Eyeing the BMI

November 2nd, 2009

The BMI does not ‘need to be scrapped,’ as some critics state. Some applications are inappropriate, but it fills an important and irreplaceable function when used properly.

b m why?

the BMI (body mass index) is a rough measure of health that takes height and weight as input, and produces a number score.

BMI = weight(kg)/height(m)2

A score between 18.5 – 24.9 is associated with lowest health risks. Lower and higher scores are considered underweight, and overweight, respectively. A score above 30 is considered obese.

The primary complaint voiced about BMI is that since extra weight can be caused by either muscle or fat, that a healthy muscular person is mistakenly categorized as overweight or perhaps even obese. It has been proposed that this index therefore be scrapped entirely.

I disagree that this is a flaw serious enough to render the index useless, and believe the BMI should remain a tool for identifying people at risk of health complications related to weight.

there is no i in bmi

A fresh set of criticisms were recently inspired by discussion brought up during US healthcare reform process. Safeway has a private health plan that offers employees an incentive for improving their BMI scores. Congress may or may not adopt a similar strategy, but in any case, criticism seems to be focused on the fact that some healthy people will fall outside the preferred range and may be penalized.

The approach used by Safeway is that if an employee is – say, deemed to be overweight with a BMI of 38 – he will pay a certain premium for being higher risk. If he drops BMI below 30, his premium will become more affordable. The first criticism is that some healthy-but-muscular people are being unfairly penalized.The second criticism is that using thresholds can reward small gains (such as decreasing BMI from 30.0 to 29.9). The typical advice when confronted with these concerns is often “so the BMI is useless.”

I actually agree with these criticisms of Safeway’s plan, but do not believe that they mean the BMI is useless – just that BMI should not be used in isolation as a single metric for health risk.

And you know what – health experts are pretty clear that BMI is not to be used in isolation as a single metric for health risk.

Safeway’s system is crude and some outliers may be treated unfairly. That’s a business decision that is misusing a health metric. It does not indicate a flaw in the BMI.

BMI is a rule of thumb, not a diagnostic tool. It should also be used with a waist circumference measure. These two quick-and-dirty measurements will provide a very good picture of health risk. In addition, the most common way to determine body fat content these days is a Tanita scale. The result is ‘percent body fat’ – they’re easy to use and getting very affordable. Guess what: almost everybody with BMI over 30 has a large waist circumference and high body fat percentage.

The number of people who are over 30 BMI because they’re built like a brick shithouse is so miniscule that I can’t imagine why somebody would think this constitutes a ‘flaw’. These people know who they are, their doctors know who they are…

To my knowledge, nobody who is lean and muscular has ever been ‘accidentally mistaken for obese’ during an annual doctor’s visit. Doctors do have eyes, and are not unthinking robots who walk back and forth between the scale and BMI wall chart. But if this is your doctor, and you are stripped down to your Speedo showing ripped abs but your doctor still tells you you are obese… the BMI is not the problem.

weight for me

There is one legitimate criticism that I can support: the threshold at 25 appears to be more arbitrary in retrospect.

At least one well-designed study actually seems to show that the BMI segment between 25 and 30 (overweight) has slightly less risk than that of the 20-25 (normal) segment. This is plausible, as the threshold of 25 was selected for being a round number. (The same study, however, did show very decisively that those with BMI above 30 were at significantly high risk.) My opinion is that we should be paying more attention to the 30 threshold than the 25 threshold.

Furthermore, my opinion is that the ‘normal’ range should be extended to a higher score, to be determined with statistics. eg: if the risk decreases until 27 and then starts to increase, that would be an appropriate score to use for goalsetting.

once a mighty athlete

My impression is that the resistance to BMI is just ordinary cognitive dissonance or denial. Middle age creeps up on us all and the body that spiked the winning touchdown in senior year has evolved gradually into a middle-aged spare tire rack – and that’s hard to admit.

We want to raise healthy kids, but it’s hard with fast food and Nintendo, so it’s easier to convince ourselves that Junior is big because he’s so healthy.

If BMI says obese, then either we’re obese… or BMI is wrong. When confronting ourselves with this dichotomy, we choose the option that presents ourselves in the best light. (It’s the same mentality that leads some of my colleagues to reassure themselves that their kid failed grade 9 because he’s too smart and the education system must be the problem.)

Another factor is the logical fallacies of ‘false dichotomy’ or ‘unreasonable expectations’ that contribute to rejecting the BMI because outliers are misjudged. Indexes are not intended to be ‘all or nothing’ – the benefit is that most people’s situation is accurately reflected.

Lastly, there’s a criticism about the fact that risk is chopped up into intervals with thresholds. A person whose BMI is 30.0 is ‘obese’ while a person whose BMI is 29.9 is ‘overweight’. Again: this is fuelled by a logical fallacy. In this case, corrupt continuum. Many health risks are sliced into intervals for simplicity. For example, normal blood pressure is between 90/60 and 140/90. Nobody is lobbying to scrap blood pressure testing.

references:

Posted in Health Fraud, Logical Fallacies, Scientific Errors | No Comments »

« Older Entries