A few weeks ago, we had a seminar with a case study (an invented case) about a woman with a BMI of 20. The medical students presenting this case described this as a "normal, healthy BMI" and then recommended that she increase her exercise as part of a list of general recommendations for improving health. The seminar leader, Dr. P, a family physician who had written these cases as composites from real scenarios, did not disagree.
A friend of mine with background in eating disorders work did say to the group that a BMI of 20 is very often neither normal nor healthy. Any woman presenting to her family doctor with a BMI of 20 deserves a screening for eating disordered behaviours rather than a blanket recommendation to increase exercise.
Another case during the seminar described a patient with a BMI of 32. At one point there was a list of behaviours that were affecting his health. Obesity was listed. I pointed out to the group that obesity is not, in fact, a behaviour. Then, as the case continued, Dr. P said that the individual's obesity would limit his ability to do his job as a construction worker. I asked why, and was told that he would not be fit enough to do his job. Dr. P did not seem to buy the argument that fitness is not the same as obesity.
The whole thing was incredibly irritating. Note that we're talking about a BMI of 32 here. In fact, every obese person described in the seminar had a BMI of 32 or 33. Apparently people larger than that are too unbelievably enormous to even consider... Thanks to Dee for the photo below, an example of a man with a BMI of 32.
Edited to Add: A number of commenters objected to my statement above about a BMI of 20 triggering screening for an eating disorder. That is not the same as assuming this patient has an eating disorder. She probably doesn't. However, given that anorexia nervosa is a relatively prevalent (0.5 to 1% prevalence means all family practitioners will see several cases) and very severe illness, I do think that every woman with a BMI on the low or low-normal side deserves a screen for eating-disordered behaviours. Screening in this case would be a minute or two spent asking questions about weight loss, amenorrhea, eating patterns, and exercise patterns.
Similarly, I screen everybody with a sleep complaint for major depression, and anybody who has had a traumatic event for PTSD. The vast majority of patients won't have depression or PTSD. But if 1/100 has a good chance of it then these are serious enough risks that I think screening is warranted. One problem with eating disorders is that it's really difficult to find prevalence rates and the studies vary widely about how common these illnesses are. I'd rather err on the side of screening too many people than miss one.
An additional issue here, unique to eating disorders, is that physicians do a pretty poor job of identifying patients until they get really, really sick. Knowing that there's a good chance I am already missing a lot of patients with these illnesses tends to lower my threshold for screening for them.
