is obesity a problem?

I started this post as a reply to the OrgTheory thread on obesity but it got long enough I decided to move it here.

With all due respect, I think the claim that “obesity is not the problem” is overstated. I suspect that this overstatement is due in part to medical sociology’s tendency to infer from the “social construction of <condition>” that <condition> is not really real (see here for more on that). (Quick disclaimer: I have not yet read Abigail Saguy’s What’s Wrong With Fat?, though I hope to soon.)

BMI is best understood as a screen for, not a measure of, obesity. Like other screens, it can be used in a clinical setting to flag possible morbidity. When Victor Oladipo shows up with a high BMI, that’s an example of a false positive in the screen, but it’s not evidence that the screen is generally false!

The structure of the claim is that obesity (measured however) is imperfectly correlated with health. That is, some healthy people are obese, and some obese people are healthy. Consider an analogy with beta-hemolytic streptococcus infection (“strep throat”). A substantial number of people–children in particular–have strep infections in the “carrier” state, meaning that a strep test reads positive, there are strep colonies in their throats, but they show no symptoms. They are strep-positive but “healthy” (in this respect). Of course, plenty of other kids don’t have strep but do have sore throat and fever. So streptococcal infection is imperfectly correlated with relevant health outcomes. Does it follow that streptococcal infection “isn’t the problem”, or that strep doesn’t cause sore throat and fever? Of course not. Similarly: my grandfather drank like a fish and chain-smoked until his death (from other causes) at age 89. It doesn’t follow that smoking and drinking are not causes of health problems.

I can think of several possible reasons why we could find imperfect correlation between obesity and health:

  1. Obesity isn’t the problem.
  2. Obesity is a problem for some subsets of the population, less so for others.
  3. Similar to the above, obesity complicates certain other conditions, so it is an additive health risk but perhaps not so much an independent health risk.
  4. The effects of obesity on health play out over the long term and so some such effects have yet to be observed.
  5. Obesity is a mediating factor between other health risks and health outcomes.

I’m sure there are others as well.

Saguy’s Time piece┬ácites a 2009 study suggesting that BMI is about as good as other measures at indicating obesity. But the study also shows “excess deaths” associated with obesity; the main point of it is that BMI is a reasonably good measure of the risk of such excess deaths! A related editorial emphasizes that their estimate of excess deaths is likely low due to exclusion of comorbidities. And another recent article suggests that the so-called “obesity paradox” is present only among “obese NHANES male participants with a wide variety of serious illnesses”–not a very compelling case for discarding the general claim of obesity having health risks!

One of the strong implications is that there is a conflict of interest: big pharma has an interest in promoting obesity as a health concern because that will promote greater drug sales. But while I don’t doubt that this pathway exists, it is also the case that lots of skilled researchers and clinicians employed by universities and funded by federal peer-reviewed sources (e.g., not big pharma) understand obesity to be a health problem. That doesn’t necessarily mean they’re right, but it ought to give pause to the conspiracy theory under which big pharma has created a fake moral panic to sell more drugs.

Finally: the question of stigma about obesity is of tremendous importance, and the medical community is far from immune. I think it’s plausible that stigma may actually help perpetuate obesity, and the question of whether the disease moniker increases or decreases stigma is also really interesting–but I don’t think it’s a settled question at all. (My wonderful former student, Michele Easter, did some great related work on genetics and eating disorders for her dissertation.) Medicalization can plausibly reduce stigma as well as increasing it, so if the concern is (appropriately, IMHO) reducing stigma then that ought to be the goal in itself, not claiming that stigma raises the question of whether obesity is in fact a health problem.

Author: andrewperrin

University of North Carolina, Chapel Hill

6 thoughts on “is obesity a problem?”

  1. I’m confused. Isn’t the 2009 Flegal and Graubard article saying that the association between BMI and mortality is weak and that BMI is not a particularly good predictor of mortality? (And that similarly the association is weak for all the other possible measures of obesity as well). For all cause mortality figure 1 shows no effect of high BMI. When looking at just cardiovascular disease, diabetes, and obesity-related cancers, figure 2 does show a significant difference, but we’re talking about 5% more deaths than the reference group. That’s not nothing, but it doesn’t strike me as a particularly strong argument for the mortality risks of obesity.

    To my mind there are two problems with focusing on obesity as the source of health problems, rather than actually looking at better indicators of risk. The first is the stigma issue that you raise, but to my mind the more important issue is that just focusing on weight completely neglects the fact that it’s possible to lose weight in ways that don’t actually make you significantly healthier. I don’t know the research well, but it’s not clear to me that the evidence suggests that obesity is really the problem. It seems more likely that unhealthy diet and insufficient exercise cause both the weight issues and the health issues. Focusing on weight at the exclusion of other indicators of risk seems like it courts the stigma problem unnecessarily.


    1. I could certainly be reading it wrong, but my understanding of the Flegal article is that it finds excess mortality due to obesity, but quite a bit less than had been reported elsewhere; and that it finds no significant variation based on how obesity is measured (BMI vs. other measures). The accompanying editorial explains that this is a very low estimate, largely because it doesn’t account for comorbidities which are very likely the main way in which obesity causes mortality.

      You’re certainly right that there are very unhealthy ways to lose weight, and that these ought not be endorsed. But that fact doesn’t mean obesity isn’t a problem, just that people do stupid things in response to that problem!


      1. My questions is fundamentally whether the issue is the weight itself that’s a risk factor for mortality or the lifestyle choices that are associated with high BMI/obesity. My very perfunctory reading of the evidence is that it’s the latter. That is to say, obesity is actually not the problem, it’s just correlated with the real cause of the health problems in question. In which case treating obesity as the problem rather than specifically addressing high fat diet, sedentary lifestyle, etc. a) encourages a focus on weight as a desired medical outcome rather than trying to actually measure health and b) focuses attention on only a subset of the high-risk population (the subset that also happens to be obese). The problem with (a) is that it encourages people to do what you term stupid things, rather than encouraging people to do things that directly affect their risk factors, even if they may not result in rapid changes in weight. The problem with (b) is that even if you were able to solve the problem of obesity you may not actually reduce rates of mortality from diabetes, heart disease, and cancer by all that much, precisely because the association between obesity and those causes is relatively weak. Add to that the fact that the stigma issue associated with obesity is going to make some people resistant to dealing with the medical community and it seems like overall focusing on weight itself may be a poor strategy for actually improving health outcomes.

        This week’s economist has a nice write-up about research into the mechanism by which obesity might cause cancer. The basic summary is that it looks like changes in the bacteria in the digestive system as the result of a high fat diet may be what causes mice exposed to carcinogens to develop liver cancer. “Ah hah!” you might say, “an explanation for why obesity is the problem.” Let’s say this relationship holds in humans. The question then becomes will losing weight–that is, treating the disease of obesity–actually change your risk factor for liver disease? Maybe, maybe not. I don’t think anyone knows yet whether those changes in the microbiome would naturally reverse upon switching back to a low-fat diet (as near as I can tell this was not investigated, even in the case of the mice). So the story with the mice appears to be that high fat diet causes obesity and changes in the bacteria in the digestive system. The presence of the bacteria drastically increases the risk of liver tumors if the mice are exposed to carcinogens. Obesity in the absence of the carcinogens does not increase risk. So two things jump out at me. One, the actual problem here is the carcinogens. Two, the high fat diet that causes the obesity also seems to cause the shift in the microbiome. Thus, if I’m a mouse doctor it seems to me the thing to address is the high fat diet. By the time a normal-weight mouse eating a high fat diet becomes an obese mouse, they may already be in trouble.

        It’s entirely unclear to me given a pretty perfunctory review of the evidence that obesity is actually the problem. It seems, at best, to be a symptom of a larger set of problems. Take, for example, this article, which finds no mortality difference between normal BMI and obese BMI, but does find a fairly large association between being sedentary and mortality at all levels of BMI.

        Part of my concern is that, as a woman, I’ve spent much of my adult life as the recipient of messages about my weight, not my health. When I hear other women talking about weight loss strategies, their focus is usually on the end number rather than actually increasing health. If weight were really the true risk factor for health problems, then that would be fine. In that case how you get the number lower wouldn’t matter much. If the issue isn’t really weight then the method matters. In that case, the Atkins diet, a Mediterranean diet, and surgical intervention with limited lifestyle changes afterwards are not actually equally good choices even if they may get you to the same outcome number. My concern is that by focusing on obesity as a disease we risk increasing the stigma of “fat” dramatically while not doing much to actually increase the valuation of healthy life-style choices.


    2. It seems more likely that unhealthy diet and insufficient exercise cause both the weight issues and the health issues.

      I am no expert in this area, but this is exactly the take of these obesity researchers, whose blog routinely covers the literature on health and weight. They, too, are concerned that weight has come between the real causes and effects of poor health outcomes.


  2. I’m not conversant with the scientific literature but there are two general issues I’m aware of about BMI and obesity discussions. One is that highly muscular people score as obese on BMI but are not. The other is that the standard of normal body fat varies substantially among different populations. By “normal,” I mean how much body fat you are going to carry and how it will be distributed on your body, assuming comparable diets, exercise etc. That men and women have different normal amounts of body fat is well known, but there are also ethnic/racial differences.


  3. @sarahliz and @tina: here are a few other mechanisms for health outcomes (not necessarily mortality). Joint problems associated with higher weight due to extra strain. Heart disease associated with higher weight, distinct from diet and exercise, due to additional strain on the heart. Similar with diabetes and various other cardiovascular disorders.

    @sarahliz: Your point about whether obesity causes disease means that losing weight will fix disease is really interesting and important — but still not the whole story. In many cases it seems like the right question is prevention, not cure, so if in fact obesity is a cause, promoting prevention is a Good Thing.

    The article you posted has a population of 70- to 75-year-olds, which seems rather narrow to deduce no morbidity or mortality increases, right?

    Again, I am very sympathetic to the need to address health, not appearance, and to actively oppose stigma, but I don’t believe there is a direct relationship between the health findings and stigma, nor any clear evidence that stigma will decrease if health evidence is refuted.


Leave a Reply

Please log in using one of these methods to post your comment: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s