2015 has arrived, and with it the tide of New Years’ resolutions that are sure to be broken or abandoned in its first few months. I’m still deciding on mine (perhaps something about procrastination), but no matter what I choose, I know I need some way to define success. Resolutions fail for many reasons, but not having any way to measure how you’re doing can quickly end your promise to be kinder this year.
Systems of measurement, and accurately communicating those measurements, are a vital part of science. A lack of understanding of concentration measurements has fueled public confusion and fear about vaccines, and even highly trained scientists make measurement mistakes. In 1999, unit conversion errors led to the loss of NASA’s Mars Climate Orbiter. Being aware of the pitfalls of measurement makes me especially uncomfortable whenever someone proposes one of the most common New Years’ resolutions: weight loss. Discussions about obesity today often focus on the “ideal weight” for your height, which, whether you know it or not, comes from the body mass index, or BMI. The World Health Organization uses BMI when it reports about worldwide rates of obesity and to define categories of being over- or underweight. All you need to calculate BMI is your mass and your height, making it easy for governments or research bodies to use to talk about countries and worldwide trends, but for individuals it’s a really poor way to talk about weight and health.
Let’s explore why, exactly, you shouldn’t use BMI to measure your 2015 fitness goals; it all starts with a sordid past you’ll find hard to believe…
Statistics, social science, and insurance salesmen
The first people to really become concerned with obesity as a health risk were 19th century insurance brokers. Prior to that, being heavy was largely regarded as a sign of good health. Fat is an advantage when food is unreliable, but carrying extra weight can increase a person’s risk of hypertension and heart failure, diabetes, sleep apnea, osteoarthritis, and limit mobility and enjoyment of life. Insurance brokers noticed a correlation between their heavier policy holders and early death, so they began developing tables of average heights and weights to use when selling insurance, an early application of statistical methods. That’s right – the earliest versions of the BMI charts we use today were first conceived as a money-saving effort by insurance brokers, not doctors.
Describing weight got a little more sophisticated in the 1830s with the help of Belgian polymath Adolphe Quetelet. Quetelet (1796-1874) was very interested in the growing field of statistics, and in how a representative sample can be used to draw conclusions about a population. He studied people, their physical attributes and demographics, and used mathematical analyses to draw conclusions about the “normal man.” His studies, published in 1835 in a book title “A Treatise on Man,” is considered by many to be the founding of the social sciences. Quetelet’s application of statistics to social conditions inspired many other scientists, including pioneering nurse Florence Nightingale, who used Quetelet’s statistical methods to argue for changes in nursing practices.
In “A Treatise on Man,” Quetelet considered the development of weight and height from childhood to adulthood. He states, “If man increased equally in all dimensions, his weight at different ages would be as the cube of his height.” In analyzing his data, Quetelet considered three ratios: weight over height, weight over height squared, and weight over height cubed. Queletet concluded that, for in the populations he had studied, body weight was proportional to the square of the height, and that ratio became known as the Quetelet Index. While Quetelet wasn’t trying to describe obesity, his Index was a useful way to describe populations and make predictions about health risks because it helped account for differences in weight due to height. This was a powerful tool for the development of modern medicine, but it wasn’t intended as a guide for an individual to use to measure his own fitness level.
All the fat that’s fit to measure
In 1971, Ancel Keys published a comparative study that considered all of the available ways to talk about relative weight and obesity. Key’s analysis looked at many different ways of measuring obesity and compared them based on their ease of use, how well the method removed the dependency of weight on height, and how well the method indicated “body fatness.” Keys considered several more direct measurements of fatness, like body density (which requires each person to be measured under water) and skin fold thickness (measured by calipers). Both methods require more work to record – no one’s going to jump in a pool to measure body density for their passport, while height and weight are easily available.
Plus, both measurements make assumptions about our bodies. Body density doesn’t say anything about the composition of your body (fat, muscle, and bone all have different densities), and skin fold thickness assumes that fat is distributed evenly over the body, which it is not. Keys looked at the same ratios of weight and height that Quetelet considered, and a few additional ones including the ponderal index (height over the cube root of weight), and applied them to several existing samples (totaling measurements from 7426 men). Of all the ratios considered, Quetelet’s Index was the least correlated with height, and was reasonably well correlated with both body density and skin fold thickness (for the studies where those values were available). Keys proposed that Quetelet’s Index be renamed the body mass index, and recommended that it be used for studying populations. BMI was on its way to becoming the default for researchers looking at height and weight data from many people.
In his conclusions, even Keys himself noted that BMI and other descriptions of relative weight and obesity can break down when applied to individuals. Commenting on the general trend for people to gain weight as they age, Keys observes, “The characterization of persons in terms of desirable weight percentage has resulted in attributing to ‘overweight’ some tendencies to ill health and death that are actually only related to age.” BMI is usually recommended only for use for people age 20 and up, who have largely finished growing in terms of height, but it also can lose accuracy for the elderly. A low BMI, rather than a high one, is more strongly correlated with mortality for people over the age of 60, and osteoporosis studies suggest that elderly people with higher BMI (“overweight” or “obese” ratings) may have a reduced risk of bone mineral loss and increased bone fragility that leads to osteoporosis.
BMI also breaks down for athletes. Muscle is more dense than fat, so athletes often rate higher on BMI because of their increased muscle mass. Take for example Arnold Swarzenegger at age 21, who at the peak of his time on the body building circuit weighed 250 pounds and stood 6 feet 2 inches tall, giving him a BMI of 32.1, or obese. Not something you want to tell Mr. Universe!
While BMI clearly has problems, in 1997 the World Health Organization endorsed its use. In 2000, the WHO issued a technical report on obesity, which explains in detail why BMI should be used and notes many of the problems that we’ve discussed here. In discussing all of its cautions for the use of BMI, the WHO calls BMI “the most useful, albeit crude, population-level measure of obesity.” The WHO report goes on to note that ethnicity is also a significant factor, and that health risks will vary based on heritage. A later WHO report, published in the Lancet in 2004, shows that risks associated with obesity, like type 2 diabetes and cardiovascular risks, didn’t line up with the cut-offs points for “normal” weight and “overweight” BMIs when considering Asian and Pacific populations.
BMI, staging systems, and you
Measuring the height and weight of many people is easy and cheap, and BMI provides a great measurement to evaluate and track large populations like a country or a continent, but there are many other factors to consider when talking about an individual person’s BMI and their health. When medical professionals evaluate whether a patient’s weight warrants intervention, there are several factors to consider beyond BMI. AM Sharma from the University of Alberta Department of Medicine, and RF Kushner of the Northwestern University Feinberg School of Medicine, reviewed the limitations of BMI and proposed a staging system for obesity treatment in a 2009 paper. Staging systems use a variety of factors to describe a patient’s condition, often including quality of life and the presence of related diseases or disorders, called comorbidities. Similar staging systems are already used for describing cancer progression, congestive heart failure, and kidney failure.
Sharma and Kushner argue that applying BMI to individuals has limited use for doctor’s decisions, stating that “individuals with the same BMI value can have an almost twofold difference in total body fat, whereas conversely, individuals with the same amount of total body fat can present with a wide range of BMI.” Sharma and Kushner advocate instead for including other risk factors when evaluating obesity, by including obesity-related risk factors (hypertension, diabetes, sleep apnea), physical symptoms (fatigue, aches and pains), psychological symptoms, and functional limitations. The paper presents a few case studies, pairing the WHO classification of weight with a stage rating ranging from 0 to 4. A woman with a BMI of 32 but no other risk factors would be classified by this system as Class I, Stage 0, and does not need medical intervention, while a man with a BMI of 36 who also has hypertension and sleep apnea would be classified as Class II, Stage 2, and intervention would be recommended.
Sharma and Kushner also warn against using changes in BMI as a measure of success, noting, “Change in obesity class does not necessarily imply improvement or deterioration in overall health or well being. Conversely, relatively small changes in weight of only 5-10%, although associated with significant health benefits, may not be reflected by changes in obesity class.”
Of course, most of us aren’t able to use this staging system on a daily basis to measure our fitness goals, but it’s a useful reminder of all the other aspects of health and weight loss beyond the number on the scale. Knowing all the ways that BMI breaks down when we apply it to ourselves is great start for changing how we think about weight loss goals. So, if you’re feeling regret after stuffing yourself on holiday meals and thinking that 2015 is going to be the year that you lose 20 pounds or that you get your BMI down to under 25, considering picking a different metric for your fitness goals. Resolving to walk 10,000 paces a day, to spend more time playing outdoors with children, pets, or friends, or to find other ways to measure your health without weight loss as a proxy can be much more sustainable and rewarding than fighting a number. Happy New Year!