Autism diagnoses have been on the rise, but experts caution that an increasingly relevant factor to consider in growing prevalence is over-diagnosis – or a misdiagnosis or diagnosis of a profile that is mild and not significantly impairing.
In JAMA Paediatrics Network, Lester Liao and Eric Fombonne review the extent and mechanisms of over-diagnosis and examine its harmful implications.
Liao and Fombonne write:
First, diagnostic substitution contributes to increasing prevalence, specifically decreasing intellectual disability diagnoses and rising autism diagnoses. While this is not necessarily over-diagnosis, it suggests diagnostic ambiguity and evolving terminology.
Second, clinicians may stretch diagnostic boundaries to access supports that would otherwise be unavailable without an autism diagnosis and understandably see this as being pragmatic advocates for their patients with less concern for diagnostic accuracy, including whether impairments are due to correlates rather than core autistic symptoms.
Over-diagnosis could occur via misdiagnosis or milder impairments.
Third, standardised diagnostic tools like the Autism Diagnostic Observation Schedule, Second Edition, and Autism Diagnostic Interview-Revised, may be used and interpreted incorrectly.
Mechanical reliance on standardised scores in diagnosis fails to account for confounding factors in a child’s presentation. For example, poor eye contact may be due to inattention or social anxiety rather than to limited social reciprocity.
The presence of emotional/behavioural problems boosts scores of autism-specific instruments in the clinical range even among children without autism.
Close to half of children who received autism diagnoses in the community did not meet autism criteria when re-evaluated by an autism research team. This group notably had higher rates of psychiatric disorders, suggesting psychiatric complexity contributes to misdiagnosis.
Fourth, different standards of screening or evaluation lead to disproportionate diagnosis across different regions.
Autism has become a prominent diagnosis, and many behaviours are now perceived as autistic traits. A child who toe walks (which may be idiopathic) and has sensory aversion to socks (which is not specific to autism) may be considered to fulfil Diagnostic and Statistical Manual of Mental Disorders (Fifth Edition) criteria for restrictive and repetitive behaviours because autism is a widely used explanatory diagnosis for such behaviours.
By contrast, it is atypical to see a tall person and refer to Marfan symptoms.
Fifth, fundamental changes have occurred in diagnostic concepts and criteria. Autism has been reconceptualised as a spectrum rather than a categorical disorder, co-occurring diagnoses such as attention-deficit/hyperactivity disorder are no longer mutually exclusive from autism, symptoms do not need to be recognised before a young age, and language delay is no longer a mandatory criterion.
This has led to decreasing severity of presentations, with concurrent intellectual disability in autism dropping from roughly 70% to 30% over 40 years.
By contrast, where symptom phenotypes have been consistent, diagnoses have increased over time.
Relatedly, perceptions of functional impairment have also shifted. In one study, reported impairment increased with later birth cohorts despite similar symptoms of autism spectrum disorder.
In other words, milder presentations are deemed increasingly impairing.
Autism and camouflaging also highlights this trend. Camouflaging presupposes a person understands acceptable behaviour and then adjusts their more natural behavioural pattern accordingly, even if it is uncomfortable. Children with profound autism may not understand the norms in the first place, let alone adjust.
Camouflaging involves a much milder form of autism, ie, an expansion of the autism concept.
Broader analyses suggest another angle to account for over-diagnosis. When certain diagnostic categories or “human kinds” are created, such as autism, they are susceptible to looping effects in which people begin to attend to certain traits of a human kind category (e.g, autistic traits), and subsequently identify with and reshape their own understanding of themselves according to such categories.
Hence, the neurodiversity movement’s identity-based contention is that one does not have autism but rather is autistic. Contemporary Western societies are “thin” cultures without many shared moral rituals and frameworks that provide rich meaning across the society.
Combined with Western existential, individualistic, and secular values, personal identity becomes increasingly important as a means of making sense of existence.
Even beyond the empirical evidence, the fundamental cultural layers in the West appear predisposed to increasing autism diagnoses.
Harmful repercussions
Naming a cluster of traits that is not impairing does not necessitate a healthcare professional and medical diagnosis. Insofar as diagnostic expansion redirects attention from those in greater need, it is harmful.
First, over-diagnosis leads to resource dilution. There are only so many public sites and professionals for evaluation of autism and therapeutic intervention. Over-diagnosis diverts resources from children with the most significant challenges. Importantly, it is often those who are better resourced, whether in terms of cultural capital or functional capabilities (e.g, the capacity to speak and self-advocate), who can navigate the medical system to ensure milder diagnoses receive supports.
This leaves the most vulnerable – those who cannot self-advocate and exhausted families caring for their children full-time – without the support they need.
Second, the expansion of the term autism risks trivialising children with profound autism. The spectrum diagnosis has benefits, but it loses precision. A child with social and behavioural quirks is quite different from one who does not speak, isolates himself, and frequently self-injures.
Furthermore, many people will encounter milder autism but never encounter profound autism. The latter presentation is rarer, and these children tend to be in segregated settings, whether in specialised education or because behavioural profiles make public settings challenging. Thus, to speak of “autism” today and bring to mind the former but not the latter child exacerbates resource diversion.
The latter child is barely even known to exist.
Similarly, medicalisation increasingly lumps typical or transient behaviours into medical categories such as autism. A self-fulfilling prophecy can develop. A child who is socially withdrawn is permitted to isolate, minimising habituation to social circumstances and decreasing social opportunities, thereby reducing social practice.
The same applies to behavioural rigidities or sensory aversions. There is a risk of attributing all the child’s troubles to autism, as opposed to, say, a circumstance, thereby reinforcing one’s concept of the child. This does not allow a child to develop their full capacity. There’s a significant difference between a child who has difficulty doing something and a child who utterly cannot.
A spectrum does not negate this. Even on the taller end, there’s still a world of difference between being 2.1m or 1.8m tall.
Conclusions
The increasing prevalence of autism may be partly explainable by over-diagnosis. Diagnostic practices geared to accessing resources, incorrect diagnostic practices, increased attention, changes to diagnostic criteria and thresholds, and fundamental cultural shifts have contributed.
This can harmfully divert resources from children with greater needs, marginalise those with autism already at the periphery of our societies, and reshape the way children develop in ways that stunt natural resilience and growth.
Over-diagnosis can be mitigated by (1) reliance on several informants or data sources, (2) supplementing the evaluation of current functioning with earlier data points that confirm a trajectory of autistic symptomatology, (3) demonstrating that functional impairment results from autism as opposed to co-occurring conditions or contextual constraints, and (4) providing a clinically informed and validated procedure to integrate all measurements.
Some alternatives to over-diagnosis are also being explored, such as de-diagnosis clinics and collaborative approaches to autism diagnoses.
Further research must begin to explore how these practices will affect care for children in the future and ensure more equitable resource allocation.
Lester Liao – McGill University Health Centre, Montreal, Canada
Eric Fombonne – Oregon Health & Science University, Portland, United States
JAMA Paediatrics article – Autism Over-diagnosis and Its Harmful Effects (Open access)
See more from MedicalBrief archives:
Autism diagnosis rates triple in past decade – US cross-sectional study
Landmark intervention may reduce autism diagnosis rates by two-thirds
Australian experts dissect proposed new autism category