Experts raised concerns about the CDC’s plan to use the Vaccine Safety Datalink (VSD) to look for an association between vaccines and autism, despite decades of data demonstrating the two are not linked.
Details of the proposed CDC study have not been formally announced, but the Washington Post said the agency plans to analyze VSD information. The VSD is a collaborative project between the CDC and several healthcare organizations across the country that relies on electronic health record data to monitor vaccine safety.
VSD records have been used to identify real-time safety signals; they identified higher rates of myocarditis or pericarditis with the COVID-19 vaccines, for example. Whether they can identify longer-term, nuanced outcomes like an autism diagnosis is less clear.
“Could such a study conceivably be done through the VSD? Yes. The question is why you would want to do that,” said Gregory Poland, MD, president of the Atria Health and Research Institute in New York City and emeritus editor-in-chief of the journal Vaccine.
“The last time I counted, I think there were 28 studies about vaccines and autism that had been done across two or three decades, across multiple countries. The two hallmarks of science are generalizability — so, in different populations, different countries, different time frames — and repeatability,” Poland told MedPage Today. “Every single one of those studies has shown no association. Not all 28 studies have been of equally high value, but some have had superb study design.”
One of the largest was a population-based study of more than 537,000 children in Denmark published in the New England Journal of Medicine, which showed the risk of autism diagnosis was similar between those who received the measles, mumps, and rubella (MMR) vaccine and those who did not.
Another study in the mix was a case-control analysis from CDC researchers using VSD data that showed thimerosal, a mercury-containing preservative in vaccines and immunoglobulin preparations, did not raise the risk of autism.
Studies examining autism and vaccines are very hard to conduct, observed Paul Offit, MD, of Children’s Hospital of Philadelphia.
“The reason is that when people don’t get a vaccine — for example, the MMR vaccine — it’s usually because of an issue of access or an issue of choice; in other words, they chose not to do that,” Offit told MedPage Today.
Those people are different from people who get vaccines, Offit said. “When you do these studies, you have to control for confounding variables like healthcare-seeking behavior, medical background, or socioeconomic background,” he said.
“You have to control for those three things, and then see whether or not there’s a difference,” Offit continued. “Because if someone didn’t get a vaccine because they were scared of getting autism or because they didn’t trust the medical community, they may be less likely to go to the doctor and less likely to be diagnosed with autism. It could look like the unvaccinated are less likely to develop autism when in fact, that’s not true.”
A rigorous study also needs to “develop a case definition for the outcome of interest — in this case, autism,” said William Schaffner, MD, of Vanderbilt University Medical Center in Nashville.
“Autism is not a diagnosis associated with a specific laboratory test. It’s a syndrome,” Schaffner told MedPage Today. “How the diagnosis of autism is defined and corroborated in a study like this is absolutely critical.”
An autism diagnosis as it appears in the VSD database may not be sufficient, Poland said. “There are lots of different autism phenotypes or subtypes,” he acknowledged.
“How are you going to define those? It would be a massive amount of work because those aren’t always coded properly in medical records,” Poland continued. “You’re going to have to get permission to access those records. It’s harder — particularly in children, who are considered a vulnerable population from a research point of view — to access mental health psychiatry records.”
Patients in the VSD are enrolled in medical care systems and ought to have extremely high vaccination compliance, Schaffner added. “It will be very challenging to find a sufficiently sizable vaccine-negative population,” he said. “And that population is likely to differ from the majority in a number of ways.”
Still, Schaffner sees possible value in yet another vaccine study. “I’m of two minds: the scientific half of my brain says this is a settled question. We don’t need to look at it again,” he stated.
“However, the public health side of my brain acknowledges the reality that there are many people who still raise questions about the relationship of vaccines to autism. If a rigorous study were done under the aegis of the current HHS secretary and it came out with the same answer, perhaps that would put this issue largely to rest,” he said.
This would have to be a very nuanced, very labor-intensive, very costly study, Poland pointed out. “That’s why I asked, why would you want to do it? I don’t think people who still believe that there’s an association will be convinced by a study like this,” he said.
“When you want to do a large, expensive, labor-intensive study, it’s because you have preliminary data that suggest this is a wise investment of material, time, and financial resources,” Poland added. “Where is that preliminary data? Or is this an ideologic belief that we’re just going to keep doing studies until we get the answer we want? That’s not science. That’s ideology.”
Please enable JavaScript to view the