BMI, Honestly: A Useful Number With Named Blind Spots

Body mass index is probably the most quoted health number in the world and the most argued about. It has survived two centuries not because it is precise but because it is almost free: a scale, a tape, one division. That trade — crude but universal — explains both its usefulness and its blind spots. This article gives BMI its fair hearing: what it genuinely predicts at population scale, the people it systematically misreads, and the two extra measurements that together correct most of its errors.

Context worth holding: BMI’s critics and defenders mostly talk past each other, because one side judges it as an individual diagnosis and the other as a population instrument. It fails the first job and passes the second, and nearly every argument dissolves once the job is specified.

What BMI actually is

BMI is weight in kilograms divided by height in metres squared — a formula from Adolphe Quetelet, a Belgian statistician of the 1830s who was describing populations, not diagnosing patients. The squaring is the clever bit: bodies scale roughly so that weight grows with the square of height across a population, which makes the ratio roughly height-independent. Insurers adopted it in the twentieth century because it predicted mortality curves in large groups better than weight alone. At that job — cheap risk-sorting of thousands — it remains respectable.

The four blind spots

Muscle is the famous one: lean, dense tissue weighs more per volume than fat, so strength athletes and manual workers routinely score ‘overweight’ while carrying little fat. Height extremes come next — the squared formula over-penalises the very tall and flatters the very short. Age matters too: older adults lose muscle and height, so the same BMI hides more fat at 70 than at 30. Finally, fat distribution: visceral fat around the organs drives much of the risk BMI is proxying, and two people with identical scores can carry it very differently. None of this makes BMI useless; it makes it a first question, not a final answer.

The two cheap corrections

Waist circumference is the strongest simple partner: measured at the navel, it tracks the risky visceral depot far better than weight does, and risk climbs notably above roughly 94 cm for men and 80 cm for women in many guidelines. The second correction is the tape-method body-fat estimate — the Navy method our calculator uses — which converts waist, neck and height into a composition figure. When BMI and the tape agree, believe them. When they argue, the tape is usually describing the individual in front of it, while BMI is describing the crowd that individual resembles.

Ethnicity, and the case for adjusted bands

Body composition at a given BMI differs systematically between populations. Research — much of it on South and East Asian cohorts — shows higher body-fat percentages and higher metabolic risk at lower BMIs than in European cohorts, which is why several health bodies use lower action thresholds (increased risk from about 23, high risk from about 27.5) for people of South Asian, Chinese and some other backgrounds. Conversely, some Pacific Islander populations carry more lean mass at the same BMI. The single universal band is administratively tidy and biologically approximate; if your background is one the research flags, the adjusted thresholds are the more honest reading of the same number.

Our calculator prints the standard bands because they remain the global reference, but the article you are reading is part of the instrument: a BMI near a boundary, read with your background in mind, is the correct way to consume it.

BMI in older age: when the band moves

For older adults the risk curve flattens and shifts: several large analyses find the lowest mortality in the ‘overweight’ band for people over 65, a pattern argued about under the name of the obesity paradox. The likely mechanics are protective reserves during illness and the confounding effect of muscle loss — a 75-year-old at BMI 21 may be under-muscled rather than lean. The practical translation is not ‘gain weight for safety’ but ‘defend muscle and function, and treat unintentional weight loss in older age as a symptom worth reporting’. BMI was never designed to give that advice; it takes a human reading alongside it, which is this page’s recurring sermon.

Children sit outside the instrument entirely — percentile charts, age and sex specific — and our calculator says so wherever it can.

BMI and the insurance table afterlife

One reason BMI refuses to die is bureaucratic: it is embedded in insurance underwriting, drug dosing thresholds, surgical eligibility criteria and public-health targets worldwide. That embeddedness has real consequences at the individual level — a point of BMI can move a premium or gate a procedure — which generates a predictable folk industry of weigh-in tactics and boundary anxiety. The mature response is double: use the number strategically where the system demands it (know yours, know it accurately, know the boundary you are near), and refuse to let the system’s bookkeeping become your self-assessment. The tape and the waist, measured at home for pennies, remain the better mirror.

If a BMI boundary is making a medical decision about you — surgery eligibility is the common case — that is precisely the moment to ask for the composition measurement behind it. Systems built on population tools owe individuals individual evidence, and clinicians will usually provide it when asked with the numbers in hand.

How to use your number

Calculate it — our BMI page also gives the healthy weight range for your exact height, which is more actionable than the score itself. Then measure your waist honestly, in the morning, tape level and not pulled tight. Record both, and re-check quarterly. A BMI drifting down alongside a shrinking waist is unambiguous good news that no blind spot can argue with. A BMI falling while the waist grows is a warning that muscle, not fat, is leaving — a distinction the single number could never have shown you.

Common mistakes

• Weighing in the evening after food and fluid and treating it as the trend
• Letting a single BMI reading make medical decisions
• Comparing your BMI with a differently-built friend’s and concluding anything at all

Questions this raises

Related reading