Significant flaws in Henry Ford study used to claim childhood vaccines cause harm

During a U.S. Senate hearing on 9 September 2025 chaired by Senator Ron Johnson, lawyer Aaron Siri presented an unpublished research paper titled “Impact of Childhood Vaccination on Short and Long-Term Chronic Health Outcomes in Children: A Birth Cohort Study”. The paper, authored by researchers at Henry Ford Health System, allegedly showed “vaccinated children had a statistically significant increased rate of various serious chronic diseases”, according to Siri’s written testimony.

Siri added that the study “finally provided hard figures on the rates at which childhood vaccines may be causing these harms”.

For context, Siri has worked for the anti-vaccine group Informed Consent Action Network (ICAN), led by television and film producer Del Bigtree. ICAN has a record of spreading vaccine misinformation, including the false claim that vaccines cause autism. Siri also works for the current U.S. health secretary, Robert F. Kennedy Jr., who has spread misinformation about vaccine safety on numerous occasions.

Following the hearing, the paper became widely shared on social media as evidence that childhood vaccines are responsible for a wide range of chronic illnesses in children. Notably, U.S. Congresswoman Marjorie Taylor Greene shared it on X, claiming it “revealed damning CRIMINAL results”. And a post on X by Nicolas Hulscher, who co-authored flawed studies claiming COVID-19 vaccines caused widespread injury and death, received more than five million views.

It’s not the first time that vaccine opponents have claimed a “vaxxed vs. unvaxxed” study showed childhood vaccines cause harm. We addressed similar claims in the past, showing how such claims relied on published studies using dubious methods, some of which have been retracted, as Science Feedback reported in an earlier review. In addition, these claims typically ignore better-performed studies that contradict them.

The recent Henry Ford study is slated to be the subject of a film produced by ICAN titled “An Inconvenient Study”. The film’s website states the study was done after Bigtree “issued a challenge to the head of infectious disease at one of the most prestigious medical institutions [to] conduct the most thorough vaxxed vs. unvaxxed study that has ever been done”.

It’s not the first such undertaking by Bigtree, who also produced the film “Vaxxed: From Cover-up to Catastrophe”, which spread vaccine misinformation.

In this review, we’ll discuss the Henry Ford study and explain why it’s not credible evidence of vaccine harms.

What did the henry ford study do and what did it find?

The study aimed to compare the frequency of certain chronic health conditions in children who received at least one vaccine and children who received no vaccines. To do this, it examined the medical records of more than 18,000 children born between 2000 and 2016, among whom roughly 1,900 were unvaccinated. It excluded children with chromosomal abnormalities and congenital illnesses to ensure that the population studied was “generally healthy”.

Chronic health conditions that the researchers looked at included diabetes, asthma, allergy, cancer, and neurodevelopmental disorders like autism. The frequency of each condition in vaccinated and unvaccinated children was then compared.

Using this method, the researchers found that certain chronic health conditions​​—specifically asthma, autoimmune disease, eczema, and neurodevelopmental disorders—were more often recorded in the vaccinated group compared to the unvaccinated group. Interestingly, the researchers didn’t find an association between vaccination and autism.

The researchers concluded that “the overall probability of being free of a chronic health condition at 10 years of follow-up was 43% in the group exposed to vaccination and 83% in the unexposed group”. They added that “in certain susceptible children, exposure to vaccination may increase the likelihood of developing a chronic health condition”, but acknowledged that these “preliminary findings cannot prove causality”.

Flaws in the study

Difference in healthcare utilization between vaccinated and unvaccinated children

In a thread on X, Jeffrey Morris, a biostatistician and professor at the University of Pennsylvania, expressed several concerns about the study, with a key issue being the degree of ascertainment bias.

In the case of this study, this bias arises because healthcare utilization wasn’t the same in unvaccinated and vaccinated groups. The researchers noted that vaccinated children had an average of seven annual clinic visits, regardless of whether they had a chronic health condition. This is in contrast to two annual clinic visits on average in unvaccinated children.

This matters because several chronic health conditions that were counted in the study, such as neurodevelopmental disorders, require multiple clinic visits to diagnose correctly. For example, attention deficit hyperactivity disorder (ADHD) commonly requires three to four clinic visits to diagnose, wrote Morris.

Therefore, the lower incidence of chronic health conditions seen in the unvaccinated group could be explained by this group being less likely to see the doctor—and less likely to have their medical conditions diagnosed—rather than because this group was healthier.

However, the researchers didn’t consider the gap in healthcare utilization to be a source of bias that was concerning, stating that unvaccinated children who did have a chronic health condition made almost five annual clinic visits. The bottom line of their argument is that there was no meaningful difference in healthcare utilization that affected their results. This defense was echoed by Siri during the Senate hearing, who claimed this showed “when [unvaccinated children are] sick, they go” to the doctor.

But the study’s other findings suggest otherwise.

Jake Scott, an infectious diseases specialist at Stanford Medicine who was a witness at the Senate hearing, noted that the study recorded zero cases of ADHD among the roughly 1,900 unvaccinated children.

“How is that possible with a national prevalence at 11%? That’s highly unlikely, unless conditions went undiagnosed,” he said.

Another hint that a difference in healthcare utilization influenced the study’s results, Scott said, was the finding that ear infections were six to eight times greater in vaccinated children, even though there isn’t evidence showing vaccines increase ear infections.

These two observations reinforce the concern that the higher incidence of chronic medical conditions in the vaccinated group isn’t because vaccinated children are less healthy. On the contrary, it is because medical conditions are more likely to be caught in vaccinated children, while they are more likely to be missed in unvaccinated children.

Comparing apples and oranges

Another issue was the significant differences in baseline characteristics between the vaccinated and unvaccinated children.

This is evident from the study’s Table 1, which shows that the vaccinated group included a significantly larger proportion of children with low birth weight, prematurity, birth trauma, and respiratory distress at birth.

These factors independently raise the risk of future medical conditions, so they need to be  accounted for when looking at the rates of chronic medical conditions. Otherwise, they create what researchers call a confounding effect, which can either create the illusion of an association where none exists, or conversely hide a true association.

In this case, the higher rates of chronic medical conditions seen in vaccinated children could be due in part to the greater frequency of birth-related risk factors independent of vaccination.

Longer follow-up in vaccinated group makes diagnoses more likely

Finally, Morris noted that the vaccinated group was monitored for a longer time than the unvaccinated group. This created “more opportunities for diagnosis—especially for conditions typically identified after school entry (≈5–10 years)”.

The study noted that the median follow-up time for vaccinated children was 2.7 years. Contrast this with 1.3 years for the unvaccinated group. This means that vaccinated children are more likely to still be followed up when they reach the age window for diagnosis:

“Differential follow-up/age distributions make vaccinated children more likely to be diagnosed within the study window, while a larger share of unvaccinated children who will eventually be diagnosed are not yet captured. This additional factor contributes to the ascertainment bias problem.”

The researchers tried to address this inequality in follow-up duration by performing several analyses only on groups with specific follow-up durations: one year or more, three years or more, and five years or more of follow-up. Because these analyses produced the same or even higher rates of chronic health conditions in the vaccinated group, the researchers concluded that the unequal follow-up duration wasn’t a significant concern.

But Morris disagreed, pointing out that their approach “actually amplifies the imbalance”, since in the groups with different follow-up durations, vaccinated children still comprised the majority of the group. Notably, there were about 20 times more vaccinated children than unvaccinated children in the five-year follow-up group.

This makes it far more likely for medical conditions to be diagnosed in the vaccinated group. This issue would likely be compounded by the difference in healthcare utilization between the two groups.

With such an imbalance, “it is unsurprising that conditions with median diagnosis >5 y appear far more often in the vaccinated—even if true risks were equal”, wrote Morris.

Sound study design important to avoid biases and erroneous conclusions

Morris called the Henry Ford dataset “valuable”, but that “major baseline imbalances, large disparities in health-care utilization, and unequal follow-up/age distributions” were “fatal flaws” in the study.

He suggested certain improvements that the researchers could make to their study design, for example using matched cohorts of vaccinated and unvaccinated children based on birth year and important confounding factors, and ensuring that follow-up of both groups was similar.

There have been several large-scale studies comparing vaccinated and unvaccinated children that were designed to account for the issues seen in the Henry Ford study. References to those studies can be found in this earlier review. Those studies didn’t find vaccinated children to be more likely to develop medical conditions compared to unvaccinated children. A more recent study published by researchers in Denmark examined the relationship between chronic diseases and exposure to aluminum from childhood vaccines^[1]. It found no association between the two.

Conclusion

While it uses a large dataset, the Henry Ford study contains several important flaws that limit its validity. Notably, a difference in healthcare utilization between vaccinated and unvaccinated groups, which makes it so that conditions are more likely to be detected in vaccinated children, while the same conditions are likely to be missed in unvaccinated children. Contrary to the claims of vaccine opponents, the study doesn’t provide credible evidence of a causal relationship between childhood vaccinations and certain chronic medical conditions. Other published studies using sound methodology haven’t found that childhood vaccination is linked to poorer health.

References:

• 1 – Andersson et al. (2025) Aluminum-Adsorbed Vaccines and Chronic Diseases in Childhood: A Nationwide Cohort Study. Annals of Internal Medicine.