One of the diagnostic criteria for anorexia nervosa is "a refusal to maintain body weight at or above a minimally normal weight for age and height (e.g. weight loss leading to maintenance of body weight less than 85% of that expected; or failure to make expected weight gain during period of growth, leading to body weight less than 85% of that expected). Most eating disorder programs compare the patient's Body Mass Index to statistical norms to determine if this criteria is met.
Few therapists or physicians realize however, that the formula used to calculate BMI is fundamentally flawed, and will tend to result in the over-diagnosis of AN especially in shorter individuals. Use of BMI to determine expected weight may explain the increasing incidence of AN being diagnosed in pre-pubescent girls (see below for explanation).
BMI also does not give an accurate esitmate of body fat, since it doesn't account for varing bone density, body type, stage of growth, or genetic variation. BMI was devised to compare norms in various population groups, and should never be used to calculate the ideal weight for an individual. Yet this is commonplace in eating disorder programs.
There are several different types of body fat assessment. Few, if any eatng disorder programs will use these methods, usually claiming these tests are too expensive (this sort of flawed thinking is typical in many hospitals, where physicians try to save a few dollars on diagnsotic tests, only to waste thousands of dollars on treatment or prolonged hospitalization due to misdiagnosis). Methods include isotope dilution, photon absorptiometry, potassium-40, hydrostatic weighing, radiography, ultrasound, nuclear magnetic resonance, total body electrical conductivity, bioelectrical impedance analysis, infrared, and skinfold thickness. Supposedly, the most reliable of the practical methods is hydrostatic weighing (also called underwater weighing). Even with these sophisticated methods, however, it is only possible to estimate a person's body fat percentage.
Flawed Formula Over-Diagnoses AN in Short Individuals
The formula for Body Mass Index in metric units is:
BMI= Weight (in kilograms) / Height(in metres) squared
The formula in Imperial Units is:
BMI = ( Weight (in pounds) / Height (in inches) squared ) x 703
One of the basic flaws in the body mass index formula is that the denominator of height is only squared, which means that a short person will have a much lower BMI than a tall person, even though they may have the same body shape and equal body fat percentages. This is because the formula ignores one fundamental law of physics, that being that volume (and therefore mass & weight) increases by the cube of the scale factor rather than the square of the scale factor.
A Lesson In Lilliputian Physics
In Gulliver's Travels, a political satire and adventure fantasy story by Jonathan Swift, a human being named Gulliver encounters several strange races of people, including the tiny Lilliputians and the giant Brobdingnagians. The Brobdingnagians were proportioned like Gulliver, but were about 10 times the height. Of course these giant people would only be possible in a fantasy world, for being ten times taller, they would be 1000 times heavier, but their bones and muscles would only increase 100 times in cross-sectional area, so their strength to weight ratio would be 1/10th that of Gulliver. They would collapse under their own weight unless their bones and muscles were made of much stronger materials.
If Gulliver was 6 feet tall and weighed 180 lbs., his BMI would be:
[ 180 / (72x72) ] x 703 = 24.4
A typical Brobdingnagian, however even though proportioned identically would weigh 180,000 lbs. and have a BMI of:
[ 180,000 / (720x720) ] x 703 = 244.4
A typical Lilliputian, however, at 1/10th of Gulliver's height, would only weigh 0.18 lbs. and would have a BMI of:
[ 0.18 / (7.2x7.2) ] x 703 = 2.44
Now, although this may seem somewhat ridiculous and irrelevant, let's try scaling Gulliver up by 19% to 7' - 2", to the size of an NBA basketball player. He would weigh 306 lbs. and would have a BMI of:
[ 306 / (86x86) ] x 703 = 29.1
Even though he would have the same proportions and body fat percentage as his 6 foot counterpart, he now would be technically classified as obese.
Now let's try scaling Gulliver down from his 6 foot size by 17% to 5' - 0". He now would weigh 104 lbs. and even though he would look proportionally the same and have the same percentage of body fat, his BMI would now be:
[ 104 / (60x60) ] x 703 = 20.3
If we scaled Gulliver further down to 4' - 2", he would now weigh 60 lbs. and have a BMI of:
[ 60 / (50x50) ] x 703 = 16.9
You can see that if Gulliver's therapist relied on Body Mass Index to determine if Gullliver was of normal weight, the 4' - 2" Gulliver would likely have been diagnosed with an eating disorder, even though he had the same proportion of body fat and looked proportionally the same as the 6 ft. Gulliver. Of course, since Gulliver is a male, and likely not too concerned about being overweight, he would not meet all the diagnostic criteria for AN. If Gulliver was a female however, and therefore likely to be concerned about her weight, she could be diagnosed with AN, except if she was still menstruating. If, however, she had a polycystic ovary or ovarian tumour that caused amenorrhea, she would meet all the criteria for Anorexia Nervosa.
The main point of this exercise is to point out that many individuals (especially short females) may be diagnosed with AN even though they may be of a healthy weight, if BMI is used as an indicator of nutritional status. More reliable methods should be used to assess body composition. Short individuals may have a low BMI yet have perfectly healthy percentages of body fat. As an example, two of my neices aged 7 and 10 are fairly petite, and both have body mass indexes of 15.7. Yet they are healthy, well nourished, and eat very well. Such is their genetic heritage.
It is entirely possible that given the subjective nature of the diagnostic criteria and the flawed methods of assessing nutritional status, eating disorder clinics may have patients ranging from short, perfectly healthy individuals with a low BMI who respond well to cognitive behavioral therapy, to severely malnourished patients with digestive tract disorders who don't respond at all to psychotherapy.
Failing to differentiate between the two may convince the therapist that because they can successfully treat the former, they should be able to achieve success with the latter, given enough time.
This may partially explain why some patients respond well to nutritional counselling, while others struggle for years, unable to eat enough to meet their caloric requirements. Since little diagnostic testing is likely to have been performed, the therapist will likely view the patient with an undiagnosed digestive disorder as a difficult case, who may take years to achieve "objective reality". Such was the case with my daughter.
What is your experience with Body Mass Index? Was BMI used to determine what your weight should have been? Contact us and let us know!