A top government research scientist I know has a theory about the “obesity epidemic.” Dr. X isn’t allowed to speak on the record to the media about X’s work without getting permission from X’s superiors (one thing that came as a shock to me when I started studying this stuff was the extent to which U.S. government scientists are censored by their employer) so X calls and emails me a lot to complain about the craziness currently engulfing X’s field.
X’s theory is that most of the current panic over fat is a product of a simple and cynical swindle: The pharmaceutical industry wants to increase the market for its products, so it spends lots of money in all sorts of creative ways in order to generate a sense of crisis, fueled by ridiculously unscientific definitions of what constitutes “diseases” that require “treatment.”
This NYT story, about medicating “overweight” and “obese” children, is a classic example. The news hook for the piece is that the American Academy of Pediatrics is now recommending prescribing statins to kids as young as eight, as a response to — of course — the “childhood obesity crisis.” (I hate having to use so many scare quotes but there’s no good alternative when the discourse you’re analyzing consists of a bunch of terms that don’t actually make any sense).
As per usual the piece includes no explanation of the current definitions of overweight and obesity in children, let alone any hint of where those definitions come from, or why they might be controversial. Instead, its sole representative of these categories is a 267-pound 13-year-old girl whose grandmother’s leg was amputated as a consequence of diabetes. The author, Stephanie Saul, then adds that kids this fat were rarely seen more than ten years ago, but now as many as 30% of America’s children are “overweight.”
The piece, also per usual, is structured around a debate between people who want to make kids thinner via lifestyle interventions and those who say we tried that and it doesn’t work so now we have to drug them. Dr. X laughs and laughs when s/he reads this stuff: “The drug companies love these stories! Yes, by all means, lets debate whether we should use lifestyle interventions or drugs to ‘treat’ ‘overweight kids.’ Guess what? Lifestyle interventions never ‘work.’ Ever! Just like with adults. So I guess it’s time for Plan B — Lipitor for eight-year-olds!”
Some history: Until about a decade ago, there was no medical definiton of overweight or obesity in children. The Body Mass Index definitons for adults (BMI 25 = overweight, BMI 30 = obese) were considered inappropriate for kids, in part because BMI correlates positively with height in children (taller children have marginally higher BMI than shorter kids). Another reason was that there was very little data on correlations between weight and health risk in children.
But in the 1990s, as panic over fat in America began to build, people at the big federal public health agencies were put under pressure to come up with some definitions. As UCLA sociologist Abigail Saguy has suggested, the current concern over body weight displays many of the features of a moral panic — and moral panics often end up focusing on children.
So definitions were created. A complex bureacratic process at the Centers for Disease Control, which was beginning to get into obesity panic in a big way, ended up spitting out the following labels: kids in the 95th percentile of body mass for their age would be declared “overweight,” while those in the 85th and up would be labeled “at risk for overweight.”
It’s important to understand that these definitions were completely statistical, rather than outcome-based. In other words, they were based on literally no data whatsoever suggesting that these cutoffs correlated with increased risk for any particular medical problem, let alone increased mortality risk. CDC simply drew a couple of statistical lines and attached labels to everybody on one side of them — thereby pathologizing the bodies of millions of American children for no better reason that someone (lets call this person “William Dietz“) thought it was a good idea to do so.
Now a question that might occur to you is, if these are the definitions, why do I keep reading stories in the New York Times about how 30% of America’s children are “overweight?” How is that statistically possible? The answer is that, for those who are profiting from the panic over an “epidemic of childhood obesity,” the great disadvantage of these definitions is that they don’t allow for epidemics, which are by definition increases over some statistical baseline.
This problem was solved by using data from the National Health and Nutrition Examination Survey . This is the best data pool available on the health of the American population: NHANES has followed a statistically representative sample of Americans for several decades now. By examining the early versions of NHANES, it was possible to compare the weight of children today to those from the 1960s and 1970s. This data shows that in the late 1990s around 30% of children had a body mass that would have been in the 85th percentile or higher in the early NHANES surveys. Bingo: Epidemic!
Meanwhile, Dietz started lobbying to get CDC’s terminology changed. He wanted to change the definitions so that the 95th percentile of BMI among children in the early NHANES data would now be considered “obese” (not “overweight”), and the 85th percentile in the same 30 and 40 year-old data would be used to define “overweight” (not “at risk for overweight”). This is a battle he seems to have won.
So we’ve gone practically overnight from a situation where 5% of America’s children were defined as “overweight” according to an almost completely arbitrary definition, to one in which around 15% are now “obese” and more than 30% are “overweight,” by even more radically arbitrary definitions — even though America’s children weigh no more than they did ten years ago.
And none of this, of course, even begins to address such basic questions as whether the weight of American children 40 years ago (when malnutrition was more common than it is today) was “better” than it is at present — an extraordinarily complex question, given the complexity of the enormous number of medical, scientific, economic, political, social, and cultural variables involved.
One thing that frustrates Dr. X no end is that, just as in the case of adult definitions of “overweight” and “obesity,” people have no idea what any of this means. “The 85th percentile of body mass in NHANES I and II is nowhere close to what people think of as a fat child,” s/he tells me. “These are ordinary-sized kids. But people hear “childhood obesity” and in their heads they see a 250-pound ten year old.” (Of course the New York Times story is a perfect illustration).
Which brings us back to statins. Consider the following:
(1) There is little evidence that statins are an effective primary prevention tool for lessening the risk of developing cardiovascular disease in adults. The tables in this story about a recent meta-analysis of the data from eight large studies show no statistically significant lessening of risk for mortality, or serious adverse medical events, among adults using statins for primary prevention of heart disease (that is, adults who don’t already have heart disease). Nor was there any lessening of risk for cardiovascular disease in women. The only group that seemed to benefit from using statins to lessen the risk of heart disease (although not mortality risk or medical risk overall) consisted of high risk men under the age of 70 — and even there the effect was marginal enough that it brings into question whether the cost of use was worth the benefit.
(2) There is no data at all on what effect statins have on children in terms of disease risk.
In other words, what the American Academy of Pediatrics is recommending is that an (enormously profitable) class of drugs that seem to be of little or no value to adults who don’t already have heart disease be given to eight-year-old children, on the grounds of the following chain of reasoning:
There is a very weak association (approximately .10) between CVD risk and obesity in adults. It’s possible that statins might be of benefit to adults who don’t already have CVD although we don’t actually have any data at this point demonstrating that. It’s possible that “overweight” kids might develop as adults the very weak association between CVD risk and heart disease seen in “obese” adults, although we have no data on that either. Therefore it’s also possible that giving statins to kids might benefit them somehow in the future, although we have less than no data on that.
Q.E.D.
