BMI is a rough estimate, not a precise measurement of health. It correctly identifies obesity about 43% of the time when checked against actual body fat measurements, meaning it misses more than half of people who carry excess fat. At the same time, it reliably rules out obesity in about 96% of people who truly aren’t obese. So BMI is better at confirming you’re not overfat than it is at catching everyone who is.
The American Medical Association acknowledged these problems in 2023, adopting a policy stating that BMI “loses predictability when applied on the individual level” and should not be used as a standalone measure of health. That doesn’t mean BMI is useless, but understanding where it breaks down can help you interpret your own number more clearly.
What BMI Actually Measures
BMI is your weight in kilograms divided by your height in meters squared. That’s it. The formula was created in the 1800s by Adolphe Quetelet, a Belgian statistician who wanted to define the “average man” for population-level research. He wasn’t a physician, and the tool wasn’t built for diagnosing health problems in individuals.
The modern name “body mass index” came later, from a physiologist named Ancel Keys. He also didn’t intend it for clinical use. He described it as a simple, easily obtainable measurement that worked well in research settings. Over time, though, it became the default screening tool in doctors’ offices because it requires nothing more than a scale and a tape measure.
Where BMI Gets It Wrong
The core problem is that BMI uses total body weight, which doesn’t distinguish between muscle, fat, bone, and water. A person with high muscle mass can easily land in the “overweight” category despite having low body fat. This is common in athletes and people who strength train regularly, but it also affects anyone who carries more lean mass than average.
The reverse problem is arguably more dangerous. When researchers tested BMI against full-body scans that measure actual fat percentage, a BMI cutoff of 30 (the standard threshold for obesity) caught only 36% of men and 49% of women who qualified as obese by body fat standards. That means roughly half of women and nearly two-thirds of men with excess body fat were classified as “normal” or merely “overweight” by BMI. These people may not receive the health guidance they need because their number looks acceptable on paper.
BMI also has a weak connection to visceral fat, the deep abdominal fat that surrounds organs and drives the highest health risks. Research measuring fat deposits directly found that BMI’s correlation with visceral fat ranged from just 0.61 to 0.69, considerably weaker than its correlation with total body fat (0.91 to 0.94). Two people with identical BMIs can have very different amounts of visceral fat, and therefore very different risk profiles for heart disease and diabetes.
BMI Varies by Ethnicity
Standard BMI categories were developed primarily from data on non-Hispanic white populations. They don’t apply equally across ethnic groups. Asian populations, for example, develop higher rates of cardiovascular disease and type 2 diabetes at BMIs well below the standard “overweight” cutoff of 25. In South Korea, the prevalence of people who are metabolically obese but normal-weight by standard BMI is nearly twice the rate found in the United States.
To account for this, Asia-Pacific countries use lower thresholds: overweight starts at a BMI of 23 instead of 25, and obesity at 25 instead of 30. The normal-weight range tops out at 22.9 rather than 24.9. If you’re of Asian descent living in a country that uses the standard WHO cutoffs, your BMI number may be understating your actual risk.
BMI Shifts With Age
For older adults, the standard “healthy” range of 18.5 to 24.9 appears to be too low. A large meta-analysis found that among people over 65, a BMI of 23 to 24 was associated with the lowest mortality risk. Dropping below 23 actually increased risk: a BMI of 21 to 21.9 carried a 12% higher mortality risk, and a BMI of 20 to 20.9 carried a 19% higher risk compared to the 23 to 24 range.
This likely reflects the fact that older adults lose muscle mass naturally, so a lower weight often signals frailty rather than fitness. For people over 65, a BMI in the low-to-mid 20s may be protective rather than something to reduce.
Metabolically Healthy at a High BMI
Not everyone with a BMI over 30 is metabolically unhealthy. In a large cohort study, about 31% of people classified as obese by BMI met the criteria for “metabolically healthy obesity,” meaning they had normal blood pressure, blood sugar, triglycerides, and cholesterol despite their weight. Among men specifically, that figure was 37.5%.
This doesn’t mean a high BMI is harmless for these individuals long-term, as some research suggests metabolically healthy obesity can shift over time. But it illustrates that BMI alone can’t tell you whether someone’s weight is actively damaging their cardiovascular system right now.
How BMI Works for Children
In children and teens, BMI is interpreted differently than in adults. Rather than fixed cutoffs, pediatric BMI uses age-and-sex-specific percentiles. A child at the 85th percentile or above is considered overweight; at the 95th percentile or above, obese. The CDC and the American Academy of Pediatrics recommend BMI screening at least once a year for all children ages 2 and older.
Percentile-based screening accounts for the fact that children’s body composition changes dramatically as they grow. A BMI that’s normal for a 7-year-old might be concerning for a 12-year-old, and vice versa. It’s a more nuanced application of BMI than the adult version, though it still shares the fundamental limitation of not measuring body fat directly.
Better Alternatives and Additions
Waist-to-hip ratio consistently outperforms BMI for predicting heart attack risk. In a major international study, the link between BMI and heart attack disappeared entirely once waist-to-hip ratio was factored in. Meanwhile, waist-to-hip ratio remained a strong predictor even after adjusting for BMI and other risk factors, with the highest-risk group facing 2.5 times the heart attack risk of the lowest-risk group. This held true even among people with a BMI under 25, people who would be classified as lean or normal weight.
Waist circumference alone also outperformed BMI, with those in the highest category showing 1.77 times the heart attack risk of those in the lowest. Researchers estimate that using waist-to-hip ratio cutoffs of 0.83 for women and 0.9 for men would capture three times as many at-risk people as BMI-based screening.
Other measures the AMA now recommends using alongside BMI include body composition testing (which separates fat from lean mass), relative fat mass calculations, and assessments of metabolic markers like blood sugar and cholesterol. No single number captures the full picture. The most useful approach combines a simple measurement like waist circumference with basic blood work to assess what’s actually happening inside your body, rather than relying on a 200-year-old formula that was never designed for the job.