The term BMI is all over health care. The abbreviation stands for body mass index, a simple number that has evolved into a ubiquitous medical tool for obesity screening. For example, a BMI from 18.5 to 24.9 indicates a “normal” weight status, according to the World Health Organization, while a BMI of 30 or higher signifies obesity, a condition that raises your likelihood of developing type 2 diabetes, cardiovascular disease, various cancers and other illnesses. The metric has become a catchall proxy for body fat percentage, nutritional status and health risks.

But now BMI should be given much less importance, the American Medical Association (AMA) says. Last month the leading physician’s group recommended that practitioners de-emphasize BMI because it can get weight-related health risks wrong, especially when used as the sole diagnostic tool. The policy announcement also noted that BMI, originally developed from data on non-Hispanic white men, has played a role in perpetuating “racist exclusion” and causing “historical harm” by misidentifying the weight status of people in many racial and ethnic minority groups.

Francisco Lopez-Jimenez, a cardiologist at the Mayo Clinic, who has studied the limitations of BMI, says that the metric, which is calculated using a person’s weight and height, has become overused by physicians. He adds that BMI is also used to determine eligibility for procedures such as bariatric surgery, establish criteria for life insurance policies and even prioritize individuals for COVID vaccine administration. But other measures exist that produce a more accurate picture of a person’s health. Lopez-Jimenez spoke with Scientific American about some of these alternative methods, BMI’s pitfalls and the reasons why the new AMA policy may begin to alter our understanding of obesity.

[An edited transcript of the interview follows.]

What are the main problems with using BMI when diagnosing obesity?

Even though BMI is useful to assess population-level trends, it isn’t an accurate measure of obesity at an individual level. It doesn’t measure body fat directly and can’t distinguish between fat, muscle and bone. Relying on BMI to diagnose obesity is like diagnosing diabetes on the basis of weight instead of testing someone’s blood sugar level. Someone who weighs more may be more likely to have higher blood sugar, but that doesn’t mean you should neglect measuring their blood sugar. Similarly, someone who weighs more may be more likely to develop obesity and its associated health risks, but it’s not a perfect substitute for measuring body fat.

Is BMI less accurate when applied to particular groups of people, such as underrepresented minority groups?

BMI suffers from both overclassification and underclassification, and it disproportionately impacts certain groups. For instance, Black people tend to have lower body fat percentages and higher muscle mass, as compared with white people with the same BMI. Consequently the standard BMI cutoffs are more likely to misclassify them as having obesity, even when their body fat percentage falls within the healthy range. Labeling patients with a condition they don’t have can then lead to other misdiagnoses.

On the other end of the spectrum, people of Asian descent tend to have more body fat than white people at the same BMI, so their risk of metabolic and cardiovascular diseases are more likely to be overlooked.

BMI is often used because it is quick and inexpensive. Are there alternative measurements that have these same benefits?

Yes. The AMA’s new policy mentions waist circumference as one alternative.

That’s essentially a tape measure around the waist. And the American Heart Association uses it to classify abdominal obesity at 35 inches for women and 40 inches for men, right?

Correct. But we should avoid repeating past mistakes by relying on waist circumference instead of directly measuring body fat. Like BMI, the standard cutoff values for waist circumference are based on white populations, so they may not be universally applicable. Waist circumference does not consider variations in height either, which is why incorporating additional measures such as waist-to-height ratio is important for a comprehensive assessment.

There is no one-size-fits-all approach for diagnosing obesity. It’s a complex disease and should be assessed using multiple measures.

How have opinions on BMI shifted in the past 20 years?

Before the American Medical Association defined obesity as a disease in 2013, our main challenge was persuading insurance companies to cover the cost of essential treatments, such as weight-loss medications and bariatric surgery. At that time, patients’ weight would be recorded, but BMI would not be easily accessible in their medical record. This made it more difficult to encourage health care providers and insurance companies to see obesity as a legitimate medical condition, not a simple consequence of personal choices and behavior. So when BMI became widely accepted as a vital sign, it was a big win. BMI was an ideal way to simplify diagnosing obesity.

But there weren’t many other reliable, well-tested techniques to measure body fat. Now we have more options. For instance, DEXA scans are a special type of x-ray that can estimate the amount of fat tissue you have. In any case, BMI or any measure of body fatness should not be used alone to address cardiovascular risks. Health care providers should also cross-check for cholesterol, triglycerides and blood pressure, among other measurements. That gives us a better picture about the health risk of a patient.

How do you think the AMA’s new policy will change clinical practice?

To be honest, the immediate impact will likely be limited. Health care providers are often more responsive to changes in clinical guidelines from their corresponding medical society. So if the American Academy of Family Physicians were to issue a statement reaffirming that BMI alone is an imperfect measure of obesity, family medicine providers could be more responsive to that than to the AMA.

Do you think the new policy will affect our understanding of obesity as a disease?

Absolutely. The fact that the AMA issued a public statement on obesity underscores that it is a legitimate medical condition. It also highlights the need for more research to understand aspects of the disease we don’t understand very well, such as how health risks vary across demographics and the genetic, environmental and behavioral factors that increase the risk of developing obesity.

When I began researching BMI and cardiovascular disease in the early 2000s, people would say, “Are you crazy? Who cares about BMI?” And just 10 years ago we had to convince health care providers to see obesity as a disease. Now we’re trying to convince them to use better measurements. That context shows how change doesn’t happen overnight, but it is achievable. And it’s crucial to advocate for improvements to our understanding and treatment of obesity.