Asymmetrical Information

The biggest problem with BMI as a metric is that it doesn't differentiate between types of weight (e.g. muscle vs. fat).

Mike Tyson, at his prime (5'10", 218 lbs.) was "obese" by BMI standards, despite a very athletic 5.5% body fat.

When you consider how many professional athletes are either "overweight" (football) or "underweight" (track and field), BMI is a pretty weak gauge for fitness.

According to the BMI scale, I'm at the bottom of the "obese" range, yet I have a resting heart rate of 60 and cover upwards of 40 miles a day on the bicycle, and often have to do things like load and unload trucks at work. Yes, I'm carrying some extra weight, but people don't look at me and think "obese."

It's well-known that BMI doesn't work for people at the outer ends of the normal distribution (less than 5'4 or more than 6'2 for men). I expect for your height you are far from any reasonable estimate off the standard tables.

According to the BMI calculators, at 6' 3" and 185 lb, my BMI is 23.1. That puts me on the high side (75%) of the 18.5 to 24.9 range of normal. Most people consider me thin and some call me skinny.

I think that the BMI by itself is not all that useful, at least when it uses height and weight as the only inputs. The BMI calculator I used did not even ask for sex (male).

I suspect that those who take refuge in BMI's inaccuracies are not subject to the exceptions. (Sadly, I am not, either.)

Body fat % is much more accurate - if you don't know whether or not you're overweight (or are just curious), get it read. It is very cheap (for an electro-current reading, skin caliper is also cheap while the more accurate water-immersion is prohibitively expensive) at any large university campus - just look for the "training center" or equivalent at the largest campus gym.

Why even bother with a measurement as flawed as BMI (a fatally simple equation that doesn't take muscle - heavier than fat - into account)?

BTW here are my "stats" from last week's free reading at my university:

174 lbs, 9.0% body fat, BMI of 22.6.

http://www.healthchecksystems.com/bodyfat.htm

That will give you "expected ranges" and anything up to 24% is considered "fit" for a woman. As a man, I am basically in the middle of "athlete" range - good enough for intramural soccer :).

I'm no expert, but it seems BMI should taken into account with other factors such as blood pressure, cholesterol (did I spell that right?), etc. in determining's the status of one's health.

Since many of us are going to die anyway as a result of a terrorist attack, natural disaster, or a communicable (did I spell that right?) disease such as SARS or bird flu, maybe we should stop worrying about BMI and how overweight may be.

BMI is based on a 19th-century metric and is simply a ratio of weight and height. Frankly, it's basically useless as anything other than a very, very rough measure. It certainly doesn't measure fitness, and it certainly doesn't measure body fat percentage. I'd skip using it for anything, at all. But, then again, I'm not one of the government's health nannies.

The BMI is proportional to the weight divided by the square of the height. Volume and weight should be proportional to the height cubed. So a very tall person will have a higher BMI than a medium-height person of the same proportions.

It's possible that the BMI is better than the above analysis suggests, if it is supposed to measure the cardiovascular load rather than fatness. Someone who gets a high BMI by being short and fat is obviously in trouble - that's a heavy load on a heart that hasn't been built up much by exercise, but it's possible that the 6 foot 8 body builder with the same BMI (and a lot more weight) is also headed towards a heart attack. Muscles and height require more work from the heart than the same weight shaped into a round lump of fat, and in the extremes it's possible that no amount of exercise can build up the heart enough to keep running that body for a normal lifespan. But that's pretty speculative, and it's quite sure that an index that doesn't distinguish between a tall fat person who never exercises and a tall strong person who exercises several hours every day is not a great indicator for individuals.

BMI isn't much good on it's own, but it is a decent shorthand for body weight.

Life insurance companies use BMI in underwriting individual applicants. However, it is just one of a number of measurements that are reviewed (blood pressure, bloodwork, smoking, family history etc).

I'm not sure who else uses it

Many have noted correctly that BMI does not work well with athletes and people on the extremes (the very tall, very small, very thin, and very fat). What most people do not know is that the BMI works terribly among elderly persons, especially inactive ones (more below).

In medicine, the BMI was used as a way to "normalize" body measurements and place persons into rough categories for looking at drug effects or disease progression. It was never designed to be a tool for monitoring individual patients or categorizing obesity or thinness. A better way to categorize body morphology is to compare lean body mass to total body weight. Lean body mass calculations can be made from a few skinfold thickness measurements. However, it is far easier to use just height and weight, so the markedly inferior BMI gets used in government standards. Grrr.

Elderly persons who are inactive, unmuscular, and have flabby abdomens filled with fat often end up with low BMIs. (Fat weighs only 80% as much as the same volume of muscle, so the patients weigh less.) They are much less healthy than more active elderly persons with "elevated" BMIs. That's one reason why the mordidity curve is U-shaped: higher death rates among the low-BMI unhealthy elderly and higher deaths rates among the morbidly obese.

I wanted to add a personal comment. When I was young and athletic, my BMI was at the bottom of the current "overweight" range. I gained weight in middle age and my BMI now is in the "obese" range. I am short (5' 6") and have plenty of muscle under my fat. Yet, I can lay down motionless at the bottom of a swimming pool. I have met very few thin persons who can do this, and no truly obese person should be able to lie underwater. Just another anecdote about the poor utility of BMIs.

Well, assuming your body to be a scaled up version of a 5'4" woman, your volume will increase with the cube of the height difference. In practice, there are reasons why you'll be even heavier, though. For example, the larger you are, the disproportionately thicker the bones that you need to carry you (this is why insects can get away with some amazing stuff: being scaled down so small they need way less stiff support structures, and so a lot less of their weight is in their skeleton (well, exo-skeleton)).

So in general, the taller that you get, the higher that your BMI will be at a healthy weight, since they generate BMIs for various heights by using a reference height and then just scaling the numbers. At least, that's how BMI calculations were explained to me.

Markm, and to some extent, ctl, get it right.

If one increases a person's proportions evenly, their weight will increase as the cube of their height, while BMI assumes that ideal weight increases as the square of the height.

In the real world, not all dimensions of people increase proportionally with height, and so the exponent used should be somewhere between 2 and 3. But that makes the math *hard*, so nobody does it. Instead, BMI fails outside a certain range of heights.

I'm at the top end of the healthy BMI range (around 24) when I'm fit. It's not an ideal measure, but it's simple, requires information that most people know or can get quickly, and it works extraordinarily effectively at a population level. Hard to believe that the whole population has suddenly got more muscular and as a result the US (and Australian) population has moved wholesale into the overweight range.

Where is the data that says that "tall people tend to die sooner"? You are tossing this off as a well known fact, but I have never heard this and I would tend to believe that it is not factually accurate.

Can we have some data please?

I find BMI to be very inaccurate when it comes to taller people. For example, many Australian footballers have BMIs of 27-28 but are very fit guys. I think even George Clooney has a BMI of about 27 or so....

Muscles and height require more work from the heart than the same weight shaped into a round lump of fat, and in the extremes it's possible that no amount of exercise can build up the heart enough to keep running that body for a normal lifespan.

Something of the sort may the case with football players. Linemen in particular can be extremely large, a 300-pound NFL tackle is considered "small," though of course they are far more muscular than most other people. And yet there's mounting evidence that retired players face a significantly elevated risk of death and serious illness in their 40's and 50's. It may be the case that being excessively large indeed puts too much of a strain on one's heart even if much of the excess bulk is muscle. There may be other factors at play, of course, but the connection cannot be dismissed.

it's possible that the 6 foot 8 body builder with the same BMI (and a lot more weight) is also headed towards a heart attack

Professional bodybuilders in fact do face serious health risks, but in most cases that's a result of the massive amounts of steroids and human growth hormone they consume.

So in general, the taller that you get, the higher that your BMI will be at a healthy weight, since they generate BMIs for various heights by using a reference height and then just scaling the numbers.

I originally had the same idea, but when I analyzed the NHANES data, I found that BMI was negatively correlated with height. In theory, tall people should have higher BMIs. In practice, they tend to be so much thinner that they actually have lower BMIs.

Waist-hip-ratio seems to be a better predictor of mortality than BMI, because people generally don't gain a lot of muscle around their waists, and because abdominal fat is a sign of insulin resistance.

Piso -- I was told it verbally by the son of an economist who does research on height and incomes.

Brandon--that makes sense to me. Tall people don't scale up exactly--we grow up more than out, so that our waists are smaller relative to our height/ribs than shorter people. I remember seeing a photograph long ago that purported to be of a giant woman walking next to a midget; I instantly knew it was fake, because the tall woman wasn't proportionally elongated the way I am.

BMI doesn't work for people who are short or tall or very muscular. Here is a link to an advocacy group that disapproves of BMI - click their calculator to see the company you are in if BMI reports you as overweight (Geo. Bush) or obese (the Governator).

http://www.consumerfreedom.com/news_detail.cfm/headline/2860

BMI is so bogus that using it contributes negatively to understanding. Boycott it.

I have a friend who is math-intensive (though I think his doctorate is in geophysics.) He calculated his BMI and got his body fat measured at the same time. Once he ran the numbers, he determined that to have anything less than an "overweight" BMI, his body fat percentage would actually have to be negative.

Heh.

Comments are Closed.