BMI is your weight in kilograms divided by the square of your height in metres. In imperial units, multiply your weight in pounds by 703 and divide by your height in inches squared. It is quick, it is free, and it was never designed to tell an individual person how healthy they are.
What the categories say
Under 18.5 is underweight. From 18.5 to 24.9 is the normal range. From 25 to 29.9 is overweight. Anything from 30 up is classified as obese.
Those bands come from population data, where they do a reasonable job of predicting risk across large groups. Across a country, BMI tracks with rates of heart disease and diabetes well enough to be useful to public health researchers.
The problem is what happens when a number built for populations gets applied to one body in a doctor's office.
It cannot tell muscle from fat
BMI knows two things: your height and your weight. It has no way of knowing what the weight is made of.
A muscular athlete and a sedentary person of identical height and weight receive the same score. Plenty of rugby players and weightlifters land in the obese category with body fat percentages in single digits. The number is not measuring what it appears to measure.
The reverse case is more common and more dangerous, because it is invisible. A person can sit comfortably in the normal band while carrying very little muscle and a fair amount of fat, particularly around the middle. Nothing about their BMI raises a flag, and their metabolic risk may be worse than someone in the overweight band who trains regularly.
It has no idea where the fat is
This matters more than the total. Fat stored around the organs, called visceral fat, carries a much stronger association with heart disease and type 2 diabetes than fat stored under the skin on hips and thighs.
Two people with the same BMI and the same body fat percentage can carry it in different places and face different risks. BMI cannot see any of that. A tape measure around the waist can, which is why waist-to-height ratio does better as a one-number screen despite being just as cheap.
It works less well for some groups than others
Age is one. Older adults lose muscle and bone density, so a stable BMI can hide a real shift in composition. Children need age and sex-specific percentile charts, and adult categories do not apply to them at all.
Sex is another. Men and women carry different amounts of muscle and fat at the same BMI, and the same cut-offs are applied to both.
Ethnicity matters too. Health risk begins at lower BMI values in some populations than in others, and several health bodies use adjusted thresholds for South Asian populations for exactly this reason. [ADD SPECIFIC: the adjusted threshold your audience needs, with the issuing body]
Very tall and very short people also get distorted results, because the formula squares height while human bodies do not scale that neatly.
What to use instead
Nothing alone. Use two or three cheap measures together and you get a far better picture than any single number offers.
Waist circumference, or better, waist divided by height. Keep it under half your height and you have cleared the most useful threshold there is.
A body fat estimate, if you want composition. Our guide to measuring body fat covers the options and what each is worth.
And the things that are not measurements at all: how much you move, what your blood pressure and blood markers look like, how well you sleep, whether you can carry your shopping up two flights without stopping.
So is BMI useless
No, and it is worth being clear about that, because the backlash has overcorrected.
BMI is a fast screening tool. If yours is 35, that is meaningful information regardless of your muscle mass, and no amount of context makes it disappear. If yours is 26 and you lift three times a week, it means very little on its own.
Treat it as one number among several. The right response to a BMI outside the normal range is to look further, not to panic and not to dismiss it. And if you want a number that actually changes what you do this week, your maintenance calories are far more actionable than a ratio of your weight to your height.