Erasing Race in Medicine

The pandemic's spotlight on inequity offers fuel for efforts to remove race as a factor in clinical equations.

By Steven Ross Johnson

Feb. 3, 2022

Critics claim clinical algorithms that factor in race when assessing a patient’s risk for a host of health conditions may result in patients of color receiving less access to care than their white counterparts.

Seminal events during the pandemic – such as the protests over the death of George Floyd at the hands of Minneapolis police in 2020 and the disproportionate impact of COVID-19 on communities of color – have brought the issues of systemic racism and injustice to the forefront of the public’s consciousness.

For some health care providers, the moment also has presented an opportunity to leverage the national conversation into new efforts to examine how existing policies and practices in medicine may be perpetuating racial health inequities.

“We’re at a pivotal time,” says Dr. Nichola Davis, chief population health officer for NYC Health + Hospitals, New York City’s public hospital and health care system. “People’s ears are open and they are willing to hear and have these courageous conversations that they maybe weren’t able to have in 2019.”

At issue are clinical algorithms that factor in race when assessing a patient’s risk for a host of health conditions, ranging from death in connection with heart surgery to urinary tract infections in children. Critics of such calculations contend they may result in patients of color receiving less access to and poorer quality of care than their white counterparts, fueling long-standing racial health disparities.

“Race was never necessary or needed” in these algorithms, says Dr. Milda Saunders, an ethicist and health equity researcher and associate professor of medicine at the University of Chicago. “I think a lot of times when we’re measuring race we’re measuring racism, both structural and interpersonal racism.”

Race-based adjustments within algorithms that help make clinical assessments have been around for decades. But while there is widespread acceptance within the scientific and medical communities that the concept of race is not based on biology and is instead a social construct, race correction continues to influence patient diagnoses and treatments.

And despite growing acknowledgment of the potential harms caused by race correction in medicine, so ingrained are these calculations in clinical assessments that the effort to overhaul them is often complicated with questions over what factors can be an adequate replacement.

“These are conversations that have been happening among certain subsets of people for years,” Davis says. “This isn’t new work for some people.”

There has been notable movement of late, however: Last May, for example, NYC Health + Hospitals announced its removal of a patient’s race from calculations used to assess kidney function – part of the launch of the health system’s broader “Medical Eracism” initiative.

Race was first introduced in the late 1990s to calculate estimated glomerular filtration rate, or eGFR, which is commonly used by clinicians to measure the kidneys’ ability to filter waste from the blood. A higher eGFR score normally suggests better kidney function.

Black patients, though, are typically assigned a race adjustment that raises their eGFR by approximately 16% or more than non-Black patients, depending on the equation used. The adjustment is tied to past findings showing higher levels of the waste product creatinine among Black people.

“Explanations that have been given for this finding include the notion that Black people release more creatinine into their blood at baseline, in part because they are reportedly more muscular,” a 2020 article in The New England Journal of Medicine states. “Analyses have cast doubt on this claim, but the ‘race-corrected’ eGFR remains the standard.”

Davis says this adjustment can negatively impact the quality of health care Black patients receive, causing delays in receiving medical interventions such as referrals to a nephrologist or dialysis. One such potential consequence was outlined in a November piece published by STAT, in which Glenda V. Roberts – a woman with kidney disease who identified as Black – wrote of how the equation delayed her evaluation for a kidney transplant.

“Unfortunately, when you look at the clinical algorithms, what many of them do is they may downplay the severity of some illnesses when you do that race correction,” Davis says.

Recent steps to remove race in the calculation of eGFR were bolstered by the release in September of a recommendation by the National Kidney Foundation and American Society of Nephrology to use a new equation that does not include race.

“The big change is that more institutions like the hospital system, hospital leadership and administrators are more on board with efforts toward equity than they were in prior years,” says Dr. Monica Maalouf, associate program director of internal medicine for Loyola Medicine, a health system based in the Chicago area.

Maalouf says Loyola is in the process of determining how to remove race as a variable within the health system’s eGFR calculations. That process has involved talks with physicians and laboratory coordinators and directors, as well as working with information technology specialists on how to remove the racial designation from the risk score’s automated reporting system.

National Kidney Foundation President Dr. Paul Palevsky says the new equation marks an important step toward efforts to achieve equity in health care by providing a more accurate, race-free approach to assessing kidney function.

“It was the most common equation used in medicine that incorporated race that was reported to be used in clinical care,” Palevsky says, estimating a race-inclusive equation for kidney function is used approximately 240 million times a year.

But while the change in the eGFR calculation signifies a substantial step forward, eliminating race correction in calculations for other health conditions has been more challenging, despite increased recognition about the potential harms and health disparities that can come from their continued use.

“The field has realized that when we are using race as a factor in our clinical assessments that we’re actually not using a precise measure like we would do for measuring sex, or age, or other things that we use to make our clinical decisions,” says Saunders, with the University of Chicago. “But we want to find the best answer, so that may involve sort of reexamining the questions and not just saying, ‘Let’s just drop race,’ but asking what was race measuring, and then substituting that measure.”

Saunders says University of Chicago Medicine recently removed race in calculating eGFR in patients, opting to adopt the algorithm endorsed by the National Kidney Foundation. She notes there also has been a broader push among clinicians in recent years to consider the impact of social drivers like income, access to food, housing status and access to transportation on patients of color when making clinical assessments.

Much of the recent attention toward removing race in clinical assessments began with the 2020 article published in the The New England Journal of Medicine, which listed about a dozen tools that included race in assessing clinical areas such as patients’ lung function, risk of in-hospital death from heart failure and survival from cardiac surgery. Such tools hold the potential for harmful consequences, the article noted, and could lead to issues such as nonwhite patients not getting health care interventions as early as white patients.

“Many of these race-adjusted algorithms guide decisions in ways that may direct more attention or resources to white patients than to members of racial and ethnic minorities,” wrote the authors, from Massachusetts General Hospital, Harvard Medical School and NYU Langone Health. “If doctors and clinical educators rigorously analyze algorithms that include race correction, they can judge, with fresh eyes, whether the use of race or ethnicity is appropriate.”

Last May, a study published by the American Journal of Obstetrics and Gynecology introduced an updated model that does not factor in patient race for use in estimating the chance of a successful vaginal birth for women who previously had a cesarean section. By lowering the estimated chance of success for women of color, a previous version of the VBAC calculator – a tool singled out in the NEJM piece – could push them toward a C-section, which is a major surgical procedure many maternal health experts believe is performed too often.

At the same time, adjusting for race in some cases may raise the question of whether doing so is useful if it results in minority groups potentially receiving more consideration.

For example, race adjustment within an atherosclerotic cardiovascular disease risk estimator – which estimates the risk of having an event like a stroke and helps clinicians determine when to start patients on cholesterol-lowering medication – can put patients who identify as Black at higher risk than white patients. Having race correction within the tool could make it more likely clinicians will intervene earlier for Black patients.

“Using race in these clinical algorithms hasn’t been all bad,” says Dr. Malika Fair, senior director of equity and social accountability for the Association of American Medical Colleges. “Their intent when they were being used sometimes was to identify health inequities.”

Fair says unilaterally removing race from every algorithm is a feasible approach, but fails to answer whether doing so would make things better or worse for patients. “We have to take the time to figure that out so that we don’t end up inadvertently creating more racial health inequities when we remove race from the equation,” Fair says.

Notably, a study published last year in The Lancet Oncology found that removing race from an eGFR equation could undercut cancer treatment in Black patients.

Dr. Megan Mahoney, chief of staff for Stanford Health Care and a clinical professor of medicine at Stanford University, says professional medical societies can play a huge part in providing and promoting guidance on alternatives that clinicians can use to assess disease risk without factoring race. Yet much of the action taken thus far aimed at eliminating race-based clinical assessment has been led by individual hospitals and health systems, which have made policy decisions after deliberation and consideration from their own experts.

Mahoney says medical societies can streamline the decision process for hospitals by serving as a trusted arbiter to review the evidence and provide guidance that most will accept.

“A particular health care system can do it, it’s just that it’s going to be harder to get buy-in from those who are actually implementing the change – the physicians on the front line – without the professional societies backing them,” Mahoney says.

Mahoney agrees it’s important to identify the root causes of health disparities for patients of color that have led to differences in risk outcomes. But she fears leaving race-based algorithms in place could reinforce concepts that tie race to biology, and could have long-term social and clinical ramifications.

“Even if there are going to be downstream changes that might put African Americans at a disadvantage, the perpetuation of race as a biological factor both in medical education and in medicine in general, the risk of that far outweighs that minor benefit some racial groups might be getting from the current state,” Mahoney says. “We are really putting our medical students and a whole generation of trainees at a disadvantage when the concept of biological race is reinforced in their training.”