Obesity : BMI's inacuracy leads to the epidemias's underesti

http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0033308

This is the actual study.

I think the most important part is this:

mean age = 51.4 (SD = 14.2).


I would argue that the conclusions of this study can be applied across the demographic upon which it was performed, but cannot be applied to to those aged 36 and younger. It is quite a leap to extrapolate an estimate of 64% obesity from a number that was derived from a mean age of 51. I am fairly sure that mean age of the general adult population is much lower than 51!

Actually, I have so many issues with their now-famous conclusion that 64% of the population is obese....

They pulled the '30%bf = obese'seemingly out of their a*se. It came from this document: http://www.asbp.org/resources/uploads/files/Accepted OOEM Final nov 1 09 Ref edits(1).pdf

The ASBP recognizes that there are several acceptable anorectic agent usage
criteria for patients including but not limited to at least one of the following
parameters:
• BMI ≥ 30.0 in a normal, otherwise healthy individual
• BMI ≥ 27.0 in an individual with associated co-morbidities (e.g. type II
diabetes, hypertension, abnormal glucose tolerance, atherosclerosis,
cardiovascular disease, stroke, hyperlipidemia, hypercholesterolemia,
osteoarthritis, gall bladder disease, breast cancer, or sleep apnea)
• Current body weight ≥ 120 percent of a well documented, long-standing,
healthy weight that the patient maintained after age 18
• Body Fat ≥ 30 percent in females
• Body Fat ≥ 25 percent in males
• Waist-hip ratio or waist circumference such that the individual is known to
be at increased cardiovascular and/or co-morbidity risk because of
abdominal visceral fat
• Presence of a co-morbid condition or conditions aggravated by the
patient’s excessive adiposity

What a mess. There is not even a reference anywhere in the document to suggest 30%bf = obese, and there is no mention of what bf% = overweight.

This is how 30%bf (women) corresponds with our well-known body fat % table:

Description Women Men
Essential fat 8–12% 3–5%
Athletes 14-20% 6–13%
Fitness 21-24% 14–17%
"Average" 25-32% 18-24%
Excess fat 32%+ 25%+

It falls into average. Not overweight, not excess, not obese. Average. It allllmost falls into 'excess'. Funny that... BMI 24 also falls within 'healthy' but almost falls into 'overweight', or excess.

If anything, this study shows that BMI is a fantastic surrogate for determining whether an individual is overweight or obese (within the well-known limitation that it places well-muscled men into 'overweight' and lean women of healthy bf% into 'underweight'.) BMI currently estimates that 28% + 35% of the population are overweight or obese... that is 65%.


/end nerd rage. I've always been warned that papers published in the health sciences are atrocious... but that is almost implausible.

I think that BMI is of extremely limited worth and probably does underestimate the problem, but the original study that they keep basing these stories on has some issues. They compared BMI to DEXA scans only for patients who were getting DEXA scans for another reason. They didn't randomly select healthy people for the study, and they didn't exclude people who had other diseases like osteoporosis, muscle diseases, diabetes, etc. This could overestimate the problem. The sample was also not diverse. 75% of their subjects were white, and the average age was 51. They also did not follow their population to determine if those newly classified as obese were at the same risk for diseases of obesity as has been demonstrated in some BMI studies. They simply say that they probably are. This may be true, but they have no data for it.

^ yes, all of this, too.

I'm happy to say that the conclusions of the study are bad, though.

The BMI might be a bad indicator for measuring obesity, due to it's inaccuracy.

American population may not be composed by 26% of obese people, but 64% according to the Dexa method. Should obesity's perception be updated ?
Link :
http://www.latimes.com/health/la-he-obesity-20120403,0,3807756.story

Neither the 26% number, nor the 64% are mentioned in the article you link to. Can you please cite the source you actually used to get these numbers?

Also note that the article reports that some people moved from being classified obese to non-obese based on this study. BMI is widely known to be a problematic measure, but no other suggested method has gained significant traction in the research community, probably because BMI is relatively easy to measure. That said, if BMI weren't measuring something of use, then it wouldn't be as useful in predicting a variety of health outcomes. It may not be ideal, but that doesn't mean it's worthless.

@ Lolli- We cross posted, but I wanted to add that this one is worse than most. Keep in mind that this is published in an online journal for peer review after publication, rather than an actual peer reviewed publication with prior editorial input, and it's a cross sectional epidemiological study for which they apparently had almost no actual funding, so it's...flawed in many ways.

wow... I did not know that was how PLoSone functioned.

Funding aside, those conclusions are smack! haha. With a sample size of >9000 they can certainly draw some interesting conclusions even within the limitations, even with little funding, just not those conclusions, haha.

Yup. There is some review for "technical soundness", but other than that, the idea is to start discussions and publish everything. For fun have a look at the editorial philosophy: http://www.plosone.org/static/whypublish.action

@kalynn

"PLoS ONE will rigorously peer-review your submissions and publish all papers that are judged to be technically sound. Judgments about the importance of any particular paper are then made after publication by the readership (who are the most qualified to determine what is of interest to them)."

It is peer-reviewed.

It's just that there is no determination of what's a hot topic from the editor.

It is, though, if you read the actual process it seems less rigorous than other journals, and I did overstate (too little coffee and too much stats rage). The emphasis is on publishing anything that is primary science, which I think is not a bad thing, though I think more "marginally acceptable" things probably sneak in than might be the case elsewhere. Only my opinion, though