Autism is everywhere – or so it seems. Reported diagnoses have risen sharply over the past few decades. (Photo: iStock) 
Autism is everywhere – or so it seems. Reported diagnoses have risen sharply over the past few decades – for instance, from about one in 150 children in the early 2000s to around one in 36 today in the US.
This apparent surge has seen the public scrambling to find a cause. But is autism really more common nowadays? If it is, it suggests that something in modern life is driving the higher prevalence. If not, why are claims about a rising prevalence even being made?
Scepticism about casual links between autism and vaccines has endured from the Wakefield days through Covid-19 to measles. And now Tylenol/acetaminophen (paracetamol in non-US settings) is deemed the latest concern.
How do we as the public make sense of positions held by apparent authorities, based on claims of scientific evidence? Let’s take a look under the hood of media claims and try to make some sense of it.
Searching for ‘simple causes’
The search for an “easy” or single cause of autism isn’t new. In 1998, the Wakefield paper claimed that 12 children developed autism or related disorders after receiving the MMR (measles, mumps and rubella) vaccine. Despite media reporting and public interest, the science behind the claims was flawed. This means that a group of children who had not received MMR were not studied to report on whether or not they had developed autism. This makes it impossible to assign the diagnosis of autism (or “case-ness”) for children who had received MMR.
Secondly, cases were cherry-picked, and the “participants” were not selected at random from the population – the gold standard of any experimental research. This selection bias results in a lack of certainty that the development of “autism” was not because of some other factors, or that children with some other known predisposing cause or even clear clinical feature were not included in the study to strengthen its claims.
Thirdly, the study simply described observations without showing a true causal link between MMR and autism. A cross-sectional study looks at only one point in time, and therefore cannot link a cause that happened earlier to an effect that happened later. We know that autism may only become evident over time. In other words, a child given MMR today does not develop autism tomorrow. This paper was eventually retracted because of serious methodological flaws.
Despite this retraction, the idea that vaccines might cause autism took hold, possibly because autism often becomes noticeable around the same age that children receive routine vaccinations, and parents naturally look for explanations when their children regress or show “divergence” from the expected normal. Over the years, a number of large, good quality, well-designed studies across multiple countries have found no link between vaccines and autism.
The reasons myths such as those involving Tylenol or vaccines persist is partly just human nature: when something as complex as autism seems to be increasing, people naturally want a simple explanation or something to blame. It’s easier to focus on a single drug, vaccine or exposure than to grapple with the uncertainty of a more nuanced reality.
The Tylenol claim: what does the evidence actually say?
President Trump recently made headlines with claims that taking Tylenol (acetaminophen) during pregnancy causes autism, urging expectant mothers to avoid it at all costs. Unsurprisingly, these statements sparked a wave of public concern about the safety of a common, widely used pain-relieving (analgesic) medication. However, when we look deeper into the acetaminophen research, past the fear-mongering and sensational headlines, we find a very different story. Research into a proposed acetaminophen/paracetamol-autism link spans decades, with studies that vary in quality and design. The most reliable, high quality, controlled studies have found no meaningful association between acetaminophen/paracetamol use during pregnancy and autism.
So why do some studies appear to show a connection? The answer lies in the complexity of research and the challenge of distinguishing correlation from causation. In ideal circumstances, scientists would conduct a randomised controlled trial, where, for example, one group of pregnant women takes Tylenol and another does not, and their children are followed closely for years, examining for the features of autism. But such studies are neither ethical nor practical (it is not possible to withhold paracetamol from pregnant women, in large enough numbers to elucidate outcomes). Instead, researchers rely on observational studies, tracking what happens naturally and comparing outcomes.
Observational studies, however, can be misleading, because of other factors influencing both the use of Tylenol and the risk of autism. For example, pregnant women may take acetaminophen to treat infections or fevers, which are conditions that in themselves are linked to developmental differences in the child. In research, this is considered a confounder, an external factor that distorts the relationship between other factors. We can compare this with how eating ice-cream appears to be linked to drowning if you look at data alone. The confounding factor is hot weather, which increases both swimming (and hence drowning) and ice-cream consumption. When you account for the weather factor by only considering data from hot days, the association completely goes away.
Even large, well-designed studies into the use of acetaminophen in pregnant women, show only a tiny, insignificant difference in autism risk. One Swedish study of nearly 2.5 million children found that 1.42% of children exposed to acetaminophen in the womb were later diagnosed with autism, compared with 1.33% of children not exposed, a difference so small that scientists consider it negligible (in other words, the finding could have occurred by chance). Importantly, this study was also able to account for genetic factors, something previous studies had not. Because siblings share about half of their genetic makeup, as well as similar upbringing and maternal health, any differences in autism outcomes are more likely to reflect the effect of the drug rather than other factors. When researchers compared siblings born to the same mother, where one pregnancy involved Tylenol use and the other did not, even this tiny difference disappeared entirely. Similar negative results have been found in a large study in Japan.
In short, when appropriately performed research is examined properly, the claim falls apart.
The harm in these claims
These kinds of widespread, media-driven claims can be extremely harmful. Pregnancy-safe medication options for relieving pain and fever are limited, and acetaminophen/paracetamol has the one of the longest and safest track records. It is the only over-the-counter medication approved to treat fever during pregnancy, which is an important consideration, as untreated high fevers in the mother can pose risks to the foetus. Alternatives such as ibuprofen and aspirin have well-documented adverse effects on foetal development, making acetaminophen/paracetamol often the safest choice. Advising pregnant women against taking Tylenol can fuel unnecessary fear at a vulnerable time. It also risks placing blame on mothers by implying that a medication choice could cause their child’s autism, leading to guilt, stress and stigma.
Beyond this, portraying autism as a preventable tragedy linked to a mother’s choices reduces a complex neurodevelopmental difference to a problem that should have been avoided. This framing fuels stigma and undermines autistic people’s right to acceptance. It also reinforces the idea that autism is something “gone wrong”, rather than a natural form of human neurodiversity. When autism is treated as a crisis caused by something external, attention shifts towards blame instead of understanding and support.
Misinformation of this kind also erodes trust in healthcare and public health and diverts research funding away from areas that could meaningfully improve autistic people’s quality of life. When leaders and media personalities amplify baseless claims, they give legitimacy to prejudice. Ultimately, the danger lies not in the medications themselves, but in the narratives that teach people to fear difference.
The real story behind the increase in autism diagnoses
So, if the rise in autism isn’t due to Tylenol, vaccines, diet or any other single factor, what is driving it? Research over the past two decades shows that the increase is largely because of changes in how autism is understood, identified and diagnosed, not because more people are suddenly developing autism.
Importantly, research looking at a specific subgroup of autistic individuals who require high levels of support – such as those with very limited verbal communication or co-occurring intellectual disability – shows that rates in this group have barely changed over the past decade. The growth in diagnoses are concentrated among those with more subtle autistic traits, that clinicians are now better able to identify. This is now referred to as “autism spectrum disorder”, or even more broadly “neuro-divergence”.
According to Johns Hopkins University epidemiologist Christine Ladd-Acosta, diagnoses in this population have increased by about 300% over the past two decades largely owing to broader diagnostic criteria, earlier and improved screening and increased public awareness. To be clear: finding and diagnosing more cases does not necessarily mean that there are more cases. It means that people who were not diagnosed before are now being detected.
The diagnostic criteria have expanded to capture individuals who would have once gone undiagnosed – girls, adults and people from more diverse racial and ethnic backgrounds. Additionally, public health initiatives have increased autism screening at children’s routine check-ups, particularly between 18 and 24 months, leading to earlier and more frequent identification. At the same time, parents, educators, health professionals and communities are now more knowledgeable about autism and its many presentations. As awareness and acceptance grow, more families and individuals feel comfortable seeking assessment.
Importantly, when prevalence estimates are population-weighted or adjusted for demographic shifts, such as changes in age structure, birth rates, and regional population growth, it becomes clear that the increase is not as large as it first appears. As Ladd-Acosta notes, this is not an “epidemic”, but a gradual, positive shift in understanding. The rise in autism reflects better recognition, not a sudden surge in autism itself.
So maybe autism is everywhere – or at least, awareness of it is. The headlines, the claims, the fear-driven narratives all make it seem like something new and alarming is happening. But as the evidence shows, the “rise” in autism diagnoses is not a sudden wave caused by a single culprit. Rather, it reflects a society learning to understand, recognise, and accept neurodiversity more accurately than before.
The challenge for the public isn’t to find a simple cause, but to navigate complex evidence, question myths, and embrace nuance.
For this to occur, we need true leadership with expertise in the relevant field to guide us through the nuance and complexities.
When we look past the sensationalism, we find that autism is not an epidemic to be feared, but a spectrum of human difference to be understood, respected, and treated with compassion. DM
Lucy Wagner is an MA Research Psychology candidate and clinical research intern at the UCT HIV Mental Health Research Unit. Alice Nuttall holds a BHSc in Biomedical Sciences and is a clinical research intern at the UCT HIV Mental Health Research Unit. Prof Stephan Rabie is a psychologist and chief research officer in the HIV Mental Health Research Unit. Prof Marc Blockman is a professor of Clinical Pharmacology at UCT and chair of the Pharmacovigilance Committee at SAHPRA. Prof John Joska is head of clinical services (psychiatry) at Groote Schuur Hospital and director of the HIV Mental Health Research Unit.