The most recent issue of The Lancet contains an interesting editorial discussing recent findings from the CDC that childhood obesity rates in low-income preschoolers are on the decline. You can check out the editorial here, or view the CDC report here.
The CDC report is an interesting read, not just for the results of their three-year study, but more for the comments they make at the end of the article, where they have a difficult time explaining why these numbers have decreased. They offer several suggestions, including the fact that the economic decline in the U.S. may have caused more families to become eligible for the low-income programs — which served as the source of the data — over the course of the study.
They also noted that the demographic mix of the participants shifted during the study, with the proportion of blacks growing while the proportion of whites and Hispanics shrunk. Whether this had any effect on the outcomes of the study is unknown.
Studying macrosociological issues such as “childhood obesity” must be an incredibly frustrating endeavour for researchers, as there are just too many variables at play for anyone to draw solid findings from the data they’ve analyzed. Things from intervention programs that vary from state to state, to global economic issues, to increased prevalence of breastfeeding of infants (and the programs that promote this behaviour) all play a role in affecting the outcome of studies like this.
But sometimes hand waving results like this are enough. (Yes, it even shocks me to say that.) I don’t think that big, BIG, issues like obesity respond to one or two variables; there’s no silver bullet to solve a problems on this scale. I think sometimes you need to take a broad look at the results of everything that is being done to combat this problem, and say “yes, we’re generally headed in the right direction” and leave it at that.
The problem with getting too specific and drilling too far into the results leads people to fall into survivorship bias, where they focus too much on the regions that posted decreases to discover what they’re doing right — instead of focusing on the regions that posted increases to find out what they’re doing wrong.
The assumption is that the regions posting decreases in childhood obesity risk are taking the proper steps to achieve these promising results — but the possibility exists that it’s more a case that they’re really seeing these results by not doing the same things as the regions who posted increases.
For instance, there may have been a general movement of some participants from an area with insufficient food systems, such as inner city areas plagued with corner stores and bodegas, to an area with better delivery and promotion of healthier foods, such as a group housing district 30 minutes away that has a proper grocery store within walking distance. Access to healthy food is something that is incredibly difficult to measure in a study such as this, but it’s one key factor in predicting — and controlling — obesity rates.
However, these sort of details fall far outside of the scope of a study like this, so it’s really best to look at this with a macro view and say “Okay, we have seen an overall decline based on everything we are currently doing for at-risk groups. Let’s either stay the course, or increase our efforts slightly, and then see what happens in another three years.” To take anything more specific from a broad study like this would be folly.