Widely used physician guidelines that ignore patients’ race and ethnicity could be doing more harm than good when it comes to catching diabetes in people of color. New research, published in the Annals of Internal Medicine on Monday, suggests that people from certain racial and ethnic groups should be screened for diabetes at lower body mass index than non-Hispanic white people — a recommendation that contradicts recent guidelines from the United States Preventive Services Task Force.
It’s an admittedly tricky proposition, to reaffirm the role of race and ethnicity at a time when medicine is trying to rid itself of race-based tools — such as an algorithm used to assess kidney function — that have contributed to the large health disparities in the United States. The paper’s authors recognized as much in interviews with STAT.
They argue, however, that using a one-size-fits-all approach to screening, when diabetes is two to four times more prevalent and more deadly in Black, Hispanic, and Asian Americans, is likely to result in underdiagnosis of the disease, and widen health gaps.
The current guidelines, issued by the task force last year, recommend doctors screen adults between 35 and 70 years old, with a BMI of 25 or more, for prediabetes and type 2 diabetes. But data analyzed by the paper’s authors show certain racial and ethnic groups should be screened at lower BMIs to detect diabetes at the same rates as in white people — a BMI of 18.5 for Black Americans and Hispanic Americans, and 20 for Asian Americans.
“That’s the case we’re making: that if you’re looking for a test that’s equally sensitive in all of the subgroups, then you take the lowest risk subgroup … which is white Americans, and then you use that as the threshold, and then adjust it to everybody else,” said Dhruv Kazi, senior author of the paper. “It’s actually possible that individuals of racial and ethnic minorities might need an even more sensitive threshold.”
That is, even at “normal weights,” non-white people are multiple times more likely to have diabetes than white people, calling into question whether BMI is an effective way of determining patients’ risk of developing diabetes.
“I feel comfortable saying that, even as we move away from race-based equations that have done more harm than good, we should embrace risk-based equations, even if a lot of that risk comes with race,” Kazi said. “There is no way to get equitable diabetes screening if we ignore this gradient in risk.”
Neil Powe, chief of medicine at Zuckerberg San Francisco General Hospital, and one of the people who made the effort to change the equation for kidney disease, said a “unified threshold” for diabetes would be desirable, one “that is not normalized to one race and does not disadvantage one group more than another. Clinicians also need to consider factors other than race, age, and BMI such as family history, history of gestational diabetes, and sedentary lifestyle.”
Lead author Rahul Aggarwal became interested in the question of racial and ethnic differences in diabetes risk after his mother was diagnosed with the disease when Aggarwal was in medical school at Boston University.
Neeru Aggarwal was a young, healthy woman with an average BMI and a regular exercise routine. The diagnosis shocked her family, causing her son to wonder if there was science underlying the high rates of diabetes, or “high sugar,” he saw anecdotally in his family members and community. He remembers how, at gatherings, loved ones would need to monitor how many carbohydrates they ate, or turn down desserts — all painful restrictions when faced with plates of mouth-watering Indian food.
“We often think that diabetes is a purely lifestyle-based disease that people get it because they have poor diets, don’t exercise enough,” said Aggarwal, a resident physician at Beth Israel Deaconess Medical Center. “But in actuality, it’s a multifactorial process.”
Kazi, also Indian American, and an associate professor at Harvard Medical School, experienced something similar when his parents were diagnosed with diabetes in their 30s and 40s, despite having thin frames.
Kazi and Aggarwal say their paper is not meant to offer a perfect solution to a complex problem. “We’re making more of a comparative argument between different racial/ethnic groups, but we’re not actually saying that this is the exact right threshold to screen. And that’s something that’s an area of important future work,” Aggarwal said.
BMI and broad race and ethnicity labels are often crude measures of risk, used as proxies for other things, but they are available tools, and the stakes of continuing to under-diagnose diabetes are dire. Diabetes can lead to heart attack, vision problems, dialysis, and a number of other health issues and complications. But early intervention can significantly reduce the risk of complications and death. The researchers hoped to offer interim guidance that doctors and patients can adopt immediately, while researchers work toward more sophisticated risk calculation.
“The harms are pretty catastrophic,” Aggarwal said. “So identifying diabetes appropriately and early can actually have a substantial impact on both morbidity and mortality for patients.”
To Licy Yanes Cardozo, a physician-scientist at the University of Mississippi Medical Center, and a practicing endocrinologist, the findings in this paper bolster the data doctors already have about unequal health outcomes, and serve as a call to action. Physicians can change their screening practices immediately, using formal guidelines as a starting point but using their judgment based on the patient. “It really brings back the power to the physicians,” she said. “Using the same rule for every single patient, it wasn’t ever a good idea in medicine.”
Cardozo, who co-authored a paper on eliminating racism in endocrinology, wants to see the formal USPSTF guidelines changed to allow for more flexibility in diabetes screening practices. Only when those physician guidelines are changed will Medicare and insurers be required to cover early diabetes screening at lower BMIs.