Childhood Obesity: Where Are We Now?

nobroccoli.jpg
The truth about childhood obesity trends. Jose Luis Pelaez Inc/Blend Images/Getty Images

One recent paper, appearing in JAMA in February 2014, suggested a slight decline in obesity rates among children in one age group (ages 2-5), spawning headlines telling us childhood obesity rates had “plummeted.”  Another, published on-line in early April in JAMA Pediatrics reported that not only had childhood obesity rates failed to decline, but that severe obesity in children was rising disproportionately.

 Such wild gyrations in reporting over such a short period of time say something about the emotionally charged and tumultuous nature of our relationship with this topic. They also beg the question: where are we now, really?

The first of these articles was published with the seemingly bland title, “Prevalence of Childhood and Adult Obesity in the United States, 2011-2012.”  Packaged within a morass of data pertaining to all age groups was a downward deflection in the obesity rate in children ages 2 to 5.  Somehow, this yielded the following headline in the New York Times: “Obesity rate for young children plummets 43% in a decade.”  If that were accurate, it would clearly be cause for celebration.

The first reality check, however, is the study conclusion in the authors’ own words: “Overall, there have been no significant changes in obesity prevalence in youth or adults between 2003-2004 and 2011-2012.”

What the investigators actually reported, tracking data over a decade, is no overall change in the population rate of obesity.  In the survey sample of just over nine thousand people, representing the population of 300 million, obesity rates were stable in most age groups, including infants under the age of 2.

  Rates rose significantly in women over age 60, and fell significantly-although barely so (p=0.03)- in children between the ages of 2 and 5.

That, then, was the tale that shook the dog in late February, spawning a deluge of hyperbolic headlines: in 871 children between the ages of 2 and 5, obesity prevalence declined while it remained steady or rose in all other age groups, including children younger still.

The 43% decline reported was not an absolute percentage; but a relative percentage. To get a handle on that, imagine a population of exactly 100 children, ages 2 to 5; and imagine that 60 of them are obese.  Compare this group to another group the same age a decade later in which only 17 of the kids are obese.  The absolute difference in the obesity prevalence here would be 43%.

Now, imagine instead that in the original group a decade ago, about 15 of the kids were obese; and currently, in a comparable group, about 9 of the kids are obese.  The absolute decline in obesity prevalence is obviously only 6% (i.e., 15%-9%).

  But what about the RELATIVE decline in obesity?  That would be 40%.  The formula for it is [(15%-9%) / 15%] = 40%.  The 6% decline is 40% of the baseline number.  That’s the difference between absolute and relative percentages.  Absolute percentages are out of 100, which is what most of us expect.  Relative percentages are out of the starting number, whatever that happens to be.  On a relative scale, a drop from 2 people per hundred having X to 1 person per hundred having X is not 1%- it’s 50%.  That would still be true if the drop were from 2 per million to 1 per million. Relative percentages are often reported in the medical literature, and routinely reported in the popular press, for the most obvious of reasons: they tend to sound a whole lot more dramatic than the much smaller, absolute numbers. 

What were the actual, absolute numbers in this case?  Obesity rates in the 2 to 5 age group fell from 13.9% to 8.4%, an absolute difference of 5.5%.  I trust you can see why the relative change made the headlines.  An announcement that “overall obesity rates are unchanged over the past decade, with some increases seen, and a possibly encouraging decline of roughly 5.5% in children 2 to 5” is accurate, and about as dull as dishwater. 

The more recent study raises the concern that even this bit of good news may not be entirely reliable.  If rates of severe obesity are rising briskly in children as they have been reported to be doing in adults, then it may no longer help us gauge the scope of the epidemic to ask how many children are overweight or obese.  We may need to start asking: how overweight or obese are the affected children?  Trend data suggest: ever more severely so.

Of course, much attention has been directed at the problem of childhood obesity over the past decade, including the signature efforts of the First Lady. There are some genuine indications of progress.

But the question at the start was: where are we now?  The answer is: with the promise of true success still a long way off, and miles to go before we sleep!

Continue Reading