What Is Body Mass Index and What Is It Used For?

Body mass index is a single number calculated from your weight and height, and it serves as a rough screening tool for whether your body size puts you at higher or lower risk for certain health problems. Doctors, insurers, and public health agencies use it to sort people into categories like “underweight,” “normal weight,” “overweight,” and “obese.” The formula is simple enough that anyone can calculate it at home, which is both its greatest strength and the root of its most serious limitations, because a number built from just two inputs inevitably misses a lot about what is actually going on inside your body.

The Formula and Where It Came From

BMI equals your weight in kilograms divided by your height in meters squared. If you use pounds and inches, multiply the result by 703 to get the same number. A person who weighs 70 kg and stands 1.75 m tall has a BMI of about 22.9. That puts them squarely in the “normal weight” range under current guidelines.

The concept dates back to the 1800s. Belgian mathematician Adolphe Quetelet studied patterns of human growth and observed that in adults, weight tends to increase proportionally to the square of height. His formula, originally called the Quetelet Index, was renamed “Body Mass Index” in 1972 by the American physiologist Ancel Keys.¹ Keys was looking for a simple way to classify weight status in large populations, and the formula stuck. It was never designed to diagnose individual health, which is a distinction that still matters today.

The Standard Categories

The World Health Organization and most national health agencies divide BMI into four main brackets for adults:

• Underweight: below 18.5
• Normal weight: 18.5 to 24.9
• Overweight: 25.0 to 29.9
• Obese: 30.0 and above (often further split into grades 1, 2, and 3)

These categories are widely used in clinical screening, insurance risk assessment, and population-level health surveys.² They are convenient for tracking trends in a population over time. But the cutoffs were developed primarily from data on European-descent populations, which creates problems when applying them globally.

How Well Does BMI Actually Track Body Fat?

The whole point of BMI in practice is to estimate whether someone carries excess fat. So the natural question is how closely the number correlates with actual body fatness when you measure fat directly, using tools like dual-energy X-ray absorptiometry (DEXA) or bioelectrical impedance.

The short answer: reasonably well in large groups, less reliably in individuals. A large study using DEXA-measured body composition found that BMI correlated strongly with fat mass index in both men and women, though the correlation was stronger in women and weaker in older age groups and in Korean participants compared with other ethnicities.³ A cross-sectional study in India found a moderate overall correlation between BMI and body fat percentage, but again a much stronger link in women than in men.⁴ Sri Lankan data showed a similar pattern, with correlation coefficients higher in females.⁵

The consistent finding across these studies is that at the same BMI, women carry more body fat than men throughout adulthood.⁶ A muscular man and a sedentary woman could share the same BMI of 27, yet one has significantly more fat tissue than the other. BMI cannot tell you which is which.

BMI and the Risk of Dying

Much of BMI’s usefulness comes from its relationship with mortality. Across large studies, the link between BMI and death from any cause tends to form a J-shaped or U-shaped curve. Very low BMI carries elevated risk, the middle range is safest, and risk climbs again at higher values, though the steepness of that climb depends on how high you go.

A massive pooled analysis published in The Lancet, covering millions of participants, found that below a BMI of 25, each five-unit drop in BMI was associated with a roughly 19% increase in mortality, while above 25, each five-unit increase was linked to about a 21% increase.⁷ That makes the “normal” range look like the sweet spot, but the details are more interesting than that.

A systematic review and meta-analysis in JAMA found that people in the “overweight” category (BMI 25 to just under 30) had slightly lower all-cause mortality than those in the “normal” category. Even grade 1 obesity (BMI 30 to under 35) showed no statistically significant increase in death risk. It was only at BMI 35 and above that mortality rose meaningfully, by about 29%.⁸ A more recent U.S. analysis using 21st-century data confirmed that mortality risk was similar across BMI values from roughly 22.5 to 30, with a substantial jump, between 21% and 108%, only at BMI 30 and above.⁹

These findings feed into what researchers call the “obesity paradox,” the observation that people who already have chronic diseases sometimes fare better if they are overweight or mildly obese than if they are at a “normal” weight.¹⁰ Whether this reflects genuine protective effects of extra tissue reserves during illness, or biases in how the data are collected, remains debated. But it is one reason clinicians have grown cautious about treating the BMI cutoffs as firm health boundaries.

Why the Cutoffs Don’t Work the Same for Everyone

Standard BMI categories were built mostly from studies of white European populations. When applied to people of other backgrounds, the same number can mean very different things in terms of disease risk.

A WHO expert consultation reviewed evidence from Asian populations and concluded that significant health risks, particularly for type 2 diabetes and cardiovascular disease, begin at BMI levels well below 25. The observed risk thresholds varied from 22 to 25 across different Asian groups, and the high-risk range started between 26 and 31 depending on the population studied.¹¹ A large U.K. cohort study put specific numbers on the gap: to match the diabetes risk that white adults face at a BMI of 30, the equivalent cutoff was 24 for South Asian adults, 25 for Chinese adults, and 26 for Black adults.¹² South Asian individuals in that study were also diagnosed with diabetes at a younger median age than any other group.

This means a South Asian person with a BMI of 25, technically “normal” by standard categories, may carry the same metabolic risk as a white European with a BMI of 30. Several countries, including India, have adopted adjusted BMI thresholds for clinical and public health use to reflect this reality.¹³

BMI in Children and Older Adults

Children and teenagers cannot be evaluated using the same fixed cutoffs as adults, because their body composition shifts dramatically as they grow. Instead, pediatric BMI is interpreted using age- and sex-specific percentile charts. A child’s BMI is compared against a reference population of children the same age, with overweight typically defined as the 85th to 95th percentile and obesity as the 95th percentile and above.¹⁴ Different countries maintain their own growth charts, sometimes with adjusted thresholds for populations with higher cardiometabolic risk at lower BMIs.

At the other end of life, the story flips in a different way. For older adults, research consistently suggests that the “ideal” BMI is higher than the standard 18.5–24.9 range. A meta-analysis of studies in people aged 65 and older found that the lowest mortality risk sat at a BMI between about 27 and 28, and that having a BMI under 21 was associated with at least 28% greater mortality risk compared with the reference range.¹⁵ Another study found that older adults with BMIs below 25 or above 35 faced higher risks of functional decline, falls, and loss of muscle strength, with the optimal range landing around 27–28 for men and as high as 31–32 for women.¹⁶

A study of Polish older adults reinforced this picture, finding a strong inverse relationship between BMI and nutrition-related risk: higher BMI was consistently linked to lower risk of malnutrition in that age group.¹⁷ The practical takeaway is that an older person with a BMI of 26 or 27 is probably in a healthier position than one with a BMI of 22, even though the standard chart would label the first person “overweight” and the second “normal.”

What BMI Misses

The most fundamental limitation of BMI is that it cannot tell the difference between fat and muscle. A competitive athlete with dense muscle tissue and low body fat can easily register an “overweight” or “obese” BMI. Conversely, someone with relatively little muscle but significant fat around internal organs can have a “normal” BMI yet face serious metabolic risks.

This matters especially in older adults, who tend to lose muscle and gain fat as they age, a shift that BMI is essentially blind to. A fat-to-muscle-mass ratio may capture these age-related changes in body composition better than BMI alone, but research on whether such ratios consistently outperform BMI is still developing.¹⁸

BMI also tells you nothing about where your fat is stored. Visceral fat, the deep abdominal fat packed around your organs, is strongly linked to metabolic and cardiovascular problems. Subcutaneous fat, the kind just under your skin, appears to have a different risk profile.¹⁹ Two people with the same BMI can have drastically different amounts of visceral fat, and therefore very different health outlooks.

The scale of misclassification can be staggering. In a study of adults in Saudi Arabia, using standard BMI cutoffs identified about 51% of participants as obese, but when body fat was measured directly using DEXA, the true figure was 92%. Roughly 41% of people were classified as non-obese by BMI when they actually had excess body fat by direct measurement.²⁰ That study was in a specific population, and the gap may be smaller or larger elsewhere, but it illustrates how much individual variation BMI can hide.

Waist Measurements and Other Screening Tools

If BMI is a blunt instrument, what else is available? The two most commonly discussed alternatives are waist circumference and waist-to-hip ratio, both of which capture something BMI does not: where fat is concentrated.

A large cross-cohort analysis published in the Journal of the American College of Cardiology found that among people classified as normal weight or overweight by BMI, having a high waist circumference or waist-to-hip ratio was associated with 15% to 50% greater risk across most cardiovascular outcomes.²¹ In other words, these measurements catch at-risk people that BMI misses. Among those already classified as obese by BMI, having a low waist circumference was linked to similar or even lower risk compared with normal-weight individuals with low waist measurements, which suggests the waist measurement adds genuinely useful information beyond what BMI provides alone.

Consensus reviews treat BMI as valuable for population surveys and initial screening but limited for predicting individual risk of chronic disease or for assessing excess fat in specific patients.²² In clinical practice, many guidelines now recommend combining BMI with waist circumference and metabolic markers like blood pressure, blood sugar, and cholesterol levels to get a fuller picture.

BMI Trajectories Over a Lifetime

One underappreciated aspect of BMI is that a single snapshot in time tells you much less than the pattern over years. A person with a stable BMI of 28 since their twenties and a person who recently climbed from 22 to 28 may have very different health profiles, even though their current number is identical.

Research on life-long BMI trajectories found that patterns of weight change over decades were more predictive of mortality than any single BMI measurement, including maximum BMI or how long someone had been at a given weight.²³ Steady upward trajectories carried different risks than stable higher weights, and rapid drops in BMI, especially in older adults, often signaled underlying illness rather than improved health. This is part of why clinicians increasingly look at weight history rather than a one-time reading.

Weight Bias in the Clinic

Beyond its technical limitations, BMI carries a social dimension that affects real patient care. A systematic review found that weight-biased attitudes among healthcare professionals can negatively affect the patient-provider relationship and the quality of care delivered.²⁴ That bias shows up in shorter visit times, different communication styles, and even differences in which treatments are offered. Patients who experience weight bias are more likely to develop stress and depression, which in turn affect chronic disease risk.²⁵

When BMI is the primary or sole metric a clinic uses, the number on the chart can dominate the conversation in ways that crowd out more nuanced assessments. A patient labeled “obese” by BMI may leave an appointment having been told to lose weight without receiving attention for the issue they actually came in to discuss. This is not an argument against measuring BMI, but it is a reason for both patients and providers to treat it as one data point among many rather than as a verdict.

In schools, BMI screening programs have been tried as a childhood obesity prevention tool. Experimental studies of such programs have not demonstrated a positive effect on students’ actual weight status, and the evidence is not strong enough to draw firm conclusions about their value.²⁶ The concern is that sending BMI “report cards” home with children may cause anxiety without spurring the systemic changes that actually influence childhood weight.

Genetics and How Much BMI Is Inherited

Your BMI is not entirely under your behavioral control. Twin and family studies have long suggested that genetics plays a significant role, and recent work with polygenic scores, which combine the tiny effects of thousands of genetic variants into a single risk estimate, has begun to quantify this.

A 2025 study in Nature Medicine tracked how well a polygenic score predicted BMI from infancy through adulthood. The genetic signal was small at birth but roughly quadrupled by age 12, and when combined with a child’s measured BMI at age 5, the polygenic score roughly tripled the amount of variance in BMI at age 18 that could be predicted, from about 22% to 35%. At even younger ages the relative boost was larger still.²⁷ This does not mean BMI is “determined” by genes, because environment, diet, activity, and socioeconomic conditions all exert powerful influences. But it does mean that some people are genetically predisposed toward higher BMIs in ways that simple willpower narratives fail to capture, and it adds another reason why the same BMI number can mean different things for different people.