Woman Nursing Her Son

Correlation, Causality, Breastfeeding, and Obesity

The distinction between correlation and causality is basic to any serious scientist. But in PLOS Genetics yesterday, scientists toss it out the window. Yanyan Wu et al found an association between breastfeeding and obesity. Then they lept to claim cause and effect. Right up front in their title, they make the bold claim. “Exclusive breastfeeding can attenuate body-mass-index increase among genetically susceptible children.”

Breastfeeding is good for many reasons. But obesity prevention is not one of them. Evidence of correlation is not evidence of causality. And other evidence tells us that breastfeeding is not very effective for preventing obesity.

A Study Over Time of Infants with Genetic Risk of Obesity

Wu et al studied a sample of 5,266 children from an ongoing study. They examined the interaction between genes, exclusive breastfeeding (EBF), and BMI from birth to 18. In this study, the children with more EBF had lower BMIs as they grew up. Even children with high genetic risks for obesity.

Based on this association, the authors make bold claims. “EBF to 5 months has substantial effect in decreasing BMI. Targeting children with high risk for obesity “should be a priority” for breastfeeding.

Correlation Is Not Evidence of Causality

Finding an association between the development of obesity and breastfeeding is not especially hard. Many factors that correlate with childhood obesity also correlate with less breast feeding. These include education, income, and employment factors. The reasons go quite deep into personal lives. Thus, it’s hard to know what the true causal pathways are. Multiple factors are working together in concert to cause obesity.

Correlation is not enough to prove causality. And careful reviews tell us three basic things about this association. First, any effect – if there is one – is small. About a ten percent reduction in risk. Next, publication bias is a significant issue. Studies that find more effects find their way into the literature more often. Finally, many confounding variables are impossible to exclude.

A Bias of Desirability

Breastfeeding is the good and right thing to do. So say our cultural norms and much research on child health. Apparently, that’s reason enough for WHO, CDC, and others. They will likely keep promoting the myth that breastfeeding is a very effective way to prevent childhood obesity.

But this strategy has a weak foundation in science. An association doesn’t prove an effect. Relying on ineffective tools for obesity prevention – even though they feel right – likely explains why we’ve made so little progress on this subject.

Four decades of weak prevention science has done nothing to reverse relentless growth in obesity rates.

Click here for the study by Hu et al.

Woman Nursing Her Son, painting by Salvador Dali / flickr

Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.


 

June 12, 2020

3 Responses to “Correlation, Causality, Breastfeeding, and Obesity”

  1. June 12, 2020 at 6:39 am, Al Lewis said:

    Here is a classic example — a diabtes company simply ignoring its own findings to shout about causation https://dismgmt.wordpress.com/2019/08/26/are-livongos-outcomes-real/

  2. June 12, 2020 at 8:48 am, David Stone said:

    “Four decades of weak prevention science has brought us relentless growth in obesity rates.”

    Speaking of fallacious cause/effect claims, no, “weak prevention science” has not been established as the cause of increasing obesity.

    No need to publish this response. Just some feedback.

    Cheers,

    DLS

  3. June 12, 2020 at 10:48 am, Ted said:

    Good feedback, David. Thanks!