Four Points About The "Leveling" Of Obesity

From The New York Times:

Obesity rates in women have leveled off and stayed steady since 1999, long enough for researchers to say the plateau appears to be real. And, they say, there are hints that the rates may be leveling off for men, too.

The researchers’ report, published online at cdc.gov/nchs, used data from its periodic national surveys that record heights and weights of a representative sample of Americans. Those surveys, said Cynthia L. Ogden, an epidemiologist at the National Center for Health Statistics and the lead author of the new report, are the only national ones that provide such data.

Dr. Ogden added that the trend for women was “great news.” Obesity rates have held at about 35 percent since 1999, convincing her that the tide had changed. “I’m optimistic that it really is leveling off,” she said.

Men’s rates increased until 2003, when they hit 33 percent and stayed there through 2005-6. Dr. Ogden said she would like to see a few more years of data before declaring that men’s rates had stopped increasing.

Here are some takes on the story suggested by Paul Ernsberger of Case Western . (Any mistakes here are probably my fault, not his.)

1) There are two government data sources being drawn on here; the frequently-updated Behavioral Risk Factor Surveillance (BRFS) survey, and the less-frequent but more reliable NHANES survey.

Why is NHANES better? First of all, because the BRFS is a phone survey, it relies on people’s self-descriptions to get height and weight data; but self-descriptions can be mistaken or dishonest. NHANES measures and weighs its subjects to get the data, which is expensive, but more accurate.

Second of all, although both surveys attempt to measure a representative sample of Americans, BRFS excludes people without phones and people who just have cell phones, making it less representative.

Why does this matter? Because the evidence that there are marked increases in obesity (BMI > 30) in every state since 1999 is based on the BRFS; as I understand it, NHANES doesn’t show such an increase. But the BRFS is less reliable.

2) Obesity rates are equal between men and women now, even though statistically weight loss diets and other weight loss methods are used much less by men than by women. Like a lot of other evidence, this suggests that weight loss methods are not successful at reducing obesity rates.

3) A huge number of people will take credit for the plateau in weight. Every purveyor of weight loss advice and programs will claim credit.

4) From the point of view of promoting health, the most important issue is to promote weight stability, and to focus on health indicators other than weight, such as blood pressure and cholesterol.

This entry posted in Fat, fat and more fat, In the news. Bookmark the permalink. 

17 Responses to Four Points About The "Leveling" Of Obesity

  1. 1
    Sailorman says:

    Second of all, although both surveys attempt to measure a representative sample of Americans, BRFS excludes people without phones and people who just have cell phones, making it less representative.

    No debates on the self-reporting issue. Hell, I don’t know what I weigh, and I’d have to guess if anyone asked me.

    But obviously, the phone thing is only an issue if the population who is excluded is significant enough in size or composition to affect the results. Do you actually think that phone ownership is significantly linked to obesity?

    From the point of view of promoting health, the most important issue is to promote weight stability, and to focus on health indicators other than weight, such as blood pressure and cholesterol.

    This is certainly true. But there’s the functional test aspect as well.

    Problem with BP is that you need a nurse/doctor/BP machine to take it for you, and a nurse/doctor to tell you what it means and/or what to do to fix it. That costs cash.

    Cholesterol suffers from the same problems. Do you know what your cholesterol is? I don’t. I probably should, but I don’t. And if I wanted to know, I’d have to get a blood draw I think, and wait for results, and see (or at least interact with) the hospital twice.

    OTOH, if I want to know whether i’m overweight, gaining, losing, stable, etc I can step on a scale. Or look in a mirror, for that matter.

    This simplicity has to have an effect, don’t you think? Even if maintaining/changing weight offered only 10% of the benefits of fiddling with BP or cholesterol, it might well be possible that the # of people who would pay attention to and attempt to modify their weight would justify the disparity.

  2. 2
    NotACookie says:

    But obviously, the phone thing is only an issue if the population who is excluded is significant enough in size or composition to affect the results. Do you actually think that phone ownership is significantly linked to obesity?

    I suspect landline phone ownership does correlate with obesity — and moreover, that the correlation has shifted. A lot of people are cell-phone only these days, and I’m pretty sure those people are disproportionately young. They may also be comparatively well-off financially, which correlates with health consciousness.

    In contrast, “no phone” adults are probably a constant slice of the public, and therefore they won’t affect the trendline in BRFS.

    If I read the report right, we’re only talking about a few percent shift — and if cell-phone-only adults are different enough from the general public, that could explain some of the trend in BRFS.

  3. 3
    Charles says:

    Sailorman,

    Total cholesterol can be done with a 10 minute wait. It does require someone who knowns how to do a blood stick (IIRC), but that is way a skill way below the nurse/doctor level. The device to do the test probably isn’t super cheap, but it is reusable. Certainly, this isn’t free, but both BP and total cholesterol is certainly within the reach of reasonable public health efforts, particularly since most people should only need to be having checked once a year as a screening effort.

    The problem with using weight as a proxy for these health issues is not so much that weight doesn’t provide some limited degree of (probabilistic) information about cholesterol and blood pressure, but that paying attention to weight provides less than zero information about how to correct cholesterol and blood pressure. Reasonable efforts to change diet and lifestyle that will be likely to change BP and cholesterol will be unlikely to change weight substantially, and radical efforts to change weight will (a) probably not work, and (b) in failing to work will cause active harm.

    Also, checking weight produces large numbers of false positives (fat people with good BP and cholesterol) and false negatives (thin people with bad cholesterol). If I paid attention to weight instead of cholesterol, I would probably die (still relatively thin) in my late fifties or early sixties of a heart attack.

    The public health costs of providing people with the means to monitor what actually matters would not be huge, and the public health costs of having people focus on a poor proxy for what matters are almost certainly larger (although born by different organizations).

    Haven’t we been through all this before?

  4. 4
    Tapetum says:

    Sailorman – I don’t know where you live, but where I live I can check my own blood pressure for free in six different places within five miles. Those lovely little sit-down pressure checks are in nearly every supermarket and drug store. Those same places check cholesterol for free or nearly free periodically – several times a year usually – and also occassional bone mass checks.

    So yeah. I’m 5’8″, 191, Total cholesterol of 182, down from 257, ratios good, except for somewhat high triglycerides, normal blood pressure… none of this should be hard information to lay one’s hands on.

    The least informative bit of information in the whole list is the number for weight. Even my height gives you more, by putting weight in some sort of context. Not to mention that my weight was the same at 257 that it is at 182 total cholesterol – and my cholesterol was actually higher at 191 than it had been the year before when I weighed 217.

  5. 5
    Mandolin says:

    Also, some number of fat women have PCoS (a fertility-affecting metabolic disorder which can cause and complicate weight problems) which is correlated with high blood pressure and high cholesterol — which in the PCoS population, unlike the rest of the population (according to my gynecologist), are not correlated with heart attacks.

    I read somewhere that the estimated percentage of women with PCoS is 10%.

  6. 6
    outlier says:

    Point 1 doesn’t explicitly state how it relates to the NYT article, so I’m not sure what the intended conclusion is. Is it that the “leveling off” hypothesis is based on numbers from the NHANES study instead of the phone survey, and therefore probably accurate? That’s the conclusion I’d drawing from your writing, but I don’t know if that is the point you intended to make.

  7. 7
    Ampersand says:

    That’s my understand, Outlier, although I haven’t read the study directly so I could be misunderstanding.

  8. 8
    Sailorman says:

    I wasn’t disputing that BP and C are far better health indicators than is obesity. I’m just noting that compliance is a big issue with any public health measure.* Because the mechanics of weight-awareness mean that it is INCREDIBLY simple to achieve, even if it is not especially helpful, the overall health effect may be notable in comparison to other “better” but harder-to-implement issues.

    *yea, in a different health system things would be different. I’m talking about now, here.

  9. 9
    Kevin Moore says:

    Why is NHANES better? First of all, because the BRFS is a phone survey, it relies on people’s self-descriptions to get height and weight data; but self-descriptions can be mistaken or dishonest. NHANES measures and weighs its subjects to get the data, which is expensive, but more accurate.

    I agree that reliance on self-reported data runs a credibility risk, but would not accounting for that lead BRFS to speculate that obesity rates are higher than this data would indicate? The assumption here is that respondents would more likely underestimate than overestimate their weight, out of vanity, embarrassment, insecurity, denial or other psychological factors. This is not to say that some folks might not exaggerate their condition, but it is less likely (IMHO) than the other way around.

    And having said all that, I should note that people are probably far more honest than we give them credit for.

    Sincerely,
    Mr. Sunshine

  10. 10
    Charles says:

    Sailorman,

    I think that your justification for focusing on weight over BP and cholesterol is a serious and unjustified stretch and I’m puzzled why you are making it. You are arguing that focusing on a poorly correlated measure that doesn’t change meaningfully over time instead of focusing on an accurate measure that changes meaningfully over time is preferable because the inaccurate and misleading measure is somewhat cheaper to measure than the accurate measure. This is particularly bizarre because BP and cholesterol are not measures that need to be made daily or even weekly during the course of attempting to adjust them.

    It is so facially obvious that the difficulty and expense of measuring BP and cholesterol are absolutely trivial and irrelevant compared to the difficulty and expense of changing behaviors that influence BP and cholesterol, that you would have to produce multiple studies to convince me that the difficulty or expense in intermittently measuring BP and cholesterol were anything other than a complete red herring.

    Compliance is certainly an issue in getting people to change their diet and activity to healthier patterns, but the difficulty of measuring BP and cholesterol is not the hurdle that makes compliance difficult. To argue otherwise is absurd and embarrassing.

  11. 11
    Sailorman says:

    Charles,

    I know we’ve run into this before, … but do you realize that in three short paragraphs, you’ve referred to my posts or statements as “bizarre,” “absurd,” “embarrassing,” “trivial,” “irrelevant,” and “a complete red herring.” those characteristics, as far as i can tell, are or should be “facially obvious.”

    This is especially notable given the circumstances: I thought that I said that BP and C were more important; I was raising what I thought was an interesting sub-issue of public health compliance as a factor in overall results of public health measures.

    And in any case, this has happened a few times between us, though in this event I have managed to refrain from writing an equally snarky reply. Perhaps we can avoid responding directly to each other’s posts? It’s probably best for Alas in general if we avoid engaging, I think.

  12. 12
    Ampersand says:

    Kevin,

    According to Ernsberger, it’s also statistically true that people tend to overestimate their height, which would have the opposite effect of underestimating weight, when it comes to measuring BMI from self-reported height and weight.

    Additionally, has the last ten years of press panic about an obesity “crisis” and how commonplace and typical obesity is changed people’s tendency to misreport weight? And if so, in which direction? Can anyone say they really know?

    How does it all pan out? We don’t know — and it’s because we don’t know that the NHAMES data, based on actual measurements rather than self-reports, should be taken more seriously when the two data sources conflict.

  13. 13
    Ampersand says:

    Sailorman,

    Rereading his most recent post, Charles’ attacks were attacks on your arguments, not on you personally. (Although I think you could argue that his last sentence added nothing, and his argument would have been stronger without it.) And you seem to be using his language choices as a reason to avoid answering his arguments.

    If you want to choose not to respond to Charles anymore, that is of course your prerogative. But I don’t think there was anything inappropriate to “Alas” in Charles arguing that a particular policy position is absurd and a red herring, and do not agree that he ought to avoid responding to you on “Alas.”

  14. 14
    Charles says:

    Sailorman,

    As Amp says, you are free to refuse to respond to my responses to you, but I’m afraid I will continue to sometimes respond to your weakest arguments.

  15. 15
    Emily says:

    I’m intrigued as to why the media haven’t yet considered the possibility that the issue of obesity is as much a question of culture as of health. If:
    1. being slightly ‘overweight’ is associated with a decreased risk of certain diseases (as recent studies have indicated); and
    2. rates of ‘obesity’ are actually decreasing; but
    3. the categories of ‘overweight’ and ‘obesity’ have historically shifted to reflect constructed paradigms of femininity and attractiveness,
    it seems logical to question whether this marks some vague progress for ‘nutrition’ or the proliferation of a polluted ideology that distorts self-understandings.

    To say nothing, of course, of the socio-economic issues that are intimately related to health.

  16. 16
    Silenced is Foo says:

    Interesting news on fat:

    http://forums.fark.com/cgi/fark/comments.pl?IDLink=3248124&ok=1

    “Fit fatties healthier than thin couch potatoes”

    Nasty headline, but it’s the NY Post so what do you expect?

  17. 17
    Silenced is Foo says:

    Ouch. I take it your blog filters out the concatenation of the word “fat” and “ties”?

    I was trying to post a link to an article on the subject of how athletic overweight people are healthier than skinny lazy people, but the headline includes some rather nasty language (it’s the NY post, so what do you expect).

    Anyhow, a link is here:

    http://forums.fark.com/cgi/fark/comments.pl?IDLink=3248124&ok=1

    Warning: Fark Thread. Not safe for eyeballs.