When a Spectrum Becomes a Vacuum: How DSM-5 Broke Autism Diagnosis and Fed a Diagnostic Epidemic

A Simplification That Simplifies Nothing

The 5^th edition, published in 2013, set out to simplify and modernise the nosology of autism-related disorders, replacing the DSM-IV’s cluster of Pervasive Developmental Disorders (PDDs) — Autistic Disorder, Asperger’s Disorder, Childhood Disintegrative Disorder, Rett’s Disorder[1] and PDD-Not Otherwise Specified (PDD-NOS) — with a single diagnosis: Autism Spectrum Disorder (ASD).

The intent was, in principle, laudable: a spectrum captures gradation and avoids splits between “high-” and “low-functioning” labels. In practice, DSM-5 produced a conceptual flattening by collapsing important distinctions and introduced criteria so under-specified they undermine diagnostic coherence – creating a set of internal contradictions that have done more to muddy than to clarify diagnosis. These conceptual problems and the structural failure have real consequences: an expansion of contested adult diagnoses — some accurate, many uncertain, a fragmentation of access to care, and the political hardening of identity in ways that can protect errors from criticism.

The DSM-5 also introduced Social (Pragmatic) Communication Disorder (SPCD) which was meant to capture people with pervasive social communication problems without the restricted and repetitive behaviours required for ASD. On paper, SPCD appears precise. In practice, the boundary between SPCD and ASD is blurred, porous, and consequentially arbitrary (Swineford et al. 2014; Flax et al. 2019). When two diagnostic categories sit next to each other with soft borders, nearly identical presentations can be assigned different names depending on the clinician’s interpretation of whether, say, a repetitive behaviour is “sufficiently” present. A person labelled with SPCD rather than ASD may be denied autism-specific services, funding, and the political weight that often accompanies an ASD diagnosis.

Instead of simplification, the DSM-5 created ambiguity that became opportunity: for misdiagnosis, for identity formation around provisional labels, and for a mismatch between need and service.

From many disorders to one spectrum

Under DSM-IV (1994) the diagnosis of autistic disorders was messy but rule-bound. Clinicians had multiple discrete categories and fairly detailed symptom structures: Kanner’s classic formulation — impairments in social interaction, impairments in communication, and restricted, repetitive behaviours — remained the organising frame (Kanner 1943; APA 1994). Those categories were not perfect, but they set out combinations of symptoms and thresholds for diagnosis. With DSM-5 social communication and social reciprocity were folded into a single cluster (A, “deficits in social communication and social interaction”) and repetitive behaviours and sensory issues into a second (B). The manual also introduced a dimensional severity rating intended to capture inter-individual variability.

When Two Systems Collide: The Grandfather Clause and the DSM-IV–DSM-5 Transition

The shift from DSM-IV to DSM-5 should have been supported by a clear and reliable method for handling diagnostic change. Instead, DSM-5 introduced a “grandfather clause”[2] that automatically converted DSM-IV diagnoses of Autistic Disorder, Asperger’s Disorder, and PDD-NOS into ASD, regardless of whether individuals still met the new criteria. This effectively preserved DSM-IV categories inside the DSM-5 framework, resulting in two operating diagnostic systems at once.

Under this clause, anyone with a prior DSM-IV diagnosis inherits an ASD diagnosis after 2013. This applies even when current symptoms better match another category – such as SPCD – or when the person would not receive any diagnosis under DSM-5. For example, many individuals once diagnosed with PDD-NOS might now fit the criteria for SPCD (Kim et al. 2014; Regier et al. 2013), yet DSM-5 requires that they retain an ASD diagnosis. Likewise, someone who would meet neither ASD nor SPCD criteria today may still carry an ASD label solely because of a pre-2013 Asperger’s Disorder or PDD-NOS diagnosis.

These inconsistencies produce significant disparities. Two people with similar profiles may receive different diagnoses – and different access to services – depending only on when they first entered the diagnostic system. A person whose PDD-NOS diagnosis was automatically converted to ASD may have broad access to autism-related services, while someone assessed after 2013 with comparable difficulties may receive SPCD and be excluded from those same supports (Bishop 2010; Dockrell et al. 2012).

Such cases are not rare, given the overlap between DSM-IV PDD-NOS and DSM-5 SPCD.

An additional complication is historical: PDD-NOS in DSM-IV - for 6 years - included a copyediting error that used “or” instead of “and” in the criteria (First & Pincus 2002; Cooper 2015), making it possible to diagnose someone without evidence of social-interaction deficits. This error was corrected in DSM-IV-TR (2000) but nevertheless it contributes to the diagnoses that DSM-5 now preserves through the grandfather clause (Cooper 2015). A typographical slip was therefore able to authorise diagnosis of ASD in some people whose earlier diagnoses were, bluntly, formal errors.

While some argue that the DSM inevitably reflects compromises among clinicians, researchers, insurers, pharmaceutical companies, and families (Cooper 2015; Pickersgill 2012; Solomon,2017), diagnostic contradictions still have serious consequences. DSM-5 is meant to be a “living document”, responsive to new evidence and conceptual critique.

This becomes crucial if the DSM-5 criteria for ASD are themselves flawed – particularly given growing concern that they may systematically misdiagnosed personality disorders, complex PTSD, and anxiety as autism. If so, the grandfather clause may not only preserve outdated categories but also cement future diagnostic errors, amplifying them across generations of patients.

Misdiagnosis as epidemic: adults, trauma, and the checklist problem

The structural problems above have fed a more socially visible phenomenon: a surge in adult ASD diagnoses, many contested. The DSM-5 criteria and the shift toward checklist-driven assessment have de-emphasised developmental verification in some clinical contexts: clinicians confronted with adults reporting lifelong social difficulties or present-day sensory and repetitive behaviours sometimes make neurodevelopmental attributions without robust childhood evidence. That matters because autism, by definition, is a neurodevelopmental disorder with early onset (APA 2013, Criterion C). When adult diagnoses lack developmental corroboration — school reports, childhood accounts, early developmental milestones — the risk that the clinician is mistaking outcomes of trauma, personality organisation, or anxiety for autism increases.

Many clinicians and critics have pointed out that complex post-traumatic stress disorder (CPTSD), borderline or avoidant personality disorders, social anxiety, and other conditions can mimic autistic traits. Critics have argued that the DSM framework has medicalised normal variation, pathologising personality diversity, trauma, and introversion (Timimi, Gardner, McCabe 2010); 2022). Clinical psychologists warn that rising autism diagnoses “overlap suspiciously” with presentations of borderline or avoidant personality disorder and complex trauma (Shedler 2023). Trauma survivors often develop social strategies that look like autistic “masking” or social withdrawal; personality disorders are characterised by pervasive patterns of relationship and self-regulation difficulties that a checklist may misinterpret as neurodevelopmental difference. The checklist approach — in which observable behaviours are taken into account rather than understood in developmental and relational context encourages category errors.

If these structural issues are combined with rising public awareness and the expansion of diagnostic markets (private assessments, online self-referral pathways, identity-affirming communities), you get a sociological feedback loop: more diagnoses generate more identification, more resources flowing to a fuzzy population, a hardening of identities around provisional labels and critique becomes politically sensitive.

When combined with growing cultural familiarity with “the spectrum” and accessible self-identification pathways, those errors can increase. The result is not only clinical misdirection but political complication: advocacy communities may understandably resist scrutiny of diagnostic standards that confer resources and recognition, making corrective reforms fraught and contentious.

The human cost: services, identity, and professional accountability

Label inflation and misdiagnosis have real victims. A person who receives an ASD label without developmental evidence may be offered interventions that do not address their underlying problems; conversely, someone who genuinely needs autism-specific supports may be left behind if they are categorised differently (Bishop 2010; Dockrell et al. 2012).

At the systemic level, inflated prevalence estimates can divert limited resources away from those with the most severe neurodevelopmental needs. At the individual level, being given – or having removed – a diagnosis can be traumatically destabilising. The ethical stakes thus include both distributive justice and existential integrity.

Professional accountability is another casualty. Clinicians practice within institutions that reward diagnostic certainty and rapid throughput. Admitting diagnostic error is professionally costly. Without systemic structures that require and facilitate re-evaluation, mistakes solidify. The grandfather clause illustrates this: it froze previous categories into rights and benefits, allowing diagnostic errors to persist long after their foundations were shown to be shaky.

What reform should look like — urgent, careful, humane

If the DSM-5’s approach has demonstrably contributed to misdiagnosis and inequitable outcomes, what should happen? The answer is not to revoke labels reflexively, nor to deny the lived reality of people who have found meaning and community in an ASD identity. Rather, reform must balance epistemic rigor with ethical sensitivity.

First, reassert developmental verification as a non-negotiable diagnostic pillar for ASD in adults. Childhood history should be substantive, not optional. When developmental records are absent, clinicians must either treat the presentation as provisional or pursue a multidisciplinary assessment that explicitly screens for trauma, personality pathology, and anxiety disorders in parallel.

Second, develop transparent reassessment protocols. Where new evidence suggests a diagnosis was made in error, healthcare systems should offer voluntary reassessment pathways trauma-informed and supportive, not punitive. Scripts for clinicians, transition support guides, and clear documentation practices (what changed, why, and what services will follow) must be standardised.

Third, reconfigure access to services around functional need rather than categorical label alone. Many harms occur because diagnosis gates services. If supports (therapy, workplace accommodations, education plans) can be allocated on the basis of demonstrable need and function – rather than provisional nosological status – the worst distributive injustices would be mitigated.

Fourth, own the institutional errors. The DSM should be treated as a scientific and political instrument that can be audited. Editorial mistakes, like the PDD-NOS typo, are not trivia; they are governance failures. A publicly accessible audit of diagnostic transitions and their outcomes — prevalence shifts, service access changes, and patient narratives — would force a reckoning (Cooper 2015).

Finally, fund research that teases apart overlapping phenomena: longitudinal studies that relate early development to adult presentation; trauma-informed research on social withdrawal; and comparative studies that assess the real-world specificity and sensitivity of diagnostic instruments in adult populations (Solomon 2017).

DSM-5 did not invent confusion about human difference. But by compressing categories, introducing SPCD as a residual class, and operating with internal vagueness (and yes, preserving earlier editorial errors), it institutionalised a particular kind of diagnostic instability. That instability fed an epidemic of contested adult diagnoses and created ethical dilemmas that psychiatry cannot defer.

We should be frank: many people have found explanation, services, and community through ASD diagnoses; taking a scalpel to diagnostic criteria risks harm if not handled with compassion. But compassion must not become an obstacle to clarity. If psychiatry is to be both scientifically rigorous and ethically responsible, it must own its mistakes, reassert developmental criteria where appropriate, reorient services around need, and create humane processes for reassessment.

Until then, the field will remain in the uncomfortable position of having turned a useful metaphor – the spectrum – into a vacuum that draws in too many different kinds of human suffering and then fails to distribute care fairly. That is not a minor bureaucratic glitch. It is a diagnostic and moral crisis that demands correction.

_______________________________

[1] Rett’s syndrome was removed from DSM-5

[2] A grandfather clause is a provision in which an old rule continues to apply to some existing situations while a new rule will apply to all future cases. Those exempt from the new rule are said to have grandfather rights or acquired rights, or to have been grandfathered in. Frequently, the exemption is limited, as it may extend for a set time, or it may be lost under certain circumstances.