The Diagnosis Arrived Late. ADHD Did Not.

What a Danish genetics study can—and cannot—tell us about South Africa’s growing population of late-diagnosed neurodivergent adults

A recent Danish study has added a new complication to the already uncomplicated and entirely calm public conversation about ADHD and autism.

Researchers examined the genetic profiles of more than 37,000 people diagnosed with ADHD or autism between 1994 and 2015. They found that people diagnosed more recently carried slightly lower average polygenic scores for the condition than those diagnosed earlier.

This is the sort of finding that can travel from a careful statistical paper to “ADHD is being handed out to everyone” before the abstract has finished loading.

That is not what the study found.

What it found was that the diagnosed population had changed. The authors interpreted this as evidence of a “lowering threshold”: earlier diagnostic systems identified a narrower group with more concentrated measurable genetic liability, while later systems admitted a broader range of presentations.

Even that interpretation requires care. A lower average polygenic score does not mean a weaker condition, milder impairment or an invented diagnosis. Polygenic scores capture only part of inherited liability. They are statistical tools, not genetic certificates of authenticity. They also perform most reliably in populations similar to those from which the underlying genetic data were drawn—which, as usual, means people of European ancestry were allowed to stand in for humanity while everyone else waited outside.

The study nevertheless raises an interesting question for South Africa. If diagnostic populations have broadened in Denmark, are we seeing the same process here?

Possibly. But not in the same way.

Denmark was asking who crossed a diagnostic threshold that already existed. South Africa must first ask who had access to the building.

A diagnosis can increase without a condition increasing

The LaBianca study included 20,111 people with ADHD and 17,071 autistic people. For every ten-year increase in year of diagnosis, the average ADHD polygenic score among people diagnosed with ADHD fell by 0.06 standard deviations. The autism score among autistic people fell by 0.07 standard deviations.

These are small population-level changes, not dramatic biological collapses. Recently diagnosed people still carried higher condition-related scores than the general population.

The wider pattern mattered more than the size of any single change. Scores associated with several genetically related psychiatric conditions also tended to decline. In the researchers’ simulations, this fitted a broadening diagnostic boundary better than either diagnostic substitution or the simple discovery of previously missed, genetically similar cases.

There is supporting evidence for part of this story elsewhere. An Australian study found that childhood ADHD diagnoses increased between two birth cohorts even though average parent-rated hyperactive and inattentive behaviour remained stable. Swedish studies have found a similar gap between growing ADHD or autism diagnoses and relatively stable population-level traits.

This suggests that rising diagnosis rates cannot automatically be interpreted as rising rates of the underlying neurodevelopmental traits. Recognition, referral, diagnostic practice and access all change who appears in the official figures.

That is where the South African story becomes rather more complicated.

South Africa has never had one diagnostic threshold

We tend to discuss diagnosis as though a neutral clinical system examines the population and decides who meets a consistent standard. It is a charming idea.

In practice, South Africans have passed through at least two filters.

The first is access.

Who had medical aid? Who lived near an appropriate professional? Whose school knew where to refer them? Who could afford multiple consultations? Who had a parent, partner or employer able to recognise that the difficulty might be clinical rather than moral? Who could pursue another opinion when the first professional dismissed the possibility?

The second filter is recognisability.

Even people who could reach professional care were not necessarily assessed through a modern understanding of ADHD. For much of recent history, the expected patient was a disruptive, hyperactive boy whose difficulties were sufficiently inconvenient for the adults around him.

The child who stared out of the window, passed through anxiety and last-minute panic, or performed well because a parent supplied the executive function was less likely to attract attention. The girl who internalised the cost was less likely to be referred. The adult was often not considered at all because ADHD was still treated as something children either outgrew or had the decency to stop mentioning.

South African research gives us reason to take these barriers seriously. Child and adolescent mental-health specialists remain concentrated in a small number of urban centres. Public-sector integration is limited, professional training remains uneven, and recent estimates suggest that only a fraction of South African children needing mental-health care can access it. Research in the Cape Metropole has also found substantial gaps in teachers’ ADHD knowledge, despite teachers often being the informal gatekeepers to assessment.

Access, awareness and visibility therefore selected the people who became diagnosable.

White South Africans, particularly those with private healthcare and well-resourced schools, were more likely to clear the access barrier than most of the population. But better access did not guarantee informed recognition. Our professional community also lagged behind changing international knowledge. Even within the comparatively advantaged population, diagnosis remained tilted toward the presentation that shouted loudest.

The result was not one hidden cohort. It was several hidden cohorts, hidden for different reasons.

Invisible does not mean mildly affected

This is where the phrase “lower diagnostic threshold” becomes especially misleading.

It encourages us to picture a line representing severity. The most severely affected people sit above it; mildly affected people sit below it. Move the line down and less impaired people qualify.

That is neat, simple and probably wrong.

Clinical visibility is not the same thing as impairment. A child can create very little trouble for a classroom while using enormous internal effort to remain there. An adult can maintain employment through anxiety, overwork and repeated private collapse. A student can appear capable because intelligence and panic repeatedly rescue what routine study never could.

None of these people is necessarily close to a mild end of anything. They are simply expensive to themselves rather than inconvenient to the institution.

I was diagnosed at 45. ADHD did not suddenly emerge in the consultation room, fully formed and carrying identification. What arrived was an explanation. Before that, the same patterns had been interpreted through character, motivation, anxiety, addiction, inconsistency and the familiar instruction to try harder—with the refreshing assumption that this had somehow not occurred to me.

There is also a less visible gate waiting for adults who eventually reach the consulting room. Having spent decades doubting their own experience, they may meet a professional who begins by doubting it for them.

Adult self-referral is too easily treated as evidence of confirmation bias: the patient has read about ADHD, recognises the pattern and is therefore regarded as suspiciously invested in the outcome. The consultation can become an exercise in disproving a claim rather than hearing a history. Reasonable differential diagnosis is essential; professional discourtesy is not. Yet adults already carrying a substantial deficit of self-trust can find themselves expected to establish their innocence before their difficulties receive serious attention.

Some of this resistance comes from professionals with genuine expertise in childhood ADHD whose model has not moved with the population now seeking assessment. They know the presentation that was historically diagnosed exceptionally well. The problem arises when mastery of that narrow cohort is mistaken for mastery of everyone the cohort excluded. The adult who compensated, internalised or collapsed later is then measured against the disruptive child who was visible early—and their failure to resemble him is treated as evidence against them rather than evidence that diagnostic recognition has changed.

Late diagnosis does not rewrite every earlier experience into ADHD. It does, however, provide a framework for examining patterns that older systems were not equipped to recognise.

If newer South African diagnostic cohorts contain more adults, women, inattentive presentations and people whose scaffolding eventually failed, they may look less like the highly visible childhood cases on which earlier diagnostic knowledge was built. That does not make them less valid. It exposes how narrow the original reference group was.

The diagnosis arrived through overseas markets. The ADHD did not.

South Africa often follows European and American health discourse at a delay. Professional training, public awareness, diagnostic language and online information reach us after they have already altered practice elsewhere.

That creates an easy accusation: if South African adults began identifying ADHD after it became prominent in American and European media, perhaps we simply imported the diagnosis.

But the distinction matters.

We imported language. We imported assessment models. We imported professional debates, some better than others. We imported social-media simplifications in which misplacing your keys before lunch apparently earns a diagnosis by dinner.

We did not import the lives being described.

The adults now seeking assessment were already living with inconsistent activation, unreliable attention, emotional dysregulation, time blindness, working-memory difficulties and the accumulated self-criticism produced by decades without a coherent explanation. The arrival of better language changed their ability to interpret those experiences. It did not manufacture the experiences retrospectively.

This does not mean every assessment is rigorous or every diagnosis correct. Increased demand inevitably attracts hurried assessment, checklist diagnosis and commercial opportunism. South Africa can experience underdiagnosis and poor-quality diagnosis at the same time. We can even have pockets of overdiagnosis inside a country where most people still have no realistic access to assessment.

Public debate struggles with this because it prefers one scandal at a time.

A two-speed diagnostic catch-up

The South African increase is unlikely to be evenly distributed.

Recognition will initially expand among people who already possess some combination of private healthcare, money, education, English-language information, internet access and enough autonomy to question an earlier opinion. In practical terms, that is likely to include a disproportionate number of white and economically advantaged South Africans.

Meanwhile, Black and Coloured South Africans, people in rural communities, and those dependent on an overburdened public system may remain subject to a much higher diagnostic gate. Overt behaviour may be punished rather than investigated. Inattention may disappear inside overcrowded classrooms. Functional difficulty may be attributed to poverty, trauma or poor discipline—as though neurodevelopmental difference politely suspends itself whenever social hardship is present.

This gives us a two-speed process.

One part of South Africa may begin to resemble the European pattern: diagnostic recognition spreading beyond conspicuous childhood cases toward adults and less stereotypical presentations. Another part may remain where the better-resourced population was decades ago, with only the most visible cases receiving attention—and many of those receiving the wrong kind of attention.

If we eventually conducted a South African version of LaBianca’s study, the results would be difficult to interpret without separating genetic ancestry from the social consequences of racial classification. A change in average polygenic score might reflect changing diagnostic practice. It might also reflect changing access among racial and socioeconomic groups whose representation in the diagnosed population had previously been radically unequal.

What appeared to be a genetic trend could partly be a history-of-access trend wearing a laboratory coat.

Validity does not require resemblance to the people diagnosed first

The most troubling use of the Danish study would be to create a hierarchy of diagnostic legitimacy: earlier cases as biologically authentic, later cases as diluted copies.

That mistakes historical selection for a gold standard.

Earlier systems disproportionately identified people whose ADHD was externally visible, disruptive and compatible with prevailing professional expectations. If we require every newly diagnosed adult to resemble that group, we do not protect diagnostic validity. We preserve the bias that excluded them.

Nor should polygenic scores become the new tribunal. They are not diagnostic instruments, do not capture all genetic contribution and transfer poorly between ancestry groups. A South African’s neurodivergence does not become scientifically doubtful because European genetic databases were assembled without adequately including them.

Validity should rest on careful clinical assessment: developmental history, persistence across contexts, functional impact, differential diagnosis and the person’s wider environment. It should remain open to uncertainty without turning uncertainty into automatic disbelief.

There is a difference between questioning the quality of an assessment and questioning whether an entire class of people could have been missed. South African history should make us particularly cautious about assuming that institutional records contain everyone who was ever present.

The question behind the increase

The question is not simply whether South Africa is diagnosing more ADHD.

We need to ask who is finally being recognised, who remains excluded, and what kinds of suffering our systems are still trained to notice.

For some South Africans, increased diagnosis may represent a delayed version of the broadening seen in Denmark, Europe and the United States. For others, the diagnostic threshold remains almost unreachable. These realities can coexist because South Africa has never operated as one population with one healthcare system and one route to recognition.

The diagnosis arrived late, and for many, it has still not arrived.

But ADHD was here all along.