Is BMI actually a useful measure?

Body mass index divides weight in kilograms by height in meters squared, producing a number sorted into categories: under 18.5 underweight, 18.5 to 24.9 normal, 25 to 29.9 overweight, and 30 and above obese. It was devised in the 1830s by Adolphe Quetelet, a Belgian mathematician studying population averages, explicitly as a statistical description of groups rather than a clinical measure of individuals, and it acquired its medical role much later largely because it is cheap and easy to calculate.

Its limitations are substantial and well documented. It cannot distinguish muscle mass from fat, so muscular athletes routinely classify as overweight or obese. It says nothing about fat distribution, and visceral fat around the organs carries considerably more metabolic risk than subcutaneous fat, which is why waist circumference and waist-to-height ratio predict cardiovascular and metabolic outcomes better. Its category thresholds were derived largely from populations of European descent and perform differently across ethnic groups, with several health bodies recommending lower thresholds for people of South Asian descent because risk appears at lower BMI values.

Where it remains useful is at population scale, tracking trends across large groups over time, which is what it was built for. Clinically, the American Medical Association adopted policy in 2023 acknowledging its limitations and recommending it be used alongside other measures rather than alone. Better individual indicators include waist-to-height ratio, with a common rule of thumb being that waist circumference should be less than half your height, along with blood pressure, lipid panel, blood glucose, and fitness measures. The practical point is that a number derived from two measurements cannot capture body composition, and treating it as a verdict on health rather than a crude screening input misuses it in both directions.