Kevin Henry was born and grew up in the US in the small, majority African American town of Mound Bayou, Mississippi. He attended John F. Kennedy Memorial High School in the 1980s before leaving to study at Mississippi State University, where he played American football. Kevin’s talent saw him drafted into the National Football League (NFL) in 1993 by the Pittsburgh Steelers, where he played as a defensive lineman for eight seasons. In 1995, the Steelers won the American Football Conference championship and advanced to the Super Bowl, where they were defeated by the Dallas Cowboys. But in 2001, at the age of thirty-three, Kevin’s sports career came to an end.
Like many other American football players, Kevin had begun struggling with what he suspected were the long-term effects of the numerous concussions he had sustained while playing over the years. “Football doesn’t give you an expiration date—you just expire,” Kevin said during an interview on the ABC News show Nightline in February 2021. He said he experiences frequent headaches, depression, and memory loss, which cumulatively leave him unable to work. In the years since Kevin first started playing professional football, multiple studies have emerged demonstrating a link between head injuries in the sport and long-term conditions affecting the brain, including chronic traumatic encephalopathy and dementia.180
In 2013, the NFL reached a financial settlement with former players who had sued over head injuries sustained while playing, agreeing to pay compensation to players who received certain diagnoses, including dementia. Kevin decided to apply for compensation through the settlement program in 2017, concerned about his inability to work and his family’s future financial situation. As part of his application, he saw a doctor and underwent a series of tests aimed at assessing his cognitive function, including his language, learning, and memory. The doctor concluded that Kevin had experienced cognitive decline consistent with mild dementia. But the NFL rejected Kevin’s compensation claim, arguing that his doctor hadn’t adjusted his cognitive function test scores for his race in a practice known as “race norming.”
Race norming is a practice in which cognitive test scores are adjusted for Black people relative to White people, founded on the assumption that Black people start with lower cognitive skills at baseline. The use of race norming means that a Black person needs to fall to a lower level of cognitive function relative to a White person in order to be considered to have experienced the same amount of cognitive decline. After race norming of Kevin’s cognitive test score, he no longer qualified for the NFL’s compensation scheme.
“I felt so betrayed and I still feel that way. Two different systems? How can that be okay? White people should be upset too. Not just me, not just Black people. Because if the shoe was on the other foot, you wouldn’t like it,” Kevin told ABC News.
In August 2020, Kevin and another former NFL player, Najeh Davenport, sued the NFL, claiming that it had deliberately manipulated former players’ cognitive test scores in a way that reduced the likelihood of Black former players qualifying for compensation under its settlement scheme. The lawsuit was dismissed by a judge in March 2021, who instead ordered the NFL and the lead lawyer in the settlement to resolve the issue through mediation. Three months later, on June 2, the NFL pledged to stop settling head injury lawsuits using race norming and to review past scores for potential racial bias.181
That isn’t the end of the story, though. On the same day that the NFL announced that it would stop using the controversial practice to settle players’ head injury claims, Katherine Possin, a neuropsychologist at the University of California, San Francisco, told me that race norming of cognitive tests was still a fairly common practice among doctors in the US. “I would say it [race norming] is fairly often used,” Possin told me. “There’s no study that I’m aware of that surveys neuropsychologists [on their practices], but these are well respected norms,” she said.
When I interviewed Possin, I was in the midst of reporting an article for New Scientist, investigating racism in medicine. During the investigation, and in the months following the article’s publication, I learned some shocking things I want to share here. We have already seen how racist behaviors exhibited by individual doctors or other health care workers—interpersonal racism within medicine—can influence patient experiences and health outcomes. Yet, as I discovered during my investigation, this is only a small part of a much larger problem. Racism in Western medicine is systemic. Whether or not an individual doctor exhibits racist behavior is almost irrelevant if we consider that all doctors and other health care workers are operating within a wider system of medical education, guidelines, and practices in which racism is deeply entrenched. The use of race norming by the NFL in the US, to which Kevin’s story drew public attention, provides a perfect example of this. Race norming of neuropsychological tests wasn’t something that the NFL came up with out of nowhere; the practice already existed in medicine.182
Specifically, the practice that the NFL was using was established in 1991 and updated in 2004 by Robert Heaton, a clinical psychology researcher at the University of California, San Diego, as a way to try to account for the fact that African American people tended to score lower than White people on cognitive tests. Subsequent research has shown that adjusting cognitive test performance to take social factors such as education quality into account significantly reduces this variance by race. This fact is even acknowledged by Heaton and his colleagues in one of their research articles, in which they note that “racial/ethnicity disparities on neuropsychological testing may well be accounted for by quality of education […], literacy, acculturation, and other background differences as opposed to any direct result of race/ethnicity, per se.” It is for these exact reasons that Possin, the neuropsychologist we just heard from, believes that race norming is extremely problematic.183
“Being Black in America is on average associated with a less privileged social experience,” Possin told me, pointing out that Black people in the US are more likely than White people to experience poverty, early life adversity, and discrimination, and to be less comfortable within a cognitive testing environment that she noted “has historically been used to discriminate against Black people.” She believes it is these factors that explain average racial differences in cognitive test scores, rather than something inherent about race. As a result, she is one of several neuropsychologists in the US calling for the development of non-race-based approaches to diagnosis. In light of these calls for change within the neuropsychology field and in the wake of the NFL’s decision to stop using race norming, I was curious to know where Heaton himself stood on the issue, a few decades after he had first established these so-called “Heaton norms.”184
Unfortunately, Heaton didn’t agree to be interviewed when I reached out to him. However, he did email me a few comments, which he said he had also shared with “others in the media,” and he offered to answer any additional questions I had by email. “I really don’t have time to enter this fray, which has become far more political than science-based, but would try to answer any questions you may have,” he wrote in his response. During our exchange, Heaton acknowledged that observed differences in test performance between subgroups of the US population might be explained by racial discrimination, stressful life experiences, a lack of consistent access to good nutrition and health care, and limited educational opportunities, but he argued that measuring these factors directly would be too difficult. “These are extremely difficult to measure, quantify and ‘correct for’ in interpreting test results,” he said. “The fact remains, that a very substantial amount of variability in the test performance of normal adults can be ‘explained’ (accounted for) by the demographic variables of age, education, sex and race/ethnicity (together), so our best available norms ‘correct’ for these characteristics,” Heaton wrote to me.
Possin disagrees. “Race is a crude proxy for lifelong social experience,” she told me. “Genetic differences in cognition do not follow these race lines. So correcting for lifelong social experience with something very crude, like race, is not precise medicine,” she said. It also perpetuates the false idea that Black people are innately less intelligent than White people, she added. “That’s a big problem.” Not only that but race norming may additionally contribute to racial health disparities by making it more difficult for Black people to get diagnosed with illnesses associated with cognitive decline, such as dementia, as in Kevin’s case.
In order for race norming to be eliminated from cognitive testing, Possin told me she thought it would be helpful if prominent organizations, such as the International Neuropsychological Society and the American Psychological Association, took a stance on the issue. But both these organizations declined to do so when I asked them about it in 2021, following the NFL’s pledge to stop using race norming. “The INS does not have guidelines nor take a stance [on] race norming,” said the INS at the time. “We are a global organization and focus on topics that are applicable around the world.” In a similar vein, the APA responded: “The APA has no official position (or ‘stance’ as you put it) on race-based norming and cognitive testing.”
I would learn that receiving these sorts of responses wasn’t unusual when I pressed medical and scientific organizations to comment on race adjustments, which are ubiquitous in Western medicine.
My younger sister is an elite 400-meter sprinter who has competed internationally for Great Britain. She and I are very close and even though we now live in different countries, we still talk almost every day. When I was still living in the UK, she was just a train ride away and we used to visit each other often. I think it was probably during one of those visits, in early 2020, that she started telling me about some blood test results she had recently received. Her GP had told her that her creatinine level was a bit higher than normal—a potential indicator of a kidney problem. That wasn’t particularly surprising to me or to her; creatinine is a waste product produced by muscles, and so athletes, who tend to be more muscular on average, commonly have higher than average levels of the compound in their blood without this necessarily being associated with kidney problems. She said she had also shared her blood test results with a sports doctor, to get an opinion from someone more experienced with athletes. He confirmed that creatinine is derived from muscle metabolism and that levels are proportional to muscle mass. He also shared with her a list of factors, which he said could be responsible for raised creatinine levels. In addition to muscle mass, diet, and certain health conditions, one of the factors he listed was “Afro-Caribbean race,” my sister explained. “Could my race be affecting my creatinine level?” she asked me.
I was about to stumble on an answer to my sister’s question. I was at the beginning of my investigation into what I now refer to as “race-based medicine”—the practice of adjusting medical tests based on a person’s race or ethnicity. I had first learned about it many years earlier, in a 2015 TED talk by US academic and author Dorothy Roberts, but I had assumed it would have been a thing of the past by this point; in her talk, Roberts rightly points out that race is a social rather than a biological construct, which therefore has no place in twenty-first-century medical tests. I soon discovered, though, that contrary to my belief, race-based medicine was—and still is—alive and well.185
My first clue came in the summer of 2021, before I had started looking into race norming or Kevin Henry’s experience with the NFL. It was instead in a US-based study I came across that highlighted issues associated with a widespread practice of adjusting routine kidney test results based on a person’s race. In brief: if a doctor wants to assess a patient’s kidney health, they will usually start with a test that measures the level of the waste product creatinine in their blood—as it turned out, the same blood test that my sister had in the UK the previous year. The doctor or lab technician conducting the analysis will then plug that blood test result into an equation that calculates the patient’s estimated glomerular filtration rate (eGFR), which is the estimated rate at which the kidneys filter waste—an important indicator of how well the kidneys are functioning. Broadly speaking, more creatinine in the blood suggests a lower filtration rate by the kidneys (this is a bit of an oversimplification, but it will do for our purposes). If a patient’s eGFR is too low, that could therefore be a sign of a kidney problem. In the worst cases, when the kidneys aren’t functioning correctly, toxic waste products build up in the body—a condition that is fatal without treatment such as dialysis or a kidney transplant.186
What I gleaned from the study was that the most widely used eGFR equations globally, including in the US, included a specific numerical adjustment or multiplier that was applied to increase eGFR values for Black people. After the application of this Black race adjustment by a doctor or lab technician, a Black patient would end up with a higher eGFR compared with a non-Black patient who had received the same exact blood test results. Not only that, but the study showed that the removal of the race adjustment could improve the accuracy of kidney failure risk prediction among Black adults at high risk. Put another way, it indicated that the use of race adjustment might be reducing the accuracy of kidney function tests for Black people, by overestimating their kidney health.
As I went over the details of the study, my conversation with my sister came flooding back to me. I remembered that she’d had a similar test done in the UK the previous year. I called her and asked her about it again. This time, she shared some photographs of her test results with me, and I could see that it was indeed an eGFR test that she’d had. Printed in brackets above her results was the phrase “If Black multiply result by 1.21.”
Staring at my sister’s test result sheet, I had so many questions. Where did the Black race adjustment in eGFR come from? How many other countries have similar medical guidance? If race is a social construct, why is it being treated as a biological one? Are there similar race adjustments still being used in other areas of medicine? Might this be causing harm to Black patients? What is the medical definition of “Black,” anyway? Is it based on a patient’s self-identified race or is it based on the doctor’s assumptions about the patient’s race? What happens if a patient identifies as Mixed?
The list went on. Fortunately, I was by no means the first person to raise any of these questions. In addition to Dorothy Roberts, kidney doctors such as Vanessa Grubbs and Nwamaka Eneanya in the US have long been calling for the elimination of race-based medical practices entirely, highlighting the lack of evidence to support them. Their calls received renewed attention in the summer of 2020, as medical institutions clamored to publish “Black Lives Matter” statements in the wake of George Floyd’s murder and worldwide anti-racism protests. Students who were still being taught race-based practices in medical school began to increasingly question their teachers. Among those asking questions was Naomi Nkinsi, a medical student at the University of Washington.187
When I interviewed her in the summer of 2021, Nkinsi patiently walked me through the history behind various race adjustments used in medicine and filled me in on the work that she and others had been doing in recent years to try to bring about change. She explained that there were two equations commonly used in medicine to calculate eGFR, both of which included adjustments for Black race. The first of these equations, called MDRD, was developed in the 1990s. “What this equation does, it looks at a variety of factors that can impact someone’s kidney functioning. And one of those factors that is built into the equation is actually Black race,” Nkinsi told me. “So, if someone is identified as being Black—and there’s no specification whether they’re self-identified or if the physician says, ‘Oh, this person looks Black’—then their eGFR calculation is adjusted by 1.21,” she explained. I immediately recognized those three digits from my sister’s test result sheet.
The three digits that constitute the Black race adjustment in the MDRD equation originate from a small US study by the same name conducted in 1999, which found that study participants who self-identified as African American had higher levels of creatinine in their blood on average compared to those who identified as White. “From this they said, ‘Oh, well if they have a higher creatinine, it must mean that Black people have a higher muscle mass,’ ” Nkinsi told me. However, the MDRD study included only 1,628 participants, only 197 of whom identified as African American. “It’s based on this one observation that they found out of a very, very small population,” said Nkinsi. Moreover, there is very limited evidence to support the secondary assumption that Black people have more muscle mass compared to White people, or to people belonging to other racial or ethnic groups for that matter. That assumption is also at complete odds with the scientific understanding of race as socially constructed. “I do not think there is any evidence that race is related to muscle mass. I think there are poor attempts at showing evidence that this is the case,” Nkinsi emphasized.188
In 2009, an updated eGFR equation—the second of the two most widely used equations—was developed based on a larger study called CKD-EPI. But the assumption that it was necessary to adjust for Black race was carried through from the MDRD study, Nkinsi explained. Indeed, both the MDRD and CKD-EPI equations for calculating eGFR contain a Black race multiplier (in the MDRD equation the multiplier is 1.212, whereas in the CKD-EPI equation it is 1.159). “So now, we have all of these aspects of medicine that are all dependent on these equations that are now being recognized to be built on faulty science,” said Nkinsi. Not just faulty, but potentially harmful.189
Around the same time I met Nkinsi, I came across preliminary research from the UK led by kidney doctors Rouvick Gama and Kate Bramham at King’s College London. Their research showed that eGFR equations with race adjustments overestimated actual GFR in Black patients, compared to results using a more invasive but more accurate method. I reached out to Gama and Bramham and they agreed to meet me over a video call to discuss the implications of their findings. Gama told me that the overestimation of kidney health due to race adjustment could have serious consequences for Black patients. “It could lead to delay in diagnosis of chronic kidney disease,” he said, and therefore delays in treatment—something that is reflected in UK health statistics. “If you’re of Black ethnicity, you’re three- to five-fold more likely to end up with end-stage kidney disease,” Bramham told me. “Almost certainly we’re not recognizing it enough,” she said.
There is a similar picture in the US, where, according to the National Kidney Foundation, Black or African American people are more than three times as likely as White people to develop kidney failure. Nkinsi told me she thought that the race adjustment was a contributing factor to that disparity. “What it means is that we’re missing kidney disease in Black people,” she said. “In order for a Black person to be identified as having kidneys that are sick, they actually need to be sicker than a non-Black person,” she emphasized. “That means that you have later access to specialized care, later access to things like transplant, later access to things like Medicaid coverage for kidney care. That leads to worse health outcomes for Black people,” she said.
In the US, it is estimated that thirteen people die every day while waiting for a kidney transplant. Black or African American transplant candidates wait longer on average than White transplant candidates for kidney and other organ transplants—and use of race adjustment in calculating eGFR has meant that their chances of getting onto the transplant list in the first place have been limited. I didn’t know it at the time, but as Nkinsi and I were talking, Jordan Crowley, a college student in New York who was born with just one partially functioning kidney, was having to wait to get onto the transplant list for that very reason. Jordan identifies as Mixed; he has one Black grandparent and one White grandparent on each side of his family. When assessing his kidney health, his medical providers racialized him as Black and applied the notorious race adjustment to his eGFR calculation. As a result, he didn’t qualify for the transplant list. Yet if his doctors had racialized him as White, he would have.
Jennifer Tsai, a US-based physician and writer, drew attention to Jordan’s case in an article published in Slate a few weeks after my first conversation with Nkinsi. Tsai was introduced to Jordan by his uncle, Jay Kaufman—the epidemiologist at McGill University. Tsai and Kaufman had been collaborating on a study, which has since been published in a scientific journal. It estimated that removal of the race adjustment from eGFR calculations in the US would have resulted in 31,000 additional Black patients becoming eligible for a transplant from 2015 to 2018.190
On October 1, 2021, Jordan filed a lawsuit against his medical providers alleging racial discrimination. His ordeal encapsulates the absurdity of using race to determine biology and illustrates how devastating the consequences can be to a person’s life. The case filing sums it up: “It has been known for decades in the scientific and medical community that the use of this medical algorithm is not grounded in science, and that race is an arbitrary and inaccurate indicator of a person’s kidney health.” Why on Earth, then, was the algorithm still being used? That was the question I wanted to get to the bottom of. The answer seemed to lie in the world of medical guidelines.191
As I learned more about the use of race adjustment in eGFR and its potential harms for Black people, I began reaching out to some of the health bodies who I thought might be responsible for setting this sort of medical guidance. That’s when I began to realize the scale of the problem, and why seemingly little progress had been made in this area despite so many medical students, doctors, and academics calling for change.
Toward the end of May 2021, I sent a media request to the CDC in the US, mainly to clarify what their current guidance was for the calculation of eGFR in adults and to ask if they recommended the use of race adjustment. I received a friendly response explaining that the CDC didn’t issue guidelines on eGFR calculation, but that the current guidelines in the US came from Kidney Disease Improving Global Outcomes (KDIGO)—“a global organization that develops and implements evidence-based clinical practice guidelines in kidney disease,” according to my contact at the CDC, who also kindly sent me a link to KDIGO’s website. I clicked on the link and downloaded the most up-to-date version of KDIGO’s guideline for the evaluation and management of chronic kidney disease. The guideline was laden with medical jargon and didn’t make for the most thrilling of evening reads, but I soon landed on the section I was searching for. I was once again greeted with the same instruction that had been printed on my sister’s eGFR results, to multiply the patient’s eGFR by a specific numerical factor “if Black.”192
My heart sank. I knew this was a global organization, so its guidance would be relied upon internationally. I drafted an email to send to KDIGO’s communications team. I asked if they could confirm that the guideline I had been looking at was their most up-to-date one, and to explain what the scientific rationale was behind their recommendation to use race adjustment in eGFR calculations. I received a prompt response, confirming that the guideline was up to date and curtly stating: “KDIGO is not in a position to comment on the rationale used to determine the adjustment in eGFR calculations.” I felt a bit deflated, but I didn’t give up.
When I had spoken to Nkinsi, she had told me about how she had questioned her teachers at medical school the first time her class was taught about eGFR. “When we were presented in lecture with, ‘Okay, so we’re using the MDRD equation, you have to adjust for race,’ I was like, ‘Wait a minute, that doesn’t make sense,’ ” Nkinsi told me. “In our physiology lecture, no one ever said that Black kidneys work different,” she pointed out. “What is the physiology?” No one could give Nkinsi a straight answer. “They were just kind of like, ‘Well, this is how we do things,’ ” she said. Nkinsi wasn’t satisfied, and started digging, which is how she discovered the concerning origin of the race adjustment in the MDRD study. “One study can change the trajectory of so much of what we do, because it’s further grandfathered in,” she sighed.
“Medicine and scientists, we are supposed to pride ourselves in gaining new information, looking at how the world is changing, getting new data and saying, ‘Okay, what are we doing?’ Nothing we do is solid, right? Everything is supposed to be able to be questioned, right? But, unfortunately, medicine is very hierarchical. It’s very much based on this kind of false meritocracy, where for the longest time, White men are running everything, you can’t question your superiors, medical students and younger physicians are supposed to just take things as they’re taught. You’re taught these equations, you’re taught these facts, and you just kind of perpetuate them,” said Nkinsi. “You’re not really taught to question where this information is coming from.”
Thankfully, Nkinsi did question things, and when she realized they were fishy, she started to demand action from UW Medicine, the medical school where she was enrolled at the University of Washington. In response to Nkinsi’s calls, in 2020 UW Medicine announced it would transition away from the use of race adjustment in eGFR calculations. They were not the first or the last medical institution in the US to make this move. Other institutions, such as Beth Israel Deaconess Medical Center in Massachusetts, Massachusetts General Brigham, Mount Sinai Health System in New York, Zuckerberg San Francisco General Hospital, and Vanderbilt University Medical Center in Tennessee also abolished the race adjustment, and in 2020 the National Kidney Foundation and the American Society of Nephrology established a joint task force to “examine the inclusion of race in the estimation of GFR.”193
Inspired by Nkinsi’s success, I continued reaching out to various health bodies in other countries regarding their stances on race adjustment. I received a response from the UK’s National Institute for Health and Care Excellence (NICE) in June 2021, regarding its guideline on calculating eGFR, which at the time recommended applying “a correction factor to GFR values […] for people of African-Caribbean or African family origin.” NICE told me it was in the process of updating that guideline and reassured me that “there was some consideration in the update about adjustment based on the characteristics you have highlighted.” In other words, NICE was considering updating its recommendation to adjust eGFR for race. My contact at NICE’s press office shared a link to the draft of the updated guidance, which was due to be published two months later, in August 2021. When I read through it, I was disappointed to see that the recommendation to “correct” eGFR for people of African-Caribbean or African family origin was still there. It was accompanied by a note saying that future research should explore the use of “factors other than ethnicity” as biological markers. At this point, I decided that the best thing I could do was put my head down and write up my report for New Scientist.194
A few weeks after my article was published, Gama and Bramham—the UK-based nephrologists—got in contact to let me know that their preliminary study, highlighting the potential harms of race adjustment in eGFR to Black patients in the UK, had been published in a scientific journal. It was August 2021, and I knew that NICE was due to publish its updated guideline soon. I forwarded Gama and Bramham’s research article to my contact at the NICE press office and asked whether NICE was aware of their findings and if it planned to change its guideline on the use of race adjustment in eGFR. After a couple of days, the response I had been waiting for finally came. NICE had made the decision to remove the race adjustment from its recommendations on calculating eGFR. It would publish its updated guideline the following morning. I was delighted, as were Gama and Bramham. “This is a really important opportunity to stop race-based medicine,” Bramham told me at the time.195
Indeed, it finally felt as if there was movement in the right direction. I wrote up a quick report on the guideline change for New Scientist, reaching out to the UK Kidney Association to get their reaction to the news. “We welcome and support this change,” said Paul Cockwell, president of the association. “Ethnicity and race are social constructs and do not match genetic categories,” he said. “Adjusting for kidney function based on ethnicity could lead to an overestimation of kidney function and potential inequality in delivery of care,” Cockwell concluded.196
On the other side of the Atlantic, there had also been some movement on this issue. A few weeks after NICE published its updated guideline, the National Kidney Foundation and the American Society for Nephrology formally established a consensus against the use of race adjustment in kidney function equations in the US. Nkinsi and I caught up afterward. “It is exciting to see that such a big change is being made,” she told me. “But the concern is that institutions will see this as, ‘Okay, now we’re not racist anymore,’ ” she added. “There’s a concern that once institutions around the country are making this change, they’re not going to go deeper and look at, what are other ways that Black patients are getting inadequate kidney care?” she explained.197
The change is unlikely to be instantaneous either. Kaufman, the epidemiologist at McGill, explained to me that while the National Kidney Foundation and American Society for Nephrology are influential organizations in the US, they don’t have any power to dictate what doctors do. “It’s not like a law,” he said. “This is just a professional recommendation.” That means it is up to individual medical institutions to adopt the change. “I’m expecting they’re going to drag their feet,” Kaufman concluded.
In the UK, it will probably also take time for medical practice across the country to catch up with the updated guideline from NICE. In January 2023, more than a year after it updated its guideline on eGFR to remove the race adjustment, my sister forwarded me her results from a recent medical checkup. Under her eGFR result was the now all too familiar phrase “If patient is black, multiply by 1.21.” Kate Bramham explained to me at the time that—at least in England where my sister is based—laboratories “are often not changing their practices until they upgrade their software.” As of February 2023, most labs in England had changed their practices, according to Bramham, so perhaps my sister was unlucky. Fortunately for my sister, her creatinine level and eGFR were considered normal given that she is a muscular athlete, irrespective of race. For others, though—such as Jordan Crowley in the US—the use of race adjustment literally means the difference between being eligible for a kidney transplant or not.
Although progress might be slow, the guideline changes in the US and UK were still meaningful and they have been followed by further positive developments. As an example: in June 2022 the board of the Organ Procurement and Transplantation Network, which connects transplant centers and develops policies on transplantation in the US, approved a measure requiring transplant hospitals to use a “race-neutral calculation” when estimating a patient’s kidney function. This measure would seem to be beneficial for Jordan’s legal case against his medical providers, which is ongoing at the time of writing.198
“This is just the beginning,” Nkinsi told me. She also pointed out that race adjustment in eGFR represented just one of many persistent race-based medical practices—a growing number of which have come under criticism both within and beyond the medical community in recent years.
Thankfully, different fields of medicine are one by one joining calls to eliminate or at least reconsider the use of race-based medical practices—some of which have unequivocally racist origins. For example, widely used race adjustments in spirometry tests for lung function originate from a suggestion by US doctor and slaveholder Samuel Cartwright back in the 1850s. He claimed that Black people had lower lung capacity than White people and were therefore only healthy when enslaved (you may recognize Cartwright from our brief discussion of “drapetomania,” a disease he invented to pathologize enslaved Black people who tried to escape). To reiterate: a racist notion promulgated by a nineteenth-century apologist for slavery is still affecting people’s health care today. But lung doctors, such as Alexander Moffett at the University of Pennsylvania, are working to change that.199
When I interviewed Moffett in 2021, he had recently presented findings from a preliminary study—on which he collaborated with kidney doctor Nwamaka Eneanya among other colleagues—at the 2021 American Thoracic Society International Conference. The study analyzed data from more than 14,000 lung function tests in the US and showed that removing race adjustments from the interpretation of the tests saw the number of people correctly diagnosed with a lung defect increase from about 60 to 82 percent. These results suggest that adjusting for race in lung function tests may underestimate the severity of lung disease among Black patients, Moffett explained to me. “We’re assuming that their lung function should be worse and using that as a way to approach the diagnosis,” he said. That assumption is flawed. “Everything we’ve learned about race in the last fifty years has invalidated this,” said Moffett. He and his colleagues are working on developing a method for interpreting lung function test results, which isn’t built upon unfounded assumptions about racial difference.200
At the time, joint European Respiratory Society (ERS) and American Thoracic Society (ATS) technical standards still recommended the use of different lung function test equations depending on a person’s ethnicity. While I was working on my article, I asked the ERS about the scientific rationale behind this recommendation. Here’s part of the long response I received, which the ERS attributed to “authors of the guidelines paper”: “Traditionally population-specific reference equations have been developed to account for observed differences between people living in different geographical areas. However, the reasons for observed differences in lung function between people around the world are multifactorial and not fully understood.”201
I also reached out to the ATS at the time, who were more forthright in acknowledging the problems with the recommendation. “Ingrained in lung function interpretation is the long-standing assumption that the observed differences across racial and ethnic populations are biologically based,” the ATS said in a statement. “There is increasing recognition that race and ethnicity are sociopolitical constructs which are more reflective of the differing social and environmental conditions across populations than representative of true biological differences,” it noted, adding that it was “committed to leading action to address racism in medicine and eliminate the misuse of race and ethnicity in clinical decision making” and that it had “convened a workshop to critically evaluate current guidelines.” In 2023, the ATS issued an official statement for clinicians explaining why race and ethnicity should no longer be considered factors in interpreting the results of spirometry. The statement was endorsed by the ERS.202
Some medical fields have moved more quickly than others when it comes to reviewing and updating race-based practices. As we discovered earlier, the VBAC calculator—a tool commonly used by doctors in the US for assessing the safety of a vaginal birth after a previous cesarean section—used to contain adjustments for African American and Hispanic women, which reduced their predicted chances of a successful VBAC. But that changed in May 2021, after William Grobman at Northwestern University in Illinois and colleagues—the researchers who originally developed the calculator—published an updated version with the race and ethnicity adjustments removed. Grobman and colleagues cited equity concerns raised in an earlier paper from 2019 as the reason they made the change. The lead author of that 2019 paper was Darshali Vyas, who was then a medical student and is now a resident physician at Massachusetts General Hospital. Reading Vyas’s paper was how I first learned about the concerning origins of the unevidenced claim made by D’lissa Parkes’s doctor in the UK, that Black women’s pelvises are shaped differently to White women’s.
I caught up with Vyas at the end of 2021 and we reflected on the sometimes fast, but more often painfully slow, changes that were taking place across different fields of medicine, starting with obstetrics and gynecology and the recent changes that had been made to the VBAC calculator. “I’ve gotten to talk to Dr. Grobman and the group and I feel very optimistic about the changes they made,” Vyas told me. “I also think it’s super powerful that it was the same group that created the first calculator, who then took a critique, really thought about it, and revalidated the tool without it and showed that it was just as robust and just as useful a tool,” she added. “I think that is meaningful. And it’s honestly what good equity work, I think, should look like.” Vyas explained that unlike with the calculation of eGFR, which depends on individual medical institutions taking the initiative to remove the use of race adjustment, the VBAC calculator is accessed via a single online source and so was much easier and quicker to change. “They’ve updated the official calculator online, so it is now fully replaced,” she said.
Like Nkinsi, Vyas was in medical school when she first became concerned about the use of race-based algorithms in clinical practice. She noticed a contradiction between how she was being taught about race academically—as a social construct with no biological basis—and how race was being used by the doctors she was learning from on the wards. She joined a group of students at her university called the Harvard Medical School Racial Justice Coalition, which is where she realized that a lot of her classmates shared her concerns. “I think a lot of my classmates felt a similar tension,” she told me. The following year, Vyas started a core clinical rotation in obstetrics and gynecology, which is where she first encountered the VBAC calculator. “Because I had this framework from social justice organizing and from conversations with my classmates before I got to the wards, I think I was very primed to pick up on this example of race correction,” she reflected.
As Vyas transitioned from being a medical student into being a doctor, she came across more and more examples of race corrections or adjustments being used across different fields of medicine. In addition to various medical fields having individual moments of reckoning on the specific race-based medical practices within their respective domains, she wanted to stimulate a much broader discussion about the problem with the way race is used within medicine overall. “I think we need some sort of guiding principles in medicine, overall, about how we should and should not think about race,” said Vyas.
Indeed, it’s clear from what happened to D’lissa Parkes that inaccurate notions about race in medicine can cause harm without even being embedded into algorithms. US-based physician Richard Garcia alluded to this in a 2004 commentary published in the medical journal Pediatrics, in which he wrote about his childhood friend Lela, who wasn’t diagnosed with cystic fibrosis until she was eight years old. “Had she been a white child, or had no visible ‘race’ at all, she probably would have gotten the correct diagnosis and treatment much earlier. Only when she was 8 did a radiologist, who had never seen her face to face, notice her chest radiograph and ask, ‘Who’s the kid with CF [cystic fibrosis]?’ ” Garcia wrote.203
According to the Cystic Fibrosis Foundation, a US-based nonprofit organization, about 5 percent of people with cystic fibrosis in the US identify as Black. Yet the condition is often inaccurately perceived as an exclusively “White disease,” which may contribute to underdiagnosis and poorer outcomes with the condition among people of color. Sickle cell disease (the group of blood disorders that includes sickle cell anemia and other related conditions) has been similarly racialized as a “Black disease,” which in addition to being inaccurate has had the added disadvantage of contributing to the marginalization and dismissal of sickle cell patients due to racism. Sickle cell is known to be associated with extremely painful episodes called crises, yet in both the US and UK there is evidence that Black patients requesting pain relief are often disbelieved by health care workers and stereotyped as drug-seeking. “I definitely feel that race does play a significant role in how patients are treated, especially in A&E. I think there is the misconception that the drug-seeking patients are back here again,” June Okochi, a sickle cell patient representative in the UK, told a 2021 inquiry into avoidable deaths and failures of care for sickle cell patients. On top of the health harms of receiving inadequate care, the stress associated with this racist treatment results in many patients avoiding health care altogether, further contributing to health inequity.204
Historically, the sickle cell gene occurred more often in parts of the world where malaria was more common, including equatorial Africa, the Middle East, and India, because of its protective effect against malaria. While that is helpful for understanding population genetics, it doesn’t provide an accurate way of diagnosing an individual patient. Like sickle cell, cystic fibrosis is also a genetic condition and, nowadays, the most accurate way to identify genetic conditions like these is to leave the racialized assumptions at the door and simply test for the presence of mutations in the relevant genes.205
In general, vague notions of race don’t provide accurate ways of diagnosing people, even in cases where they may have been legitimized through their incorporation into medical algorithms. In these cases, race is often being used as a proxy for something else, such as muscle mass in the case of the eGFR equations. Nkinsi summed up the problem: “We’re using race as a proxy for other things, instead of measuring those things directly.”
After my article about the problems with race-based medicine was published, a few readers wrote to New Scientist with critical comments and questions. The question I think most of them were trying to get at was: how should doctors consider race? Surely, they couldn’t be expected to ignore race entirely, since it is clearly associated with people’s health outcomes. I definitely don’t think the movement away from race-based medicine should be a movement toward color-blind medicine. But I do think there is a big difference between using race as a proxy for health and acknowledging racism as an underlying process contributing to health inequity.
When I put these points to her, Vyas argued that “I would actually say that putting race into tools is a form of race-blind medicine, because you’re hiding disparities.” She went on, “You’re sort of shaping your tools around existing disparities, rather than actually addressing them at their root cause,” and further, “I firmly believe that we should always continue to study race and its effects on health outcomes. We should study and understand how racism affects health outcomes. And we should really be quite explicit in understanding how racism is impacting our patients.” In the end, “Where we draw a distinction is between studying racial inequities and building our tools in a way that could just exacerbate them or hide those disparities completely,” she concluded.
Nkinsi echoed these comments. “We want people to practice antiracist medicine, not color-blind medicine,” she told me. “What that means is that you recognize that race isn’t biological, it’s a social construct. And that these are very real social dynamics in how people are treated and how they experience life,” she said. As a physician that means thinking about racism and how that affects your patients, rather than categorizing them based on pseudoscientific assumptions about race.
Racism has a profound effect both on people’s health and on the way they are treated within the health care system when they become ill. The persistence of race-based medicine in the form of algorithms is testament to just how deeply racism is entrenched in Western medical practice. In August 2020, just over a year after her paper about the VBAC calculator was published, Vyas coauthored another paper in the New England Journal of Medicine, which detailed examples of race-based algorithms from various medical fields and highlighted their harms. “By embedding race into the basic data and decisions of health care, these algorithms propagate race-based medicine,” Vyas and colleagues argued. Their paper, titled “Hidden in Plain Sight—Reconsidering the Use of Race Correction in Clinical Algorithms,” played a key role in helping to catapult the issue of race-based medicine into mainstream medical and scientific discussions. It has since received more than six hundred citations, including from US senators Elizabeth Warren, Ron Wyden, and Cory Booker, and Congresswoman Barbara Lee, who a month after its publication requested a national review of the use of race-based medical algorithms. “In order to reduce health disparities among communities of color, we must ensure that medicine and public health organizations take a staunchly anti-racist approach to medical care and reevaluate the ways in which current practices, including the use of race-based algorithms, could be worsening outcomes for people of color,” the four politicians wrote in a letter to the Agency for Healthcare Research and Quality in September 2020.206
“I think many people would accurately say this is overdue and certainly it’s important to highlight there have been scholars, Black scholars, Black doctors, Black activists, who have been sounding the call around race-based medicine for a long time,” Vyas told me. “It might be something about the past few years—a global pandemic that has really highlighted these disparities, as well as mounting police violence, and police killings of Black people—that maybe has cast this in a new light that has gained more momentum, but certainly these conversations preceded me,” she acknowledged.
Vyas’s 2020 paper lists many more examples of race-based algorithms than I have space to write about—and during the process of writing this book I have continued to discover more. After my New Scientist investigation was published, I received a message via social media from Dipesh Gopal, a GP at Queen Mary University of London. He told me he had been pleased to see my article about ethnicity-based kidney test adjustment being removed from UK medical guidance issued by NICE, and he mentioned that he had been trying to encourage NICE to review another one of its guidelines with less success. We met over a call so that he could elaborate.
He explained that the NICE guideline on the treatment of hypertension or high blood pressure recommends different treatments for people depending on their ethnicity. As of 2022, the guideline suggests that doctors should prescribe drugs called ACE inhibitors to people under the age of fifty-five with high blood pressure—unless they are of “black African or African-Caribbean family origin,” in which case it recommends prescribing different drugs. Gopal told me he and his colleagues had written to NICE twice requesting that the guidance be urgently reviewed. NICE declined in both cases, responding that evidence indicated there were “clinically meaningful differences in the effectiveness of treatments for individuals in these family origin subgroups.” However, Gopal and other doctors and academics—including epidemiologist Jay Kaufman, and physician and writer Jennifer Tsai—dispute this evidence.207
Kaufman pointed out several flaws in this evidence in an email sent to Gopal and others following NICE’s decision not to review its guidance. He noted that while it was true that some studies had found that Black people on average have less activity of an enzyme called renin in their blood—an indicator that they may respond less well to treatment with ACE inhibitors—those studies also found a large amount of overlap in renin activity levels between people from different racial and ethnic groups. Based on data from one such study, if you select any Black or White patient at random, there is a one-in-three chance that the Black patient will have a higher plasma renin activity than the White patient, said Kaufman. This seemed to run counter to another argument made by NICE in its response to Gopal and colleagues, that ethnicity provided “a pragmatic means of estimating a patient’s renin status.”208
When we spoke, Gopal also noted that it isn’t clear from the guideline how doctors are expected to determine a patient’s ethnicity in practice, or what to do in cases where people have more than one racial or ethnic identity. Tsai and I decided to write an article to highlight this issue—and to call for NICE and other medical institutions globally to systematically review race-based recommendations across their guidelines once and for all. In response to Gopal and his colleagues, and to the content of our article, NICE acknowledged that “there is not a clear-cut biological and genetic homogeneity amongst all Black and White people” and that “the guideline does not account for people with mixed heritage.” But it said performing the relevant tests on everyone wasn’t possible due to “the expense, and the additional time.” Our article was published in New Scientist in November 2021 and, unfortunately, remains relevant.209
A few weeks before our article was published, in October 2021, the House Ways and Means Committee—a US congressional committee—released a report probing what it described as the “misuse of race in clinical algorithms,” citing Vyas’s 2020 paper among others. The report noted that although there is desire within the medical community to improve current practices, “divergent opinions on the appropriate use of race and ethnicity in CDSTs [clinical decision support tools] make it difficult to develop broad consensus that would uniformly transform behavior across the medical community.”210
I am slightly more optimistic. In the process of writing this book, I have gotten to know so many inspiring patient advocates, medical students, doctors, academics, and other activists who have dedicated themselves to advancing health and medical equity, and I am excited and passionate about the vision of anti-racist medicine that they and so many others are working toward. I understand that racism can be very sticky and therefore slow, painful, and difficult to excise, including from the institution of Western medicine. But I believe it is possible, and we have to start somewhere. As Nkinsi said, this is just the beginning.
In addition to medicine, racism in science also contributes to health inequities. In particular, biases and data gaps in medical research may be entrenching existing racial and ethnic health gaps. And that begins with what is missing—or, rather, who is missing—from health data and medical studies.