Does breastfeeding make babies smarter? I don’t know and neither do you

This is a science fiction starring Stuart Ritchie, a newsletter for i. If you’d like to receive this straight to your inbox every week, sign up here.

Here’s one thing for sure. If you take a large sample of children, sort them by who is breastfed and who is not, and then have them take an IQ test, you will see that breastfed children score higher.

This is the conclusion of almost all research in this area, and based on it, many people (and some health authorities, including the World Health Organization) have added “makes babies smarter” to the ever-growing list of supposed benefits of breastfeeding.

case closed, right?

Actually the case No Closed. There are several serious problems with this type of study. First, at least in the West, wealthier mothers are more likely to breastfeed. This may be due to additional support, opportunities or training; this may simply be because breastfeeding is more of an expected social norm in the upper social classes.

critical confusion

This introduces some really important confusing for comparison. Because there is such a large “socioeconomic gap” with breastfeeding, we don’t know if breastfeeding was the cause of the difference in IQ or if it was socioeconomic status—life in a nicer, less polluted part of town, better food, more time. and opportunities. money helped children to pass the test better, and also forced the mother to breastfeed more often.

And there’s an even bigger problem that affects almost all research into how parenting affects a child’s development: genetics. As I said earlier in this newsletter, parents (at least by birth) are not just parents: they also pass on their genes. And since genes can influence your behavior to a greater or lesser extent, it could easily be that any behavioral similarities you see between children and their parents are due to heredity and nothing else.

Why do educated parents have children who do well in school? Of course, to some extent, we are talking about opportunities and the environment that parents create. But a big part of the reason is that they passed on their genes: perhaps genes related to being smarter, being motivated, being able to control bad behavior, and so on.

The same reasoning applies to breastfeeding. Smarter parents will always have smarter children: and if smarter parents breastfeed more often (perhaps because they belong to the appropriate socioeconomic category where breastfeeding is “something”), it will be – wrong – to look like breastfeeding is to blame for the fact that children are smart. An association, to use the above term, would become embarrassed.

control variables

Scientists conducting observational studies of breastfeeding may try to avoid confusion by including inspect Variables in their study: These may include parental income as a control for socioeconomic aspects. You can add other aspects of health to try and keep them all the same. And then we may wonder if, apart from these destructive factors, there is a connection between breastfeeding and intelligence.

You can think of it as a circle getting smaller and smaller. The first circle is the main connection between calmness and intelligence, which seems to be quite large. But as you add controls, the circle gets smaller and smaller. It does not always come down to zero, but it is not always the case that the study is analyzed. All mixed variables we want.

Thus, “immeasurable confusion” occurs. And then there is the “underestimated” confusion. How Away If you don’t accurately measure your control variables, you’ll miss out on a lot. For example, income may not cover everything important about socioeconomic status, and the circle narrows even more if you can somehow accurately measure socioeconomic status.

All of which brings us to a new study published in the journal PLOS medicine last week.

IQ values ​​were not taken into account Necessarybut with learning difficulties. If breastfeeding affects the brain enough to boost your IQ score, it can also lower your chances of being diagnosed with dyslexia, dyspraxia, or even socio-emotional problems.

Researchers have accessed data from tens of thousands of children in Scotland and linked their medical records to their educational records. From this, they could see who was breastfeeding, who was being bottle fed, and whether they were diagnosed with special educational needs — different types of learning difficulties and disabilities — between the ages of five and nine.

The key finding is that children who were exclusively breastfed for the first few months of life were less likely to be diagnosed with a range of special educational needs: learning, communication and behavioral difficulties (although they are no less likely to have autism or other medical conditions). mental illness). The results for “mixed feeding,” where the baby was bottle fed with some breast milk, were somewhere between breastfeeding and bottle feeding.

shrinking circle

But the narrowing circle was clearly visible here. At first the results were strong, it seemed that breastfeeding really had a significant impact. Adding controls related to the child (e.g. their gender, their specific age when measured) made it slightly smaller, and then adding controls related to the parents (e.g. how old was the mother, did they smoke, how did they give birth) . it shrinks even more.

There was no control over the parents’ education or their abilities – or even over whether they were themselves learning disabilities. Also, as with almost all of these studies, there was a lack of underlying confounding of genetics.

You can’t blame the researchers – it was an impressive attempt to match datasets that must have been painstakingly obtained by government bureaucrats (just imagine how many forms they would have to fill out!). But when you use this type of data in your research, you miss out on many variables, including variables that are practically essential to interpreting the research.

And just take a step back – the socio-economic divide here was very clear. 36% of formula-fed children belonged to the “most needy” group in terms of socioeconomic status; only 14 percent of babies are breastfed. Ten percent of formula-fed children were categorized as “least needy”; 25 percent of those who are breastfed. 34 percent of formula-fed parents were married; 62% of parents breastfed. In other words, there were really big differences between these groups—differences that are not easy to control statistically.

So, if you add the immeasurable hindrances I just mentioned, AND have perfect control variables, AND somehow genetically controlled (easier said than done, but only hypothetically), will the circle narrow? I’m absolutely sure it will. Will it drop to zero? We have no idea.

Never fully answer a question

And that’s disappointing in these kinds of studies. It may be time to stop this kind of observational breastfeeding research, as it never happened. quicker Answer the question we need to answer. Instead, can we pool resources that we would otherwise spend on additional observational studies and run a randomized trial that eliminates all these “confusion” issues?

It’s not impossible, although to my knowledge it’s only happened once. In Belarus in the mid to late 1990s, a study called PROBIT (Advancing Research on Intervention in Breastfeeding) provided additional breastfeeding support to mothers in randomized hospitals by successfully increasing the number of breastfeeding sessions.

Initial results showed that children in the randomized breastfeeding intervention performed better on IQ tests. One point for breastfeeding? Not entirely true—a later observation found that by age 16, breastfed children performed just as well on cognitive tests as those who didn’t. Of course, this was just a test. Belarus in the 90s was a very different place than Britain in the 2020s – it’s easy to argue that it’s time to hold a new trial as soon as possible.

We often hear that a non-randomized observational study is just the beginning: once we get a hint that an effect might be real, we can follow it up in randomized trials to really identify it and avoid confusion. But, with a few exceptions, research on breastfeeding and its impact on a child’s development seems to have stalled at the most basic level. If we really want to know if breastfeeding makes babies smarter—and, by the way, if it makes them healthier in many other ways, as is often claimed, including from convoluted correlational studies—then it’s time to move on to a better class. evidence. .

Other things I’ve written lately

For me, it was fertility week. First, I wrote about the “timeless chemicals” and how the latest scary story about their effect on female fertility is based on some pretty dubious research.

Then I wrote an article about zebrafish research, this time arguing that starvation can negatively impact fertility. Bottom line: you can do research solely on fish, and the whole story doesn’t always have to go back to humans!

Science Link of the Week

Luckily, living in the UK means we don’t have to worry too much about snake bites. But as this insightful article shows, thinking about snakebite can help us understand one of the hardest problems in global development research: getting the data right.

This is a science fiction starring Stuart Ritchie, a newsletter for i. If you’d like to receive this straight to your inbox every week, sign up here.