Dear Reader,
You have probably heard United States President Donald Trump alleging that the usage of paracetamol by pregnant mothers causes autism in their children.
The idea that such a common treatment may have caused parents to give birth to autistic children will doubtlessly cause alarm among parents and distress among the autistic community. As an autistic person, I am angry that my existence and that of my community must constantly be justified, debated and thought of as a problem that needs to be dealt with.
I am also disappointed in the continuing trend of accusing mothers of making their children autistic by bad parenting, a topic that has persisted in public discourse since the 1950s and ’60s.
However, this letter is not meant to be an outlet for my emotions.
My hope is that by reading this letter, you will have gained an understanding of the current findings around paracetamol and autism. I will not be discussing everything on the subject. To do so would require far too many pages.
Instead, I will look at the most relevant piece of literature investigating whether autism is linked to paracetamol use in pregnancy.
Paracetamol — also known as Panadol, Tylenol or Acetaminophen — is a common over-the-counter painkiller used to relieve pain and fever. It is a key part of treating minor scrapes, fevers, injuries and many other conditions.
The paper that President Trump may have based his statement on is a 2024 article in the Journal of the American Medical Association (JAMA) entitled "Acetaminophen Use During Pregnancy and Children’s Risk of Autism, ADHD, and Intellectual Disability".
The paper contains a lot of jargon that can be off-putting to people who don’t have the tools to interpret it. I do not speak for the authors, but I hope that I can share some insights from my university education to help make sense of the paper’s findings.
The authors studied just under 2.5 million children in that paper. About 185,000 of them (7.4%) were exposed to paracetamol in the womb; they called this group the "exposed" group. The rest were the "non-exposed" group.
Of the non-exposed group, 1.33% of children were autistic. Of the exposed group, 1.53% were autistic.
The percentage of autistic people is slightly higher in the group exposed to paracetamol. Similar trends were identified for ADHD and intellectual disability.
This may appear as proof that paracetamol causes autism, but it is important not to jump to conclusions here. The authors knew this and had a closer look at the groups.
A lot of these children are part of a set of siblings, and it is known that children with autistic siblings and/or parents are more likely to be autistic themselves. If paracetamol truly did cause autism, then having an autistic sibling or parent would not change the effect of paracetamol.
However, the effect did change. The difference in autism occurrence completely disappeared.
Looking at all the sibling sets containing both exposed and non-exposed children, the difference became so small that the authors could not tell which group had more autistic children.
This paper is but one example of an important lesson I was taught in high school: correlation does not equal causation.
The classic example of this is ice-cream and sunburns. When ice-cream sales increase, sunburns become more common.
But it is not ice-cream that causes sunburns, nor is it sunburns that cause one to eat more ice-cream. Both are just side effects of a third factor, the summer sun.
An outside factor making two linked things look like one causes another is very common, and not just in research. Scientists call these outside factors confounders.
In the ice-cream example, the confounder was the sun. Confounders are not always as obvious as the sun, so it is important to look for them whenever claims are made about something causing something else. Correlation does not equal causation.
With this lesson in mind, I present a hypothetical.
Imagine a pregnant person with autism. Already this mother is more likely to have an autistic child.
Autistic people tend to have different sensitivity to pain, and thus may be more likely to take paracetamol during pregnancy. This mother is more likely to have an autistic child than a non-autistic mother, even before taking paracetamol.
Now imagine this mother goes on to have a second child. They may have expectations of pain levels from their first pregnancy, and may take less paracetamol, or none.
But, as with their first child, their second child is still more likely to be autistic before considering paracetamol use.
In this situation, it is implied nowhere that this parent gave their child autism by taking paracetamol. There is a confounder here, having an autistic parent and/or sibling. Correlation does not equal causation.
I hope that by reading this letter, I have shared some insight into understanding what links and correlations between two things mean. If you take nothing else from this letter, remember this lesson: correlation does not equal causation.
Unfortunately, it is a lesson that appears lost among some of the most powerful people in the world today.
By jumping to conclusions and ignoring confounders, President Trump has portrayed the autistic community as problematic, unwanted and less human than others.
The blame is flicked on to their families who must live with the worry that they changed their child irreversibly.
Remember that correlation does not equal causation. Recall that lesson when reading news releases and articles. Look for confounders wherever two things are linked.
By doing so, you have an advantage not many others have.
• George Warren is a PhD candidate in the department of microbiology and immunology, University of Otago. He is also autistic and unashamed.
