Both developments risk conflating autism with general social dysfunction and undermining the validity of related clinical constructs.  

Irish neurologist Suzanne O’Sullivan, in her book The Age of Diagnosis, describes this expansion as “diagnostic creep.” She points out that concepts like “masking” allow those with subtle difficulties to claim the label and feel validated there is little evidence that such labels consistently improve everyday functioning. Worse, she argues, it risks becoming a self-limiting identity that may constrain rather than liberate.

At the same time, we must acknowledge the powerful cultural forces at play. Autism has moved from being stigmatised to being actively sought as an explanation, even an identity. On social media platforms like TikTok, autism is often presented in broad and simplified terms, encouraging many young people to self-identify with the condition. This trend has further loosened the boundaries of what autism “is,” contributing to the impression that almost any social difference might qualify.

The question, then, is not whether autism exists – it clearly does – but how to recognise it without stretching the diagnosis so far that it loses meaning. Mislabelling matters. A child wrongly identified as autistic may miss interventions better suited to their true difficulties, whether those are anxiety disorders, ADHD, learning difficulties, personality disorders or something else entirely. Overdiagnosis also makes research more difficult, since studies based on heterogeneous samples struggle to find consistent biological or psychological patterns.[3]

Some argue that overdiagnosis is a lesser evil than underdiagnosis, since services and supports are often tied to receiving a label. There is truth in this. But as Bishop and Lord (2023) caution, the solution cannot be to treat every individual who struggles socially as autistic. The real task of clinicians is to weigh the evidence from multiple sources, consider the developmental history, and exercise the judgment that comes only with experience. Autism is, at its core, a neurodevelopmental condition—its symptoms may look different across the lifespan, but they usually leave traces in early childhood. Recognising this trajectory, and distinguishing autism-related impairment from that caused by other psychiatric disorders, is critical to making the right decision.

Ultimately, what matters most is not the label itself, but what follows from it. Accurate diagnosis allows children and families to access the right kinds of support, without unnecessarily narrowing their identity or path. Overdiagnosis, by contrast, can create confusion, dilute the concept of autism, and lead to mismatched care. As O’Sullivan suggests, taking overdiagnosis seriously means looking more critically at our expanding definitions, questioning whether every new screening approach truly helps, and remembering that the goal is not to increase the number of people labelled but to improve lives.

______________

[1] Autism spectrum disorders (ASDs) are a heterogeneous group of neurodevelopmental disorders of genetic and environmental etiologies. Some ASD cases are syndromic: associated with clinically defined patterns of somatic abnormalities and a neurobehavioral phenotype. Many cases, however, are idiopathic or non-syndromic. Such disorders present themselves during the early postnatal period when language, speech, and personality start to develop. ASDs manifest by deficits in social communication and interaction, restricted and repetitive patterns of behaviour across multiple contexts, sensory abnormalities across multiple modalities and comorbidities, such as epilepsy among many others (Carroll et al. 2020).

[2] For example: the Diagnostic Interview Schedule for Children (DISC) and semi-structured ones such as the Child and Adolescent Psychiatric Assessment (CAPA); the Autism Diagnostic Observation Schedule (ADOS).

[3] Besides, recruitment for studies often relies on self-reported diagnoses or online checklists, which may introduce non-autistic participants into “ASD” groups.