Your BMI number, explained — and why it's often misleading

The Body Mass Index was invented in 1832 by a Belgian astronomer, not a doctor. He was trying to describe statistical averages across populations, never individuals. Two centuries later it's the first thing most clinicians write down — a single number that ignores muscle, bone, body composition, age, and ancestry. Understanding what BMI actually captures helps you read it for what it is: a starting point, not an answer.

FOLIOKIT · MAY 2026

The man who never meant for any of this

Adolphe Quetelet was a 19th-century Belgian polymath — astronomer, mathematician, sociologist. In 1832 he published work on what he called l'homme moyen, the "average man." He was attempting something new: applying statistics to human populations the way they'd just been applied to celestial mechanics. Quetelet noticed that, across populations of healthy adults, weight scaled roughly with the square of height. So he proposed dividing weight by height squared as a way to compare populations.

He wrote, repeatedly and explicitly, that the index was a population-level statistical descriptor. It was not, he insisted, an individual diagnostic measure. He'd be surprised to see it on the back of a doctor's clipboard.

The formula sat unused outside academic circles for 140 years. In 1972 the physiologist Ancel Keys — investigating heart disease in the famous Seven Countries Study — needed a quick proxy for body fat. He compared several formulas, picked Quetelet's, and renamed it the "Body Mass Index." That paper named the index. It did not validate it as a diagnostic.

The math

The 703 in the imperial formula is just the unit conversion. There's no biological insight in it. Different countries kept different units; the math is the same.

The WHO categories

The World Health Organization adopted these cut-offs in the late 1990s. They came from epidemiological studies on Western European and North American populations:

The boundaries are not magic numbers. They were chosen as round figures that produced reasonably continuous correlations with cardiovascular outcomes in the populations studied. They are guidance — and the WHO's own technical report acknowledges they should be interpreted differently across populations.

What BMI does not see

BMI looks at one ratio. It cannot distinguish between two bodies of the same weight and height that are very different on the inside. The well-known limitations:

Muscle vs. fat

A pound of muscle and a pound of fat weigh exactly the same — but muscle is denser. A muscular person and a sedentary person of identical height and weight have identical BMIs and very different bodies. NFL running backs and Olympic rowers routinely register as "overweight" or "obese" by BMI. So do many serious recreational lifters.

Body composition

Two people can have a normal BMI and very different fat-to-lean ratios. A thin, sedentary person with little muscle (sometimes called "skinny fat" in casual use, "normal-weight obesity" in literature) can carry significant cardiometabolic risk that BMI completely misses.

Age

People naturally lose muscle and bone density with age. An 80-year-old at "normal" BMI may have far less lean mass than the formula assumes. Geriatricians often advise different cut-offs for older adults; some research suggests slightly higher BMIs are protective in later life.

Ancestry

The cut-offs were derived from European and North American populations. People of South Asian and East Asian ancestry tend to develop metabolic risk at lower BMIs; the WHO has issued separate Asian-population guidelines suggesting "overweight" begins closer to 23. People of Black African ancestry tend to have higher lean mass at a given BMI; their cardiometabolic risk profile at any given BMI is different again.

Height

The square-of-height formula slightly under-estimates BMI for very tall people and over-estimates it for very short people. There are alternative formulas (Trefethen's revised index uses height to the 2.5 power) that correct for this — but they haven't displaced the original.

What clinicians actually use

BMI is on every chart because it's free, fast, and only needs a scale and a height ruler. But a clinician evaluating real metabolic health will look at additional measures, often together:

• Waist circumference — visceral fat (the kind around your organs) carries far more risk than subcutaneous fat (under the skin). A waist measurement captures this; BMI doesn't.
• Waist-to-hip ratio — body shape matters; "apple" distribution carries more risk than "pear."
• Body fat percentage — measured via DEXA scan, hydrostatic weighing, or bioelectrical impedance. More accurate than BMI, less accessible.
• Blood markers — fasting glucose, HbA1c, lipid panel, blood pressure. Two people of the same BMI can have wildly different metabolic profiles, and the labs tell the truth.

How to read your number, then

BMI gives you one signal. If it sits comfortably in the "normal" range, that's mild reassurance — but not a guarantee everything's fine, and not a complete picture. If it's outside that range, that's a prompt to think more carefully and possibly talk to a clinician — not a diagnosis on its own. The number is most useful as a tracking metric over time for one person (does it move much, in which direction, alongside what other changes), and almost useless as a comparison between two different bodies.

Quetelet would have agreed. He wasn't trying to tell anyone what to do.