BMI Correction

[mistermack]

I've known for years that the BMI charts are miles out for me. There are various adjustments for "heavy build", "race" etc. But I haven't seen the one that should apply to me. 

My build is quite Neanderthal. Long body, short legs. And it's the ratio of legs to height that I'm posting about. 

I'm 5ft 6in in height. (1.68m approx). I used to have a friend who was 6ft 3in  (1.91m approx). Sitting down we were the same height. But standing he was 9 inches taller.  His extra height was all leg, which weighs a lot less than body, but the BMI only goes on overall height. I would expect his ideal weight to be a bit more than mine to reflect the extra leg length, but the BMI just assumes that everyone has the same leg/height ratio.  

My ideal weight, according to a BMI chart, is 135 lbs. His was 170 lbs. In actual life, we were both about equally fit back then, and he was about 10 to 15lbs more than me, not the 35lbs that the BMI chart suggests. 

When I was the fittest of my entire life, at age 21, after six months roaming the US and Canda on foot and by thumb, I remember I weighed about 157 lbs. I really was super fit at the time, but according to the BMI chart, I was classed as overweight even then. (and 22lbs over my "ideal"). So the BMI chart is quite a long way out for someone with my body/leg ratio. 

My way of adjusting is simply to look up the ideal height for 157lbs, my fittest ever weight, and it's six foot.  So now, if I want to know how I'm doing, I look up my weight for a six foot man, and that gives me what I think is a realistic corrected BMI for my build. ( It still tells me I need to lose some though )

So for people like me, BMI is overstating the situation, but for my long-legged friend, it would actually be understating the degree of overweight.  

It would probably be fairly easy to work out a mathematical correction to be applied, taking the leg - height ratio into account. Easy for a mathematician but not for me. 😟

[John Cuthber]

BMI is a system by which a doctor can tell a patient  "You are too fat" while causing less offense.
Reading anything more into it is fishy.
It's a bit like IQ.

 

 

[exchemist]

16 minutes ago, John Cuthber said:

BMI is a system by which a doctor can tell a patient  "You are too fat" while causing less offense.
Reading anything more into it is fishy.
It's a bit like IQ.

 

 

A friend at work once went for a medical and returned to say, with a grin, he'd been told he was 1kg short of being "technically obese".  We all thought "technically obese" was a wonderful expression.

[TheVat]

BMI is just another failed attempt to categorize people with an algorithm.  The idea that 157 would be an ideal weight for a six foot male, in this part of the country would be preposterous.  Lots of Scandinavians, Germans, Ukrainians, and others from northern European peasant stock who are built heavier and would be nearly emaciated at that weight. 

One thing I notice, related to that, is that the younger generation now seems to be averaging a slighter build, which leads me to wonder about the endocrine disruptors in our environment and the junking up of the American diet in recent decades.  There could also be developmental effects for a generation that largely lives in an electronic coccoon, which should be an area for active research.  

 

 

[Endy0816]

I always found Military ridiculous for that. Big, strong guys would often be outside the standard. It forced a number into unhealthy weight loss routines.

[Sensei]

The density of muscle is ~1.1 g/cm^3.

The density of fat is ~0.9 g/cm^3.

1.1/0.9 = ~1.22 ratio (+22%).

A person with the same height, same weight, but an athletic body, can have a much smaller body volume than a person with the same properties but with fat instead of muscles. BMI does not include this information. Therefor the BVI (Body Volume Index) was introduced.

https://www.google.com/search?q=body+volume+index

"How do you calculate BVI? BVI was defined as total body volume/height^2 (L/m^2)."

 

[mistermack]

The replies really indicate how off the BMI index can be. If you are Mr or Ms average, of european ancestry, it's probably spot on. It's accepted that Asian ancestry means lighter weights for the same height on average. Don't know how that works. 

The figure I've added shows crudely what I'm getting at, about leg proportions. 

[John Cuthber]

4 hours ago, TheVat said:

Lots of Scandinavians, Germans, Ukrainians, and others from northern European peasant stock who are built heavier

How do I distinguish that from being fat?

Also, it may have escaped your notice but essentially everyone is "peasant stock".

However, that's enough time wasted on a random number generator.

[mistermack]

5 hours ago, TheVat said:

One thing I notice, related to that, is that the younger generation now seems to be averaging a slighter build

And they seem to be getting younger too.  

[CharonY]

5 hours ago, TheVat said:

BMI is just another failed attempt to categorize people with an algorithm. 

I rather think that similar to pretty much all categories it has a (limited) area of application, but obviously falls short outside of it. Another issues is that BMI was developed based on mostly European cohorts, so the model is inherently biased.

What it has been somewhat effectively is (again, for European cohorts) to find groups at risk for certain conditions, including diabetes and certain cardiovascular diseases. As such a rough indicator it works somewhat well (at least compared to other risk factors).

 

[mistermack]

13 hours ago, CharonY said:

As such a rough indicator it works somewhat well (at least compared to other risk factors).

What I've read is rather different. The BMI as a risk guage is far less reliable than the waist measurement. That's from memory, I can't provide a linkwithout searching, but I've seen it stated in many different articles. It might have been waist/height ratio, or something like that, but it was in medical articles.  

[Prometheus]

This falls firmly under 'all models are wrong, some are useful'. BMI is used as a rough marker in certain settings, no professional is (or at least should be) using this single metric to make any clinical decisions. If fat distribution is particularly important a more thorough method would be used. Bioimpedance is common in nutritional studies. 

First you should decide what you actually want to measure, then decide if BMI is good enough to that end. In some cases it is.

[CharonY]

3 hours ago, mistermack said:

What I've read is rather different. The BMI as a risk guage is far less reliable than the waist measurement. That's from memory, I can't provide a linkwithout searching, but I've seen it stated in many different articles. It might have been waist/height ratio, or something like that, but it was in medical articles.  

You are not wrong and I should have added qualifiers, but the waist/ waist ratios has also some issues. BMI has been around for a very long time and I should have said that BMI worked better than other indicators investigated for a a long time. 

Now with regard to waist circumference and ratios, if I remember correctly they did not perform consistently better. In leaner cohorts they were generally superior, but when heavier folks were mixed in, it becomes more inconsistent. Central adiposity, as tracked by waist measurements, seems to be a good marker for risk in otherwise seemingly healthy folks. However, after a certain weight gain other factors contribute equally or more.

1 hour ago, Prometheus said:

This falls firmly under 'all models are wrong, some are useful'.

Exactly.

[mistermack]

The BMI might be of minor use to doctors, to bring up the subject that you need to lose weight. But I think that my own system, of comparing my weight to my fittest ever weight, is all I need to know. 

Of course, a doctor or trainer has no way of knowing what you weighed, when you were at your best-ever fitness. And of course also, some people have never been fit, so can't really compare. 

At peak fitness I was 157 lbs, when I was 21. Now I'm 71, I'd settle for about 170. (or be over the moon with 170 in reality)😀

[CharonY]

47 minutes ago, mistermack said:

The BMI might be of minor use to doctors, to bring up the subject that you need to lose weight. But I think that my own system, of comparing my weight to my fittest ever weight, is all I need to know. 

Of course, a doctor or trainer has no way of knowing what you weighed, when you were at your best-ever fitness. And of course also, some people have never been fit, so can't really compare. 

At peak fitness I was 157 lbs, when I was 21. Now I'm 71, I'd settle for about 170. (or be over the moon with 170 in reality)😀

There are age-adjusted charts and roughly I think in either case you fall in roughly around low- moderate risk bracket. Whether that actually means anything is a different matter, of course, it is just based on comparison with age-adjusted cohorts.

[mistermack]

2 minutes ago, CharonY said:

There are age-adjusted charts and roughly I think in either case you fall in roughly around low- moderate risk bracket. Whether that actually means anything is a different matter, of course, it is just based on comparison with age-adjusted cohorts.

I consider being 71 as at least a moderate risk bracket, whatever my weight. The best incentive, for me to keep my weight down, is the danger of diabetes and strokes. Dying is bad enough, but living in misery is even worse. 

[CharonY]

Just now, mistermack said:

I consider being 71 as at least a moderate risk bracket, whatever my weight. The best incentive, for me to keep my weight down, is the danger of diabetes and strokes. Dying is bad enough, but living in misery is even worse. 

Well, in absolute terms that is certainly true, we just eventually fall apart. The moderate is just referring to risk associated with a certain BMI in a certain age bracket. I.e. relative to other 71 year-olds your BMI-related risk is still comparatively low (but as you mentioned, does not take other health markers into account). And I agree with both, the issue with stroke is that it is just so bloody difficult to predict.