As a side note, BMI is a terrible measure of obesity for outliers. I have several first hand experiences where the BMI indicates obese or even morbidly obese, while the DEXA scan would show fat in single digits.
BMI is a terrible measure in general, not just for outliers. It doesn't normalize well to fairly obvious factors. You could arguably get a better reading just by looking at someone and visually estimating how fat they are.
Waist measurement is a stronger measure but still flawed.
Yea, ask many athletes at the US military academies about this. All of the football players basically had to have a waiver for BMI for the military standards or had to get taped (body dimensions measured) to do a more-complex calculation to determine a more accurate fat content. They also had to make standards in off-season, so the big linemen were constantly eating before the season to put on weight, then working hard at the end of the season to lose it.
That's basically tautological. Outliers are outliers because they don't fit the model.