MW It's time to rethink how we define obesity as millions turn to GLP-1s
By Jaimy Lee
Is BMI pure B.S.? It's 'definitely not a measure of fatness,' and it doesn't accurately define what it means to be overweight.
Body mass index isn't always the best predictor of health.
In the age of weight-loss drugs, there has to be a better way to diagnose obesity. Scientists are now locked in a fierce debate over the body mass index, and that could change who qualifies for a GLP-1 medication.
BMI has long been the standard to assess whether someone is at a healthy weight. It uses an individual's weight and height to calculate a figure that fits into one of four categories: underweight, healthy weight, overweight and obesity.
"The reason we keep using it is because it's simple and it's everywhere," said Holly Russell, a family-medicine physician at the University of Rochester Medical Center.
BMI has traditionally been incorporated into medical records and life-insurance policies. It's also used to track obesity rates at a national level. In the GLP-1 era, however, BMI has become hugely important. When the Food and Drug Administration approved Novo Nordisk's (NVO) (DK:NOVO.B) Wegovy in 2021 and then Eli Lilly's $(LLY)$ Zepbound in 2023, it cited BMI, saying both GLP-1 injections should be used for people with a BMI of 30 or higher and a BMI of 27 and a weight-related condition like Type 2 diabetes or hypertension.
But BMI largely fails to take into account the details that tell us whether someone is healthy even if they carry extra weight. It doesn't distinguish muscle from fat, meaning athletes often have high BMIs because of their muscle mass. It also doesn't address the well-documented racial distinctions in body mass; Asian adults have a higher risk of weight-related conditions even with a BMI in the healthy range, while some Black women are healthy at higher BMIs. And then there's the fact that not all obese people are unhealthy.
"We acted like we know all these answers, but actually we don't," said Katherine Flegal, a former scientist at the Centers for Disease Control and Prevention who has studied BMI for 30 years. "Most of this stuff is approximations. And it's almost like pseudoscience in a way."
There are now a handful of alternative calculations that aim to better define healthy and problematic weight. This includes the body roundness index, the waist-to-height ratio (which is now being used by the military) and a new model that uses BMI as the first of several steps in diagnosing obesity. Even Lilly quietly dropped BMI from Zepbound's label.
It's unclear whether any of the new methods will replace BMI - that will require buy-in from physicians, insurers and the government - but researchers say it's time to rethink how we diagnose obesity as millions of Americans turn to GLP-1 drugs.
"I would hope that insurers could use our definition to make sure that the folks who get this are the ones who are sick from having too much body fat, and therefore would benefit most from losing that body fat and the illness that goes with it," said David Cummings, professor in the metabolism, endocrinology and nutrition department at the University of Washington. GLP1s shouldn't "end up going to athletes who just want to look a little bit thinner," he said, "or Hollywood stars who just want to look better for a role, or rich people who can just plain afford it."
Fixing BMI
Obesity rates steadily increased until 2023, when they dropped for the first time since 2011. Much of that success can be attributed to widespread use of Wegovy and Zepbound and the off-label prescription of GLP-1s approved for diabetes. But the runaway success of these drugs -some 12% of Americans reported taking a GLP-1 drug as of November, according to a KFF survey - has also raised concerns about who should get access to the pricey medications and who is responsible for paying for them.
"We need a triaging system," said Cummings.
He is one of the roughly 80 researchers worldwide who last year came up with the new obesity definition, which builds off BMI and then factors in an anthropometric measure like waist-to-height ratio and symptoms like a weight-related condition or limitations on daily activities. Let's call it BMI-plus.
In an interview, Cummings pointed to research published in 2024 that found more than 50% of U.S. adults at that time qualified for semaglutide, the drug marketed as Wegovy for weight loss and Ozempic and Rybelsus for diabetes. That's far more people than those who are eligible to take statins.
'We acted like we know all these answers, but actually we don't. Most of this stuff is approximations. And it's almost like pseudoscience in a way.' Katherine Flegal, former CDC scientist
That's worrisome to him, given the cost of GLP-1s - both to patients and the system - and that not all people who are currently considered overweight or obese are unhealthy. For people who fall under the "preclinical obesity" category in the BMI-plus model, Cummings would recommend lifestyle changes or a less expensive oral GLP-1. Even Lilly CEO David Ricks recently called out BMI thresholds as one way that insurers limit which patients can access Zepbound, and the Medicare program set to start this summer requires people without other conditions to have a BMI as high as 35 to start treatment.
"Being labeled with obesity has a lot to do with what things you can get paid for by insurance," Cummings said. "If you overdiagnose it, then you may be allocating unlimited resources to people who perhaps aren't the best candidates for them. If you underdiagnose it, then you may deny those resources for people who deserve them more compellingly."
A study published in October predicts that its adoption would likely change the mix of patients who qualify for weight-loss drugs, though it would not dramatically increase or decrease the number of prescriptions.
No longer a back-of-the-envelope calculation
BMI has always been somewhat controversial. The calculation was created in the 1800s by a Belgian astronomer, and then a doctor in the U.S. renamed it the body mass index in the 1970s. As Americans became heavier and increasingly worried about weight, it became the standard - without any scientific rigor to back it up.
"BMI is definitely not a measure of fatness," said Flegal, the former CDC scientist.
Flegal has faced criticism for research that found people with a BMI between 30 and 34.9, which is considered obese, had lower mortality rates than people who were considered overweight. This is when clinicians started to describe instances in which people who are metabolically healthy and obese as the "obesity paradox," a concept that Russell, the physician in Rochester, describes as "bonkers."
"When we're doing these things, we have to wonder: Why are we doing them, and what are we looking for?" Russell asked. "What are we trying to measure, and what are we trying to track?
That's one approach to solving for BMI's gaps, to question the hunt for an obesity definition. Others are trying to modernize the idea behind measuring BMI.
Back in 2013, Diana Thomas, a mathematician at the U.S. Military Academy at West Point, and another scientist introduced a geometric calculation called the body roundness index, or BRI, as an alternative to BMI. Thomas had become increasingly interested in developing models of the body after gaining and losing weight before and after a pregnancy.
Several studies, including one published in 2024, have found that the BRI does a better job than BMI at predicting mortality. But what Thomas is increasingly excited about isn't another equation. It's using data gathered from body-scanning machines commonly used in gyms, continuous glucose monitors, FitBit and Whoop devices, and electronic health records to come up with a smarter approach to predicting who is healthy, who is not and who is at risk.
"We don't need to do a back-of-the-envelope calculation. Just put it in the machine," she said. "All this stuff is coming together, and you have this personalized prediction right there."
-Jaimy Lee
This content was created by MarketWatch, which is operated by Dow Jones & Co. MarketWatch is published independently from Dow Jones Newswires and The Wall Street Journal.
(END) Dow Jones Newswires
March 18, 2026 11:42 ET (15:42 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments