Last updated: 2026-05-22

Body Mass Index

Body Mass Index (BMI) is a screening index defined as weight in kilograms divided by height in metres squared. It appears on physical exam reports, insurance underwriting forms, public-health dashboards, and clinical trial inclusion criteria. BMI is best understood as a population-level screening tool: inexpensive, reproducible, and useful for identifying trends — but imprecise at the individual level.

This calculator returns BMI from height and weight, a reference weight (defined as BMI = 22, the JASSO value associated with the lowest morbidity risk in Japanese cohort studies), and the WHO classification.

Formula

BMI is defined by a single ratio:

The formula was published in 1832 by Adolphe Quetelet, a Belgian astronomer and statistician who was looking for a quick way to characterize the "average man" in population studies — not to diagnose individuals. The metric was rebranded "Body Mass Index" in 1972 by physiologist Ancel Keys, who validated it as a reasonable proxy for body fatness in large populations. The WHO adopted it as an obesity screening standard in 1995.

WHO Categories

These thresholds are most commonly cited from US/European data and reflect cardiovascular and metabolic risk in those populations.

Population-specific cutoffs

The same BMI corresponds to different health risks across ethnic groups. South and East Asian populations tend to develop diabetes and cardiovascular disease at lower BMI values than European populations. The WHO recognizes this and offers an alternative set of cut-offs for Asian populations: overweight at 23, obese at 27.5. Singapore, China, and Japan use modified thresholds in their national guidelines. Because of this dependence, a BMI value is most meaningfully read against the table for the population it was derived from.

Limitations

BMI is a screening index, and several well-documented factors limit its accuracy at the individual level:

• Muscle vs. fat blindness. A 90 kg bodybuilder and a 90 kg sedentary office workerA 200 lb bodybuilder and a 200 lb sedentary office worker of the same height have the same BMI but radically different health profiles. Athletes routinely score "overweight" or "obese" by BMI while having body-fat percentages in the single digits.
• Fat distribution matters. Visceral fat (around the organs) is more metabolically dangerous than subcutaneous fat. BMI cannot distinguish them. Waist circumference, waist-to-hip ratio, or DEXA scans add information BMI lacks.
• Age effects. Older adults tend to lose muscle and gain fat at the same total weight (sarcopenic obesity). BMI may stay the same or even drop while body composition worsens.
• Children use different metrics. Pediatric BMI is interpreted via age- and sex-specific percentiles, not the adult cut-offs above.
• Pregnancy. Standard BMI categories do not apply during pregnancy; obstetric weight-gain guidelines use pre-pregnancy BMI and are tracked separately.

Because of these gaps, BMI alone is an inadequate basis for a clinical decision about overweight or obesity. It functions best as a prompt for further assessment — waist circumference, body composition measurement, and clinical history — rather than a final verdict.

Applications

Personal orientation

A BMI reading places an individual's weight-for-height on the WHO scale and, where applicable, shows the gap to the reference weight (BMI = 22 or the healthy-range bounds). This snapshot is most useful as a starting point before discussing body composition with a clinician.

Tracking weight changes over time

A single BMI reading carries substantial day-to-day variability; a trend across months or years is more informative. Recording BMI alongside weight normalises comparisons across people of different heights — useful in fitness programmes, family discussions, or longitudinal personal logs.

Administrative forms

Health-insurance applications, blood-donation eligibility, and clinical trial enrolment criteria frequently specify BMI thresholds. For legal or clinical purposes, values measured in a supervised setting are what count — not self-reported figures.