Childhood obesity is falling

Federal health authorities on Tuesday reported a stunning 43 percent drop in the obesity rate among 2- to 5-year-old children over the past decade, the first broad decline in an epidemic that often leads to lifelong struggles with weight and higher risks for cancer, heart disease and stroke.

The drop emerged from a major federal health survey that experts say is the gold standard for evidence on what Americans weigh. The trend came as a welcome surprise to researchers. New evidence has shown that obesity takes hold young: Children who are overweight or obese between age 3 and 5 are five times as likely to be overweight or obese as adults.

There is more here, via Charles C. Mann.


Does anyone else find this hard to believe? The reported size of the change seems incredible. Perhaps there is something going on with the way things are measured?

It's not really such a big drop if you think about it, changes in childhood obesity don't have to happen gradually - why would they? Suppose everyone were to switch to raising their kids on a healthy diet, then we would expect to see a big difference in obesity rates from one generaton to the next.

I have to agree with ed. Sudden changes can happen, but this is very hard to believe without some major identifiable explanatory factor. I don't see it. Maybe there has been some incremental change in attitudes, but having kids eat healthy was a big deal 10 years ago.

To me the immediate conclusion is that there is reason to suspect the data.

It was a fall from 14% to 8% in the obesity rate from 2 to 5 years old children. Take a seat and compare again when this children are be 12-15 years old.

I know it's ages 2-5 that's exactly why I am not surprised by the size of the drop. I'm not saying this means anything for the overal obesity rate, I'm only commenting on this number for this particular group.

Sure, compare again at 12-15, but that's still a massive change to me that deserves an attempt at explanation.

Even if they wind up regressing as they get older (as they likely will), it's still surprising that these gains existed at all and one has to wonder why.

Blame Michelle Obama - after all, Palin et al already have.

And grandmother Palin is someone with earned experience concerning the failure of educational goals and models.


Such sudden changes in anything correlated with human bodies are often result of criteria change.

And the administration can boast success fighting the obesity epidemics.

10 years ago, people thought that fruit juice was healthy instead of metabolically similar to sugar-water.

10 years ago, people thought soda was in moderation instead of slightly-better-than-drinking-anti-freeze.

10 years ago, humanity had yet to invent the catchy portmanteau "Diabesity."

It's difficult to be aware of how much awareness has risen.

Yeah, we're not talking about a scenario where obese people are losing weight (which does take time) we're just talking about one where toddlers aren't becoming overweight to begin with. In that cases these changes can occur quickly if parents are aware - it is not so difficult for a parent change what they decide to feed their 0-5 year old kid.

> it is not so difficult for a parent change what they decide to feed their 0-5 year old kid.

This depends greatly on the parent and the kid. It's almost like saying it's not so difficult for fat people to eat less and exercise.

So has pet obesity fallen also. It arose along side child and human obesity.

Also, take into account that it's not that we need parents to change their opinions on nutrition: When you wait 10 years, the parents themselves are different.

Look at the difference in attitudes towards gay marriage in the last decade. Are they really any different in scope than this? It's pretty much the same model of replacement, but far faster. When it comes to feeding children. 50 year olds can only vote when they are babysitting their grandchildren

Last time I stopped by a 7-11 you could still by 64 oz (1.9 liters) fizzy drinks. I love the Texan flavor of awareness =)

I was thinking about this as well. The medical science has been around for a long time that simple sugars in too high of quantity are terrible for you but for a variety of reasons the US medical community completely failed to communicate that information to the public and instead published incredibly stupid guidelines. Now information about this kind of thing is finally becoming widespread.

Couldn't we get an effect like that if we went from having many toddlers just over the cutoff to just under? They could be in the overweight category now.

Also, as someone else noted, this doesn't involve kids losing weight. It's an entirely different set of kids.

That's such a jump -- does it even correlate in any way with any previous jumps in the rise of rate?

Juice (and other hyper-fast sugars masquerading as health foods) could be a key, but I'm sorry, even though marketing pumped it up parents have always given their toddlers apple and grape juice. And many still do. It's not like we went from none to tons to none again.

These are 2 to 5 year olds. Kids that have spent half to a quarter of their lives on formula and/or baby food, or breast milk. It seems unreasonable to think you could get such a ginormous change from food changes only, since the food changes can't have been that ginormous. Soda, I'm sorry, is not a big factor -- I've never known moms who thought it was o.k. to let a four year old have a soda a day. I'm sure there's some population somewhere that lets their toddlers drink soda, but it can't be substantial. Different if we were talking the 5 to 15 year range.

Again, if this stat is accurate, it seems more likely that obesity is then linked to some environmental factor or a microbial factor. You can get big, swift changes when something is infection related, or if there's an underlying condition created by a substance (e.g. you could go from a high leukemia rate to a low one in one generation if you stopped feeding a carcinogen).

But it seems more likely to not be accurate. If the admin had not made childhood obesity its pet project, I might be less skeptical.

"I’ve never known moms who thought it was o.k. to let a four year old have a soda a day."

I've seen plenty that give their kids chocolate or strawberry milk which, honestly, isn't much better. But I haven't seen that subside much either.

"But it seems more likely to not be accurate. If the admin had not made childhood obesity its pet project, I might be less skeptical."

I'm skeptical of the effect-size at well but malice or fraud seems a pretty big leap. Falsifying data is usually pretty easy to spot and if you're going to lie it makes sense to make the lies moderate and believable rather than spectacular and noteworthy. Maybe some young data analyst put a period in the wrong place?

"Maybe some young data analyst put a period in the wrong place" ;)

I doubt fraud or malice. What I think is error, but when you err in a way that makes your bosses look good you are less likely to look twice at your conclusions. Although, frankly, it would be weird if a government agency that is putting out data to be publicly scrutinized wouldn't check strong outliers like that before putting out the press release, no matter what. I guess once the announcement has been made, it does your cause some good even if later you have to take it back? Who knows.

Maybe a big part of it is from longer-term nursing?

(I recall a big push on the "breast is best" front...)

@Fallibilist, it's possible, I guess the study doesn't control for that sort of thing, not that kind of study.

But my bad memory tells me the push way pre-dated the kids with these results.

There's also been a huge increase in food problems with kids, sensitivities or allergies or autoimmune, and maybe that has pushed some moms to avoid early grains more often, or to introduce fewer or no solids until later, which could influence, I guess.

It's also possible the formula companies changed their products, which I bet could have a huge effect on kids this age because bottle fed kids sure get a ton of the stuff by the time they're in this age group? But that wanders into my tin foil hat thinking, because I've often wondered about the obesity "epidemic" and how it seems to track with formula feeding.

"I’ve never known moms who thought it was o.k. to let a four year old have a soda a day."

I've lived in a small town in the American South for about the last 20 years, and in my experience, it's very common here...particularly correlated with parent's socioeconomic status.

My sister is a social worker in the Northeastern US, and regularly sees mothers put toddlers to bed with a bottle of "juice" (Kool-Aid) or even soda.

I've definitely seen the phenom of thinking anything with a blueberry on the label picture is healthy -- not just feeding kids juice, but feeding kids punch because it's good for them-- I've wrangled with teachers who thought I was nuts because I didn't consider fruit snacks healthy.

I am very surprised you've run into soda being used by toddlers, but not fruit punch, etc. I lived in the South and never saw this, not at that age, but that was a long time ago, I guess times have changed.

I was just about to comment that I find that hard to believe.

Obesity rates are falling

at the same time stocks are surging up, Facebook is up, BitCoin despite the problems at Mt. Gox is up, property values are up all over the Bay Area and NYC. The US is running victory laps.

Dude, high five me- way up here!

Ok, now full chest bump. U ready bro?

It's interesting that while most age groups are trending flat-ish, it's only the 2-5 age group that's trending down. I agree w/ Axa, let's see what happens to this cohort over time, but it's no doubt an interesting data point. Maybe in 10 yrs there will be better data as to why.

Package sizing has decreased substantially over the past decade (candy bars are smaller, packages of cookies weigh less, the "half gallon" ice cream box became 1.75 quarts, then down to 1.5 quarts, etc.) If this were to have an impact, it would likely show up first among the young, since as was pointed out above in the comments, losing weight is a much slower process than not gaining it.

Didn't the recession suppress the birthrate. Also, I think teen pregnancy has been trending down. Maybe the less affluent are putting off having kids like they are putting off getting a divorce because they can't afford it. This would change the composition of the population. Perhaps young kids are more likely to have parents that shop at Whole Foods today than 10 years ago. I have no idea if this is really that significant but the recession did change behavior.


Assuming it's not an error or reclassification, the next thing to look at is "who's having these kids?" That could easily swing in a bad recession. But the drop is huge, so I find it hard to believe that's the whole explanation.

There's another effect to consider: There must be significant negative evolutionary pressure on obesity, given what it does to fertility, let alone your prospects for marriage. But I don't see why that effect would show up suddenly. A decade probably counts as "suddenly," but maybe not.

Good point. It's possible that a poor economy has differing fertility affects in different socio-economic groups. Perhaps at lower income levels, the wealth effect predominates, so you can afford to have fewer kids than you otherwise might. At higher levels, perhaps the work-family/leisure substitution effect is more significant than the wealth effect, and women are more likely to take time out or scale back their careers to have kids, since career/business options are less attractive.

Heard in a market ... Dad: "should we get juice?" Mom (with kid in cart): "no, juice is terrible." Rock on, Mom. (US marketing and food programs both have encouraged sugary and low nutrition "juices.")

“In 2011-2012, 8.1% (95% CI, 5.8%-11.1%) of infants and toddlers had high weight for recumbent length, and 16.9% (95% CI, 14.9%-19.2%) of 2- to 19-year-olds and 34.9% (95% CI, 32.0%-37.9%) of adults (age-adjusted) aged 20 years or older were obese. Overall, there was no significant change from 2003-2004 through 2011-2012 in high weight for recumbent length among infants and toddlers, obesity in 2- to 19-year-olds, or obesity in adults. Tests for an interaction between survey period and age found an interaction in children (P = .03) and women (P = .02). There was a significant decrease in obesity among 2- to 5-year-old children (from 13.9% to 8.4%; P = .03) and a significant increase in obesity among women aged 60 years and older (from 31.5% to 38.1%; P = .006).”

So there is no change in the 2-19 year old group, no change in the infants and toddlers group, but a 43% change in the 2-5 year old group, P=.03. I think it is pretty clear that this is a statistical artifact.


Ding ding, we have a winner!

So if P=.03, the standard error of the difference is 2.9%, so a 95% confidence interval includes declines as small as 0.3 percentage points, which would represent only about a 2% decrease. And even that is a decline that was cherry picked.

Once again, statistical practice is terrible and reporters are credulous.

Yep. I also would guess you could play around with the age groupings and get similarly "interesting" results.

Before jumping for joy, take note of the fact that the CDC previously reported that: "Among preschool children aged 2–5 years, obesity increased from 5.0% to 12.1% between 1976–1980 and 2009–2010".
So CDC data indicate that the 43% decrease over a decade reported in this article consists primarily of a 33% decline between 2010 and 2012. At that pace, we would expect obesity of the very young to be a thing of the past quite soon.
CDC data show a long and gradual increase in obesity among the very young followed by two years of dramatic decline. Obviously, this sort of data is questionable on its face, particularly because the 2 and 3 year olds of the 2010 study are the 4 and 5 year olds of the most recent study.
If I were a scientist at the CDC, I'd look a lot closer at my data for the last two years.

Link broken. This one works:

"the first broad decline in an epidemic …": if it is an epidemic then a rapid decline is much what you'd expect, rather like the one for heart attacks. But it would still be useful to identify the micro-organisms responsible. That might explain the timing of the decline. On the other hand, so much about medical "facts" is just an artefact of changing definitions and mere fashion, who knows what to believe?

Two explanations:

1. BMI is a poor metric for toddlers. Their bodies are still developing, and not necessarily in the same consistent proportions as adults. A 2-5 year old found to be obese under BMI is not more likely to be obese as an adult. At that age adult obesity is predicted by how many obese parents the child has, and by the age 12 obesity is predicted by the child's own weight alone.

2. Depending on your nature versus nurture assumptions, it might be that obese parents are having fewer children,either because of increased societal hostility or effects from the economic downturn 2 to 5 years ago.

Re: 2, Google obesity cause infertility

That could only explain recent changes if adults are becoming more obese. The strategy would be working then?

It would lag adult obesity gains by perhaps 10 or 20 years, right?

You could look at causes of obesity, too -- PCOS, for example, causes insulin resistance which can lead to obesity, is often undiagnosed, and also causes infertility.

I agree. I don't think it would matter if obesity caused infertility or something else caused both obesity and infertility, for this purpose.

First, a change from 13.9% to 8.4% is a 39.6% drop. So you get 3.4% of the 43% just from rounding.

To get some perspective, consider that NHANES samples 7,000 people. Children 2-5 years old make up about 5% of the US population. Assuming the NHANES sample is representative, roughly 366 of its subjects are in this age group. Ten years ago, 51 of these subjects were obese, and now 31 of them are.

I wouldn't bet the farm on this.

I guess the gold standard isn't what it used to be.

This thread is a fitting epitaph for America. We have the absurdity of measuring obesity rates amongst toddlers followed by a serious debate over it.

I'll bite - why is measuring obesity rates among toddlers "absurd"?

Body fat is the direct result of a specific behavior. That behavior is over eating. It is either a learned behavior and therefore modifiable or it is genetic and therefore not modifiable. The latter is the most likely answer, but the CML denies genetics so we are left with modifiable behavior. Toddlers have no agency. They will eat whatever they can because they are growing like a weed and need the calories. The very concept of an obese toddler is ridiculous enough. Thinking you can "teach" a two year old good dietary habits is proof of insanity.

The imaginary obesity problem is just an excuse for the CML to nose around in the lives of others and impose their morality on the rest of society. If the Walmart crowd suddenly adopts dandelion salads and gets super thin, you can bet the beautiful people will embrace blubber as the new black.

What a load of horseshit

This is hilarious. Do you actually have evidence for your argument that diet and food preferences are genetically determined and unmodifiable?

Uh, this is so incorrect, it isn't even wrong. (c) Wolfgang Pauli

Pretty much the only section of your comment that makes some sense is the suspicion that chronic busybodies will use anything as a reason to meddle with other people's lives.

Considering that most of this decrease seems to have occurred in just a two year period, one can probably safely conclude the data is flawed, or the the way obesity is measured has changed- either by how it is done, or by whom it is done.

Basically what I've gotten from the comments is that people are willing to believe any explanation *except* that this is good news.

Many of us disagree that its news.

I also disagree that focusing on caloric consumption of pre-schoolers is smart, and may be harmful.

My (nondisprovable) hypothesis is that if the story was "obesity rates in 2-5 years jump 50% over last decade" the reactions would've been completely different.

Possibly. But if my doctor tells me tomorrow I have advanced heart disease, I am likely to believe him. If he tells me tomorrow my advanced heart disease has disappeared, I'm going to suspect an error. I'm going to ask for a plausible explanation for this spontaneous remission, and be skeptical until I am given one.

I've seen these exaggerated claims about the importance of early childhood obesity before and they usually track back to this study:

“Obese children under three years of age without obese parents are at low risk for obesity in adulthood, but among older children, obesity is an increasingly important predictor of adult obesity, regardless of whether the parents are obese. Parental obesity more than doubles the risk of adult obesity among both obese and nonobese children under 10 years of age.”

Predicting Obesity in Young Adulthood From Childhood and Parental Obesity

The authors advise against any interventions before age 3 because of risks outweighing the benefit, as well as before the age of 10, unless one or more parents is obese.

+1 to Urso

Biologically taken, the two claims are asymmetric, even though they differ in arithmetic sign only. For example, introduction of new, cheaper and sweeter child foods could produce a swift growth of obesity among the target group.

OK, so looking at the actual statistical evidence, and applying a little Bayesian reasoning and/or cautious interpretations when dealing with multiple comparisons, the headline should have been: "Some modest evidence that there may have been some decline in early childhood obesity." Quite different than the headline as written.

The study admitted that there was no adjustment for multiple comparisons and that these results should be "interpreted with caution". This is not headline-worthy news. The interesting question now is why no such adjustment was made - but other than Razib Khan and friends on Twitter, no one seems to be asking it.

See Razib's comments here:

Yeah, maybe a fluke, or maybe a real decline that is much smaller (and more plausible) than 43%.

Excellent link, thanks.

Looks like the study probably wasn't flawed, just inconclusive on some fronts, and that the CDC press release is where the fault lies. Although I'm not sure how to reconcile the report saying this:

"Overall, there have been no significant changes in obesity prevalence in youth or adults between 2003-2004 and 2011-2012. Obesity prevalence remains high and thus it is important to continue surveillance."

and the report "author" saying this:

“This is the first time we’ve seen any indication of any significant decrease in any group,” said Cynthia L. Ogden. . . “It was exciting.”

Even researchers seem unable to distinguish between "significant" (in the ordinary sense) and "statistically significant."

Yeah, her enthusiasm is odd. Ms. Ogden is a veteran - been at CDC at least 20 years. Maybe she was just a supervising author and not really aware of the caveats in these details.

Wow, I bet their kids aren't fat, if they have any!

That's actually a good news.

Liu was one of the 31 migrant workers elected to the National People's Congress last year. She is a foot masseuse in southeast China's Xiamen City. A native of south China's Anhui Province, Liu was a school dropout at the age of 14 and worked to support the schooling of her siblings.

Comments for this post are closed