How do we know that breastfeeding DOESN’T have the benefits claimed for it?

Group of People Message Talking Communication BENEFITS Concept

It’s very challenging to explain science to those with low scientific literacy.

Consider my oft repeated claim that the predicted benefits of breastfeeding (based on mathematical models) have never been shown in real populations. Lactation professionals do not deny this. Indeed, in a memorable exchange I had with Dr. Melissa Bartick, she actually confirmed it.

The same way we know that routine episiotomy doesn’t have the benefits claimed for it.

In the comment section of a piece on the Academy of Breastfeeding Medicine blog, I asked Dr. Bartick directly:

Where is the evidence that term babies lives has been saved? Where is the evidence that the diseases you insist are decreased by breastfeeding are actually decreasing as a result of breastfeeding? Where are the billions of healthcare dollars you claimed would be saved as the breastfeeding rates rose?

Her response:

… To my knowledge, no one has actually dug it up yet.

Even Dr. Bartick acknowledges that the evidence simply doesn’t exist.

But low science literacy lactivists like Lisa Bridger of the FABIE Facebook group (Fed Ain’t Best, It’s Expected) are struggling mightily to understand.

Layne Sullivan, a member of Bridger’s group, asks:

How could it be studied without mathematically modeling?

And:

[H]ow could you possibly remove every confounder that predicts health? There are hundreds. It is not possible.

Bridger, demonstrating a different aspect of low scientific literacy, writes:

Sure more women in the US are initiating breastfeeding but less than 25% are actually meeting the world health organisation recommendations for exclusive breastfeeding for 6 months. So how can she demand results, when the foundation hasn’t been achieved??? Not a single country in the world is achieving the WHO recommendations, yet she sees her perceived lack of data as a slam potato dunk

How can I explain science to those whose understanding of science (and math and statistics) is so low that they make such nonsensical claims? I’m hoping I can explain it by analogizing to something they already believe:

How do we know that breastfeeding doesn’t have the benefits claimed for it? The same way we know that routine episiotomy doesn’t have the benefits claimed for it.

Serious vaginal tears had posed significant health hazards for women since time out of mind. They could lead to permanent urinary incontinence, permanent dribbling of stool from the vagina and permanent sexual dysfunction.

Tears occur when the diameter of the baby’s head exceeds the capacity of the vaginal opening to stretch to accommodate it. Doctors reasoned (wrongly as it turned out) that by cutting an episiotomy to accommodate the baby’s head they could avoid jagged tears and injury to the nearby bladder and rectum.

Why don’t doctors cut routine episiotomies any more? Canadian obstetrician, Michael Klein, decided to find out if the predicted benefits actually occurred in real populations. Despite the fact that everyone “knew” that episiotomies prevented severe vaginal tears, Dr. Klein showed that women who underwent episiotomies were MORE likely to experience a severe tear.

Dr. Klein did not “model” the impact of episiotomies, he looked at what actually happened when women were cut. He compared the predicted benefits of episiotomy to the actual benefits of episiotomy and found out that the predicted benefits did not exist.

Contrary to Bridger’s misunderstanding of research, Dr. Klein did not need to investigate what would happen if 100% of women had episiotomies. He didn’t have to reach any specific threshhold. He merely had to compare what the model predicted for ANY given episiotomy rate and the actual outcome at that episiotomy rate.

Layne Sullivan also misunderstands what it required for proof. Real world evidence is far more important than mathematical models.

Population based data shows that episiotomy not only doesn’t reduce the incidence of severe tears; it increases it. Real world breastfeeding data — as Dr. Bartick acknowledges — fails to show any reduction in term infant mortality, severe morbidity or healthcare costs. Dr. Bartick’s models are wrong.

How about confounding variables? They can never be eliminated entirely, but science does not require that they be entirely eliminated. Advanced statistical methods can correct for the most important confounding variables. If a benefit no longer exists after correcting for confounding variables, it wasn’t a real benefit in the first place.

The bottom line: we know that breastfeeding doesn’t have the benefits predicted for it the same way we know episiotomy doesn’t have the benefits predicted for it — by looking at population data and correcting for confounding variables.