In reply to Reach>Talent:
From that BBC article, it seems odd to me that he talks about the problems with BMI being a single number that doesn't describe individuals at all well, and then proposes a slightly different single number that doesn't describe people at all well. "I've identified the main problem here, and come up with a solution that doesn't bother to address that at all."
Odder, further down there's a mention of MRI and DEXA scans with regard to body fat percentage, and then nothing else as if those very precise (and very expensive) methods are the only way to even vaguely measure it.
If you wanted to change the way BMI works to make it reflect the physique of an individual better, why not chuck in another measurement, like waist circumference, to get an idea of body shape? Or keep the existing BMI (which seems to work just fine for considering populations) and add a simple approximate measure of body fat (like the "US Navy" approach) when dealing with individuals?
For medical professionals dealing with individual people though, the best advice probably comes from The Mash:
http://www.thedailymash.co.uk/news/health/body-mass-index-replaced-by-just-...