BMI: What It Actually Measures, and What It Completely Misses

Body Mass Index is the most widely used health metric in the world and one of the most widely misunderstood. It is useful - but only if you know precisely what it does and does not tell you.

The formula

Metric:    BMI = weight(kg) / height(m)²
Imperial:  BMI = 703 × weight(lb) / height(in)²

That is the whole thing. Two inputs. No age, no sex, no body composition, no ethnicity, no measurement of where fat sits on your body.

Standard WHO categories:

It was never designed for individuals

BMI comes from Adolphe Quetelet, a Belgian mathematician working in the 1830s. He was not a physician and had no interest in diagnosing anyone. He was developing statistical descriptions of populations - the "average man."

The index entered clinical use in 1972 when Ancel Keys published a study comparing several weight-height ratios and found Quetelet's index correlated best with body fat at the population level. Keys explicitly warned it was inappropriate for individual diagnosis. That caveat did not survive the next fifty years.

The reason it spread is entirely practical: it needs a scale and a tape measure. Nothing else comes close on cost and scalability, which is exactly what public health surveillance requires.

The four real failures

1. It cannot distinguish muscle from fat

BMI measures mass, and muscle is denser than fat. A 6'0", 220 lb athlete at 10% body fat has a BMI of 29.8 - "overweight," bordering obese. A 6'0", 220 lb sedentary person at 35% body fat has exactly the same BMI and a completely different health profile.

This is not a rare edge case. It systematically misclassifies athletes, manual laborers, and anyone who lifts.

2. It ignores fat distribution

This is the more important failure. Visceral fat - the fat packed around your liver, pancreas, and intestines - is metabolically active and drives insulin resistance, inflammation, and cardiovascular risk. Subcutaneous fat on hips and thighs is comparatively benign.

Two people at BMI 27 can have opposite risk profiles depending on whether that mass sits around the waist or the hips. BMI cannot see the difference. Waist circumference can.

3. It scales wrongly with height

Dividing by height squared is a mathematical convenience, not biology. Human bodies scale closer to height cubed in some dimensions and to a fractional exponent in others. The practical result is that BMI systematically over-reports for tall people and under-reports for short people. Some researchers propose height^2.5 as a better fit.

4. Thresholds are not universal across populations

Cardiometabolic risk appears at lower BMIs in South Asian, Chinese, and other Asian populations. The WHO and many national bodies recommend lowered action points - commonly 23 for overweight and 27.5 for obesity - for these groups. Using the standard 25/30 cutoffs under-diagnoses risk substantially.

There is also the "normal weight obesity" phenomenon: people within the normal BMI range who carry high body-fat percentage and low muscle mass, and who show elevated metabolic risk that BMI reports as perfectly fine.

Better measurements, ranked by effort

Waist-to-height ratio - best value for zero cost. Measure your waist at the navel. Divide by your height in the same units.

• Below 0.4 - possibly underweight
• 0.4 to 0.49 - healthy
• 0.5 to 0.59 - increased risk
• 0.6 and above - high risk

The rule of thumb: keep your waist under half your height. Multiple studies find this predicts cardiometabolic outcomes better than BMI, and it needs only a tape measure. The UK's NICE guidelines now recommend it alongside BMI.

Waist circumference alone. Risk thresholds commonly cited are >40 inches (102 cm) for men and >35 inches (88 cm) for women, with lower thresholds for Asian populations.

Body fat percentage. DEXA scans are the practical gold standard. Bioelectrical impedance scales are cheap but vary meaningfully with hydration - useful for tracking a trend over months, not for a single reading.

Blood markers. Fasting glucose, HbA1c, triglycerides, HDL, and blood pressure describe metabolic health directly rather than inferring it from shape. If you only get one thing from this article: these numbers matter more than any body measurement.

So is BMI useless?

No, and the backlash overcorrects. BMI remains genuinely useful for:

• Population studies. Tracking obesity prevalence across millions of people, where individual error averages out.
• A cheap first screen. BMI over 35 is meaningful information regardless of muscle mass. Nobody reaches 40 through lean tissue.
• Tracking your own trend. If your height is constant and your training is constant, BMI change is weight change, which is worth knowing.
• Clinical thresholds. Some medication dosing, surgical eligibility criteria, and insurance calculations use it, so the number has practical consequences whether or not it deserves them.

The correct framing is that BMI is a screening flag, not a diagnosis. A concerning BMI is a reason to measure your waist and get bloodwork. It is not itself a conclusion.

A five-minute self-assessment

1. Calculate your BMI with the BMI calculator - note it, do not react to it
2. Measure your waist at the navel, standing relaxed, exhaling normally
3. Divide waist by height - is it under 0.5?
4. If either number is elevated, book bloodwork: fasting glucose, HbA1c, and a lipid panel
5. Track all of it every three months, not daily

The trend across quarters tells you far more than any single reading.

Key takeaways

• BMI is a 190-year-old population statistic that was explicitly not designed to diagnose individuals
• It cannot separate muscle from fat, and it is blind to where fat is stored
• Visceral fat around the waist drives most of the metabolic risk, and waist-to-height ratio captures it for free
• Keep your waist under half your height - a better single indicator than BMI
• Asian populations face elevated risk at lower BMI; thresholds around 23 and 27.5 are more appropriate
• Blood markers describe metabolic health directly and beat every body-shape metric

General health information, not medical advice. Discuss your numbers with a qualified clinician.