Photo: FG Trade/Getty Images
As the U.S. grapples with the effects of systemic racism, some in the medical community are questioning whether the tools they use to assess patient health may be contributing to racial health disparities.
That debate is playing out most prominently in the world of kidney medicine. Black people are almost four times more likely to suffer from kidney failure than non-Hispanic whites. And once they get to that stage, Black patients spend months longer waiting for a kidney transplant than white patients.
Now, some doctors are asking whether a diagnostic formula most commonly used to assess the health of patients with chronic kidney disease may be unintentionally contributing to those poor outcomes — and reinforcing racist thinking.
The tool in question is a formula used to estimate GFR, or glomerular filtration rate. It's a measure of how fast a person's kidneys filter blood. Lower kidney filtration rates suggest worse kidney function.
The gold standard for measuring GFR is a burdensome process that involves urine collection over a 24-hour period as well as a blood sample. So instead, doctors and labs routinely estimate kidney function by measuring blood levels of creatinine, a waste product filtered by the kidneys, and then doing a calculation that also factors in the patient's age and sex.
But if a patient is African American, the person's race also plays into this calculation. Doctors and labs will routinely apply what's called a "race correction" or "race adjustment" to their estimated GFR number so that Black patients with chronic kidney disease end up with higher values.
Critics say the practice is based on flawed scientific assumptions tinged with racism. And because higher filtration rates suggest better kidney function, critics argue that correcting for race may delay critical referrals to specialists, potentially leading to worse outcomes.
"Why are we on the side of overestimating [filtration rates] if it could result in Black patients getting less care, to put it bluntly?" says Dr. Mallika Mendu, a kidney specialist and executive medical director of clinical operations for Brigham Health.
In fact, this June, Mendu's hospital system, Mass General Brigham, became one of a growing number of medical institutions in the U.S. to abandon the use of race in estimating GFR, amid a movement driven by medical students.
But other kidney specialists argue that while the use of race in kidney medicine is flawed, the rush to abandon it might cause more harm than it cures.
Is the Use of Race Exacerbating Disparities?
Mendu co-authored a study published this fall that found that if the race correction were removed, up to 1 out of every 3 Black patients would be reclassified as having a more severe stage of chronic kidney disease. And up to one-quarter of Black patients would have been reclassified from stage 3 to stage 4 of the disease — the final stage before kidney failure, which can trigger more advanced care.
The findings suggest "there is real potential for exacerbating disparities" when a race correction is applied, Mendu says.
To conduct the study, the researchers turned to a registry of more than 56,000 patients with chronic kidney disease created by the Mass General Brigham hospital network. Of those, 2,225 self-identified as African American. The researchers then recalibrated the Black patients' estimated GFR values to see what they would be without the race correction.
Mendu says perhaps the most striking finding was that when the race correction was removed, 64 Black patients had a kidney filtration rate lower than 20 — the threshold at which patients are referred for a kidney transplant. But Mendu says records showed that none of those 64 patients had actually been referred or evaluated for a transplant, because with the race correction applied, their GFR values were hovering above that threshold.
That's a really big deal, she says, because that means those patients lost critical time on the waiting list for a new kidney.
"The more time you spend on a waitlist, if you're waiting for a kidney, the more likely you are to get a kidney," Mendu says.
Given these existing disparities, Mendu says, kidney specialists should rethink the use of the race correction.
Dr. Neil Powe, a kidney specialist at the University of California, San Francisco, is among those who caution against abandoning race in estimating GFR too quickly. He points out that racial disparities in outcomes for patients with chronic kidney disease existed long before the use of race to calculate kidney function became widespread two decades ago. And in the study from Mendu and her colleagues, he notes that 80% of the Black patients who did have an estimated GFR of 20 or lower with the race correction were not referred for a kidney transplant, either.
Powe says that this suggests the race-based equations "are a smaller part of what's causing disparities in African Americans getting waitlisted [for a kidney transplant]. And those other things that cause disparities could be racism as well."
First, Do No Harm
The use of race-adjusted algorithms to estimate kidney filtration rates dates back to 1999. It was based on a study that included 1,304 white people and 197 Black people. The researchers found that, on average, Black people in the study had higher kidney filtration rates than white people at the same blood creatinine concentrations, suggesting that the formula then used to estimate GFR was underestimating kidney function in Black people. So the authors introduced a race correction to the formula that better fit the data.
Another much larger study, published in 2009, revised the equation used to estimate GFR but also found that it more accurately reflected measured GFR when it adjusted values for Black patients.
Dr. Vanessa Grubbs, a kidney specialist at the University of California, San Francisco and a longtime critic of using a race-adjusted GFR, says the originators of race adjustments in GFR algorithms were unable to explain why Black people might produce and clear creatinine from their bodies differently than white people do.
"The suggestion that Black people and only Black people are different than every other human on the planet is just ludicrous," she says.
One explanation that is often cited is muscle mass, since creatinine is a breakdown product of muscles. In the 1999 study, the authors observed that three previous studies had found Black people on average have greater muscle mass than white people. But as Grubbs noted in a recent paper, those studies, now decades old, were small and did not measure muscle mass directly. Yet the use of race as a proxy for muscle mass reinforces the notion that "black bodies are biologically different than white ones," she wrote.
Mendu notes that in the 2009 study, the majority of all patients — not just Black ones — had a sizable difference between their measured and estimated GFRs, suggesting that GFR is an imprecise variable.
And she says that race itself is a poor marker of biological difference. "We know that there is more diversity within Black patients than there is genetic diversity between a Black person and a white person," she says. "To say that somebody being Black is somehow a monolithic thing when it comes to genetics, when it comes to ancestry, I think is challenging."
How, she asks, would you apply the race adjustment to someone biracial, like former President Barack Obama?
A Movement for Change
The debate in kidney medicine comes amid a broader examination of the use of race in clinical diagnostics in the medical community. That reckoning has been brought about in large part by medical students at institutions across the country who have questioned the scientific evidence justifying the use of race in diagnostic formulas and whether that might be perpetuating inequalities.
"From our social science and genetics faculty, we were getting a message loud and clear: Race is a social construct, and it's not a reliable proxy for genetic difference. And then, on the other hand, our clinical faculty were turning around and teaching us that race is being used as a proxy for genetic difference every day in clinical medicine," says Dr. Leo Eisenstein, a second-year medical resident at New York University and Bellevue hospitals.
As a medical student at Harvard University, Eisenstein was part of a coalition of students whose research and lobbying efforts convinced Beth Israel Deaconess Medical Center in Boston to abandon the use of race in GFR in 2017. The students zeroed in on race-adjusted estimated GFR, he says, in part because it seemed to be systematically correcting Black patients to a healthier level in a way that might be less protective.
Since then, Eisenstein and some of his former classmates have advised medical students at other universities who are seeking to convince their institutions to abandon race-based GFR as well. This summer, the University of Washington health system and Vanderbilt University Medical Center also dropped race from their estimated GFR equations after students teamed up with faculty to examine the strength of the evidence behind the use of race adjustments.
Eisenstein says for younger generations of medical students who see the world with a racial justice lens, the issue was clear — the race correction had to go.
"We're possibly perpetuating or worsening racial health disparities without anyone intending to do so," says Eisenstein.
Seeking a New Standard
In August, the National Kidney Foundation and the American Society of Nephrology formed a task force to debate the pros and cons of using race in estimated GFR. The group is expected to issue its interim recommendations in January 2021. Powe is co-chair of the panel; Mendu is on it as well. Both agree that if doctors do continue to use race-adjusted GFR, they need to be transparent with Black patients about it and they should not rely on GFR alone to make decisions about patient care.
Another member of the panel, Dr. Lesley Inker, is a kidney specialist at Tufts University who helped develop the revised 2009 GFR algorithm that includes a race correction. She too thinks that the reasons behind the observed differences in GFR values for Black and white patients are not well understood and that the use of race in such calculations has limitations.
"I think it's appropriate these questions keep getting asked," she says.
But Inker warns that moving to abandon the use of race in GFR too quickly could have widespread unintended consequences and could potentially lead to less care for Black patients.
Without the race correction, Inker says, Black patients' kidney function might look worse than it actually is. For patients with other medical conditions, she says that could mean less access to treatments, clinical trials, and medications that they would otherwise have qualified for. For example, metformin is the first drug of choice to treat diabetes in patients with chronic kidney disease, but those with a GFR of 30 or below cannot use the drug, which means they might have to turn to other medications with more side effects.
Powe notes that it could even affect Black patients' ability to secure life insurance.
Powe says he sees why the use of race to estimate GFR is problematic, but when the data show actual racial and ethnic differences in kidney function, he asks, "Do we just ignore them?"
"There's benefits and disadvantages on both sides," Inker says. Her research group is working on a more precise formula to calculate GFR without the use of demographics, and it's analyzing how eliminating the race correction could affect patients.
Ultimately, Inker says, doctors should be asking, "What's the best outcome for each individual patient?"
Everyone interviewed for this story agrees that an ideal solution would be to use another biomarker to measure kidney function that does not rely on race. But Powe worries that doctors will start dropping race-corrected estimated GFR before the broader kidney specialist community agrees on what that biomarker should be. "We want to have a standardized approach so that we don't have chaos in the medical community," Powe says.
Meanwhile, the use of race in other clinical diagnostic tools has come to the attention of lawmakers. In September, the House Ways and Means Committee asked medical professional associations to reexamine the use — and misuse — of race in clinical care.
Their inquiry was prompted in part by an article published this summer in the New England Journal of Medicine — co-authored by Eisenstein — that analyzed 13 clinical algorithms that incorporate a patient's race in various specialties, from kidney medicine to pulmonology, obstetrics, urology, and cardiology.
All of the examples cited had the potential to affect the quality of care that people of color receive — for example, by underestimating the risks of heart failure in hospitalized Black patients or by steering more pregnant women of color toward cesarean sections if they'd had one in the past.
"That's perverse. The minorities are the ones who have the worst health outcomes," says Dr. David Jones, a physician and medical historian at Harvard and a co-author of the NEJM article.
Jones stresses that he and others aren't calling for medicine to abandon the collection of race data altogether, because it's necessary in order to understand the racial health disparities that exist in the United States. Instead he says: "We're calling to take a really close look at predictive uses of race, especially ones that exaggerate or accentuate health disparities."