The historical track record of major U.S. Presidential pronouncements on health is as abysmal as one might expect given their prior occupations (27 lawyers, zero physicians or medical researchers). In 1971, Richard Nixon famously declared a “war on cancer”; 54 years later, not only is cancer still very much with us, but the Trump administration is now waving the white flag and pulling back on research investments. In 2000, Bill Clinton announced the completion of the Human Genome Project, forecasting that sequencing the human genome would lead to all kinds of breakthroughs in preventing and treating genetic diseases. A quarter-century later, notable progress has been made on many conditions, but genomic insights have yet to transform medicine as initially promised. In 2016, Barack Obama launched the Cancer Moonshot as part of the 21st Century Cures Act, and after his Vice President, Joe Biden, ascended to the Presidency in 2021, he re-committed the U.S. to accelerating progress toward cancer cures. The jury is still out on this one, but the current Health and Human Services Secretary’s blanket opposition to mRNA vaccines (including those for cancer) has stacked the deck against it.
So when President Donald Trump made a “major announcement” on autism at the White House earlier this week, the former real estate developer and reality TV show host was following in the ignominious footsteps of his predecessors who, to put it bluntly, should have stayed in their lane. Taking his cue from Robert F. Kennedy, Jr., who for years led a nonprofit organization that opposes routine childhood vaccinations, Trump blamed infant shots and acetaminophen (Tylenol) in pregnancy for the increased prevalence of autism diagnoses in the U.S. and around the world. Dismissing the pain and discomfort that often accompanies being pregnant, he implored women to “tough it out” rather than take a pain reliever that is considered by every major medical organization to be safe in pregnancy. (And yes, there absolutely are downsides to avoiding Tylenol, given the clearly established harms of alternatives for pain and fever.)
I wanted to simply dismiss what Trump said, but after two days of seeing patients and responses on my social media, it’s clear that many people are taking him and RFK Jr. quite seriously. To explain why they (and perhaps you) shouldn’t be concerned about your child’s vaccines or taking Tylenol when needed, join the first-year medical student class in evidence-based medicine that I taught at Georgetown for many years. The question we would examine in our first meeting was: does radiation from cell phone use cause brain tumors? (TL;DR – although there isn’t any way to prove without a doubt that cell phones don’t cause cancer, most of the evidence suggests that the answer is no.)
But think like a researcher for a moment. How would you study this question? You could do what’s called a case-control study and compare the cell phone use of persons with brain tumors to persons without them. (This type of study would have been easier to do in the days before cell phones were ubiquitous; I didn’t purchase my first cell phone until 2002.) In addition to asking about ownership, you could ask people to recall how long they spent talking on the phone on average, and which side of their head they pressed it to their ear. (Again, easier to do in the days before hands-free earbuds.) But there’s a big problem with this type of study: recall bias. Human beings have a tendency to come up with plausible explanations when bad things happen, and a brain tumor certainly qualifies as a bad thing. Just as a recently administered MMR vaccine is a convenient explanation for the subsequent diagnosis of autism, a cell phone is a convenient explanation for cancer.
So let’s say you perform a better type of study, a cohort study where you compare two groups by a more objective measure of cell phone exposure: cell phone subscriptions and number of minutes used each month. Obviously this design poses complications as family and business cell phone plans may not reliably identify who was actually using the phone, even setting aside issues of privacy and phone companies allowing researchers to access granular data. But if you find an association between increased cell phone use and risk for brain tumors, you can feel more confident that it’s a true correlation. Similarly, some studies have shown associations between Tylenol use and neurodevelopmental disorders, though others have not.
But it’s a huge leap from showing correlation to proving causation. The latter requires systematically eliminating confounding factors that may affect both the exposure and the outcome. For example, perhaps the true correlation is that women carrying pregnancies with children who are genetically predisposed to develop autism are more likely to experience fever-causing infections or musculoskeletal pain. Naturally, they will be more likely to seek fever or pain relief from acetaminophen, making it appear – incorrectly – that exposure to acetaminophen caused the outcome. We do not know if this is happening, as the FDA made clear in its news release.
In the meantime, I am not going to change how I counsel patients about vaccines in childhood or Tylenol in pregnancy. These medications have clear benefits (preventing serious diseases and relieving fever and pain in pregnancy), and the burden of proof rests on proponents of hypothetical negative effects, including Trump and RFK Jr. Finally, I think it’s unconscionable for the President to put a “I took Tylenol in pregnancy so maybe I gave my baby autism” guilt trip on mothers without ironclad proof of either correlation or causation.
