This "can't distinguish between muscle and fat" thing is repeated quite often, but... I think it's pretty clear that the exceptions are reasonably obvious. As a proxy measure, I personally haven't written off BMI. Yes, the man or woman with a visible six-pack and muscular arms and legs with a BMI of 27 is probably fine, even though they're considered "overweight". But, no, the software developer whose exercise consists of walking from the car to his chair in the office with a BMI of 30... they're probably not exempt.
The linked article talks quite a bit about waist fat being explicitly bad. That's an interesting measure for me (the waist/hip ratio) and as someone who's currently hovering around a BMI of 26 (and is average fitness), I plan on making that waist/hip measurement when I get home tonight and seeing how measure relative to their research.
I've been planning to start a bit of a weight loss program (primarily food focused with maybe a 10-20% exercise increase) and if there's specific concerns related to waist fat then I'll do some more research on how to address that.
Even for large populations BMI is a pretty mediocre metric since it is heavily influenced by length.
The formula is kept simple `weight / height^2` to help 19th century doctors. A more accurate growth rate for healthy weight is something like `weight / heigth^2.5` [1]
This oversimplification causes standard BMI treshholds to underestimate obesity for short people (<1.6m), and overestimate obesity for tall people (>1.8m).
That changed drops my BMI two points, and as a tall person, really seems more in line with where I'm at weight wise. The goal for a normal weight is actually realistic and attainable. I've noticed that tall people only seem to be within the normal BMI range if they are the slender sort - if you look a more normally proportioned tall person you're automatically overweight, and very little extra weight put you in the obese category.
You can weight 108 kg (238 lbs) instead of 100 (220 lbs) for a BMI of 25 (top of the normal range). 18 lbs is quite a lot of difference, even at that weight. I stand at 198 cm's and hit 238 lbs last summer. At that point I had very little belly fat - 220 lbs would probably get me close to visible abs (the top few at least), which I don't think should be the top of the "normal" category.
It's an imperfect but useful proxy. It isn't valid as a single number that defines your health, but it correlates very well with obesity. Waist size is probably a more useful metric because of what we've learned about the health implications of visceral fat.
The number of people with a BMI over 30 who aren't obese is extremely small and they're very easy to spot. Arnold Schwarzenegger and Lee Haney won Mr Olympia with BMIs of around 30. Dorian Yates and Ronnie Coleman had off-season BMIs of over 40, but they were the most muscular men in recorded history.
IMO, "BMI is meaningless" is the new "I'm just big boned". The select group of people who really are lean at >30 BMI know that they're a healthy weight, while a much larger group of people are simply using the imperfection of BMI as a means to delude themselves. I'm happy to be corrected, but I haven't heard of a single incident of someone being incorrectly categorised as obese and advised to lose weight by a medical professional simply because of their BMI.
The linked article talks quite a bit about waist fat being explicitly bad. That's an interesting measure for me (the waist/hip ratio) and as someone who's currently hovering around a BMI of 26 (and is average fitness), I plan on making that waist/hip measurement when I get home tonight and seeing how measure relative to their research.
I've been planning to start a bit of a weight loss program (primarily food focused with maybe a 10-20% exercise increase) and if there's specific concerns related to waist fat then I'll do some more research on how to address that.