A wild vaccine chase

If people still have questions about measles vaccine safety, more studies of safety aren’t the answer

In 2004, I was lined up to give a vote of thanks to Andrew Wakefield. He was an alumnus of my school, and had returned to give a talk to sixth-form students. It had initially fallen to me, as the current head boy, to thank Wakefield, himself an ex-head boy, for coming to speak.

This was a time when in the public – and my own naive – consciousness, Wakefield was merely ‘controversial’ because of his statements about the measles, mumps, and rubella (MMR) vaccine and autism. Although he had come under growing criticism from other scientists, he wasn’t yet widely recognized as having made false claims that led to the retraction of his work and his being struck off the medical register (which would happen in 2010, following a General Medical Council investigation).

Thankfully, a late tweak to the schedule got me out of giving the thanks. But that evening has always stuck with me. The talk was delivered with the confident authority of a senior medic. It was data-heavy and earnest. Scientific papers in big journals were cited, establishment dogma criticised. Wakefield had found something that looked troubling, it seemed, and wanted to get to the bottom of it.

Except we now know this was far from the real story. Thanks to the efforts of investigative journalists like Brian Deer¹, the public has since heard much more what went on behind false claims about the MMR vaccine in the late 1990s and early 2000s: the flawed scientific results, the undisclosed financial conflicts, and the serious ethical concerns surrounding Wakefield’s research on children.

But it appears that many have not yet heard enough about these events, and the murky origins of claims about the MMR vaccine and autism. In the US, there is a renewed effort to undermine trust in vaccines, as long-term opponents of these valuable tools have found themselves in positions of power. Last week, it was even announced that the US Centers for Disease Control was planning a large new study looking at the possibility of a link between vaccines and autism.

Just one more study

It might be tempting to think that, in an era of misinformation and mistrust, further studies are justifiable or at least excusable, because they could add more weight to the already weighty evidence against any potential link between vaccines and autism. But this argument assumes – naively – that a bit more scientific evidence will finally tip the scales against whatever claim is currently being made.

Suppose, for example, someone were to (incorrectly) have a theory that a measles vaccine increases the risk of developing autism by more than 10% relative to the baseline risk (i.e. from other, non-vaccine causes). All we have to do, goes the logic, is design a study that can test this claim. If the resulting study shows the theory is false, we can disregard it and move on.

But there’s a problem with this logic: the study has already been done. Actually, several studies have been done. Previous analyses of medical records across multiple countries, which compared the vaccine history of individuals who have and haven’t received an autism spectrum diagnosis, found no evidence MMR is linked with an increased risk of autism – let alone an increase as large as 10%.

In other words, this particular theory – if someone were to propose it – has been falsified already. And even if we were in a situation where the theory hadn’t been falsified yet, a 10% increase over baseline risk wouldn’t be enough to explain the tripling in autism diagnoses over the past two decades, which vaccine opponents regularly point to.

So if even a weak claim about the relationship between vaccines and autism – which wouldn’t explain the observed rise in diagnoses – has been debunked, what would anyone, let alone the taxpayer, gain from doing more studies?

In general, there are plenty of bad faith reasons for casting doubt on vaccines, from financial conflicts to attempts to ingratiate with those in power. There are also more organic reasons for doubt, such as mistrust of institutions among certain parts of society. Not only is it implausible that one final study will get the evidence over the finish line and convince everyone, there is also the question of what that one study should look like.

If we want to test a broad scientific hypothesis, we first have to convert it into a specific form that can be tested against data. A claim that vaccines can cause autism can theoretically take a very large number of testable forms. Which country do we study? Which clinical definition of autism? Which specific vaccine? How big a potential risk are we trying to rule out? No matter what is studied, however detailed or high quality the research, and however high the level of confidence in the results, there is always scope to argue that another, slightly different version of the theory needs to be tested to be totally sure.

Some anti-vaccination groups have argued that the only way to really check whether vaccines don’t cause autism is to run a randomised study. This would involve withholding lifesaving vaccines from children – at a time when the country faces increasingly deadly measles outbreaks – and waiting years to follow these children up and see what happens in their development. Effectively, the only hypothesis testing method they seem content with is the one that is too unethical to be feasible.

Which brings us to the second problem. Even if the ‘we need more studies’ crowd were to agree on a specific hypothesis and methodology, it doesn’t necessarily mean they’d accept the results. Perhaps the resulting data is flawed, they might argue, or there is an undetected bias somewhere. I suspect the only new result that anti-vaccination politicians would ever accept is one that finds harm. Like a gambler hunched over a roulette table, they want to try again and again until they get their result purely by chance. After all, if we slice data in enough different ways and look at enough different groups, eventually we’ll hit upon a coincidence. A 1-in-100 fluke becomes a lot more likely if you try 100 things.

From more studies to more explanation

Some of the loudest anti-vaccination groups seem to have such strong prior beliefs against vaccines, and have careers so intertwined with vaccine misinformation, that I suspect no study will ever satisfy them. It is ironic that a US administration focused on ‘efficiency’ would fund additional studies into an already debunked claim, to satisfy an audience with little interest in the results. But for them, the scientific outcome is likely to be less important that the journey: the priority seems to be an era of manufactured doubt and the endless asking of questions that can never be answered.

Fortunately, the wider public isn’t necessarily as entrenched in these beliefs. Their prior assumptions aren’t always as rigid, their incentives not as misaligned. Given the scientific community has concluded that the measles vaccine does not cause autism, scientists should therefore work to explain why this conclusion has been reached. That means showing which questions the existing evidence has already answered and identifying the logical sleight-of-hand used by critics of existing work. And, crucially, reiterating where these false claims came from in the first place, and the motivations behind them.

When talking about potential risk – or a lack of one – we should also focus on clear, testable hypotheses rather than vague or shifting claims. If someone wrongly believes that measles vaccines are the main driver of the rise in autism diagnoses, this would have created an extremely strong risk signal, which would have shown up in one of the many large epidemiological studies already conducted.

One area of research that does need expanding is our understanding of why people are hesitant to take vaccines in the first place, and what can be done to alleviate the unevidenced concerns that anti-vaccination groups seek to exploit. There have been some initial efforts in this area. A couple of years ago, some of my colleagues analyzed strategies for countering COVID vaccine misinformation. They found that use of scare tactics and overstating certainty often backfired; it was more effective to clearly communicate the weight of the evidence and the scientific consensus. Approaches like using humor and preemptively warning people they might encounter misinformation also showed promising results. Unfortunately, follow-up research will be increasingly hard to come by: it was announced earlier this week that the National Institutes of Health is planning to cut the very work needed to help curb the deadly measles outbreaks now appearing in the US.

These are the kinds of topics – communication, evidence interpretation, and public trust – that scientists should be focusing on now, at a time of rising measles deaths and growing attempts to undermine the tools that could prevent them. What we shouldn’t do is indulge unnecessary calls for further studies into safety from groups who have no intention of accepting the results.

---

If you’re interested in scientific evidence, misinformation and the search for certainty, my new book Proof: The Uncertain Science of Certainty is out next week in the UK and available to pre-order now.

1. I’d recommend his book ‘The Doctor Who Fooled the World’.