Doctors still use BMI because it’s free, takes seconds to calculate, and despite real limitations, it correlates with health outcomes across large populations better than most people assume. That doesn’t make it a perfect tool. The medical community increasingly acknowledges its flaws, and in 2023 the American Medical Association formally stated that BMI alone is “an imperfect clinical measure.” But no single replacement has proven practical enough to take its place in everyday clinical settings.
BMI Was Never Meant to Measure Health
The formula behind BMI, weight divided by height squared, was created in 1832 by a Belgian mathematician named Adolphe Quetelet. He wasn’t a doctor. He was studying population-level patterns in how weight changed relative to height over a lifetime. The index was a statistical tool for describing groups, not diagnosing individuals.
It stayed that way for over 150 years. Then, in 1985, the U.S. National Institutes of Health and the World Health Organization adopted BMI as a standard classification system. The categories most people know today (under 18.5 is underweight, 18.5 to 24.9 is normal, 25 and above is overweight, 30 and above is obese) became the default way to screen for weight-related health risks. A population-level shortcut became a clinical tool applied to individuals, and that gap between original purpose and current use is where most of the problems live.
It Does Predict Risk, at Scale
The strongest argument for keeping BMI around is that it works reasonably well as a population-level risk indicator. A major pooling project analyzed data from 1.46 million adults and found a clear, graded relationship between BMI and mortality. Compared to people with a BMI of 22.5 to 24.9, those with a BMI of 30 to 34.9 had a 44 percent higher risk of death during the study period. At a BMI of 35 to 39.9, the risk jumped to 88 percent higher. At 40 to 49.9, the risk was 2.5 times higher. For every five-unit increase in BMI, overall mortality risk rose by 31 percent.
These are not small numbers, and they’re drawn from enormous datasets. When a doctor sees a patient with a BMI of 38, the statistical picture is genuinely concerning, even if that number doesn’t tell the whole story for that specific person. BMI gives clinicians a fast, rough signal that something may warrant closer investigation.
The Practical Advantages Are Hard to Beat
BMI requires a scale, a height measurement, and a calculator. That’s it. No specialized equipment, no trained technician, no appointment at an imaging center. The gold-standard methods for measuring body composition, like dual-energy X-ray absorptiometry (DEXA) or MRI, require expensive machines and trained staff. As the Obesity Medicine Association has noted, these techniques are valuable in research but haven’t found their way into everyday exam rooms because of cost and accessibility barriers.
BMI is also reproducible and easy to track over time. If your BMI goes from 27 to 32 over three years, that trajectory tells your doctor something useful regardless of the number’s limitations at any single point. In a healthcare system that needs to screen millions of people quickly, a free tool that takes five seconds has an enormous structural advantage over a $200 scan that requires a 30-minute appointment.
Insurance and Billing Require It
One reason BMI persists that rarely gets discussed: the healthcare system is built around it. Medicare requires specific BMI diagnostic codes for coverage of bariatric surgery. Insurance companies use BMI thresholds to authorize treatments, from weight-loss medications to surgical interventions. The billing infrastructure has BMI baked into it, with ICD-10 codes for BMI ranges as narrow as single-digit increments (35.0 to 35.9, 36.0 to 36.9, and so on).
Even the AMA, while calling BMI imperfect, recommended that it should not be used as a sole criterion to deny insurance reimbursement. That phrasing acknowledges the reality: right now, it often is. Replacing BMI in clinical practice would mean overhauling coding systems, insurance criteria, and coverage policies across the entire healthcare system. That institutional inertia keeps BMI firmly in place even as the science evolves.
Where BMI Gets It Wrong
BMI cannot distinguish between muscle and fat. In a study of 172 collegiate athletes, BMI classified 35.5 percent as overweight and 4.1 percent as obese. But when researchers measured their actual body fat, 89 percent had healthy levels. Overall, BMI and body fat percentage agreed only 59.3 percent of the time. The most common error was BMI labeling athletes as overweight when their body fat was perfectly healthy. This isn’t a niche problem. Anyone who carries significant muscle mass, from construction workers to recreational weightlifters, can be misclassified.
The tool also misses people in the other direction. About 35 percent of people classified as obese by BMI are metabolically healthy, showing no signs of the blood sugar, blood pressure, or cholesterol problems typically associated with excess weight. Meanwhile, people with a “normal” BMI can carry dangerous amounts of visceral fat around their organs without it showing up in the calculation.
Ethnicity Changes the Risk Thresholds
The standard BMI cutoffs were developed using data from predominantly white populations, and they don’t translate evenly across ethnic groups. A meta-analysis published in Circulation found that for South Asian populations, the BMI associated with the same diabetes risk as a BMI of 30 in white populations was just 23.3. That’s a number the standard system would label “normal weight.” South Asian men showed elevated risk for high blood pressure starting at a BMI of 23.3 as well, while South Asian women saw elevated risk beginning around 24.0.
This means a South Asian person with a BMI of 24 could face cardiovascular risks similar to those of a white person with a BMI of 30, yet the standard chart would tell them they’re fine. The AMA has acknowledged this, noting that BMI cutoffs don’t account for gender or ethnicity and that body composition varies significantly across racial and ethnic groups.
Better Measures Exist but Haven’t Replaced BMI
Waist-to-hip ratio consistently outperforms BMI in predicting mortality. A large genetic study found that waist-to-hip ratio had a stronger and more consistent association with death from all causes than BMI did. Each standard-deviation increase in genetically determined waist-to-hip ratio carried a 51 percent higher risk of death, compared to 29 percent for BMI. Perhaps more importantly, waist-to-hip ratio’s predictive power held steady regardless of a person’s BMI, while BMI’s own predictive accuracy varied depending on where someone fell on the scale.
Waist circumference and measurements of visceral fat also capture health risks that BMI misses entirely. The AMA now recommends using BMI alongside these measures rather than on its own. But none of these alternatives have fully displaced BMI in routine practice, partly because they require slightly more effort (a tape measure positioned correctly, or an imaging scan) and partly because the systems of insurance, billing, and clinical guidelines haven’t been rebuilt around them yet.
What Your Doctor Should Be Doing With It
The emerging consensus isn’t to throw BMI out but to stop treating it as a verdict. It’s a screening number, one data point among many. A useful checkup incorporates your waist measurement, blood pressure, blood sugar, cholesterol levels, family history, and a conversation about how you actually live. If your doctor fixates on BMI alone, or if a BMI of 26 triggers a lecture without any other context, that’s a gap in practice rather than a gap in the science.
BMI persists because it fills a role no other tool currently fills as cheaply, quickly, and universally. Its limitations are real and well-documented, but the healthcare system hasn’t yet built a practical alternative that works at scale. The best approach, for now, is knowing what BMI can and can’t tell you, and making sure your care doesn’t stop at a single number on a chart.