Vaccines, autism and America: A stress test for public health standards
The CDC recently overhauled its “Autism and Vaccines” webpage, telling readers that the familiar claim that “vaccines do not cause autism” is not evidence based because studies have not ruled out the possible effect from infant vaccines. It goes on to note that “correlation does not prove causation, but it does merit further study.”
This establishes a standard for evidence. Before evaluating its applicability to vaccines, it’s worth asking: what happens when we apply this standard consistently?
Precedent. The CDC website frames its revision as required by the Data Quality Act, which mandates that federal agencies ensure “quality, objectivity, utility, and integrity” of public health information. The argument is that stating “vaccines do not cause autism” violated this standard because science cannot definitively rule out causation.
Applying the standard
In other words, if “cannot rule out” and “correlation merits further study” are the rules of the game, what else might qualify as a potential cause of autism, a condition whose diagnosis rate has dovetailed with rising awareness of the condition coupled with more systematic screening.
The first clear signs of autism usually emerge between 12 and 18 months of age, which is also when children receive several routine vaccines. That overlap in timing is easy to notice and misinterpret, but it is not evidence of a causal link.
Statistics formalizes the placement of the burden of proof with the null hypothesis. That is, the default assumption is that there is no association between vaccines and autism unless and until data show otherwise. Researchers then test whether the observed data are so inconsistent with that “no association” model that it should be rejected. In the vaccine–autism literature, large, well powered studies have repeatedly failed to reject the null, which is why bodies like the National Academy of Medicine and the American Academy of Pediatrics state that vaccines do not cause autism.
The National Academy of Medicine reviewed more than 1,000 studies in 2011–2012 and concluded the evidence does not show a link between vaccines and autism.
Regulatory asymmetry. Pharmaceutical development operates on the principle that efficacy claims require supporting evidence, not merely the absence of evidence against them. No sponsor has ever received FDA approval on the grounds that “studies have not ruled out the possibility that this compound works.” Yet the CDC now suggests that “studies have not ruled out” a causal relationship constitutes meaningful scientific uncertainty about established safety findings.
Attacks on the absence of evidence standard are not new. Philosopher Bertrand Russell once asked readers to imagine that someone claimed there was a small china teapot orbiting the sun somewhere between Earth and Mars. The teapot would be too small for any telescope to see, so no one could prove it was not there.
The analogy is now used in science and policy debates to push back on arguments that rely only on the absence of disproof. “You cannot prove it is false” is not the same as “we have evidence it is true.” In a regulatory context, that means a claim like “studies have not ruled out X” does not, on its own, carry the same weight as positive data showing X is likely.
If this standard migrates, it creates problems. What happens when a compound’s Phase IV surveillance finds a correlation with some adverse outcome? Correlation is easy to find in large datasets. Shark attacks are correlated with ice cream sales. And deaths by swimming pool drowning align with Nicholas Cage film releases. Such spurious correlations are among the many examples used to show the danger in equating correlation with causation. Under the CDC revised standard, the question becomes whether evidence rules out the claim, a bar that cannot be cleared.
Fragmented guidance. Some officials are taking matters into their own hands to create their own guidance. In that vein, several Northeastern states and New York City have announced the Northeast Public Health Collaborative. A West Coast Health Alliance: of California, Oregon and Washington has launched a parallel structure for vaccine and public health guidance, using its own scientific advisors to set recommendations for COVID-19, flu and RSV vaccines. California has passed legislation explicitly allowing the state to reject future CDC recommendations and substitute its own.
In addition, some 15 governors have launched a Governors Public Health Alliance with a similar mission, sharing surveillance data, developing joint guidance and working with outside experts on vaccine policy. Former CDC director Mandy Cohen now advises that group. In addition, Northeastern states have organized their own regional collaborative. In total, 16 states are now part of at least one health alliance.
New alliances, splintering rulebooks
While none call themselves a new CDC, in practice, they are building parallel structures for setting schedules and recommendations at the state level. If this trend continues, the U.S. could end up with multiple de facto public health organizations, where a child’s vaccine schedule and a clinician’s reference standard depend on which alliance their state belongs to.
This state by state confusion goes both ways. Florida has become the first state to remove all vaccine mandates for school. Louisiana, Idaho and Texas have all become targets of Make America Healthy Again (MAHA) advocates who are bolstered by the news from Florida and recent victories in state legislatures rolling back vaccine laws. Texas lawmakers have filed 20 MAHA-backed bills to roll back vaccine mandates.
For the pharma industry, the CDC’s last positioning poses questions for everything from guidance governs labeling to which schedule constitutes “standard of care” for trial protocols. There is also the question of what medical affairs teams, patients and doctors consider authoritative sources of information. If the phrase “cannot rule out” is enough to unseat settled safety conclusions, the result could be a degree of confusion about how to handle public health matters writ large.
Filed Under: Infectious Disease
