Infant feeding practices and obesity study: an example of how people react to scientific data
I read an interesting news report today by Liz Szabo in USA Today. It was about a recent article in Pediatrics, the official journal of the American Academy of Pediatrics. The article described an apparent association the authors found between early feeding of solid food (age younger than 4 months) and obesity at age 3 years. The association was only present in children who were formula-fed and not breast-fed. To restate: the authors found that, in formula-fed babies, early introduction of solid food was associated with being overweight at age 3. Early introduction of solid food in breast-fed babies had no effect.
What I found most interesting about Ms. Szabo’s article, although it was quite good, was the comment trail; it showed how most folks really don’t understand how to interpret medical studies. Some of the commentators denied the possibility of such an association because of their own experience with their children. Other commentators immediately leaped to the conclusion that the study authors were claiming feeding solids before the age of 4 months to your formula-fed infant would make them all fat. Still other commentators decried the “breast-feeding Nazis” who insist mothers who choose not to breast feed are negligent and try to make them feel guilty. If any of the commentators took the time to read the full study (it’s available free online here), they would have found that the authors make no such sweeping claims.
First of all, the study is observational. That means that the authors merely collected information about mothers and babies. There was no intervention, such as convincing mothers who chose not to breast-feed to nurse their babies, or vice-versa; the mothers chose, and the investigators merely watched what happened over the next 3 years. This approach leaves any study like this wide open to selection bias — the possibility that the 2 groups of mothers differed in some other way than feeding choice, possibly in a way that would influence that choice and future obesity in the children. The authors did examine a few possible confounders like this, education and family income, but there are many other possible ones.
What did the authors find, really? Well, they studied a total of 847 infants — 568 breast-fed, 279 formula-fed. Within those groups, 43 of the breast-fed babies started on some solids before 4 months (7.5%). In contrast, 91 of the formula-fed infants had started solids before 4 months (33%). So clearly mothers of formula-fed babies were more likely to start solids sooner, for whatever reason. That might matter for the ultimate results or it might not — there’s no way to tell.
At 3 years of age, 3 of the 43 breast-fed babies who had early solids were obese — 7%. In contrast, 23 of the 91 formula-fed babies were obese — 25%. To a statistician, that’s a significant number. It means there’s an association between 2 things. But it does not prove causation of anything. And note that 75% of the formula-fed babies were not obese at age 3, so personal anecdotes from commentators don’t mean much.
The bottom line to me is that this is an intriguing study, but it is far from the last word on it. And most of the irate commentators to the USA Today article complained about things that the authors of the article didn’t claim. So, whenever possible, it is good to read the original study before you decide anything — or get upset about it.