(shorter version published in Nautilus magazine.)
The Adjustment Bureau
For decades, physicians have assessed, and treated, Black patients differently, citing both epidemiological evidence and physiological differences as a ‘correction for race.’ In the quest to provide ‘equitable’ care, has this ‘separateness’ perpetuated the very inequities it hoped to ameliorate?
Amitha Kalaichandran MD
‘Race is child of racism, not the father.’ – Ta’nehisi Coates
“So, what prescription management would you recommend?” My attending asked me. We had just admitted a Black male patient to our inpatient ward at a large teaching hospital. He had been in hypertensive crisis, secondary to kidney dysfunction – all as a result, ultimately, of poorly controlled Type 2 diabetes. After the initial crisis was managed, we needed to step him down to a medication he could take at home. The standard first-line medication was an ACE (angiotensin converting enzyme) inhibitor or an ARB (angiotensin receptor blocker). Every patient I had seen before then was on one of these medications.
But this was our first patient of African descent, which meant the initial intuitive answer – an ACE inhibitor – was wrong.
This patient would instead be prescribed a CCB – a calcium channel blocker.
Why? Because a series of studies in the early 2000s, which ‘adjusted’ for race (“Black” vs “non Black”) found that Black research participants had a suboptimal response to first line ACE inhibitors. Instead, CCBs appeared to offer Black participants better blood pressure control.
I knew this well, of course, from graduate school training in epidemiology. It was called the “correction for race.” In other words: an adjustment to the final risk ratio (or odds ratio) based on the variable of race, which was typically binary (Black or not Black) and either self-reported by the research participant or determined by the researcher. The “correction” was a statistical adjustment meant to account for the confounding factor (the factor that can impact an association between the primary variable and outcome) of race. Other confounding factors included gender, body mass index, and socioeconomic status. In some instances, race was assigned to be categorical (meaning categories such as ‘Asian,’ ‘White,’ ‘non-White Latino,’ and ‘Black’) or binary (‘Black’ and ‘non-Black’).
In graduate school it was about numbers and statistics. But in medical school, coming face to face with a patient, and choosing a different approach, brought with it many questions. For one, the studies were based on African Americans. So why would a Black patient – an immigrant from West Africa in the case I described – attending a Canadian hospital, be treated the same way? Was the assumption, that there was some sort of biological difference, behind the patient receiving a different approach to care?
In 2004, the results of the African American Heart Failure Trial (A-HeFT), which randomized African American patients to the combination drug BiDil (hydralazine and isosorbide dinitrate) or placebo (though most were already on an ACE inhibitor) were published. The drug reduced the risk of death secondary to heart failure among self-identified Black participants. This led to the FDA recommending in 2005 that it be approved specifically for treatment of Black patients. Here’s the rub: only Black participants were included in the study, so whether patients of other ethnicities could benefit was unclear.
Race is a key social construct, and there’s no clear association with specific genes according to geneticists. Companies like 23 and Me and Ancestry.com misleadingly advertise their services as a way to determine ethnicity and ancestry (two separate but related ideas). They rely on a large pool of individuals who indicate in a survey where they believe their ancestry to be. That pool also provides DNA. Anyone that submits their own DNA becomes compared to the existing DNA that exists to determine their ancestry and ethnicity. There’s no comparison to a specific gene that marks African, Caucasian, Asian (and so forth) ‘race.’ An equivalent would be creating a study that obtained DNA and asked two questions: whether you take your shoes off when entering your house and your geographic origin. This would lead to a library of individuals who provided these three pieces of information. A cluster of ‘shoes off’ people in an imaginary country (geographic region) called Pandora might be found, along with the gene or allele frequencies of that country. Then, every-time someone else provides their DNA and one piece of information – whether or not they take their shoes off – their DNA gets compared to this larger library – with the results showing they are likely ‘from Pandora’ if they respond they take their shoes off. There is no biological reason per se, it’s simply an approach to categorizing an individual based on pre-existing geographic patterns.
As it turns out, later on, BiDil was found to have a potential genetic basis for action, so it’s likely that anyone with that particular gene variant may benefit from that drug. In other words, it should never have been a recommendation for one specific race. Since then, further ideas around ‘race-specific’ medications were disregarded.
We now know that, years after I cared for that patient, from a study published in early 2022, that prescribing a different first-line medication to Black patients may be linked to worse blood pressure management. As well, less access to the right medications for blood pressure is likely a bigger problem than the specific medication used. Another way to look at it: discrepancies are related to inequity, not race.
Underlying all of this are larger questions: how much were the differences found by those studies, and others, that inform the clinical decision for a given patient actually based on social factors (the social determinants of health, which include racism, lack of housing, lack of education, low income, and so forth, as opposed to simply what appeared to be surface level factors like race?
Indeed, is using race in a clinical decision-making algorithm doing more harm than good? To answer these questions, we must reconcile two eras as we enter a ‘post-COVID’ period:
The first is the era of ‘precision medicine,’ where we understand that personalized medical management – based on genetic factors is the future of medicine. Precision medicine asserts that biology as well as the environment that interacts with that biology (‘epigenetics’) is crucial.
And the second is the era of equity – ensuring ‘racial equity’ in everything from vaccine and pharmaceutical trials to the clinical setting. The era of equity brings unique calls to action and is an entirely different problem from the one that involves biology. We know for instance that Black patients face immense barriers to accessing healthcare – everything from pain management (something even celebrities like Serena Williams faced when giving birth), to how stroke prediction models are used, even how AI is used in hospital EMRs to predict risk of in hospital death (which I wrote about previously). We even know that there is in built racial bias in how pulse oximetry tools measure oxygen saturation – a key element of COVID risk assessment and management. Even a large proportion of U.S. medical students and residents, as recently as within the last eight years, have reported believing harmful ideas such as that Black patients have ‘thicker skin.’
If the intention behind ‘race correction’ is to provide better, more equitable care, why does it seem as though it has led to the opposite? To understand this we must first journey into the history of how these adjustments came to be, and more importantly ‘who’ pioneered them.
Medical researchers in the US have collected data on race/ethnicity in research studies since the 1970s as a way to measure health disparities but also ensure that studies were more representative of the wider population so that any findings could be generalized to the public (what epidemiologists term ‘external validity’). Then, using the variable of race (as well as gender, education, socioeconomic status), analyses are performed to look at how a specific treatment or therapy impacted a subgroup. Here’s an important distinction though: when research studies lead to clinical practice guidelines, no guideline suggests using one approach or therapy over another based on gender, education, or socioeconomic status. No ‘correction’ exists for these variables either. Race stands alone: a social construct that’s misused as a biological one.
This approach to race has a deeply shameful history in Western-based medicine. Perhaps among the earliest on record was from Thomas Jefferson’s Notes on the State of Virginia, in which he wrote about ‘dysfunctional’ Black lungs, justifying slavery as a way to improve blood supply to further develop these lungs. Jefferson was suggesting that slavery was beneficial to Black physiology without noting that any lung differences likely had a social cause that was due to slavery and over-exertion.
Then came Dr. Samuel Cartwright, a Louisiana-based physician who presented “A Report on the Diseases and Physical Peculiarities of the Negro Race” in March 1851, at the annual meeting of the Louisiana Medical Association. Documenting ‘diseases’ such as ‘drapetomania’ (the tendency for an enslaved person to attempt escape) and ‘dysaesthesia Ethiopia’ (‘laziness’ among the enslaved), Cartwright would present pseudoscience guised as medical observations, alongside previous theories around brain size to explain why ultimately those who were enslaved deserved to be. His work would also lend further credence to racist scholars at the time who believed that Black people possessed a higher pain tolerance, in order to justify not using anesthesia. Among them: the ‘father of modern gynecology,’ Dr. J Marion Sims, whose experimental vesico-vaginal fistula repairs were done without anesthesia on enslaved women, even when white women were offered anesthesia for the same. On the other side of the pond, in 1872, Charles Darwin would approach an analysis of emotions similarly. In his book, The Expression of Emotions in Man and Animals he argued, inter alia, that non-White races were more subject to emotions that they were too ‘primitive’ to control. His evolutionary theories, as we know, supported ‘scientific racism.’
There exists a direct line between these 18th and 19th century beliefs and modern-day medicine. As previously mentioned, perceptions around Black pain can be traced to Sims and others. How medicine tests lung function , and even how its measured (spirometry), can be traced to Jefferson.
Years later, others would theorize as to the potential genetic underpinnings of this “correction” physicians make for differences for things like blood pressure, an attempt to find biological plausibility (part of Hill’s criteria of causality), treading on infamously dangerous territory previously prodded on by myriad biological-anthropological scholars who, taking the baton from people like Cartwright, Sims, and Jefferson, had attempted to perpetuate the superiority of Whites. But nothing has been found. In fact, in a New England Journal of Medicine article, prominent geneticists clearly state: “Meaningful biologic differences simply do not exist between different races in a correlation or association.”
Indeed, any differences may be explained by sociocultural differences, much like how taking ones shoes off before entering a house is unique to some cultures over others: an adaptation that cannot be seen in biology or in our genes, and where an association will never be 1:1 with race or ethnicity. In 2015, University of Pennsylvania’s Dr. Dorothy Roberts give a cogent TED Talk where she posited that race is entirely separate from genetics. Years before, writers such as Malcolm Gladwell (in his New Yorker essay Black Like Them) have written compellingly about how race is a poor proxy for culture and experience.
This is distinct from seeing the variable of race as a way to ensure inclusivity. This magazine has for instance covered the issue of race in neuroscience but it’s clear that any anatomical differences are not secondary to biology, but rather that it’s sound practice to include reference images that span the globe, given that patients are culturally diverse. Indeed, neuroimaging studies have only found cultural links (not racial ones) to brain appearance. As Nigerian-British writer Chimamanda Ngozi Adichie has expressed, the dangers of applying a single story to a group, to assume each individual has a similar experience are clear. When these ideas are applied to how we approach health though, it’s catastrophic.
Yet it persists. One of the most prominent examples involves an “adjustment” to kidney function. As described in a paper in 2020, the eGFR (glomerular filtration rate – the glomerulus is part of the kidney) is a rough indicator of how well kidneys are able to do their key task of filtering the blood of proteins and electrolytes. Creatinine is a protein that is a byproduct of muscle breakdown. High creatinine is a rough marker for damaged kidneys. It has been presumed that Black patients have inherently higher muscle mass compared with other races – regardless of BMI or physical habitus. This correction has been built into how eGFR is calculated (the 2009 Chronic Kidney Disease Epidemiology Collaboration creatinine equation), but perceptions have persisted in several clinical guidelines. As such, advocates have continued to advocate against race being used in kidney treatment, which can position Black patients as a lower priority when it comes to dialysis or transplant.
These “corrections” are even being used in the tiniest in society – children. In a viewpoint in 2020 for JAMA Pediatrics, citing a reference to a 2011 urinary tract infection (UTI) clinical guideline, the authors posited that given race is neither binary nor biological, using different cut-offs for determining risk based on race both increased the likelihood that UTI in Black children could be missed or undertreated and that ‘Blackness’ conferred protection against UTI. The author, Assistant Professor of Emergency Medicine and Pediatrics at New York-Presbyterian Weill Cornell Medicine, Dr. Rachel Kowalsky, was prompted to explore the issue after noticing the 2011 guideline used race, whereas the earlier 1999 version, did not.
“It seemed so obvious to me, as a multiethnic person, that the categories White and Nonblack could not be operationalized in a place like New York City, where many people are multiethnic and multiracial. People with roots from all over the world would be thrown together into these categories based largely on skin color, and it didn’t seem possible that they could all share a common vulnerability to UTI, or common protective factor against it,” Kowalsky told me. “Race was being inappropriately used as a proxy for biology. And, because the AAP fever and UTI guideline applied to the majority of infants 2-24 months old coming to the ER or clinic with fever, an enormous number of children could be negatively affected by its use.”
After discussions with a close friend, a sociologist, who believed it was an unusual way to approach race, Kowalsky decided explored it through the viewpoint. The initial submission, to a very prominent pediatrics journal, was rejected. The killing of George Floyd invigorated her to resubmit.
“Exposure to racism and colorism has an enormous impact on health,” Kowalsky says, “In clinical research, there are many markers used for exposure to racism, from zip code to self-reported race to self-reported experience of racism.”
In their seminal 2020 commentary in the New England Journal of Medicine, Harvard Medical School Professor of the Culture of Medicine, Dr. David Jones, and colleagues carefully lay out several clinical practice guidelines that use the variable of race – typically ‘Black’ and ‘non Black’ – to decide on medical management.
“More and more people are realizing that tools that rely on dumb categories (e.g., a Black vs. White dichotomy) aren’t really scientific or evidence-based: it’s not defensible to put all humans into one of those two buckets. So, all of these tools are going to have to be reconsidered eventually,” Jones told me.
Examples run the gamut from endocrinology to pulmonology to oncology to urology to obstetrics and gynecology, to cardiology, cardiac surgery, and nephrology. In each of these fields there runs the risk of contributing to inequity. Jones and his colleagues write plainly: “By embedding race into the basic data and decisions of healthcare, these algorithms propagate race-based medicine.”
In 2022, a study in JAMA Pediatrics aimed to take a quantitative approach and looked at 414 clinical guidelines in pediatrics. Of the 30% used race or ethnicity as a term, a potential negative effective was seen in 49.7% (a potential positive effect, meaning a lower threshold for care, was viewed in 28.6%). The authors concluded inequities could indeed be exacerbated more often than helped, by using race.
If the intention behind these “corrections” was to lead to more equitable care, how did it end up doing the opposite? As such, should they not be halted?
Indeed, over the last few years, these guidelines been called into question – first with lung function, and then for kidney disease and even for how computer (electronic medical record) algorithms are also skewed to incorporate these ‘corrections,’ for a variety of conditions. The pediatric UTI guideline Kowalsky referenced was retired in 2021. Yet , several current ‘best practice’ clinical guidelines still incorporate these ‘corrections.’ So, is this the right decision, or are we missing another, more sinister issue?
In 1954, in the Brown vs Board of Education of Topeka decision to desegregate public schools, the Supreme Court held that facilities that were ‘separate but equal’ were de facto unequal and violated the fourteenth Amendment, of which one pillar of said amendment was ‘equal protection’ as:
All persons born or naturalized in the United States, and subject to the jurisdiction thereof, are citizens of the United States and of the State wherein they reside. No State shall make or enforce any law which shall abridge the privileges or immunities of citizens of the United States; nor shall any State deprive any person of life, liberty, or property, without due process of law; nor deny to any person within its jurisdiction the equal protection of the laws.
Over two decades later, in 1978, legal scholar Archibald Cox defended the policy of affirmative action (Regents of the University of California v. Bakke) by resting on the argument that a diverse student body was inherently a positive force that acted to ‘cure’ historical wrongs (as Thurgood Marshall once said “They owe us”) but mostly that it would help all students navigate a more pluralistic world. This case then resulted in sweeping changes to admission policies several elite Universities across the U.S., despite the idea of “The Negroes Have Arrived,” as described in Plessy vs Ferguson.
Then in 1996, a Black businessman named Ward Connerly became a prominent backer of Proposition 209, banning affirmative action in California.
This June, the Supreme Court is expected to render a decision that will likely ban affirmative action nationwide. The central argument against it is that one applicant cannot be treated differently for the benefit of ‘diversity’ as a value in itself.
In other words, equity – that is equal access (which acknowledges a different starting point and various advantages and disadvantages that come with demographic factors such as gender or race) – to higher education, cannot be attained by impinging on the equity of others. The path to fairness cannot be through unfairness. Even if diversity itself is a social good, how diversity is defined becomes key. Is it simply a matter of race, or is it a matter of social factors or experiences of adversity or diversity in views? In matters of race, many underrepresented students that benefit meaningfully from affirmative action have been those from the most affluent of society, but the idea that this is the majority is a myth. Race and social adversity don’t have a 1:1 association in the same way that a given gene is not 1:1 with race.
Might social adversity, not race, be a better marker of pluralism, if the goal is to prepare students for a pluralistic society (as Cox has argued). This appears to be the preference among Americans polled.
So, ridding ourselves of the last remnants of ‘race-based medicine,’ much like pushing for ‘race neutral’ college admissions policies makes sense right?
This is where things get a bit more complicated.
Precision medicine is an approach that aims to use genomics, and specific gene markers to better manage medical care. Unlike in classic epidemiology, Genome Wide Association Studies (GWAS) use the marker of race to ‘reduce the noise’ of other associated variables, and it helps ensure random differences are accounted for (e.g. height). This type of study is behind understanding immunology differences according to human leukocyte antigen (HLA) type, which can be associated with geographic region. HLA differences occur due to genetic drift secondary to exposure to different pathogens over the centuries (and why vaccine trials should be diverse geographically to account for this).
Markers like the MTHFR gene (which is involved in converting folate) are found across many geographies and ancestries. The same applies to the Apo-E4 gene variant which increases the risk for Alzheimer’s dementia. Even diseases with known gene mutations, such as sickle cell anemia or cystic fibrosis, lack a 1:1 association with Black or White race specifically. Precision medicine offers a way to do effectively what an optometrist would do before prescribing eyeglasses: test first and then provide a personalized approach as opposed to the alternative – one pair of glasses based on a superficial factor like age. Yet it too is limited.
Dr. Genevieve Wojick is an Assistant Professor in the Department of Genetic Epidemiology at Johns Hopkins Bloomberg School of Public Health who researches diverse populations and genetic associations. She believes that the variable of race is tricky when larger studies are translated into clinical decision-making.
“Race in most contexts is used as a proxy and disentangling it from what we really want to measure is a challenge that would require research to distill these variables down into their sort of root causes…for example, is it race we want to measure or the effects of racism?” Wojick asks. ““I always find very ironic when it comes to biomedical research that there’s a lot of emphasis on precision medicine. But what they think is precise is only a small subset of that equation, right? There’s acceptance of some kinds of imprecision, such as race, which is a very imprecise variable. Ideally, if you think genetics play a role, if you would just test their genetics and treat accordingly. But this is not an option for most people for a variety of reasons. And so, providers are stuck with these poor proxies.”
Wojick added that most research trials assume the default of “White race,” and that the problem persists. “At the root of it is this long legacy of scientific and medical racism which has persisted in some ways to the present day, with some people genuinely believing that there are biological differences between racial groups…instead of understanding that there’s no actual boundaries between people, right? That’s just not how it works.”
Dr. Joseph Wright, professor of Pediatrics and Health Policy & Management, and Chief Equity Officer of University of Maryland Schools of Medicine and Public Health believes that the precision medicine focus can run the risk of not being inclusive. “Consistent under-representation of diverse communities in trials that drive precision medicine approaches continues and has not been resolved. Plus, the growing acceleration of genetic admixture across the world renders the underlying promise of precision medicine less reliable,” Wright told me.
In a 2022 article, Wright and his colleagues describe the example of atherosclerotic risk as a case where even if Black race is used ‘favorable,’ to early identify hardened arteries, it can still be problematic. They write: “One might argue that the ASCVD Risk Estimator may be protective in terms of potentially skewing early cardiovascular care toward Black patients. However, the inherent danger is directing differential treatment to Black versus White patients on the basis of a flawed phenotypic signal in the face of what might otherwise be identical underlying risk profiles. Incorporating race as proxy for the biological effects of differential lived experience is misplaced in this example.”
At this point in my reporting, I had formulated a clear sense of why race should be omitted from clinical practice guidelines. It seemed obvious, given the lack of any meaningful biological reason.
But that was before I met Dr. Gregory Hall at a medical conference this past February. Hall is a primary care physician based in Cleveland, the Medical Director of University Hospital’s Cutler Center for Men, and & associate professor in Internal Medicine at Northeast Ohio Medical University College of Medicine. He wrote the book “Patient Centered Clinical Care for African Americans: A Concise Evidence-Based guide to Important Differences and Better Outcomes,” and his patient population is predominantly African-American, an ethnic group he also self-identifies with. His book emphasizes the key of establishing rapport and trust, and the role of equity. It also discusses differences in care for various systems – everything from cardiovascular care to diabetes and obesity, cancer care, renal disease, rhematic diseases, pulmonary and hematology disease.
In his conference presentation, he noted that African Americans as a group have the highest mortality, longest hospital length of stay, and low adherence to medications, all while being the least trusting of medical professionals. His conference presentation described a study that looked at the roots of prostate cancer. The study found prostate cancer risk was ten times higher in African American men compared to West African men, and so the cause was likely not along race lines but another factor: inequity.
Yet Hall disagrees with the principle of dismissing race in clinical decision making. He gives the example of how the biology may be involved or different, citing higher prevalence of HER2 negative cancers in breast cancers and the higher likelihood of aggressive prostate cancer that metastasizes to bone. Even if a specific gene is not known to be involved, he points out, the interaction of biology with the environment impacts disease. Some health organizations in his state of Ohio have stopped collecting race data in efforts to be race-neutral, but this worries him as it prevents an understanding of which communities may need more attention.
“What’s the biological impact of being oppressed?’ If we had a lab experiment where we oppress or subject to harsh treatment, any animal you pick would not thrive as well as those who aren’t oppressed – their cortisol is higher, for instance, which can increase their risk for a number of chronic diseases,” Hall told me. “The oppression and stress of being ‘not preferred’ may be causing some of these outcomes, and at its core, every action and reaction in humans is the result of a chemical reaction. It’s all biology.”
He shares how the experience of race is important in the clinical setting, and governs care. It relates to trust and empathy. “If I have an African American male in the exam room, I know I’m offering a different approach versus a small Asian American woman. The conversation is different,” Hall says. “The approach to starting medications is also different.” Some of the broader discussion around removing race/ethnicity as a variable might be fine, but I’m more focused on the patients in front of me and what they need . . . and what makes them comfortable with what I prescribe.”
Hall recalls how things like essential neutropenia (low white blood cell count) often seen in healthy Black patients can result in being denied chemotherapy, and how indeed some African Americans may have a creatinine that falls just outside of the upper threshold of ‘normal’ range, which suggests the ranges should be expanded.
“My worry is if we take race totally out of things, racial/ethnic communities that have serious health problems that need targeted attention will not be readily identified. Biologically we are 99.9% the same, and based on that alone, race really is a social construct, but that social construct has created the health landscape that we currently live in,” Hall told me. “If we completely remove race as a consideration in our clinical thinking and how it impacts patients’ predispositions, treatment, health, and access to services, we will be ignoring some very critical aspects of their overall health.”
Hall’s perspective led me to understand that a deeper issue was indeed at play. It wasn’t simply about whether to remove race from a series of clinical decision-making algorithms.
To understand it, we must yet again turn back to the Brown vs Board of Education decision to identify another sinister force. That decision rightly pointed that separate is prima facie ‘unequal.’ This applies to medical care – treating one race differently is a breeding ground for inequity. But an African American legal scholar, based at Duke University, named Charles Payne uncovered why focusing on this legal precedent alone can be misleading.
In his 2004 essay “The Entire U.S. is Southern: Brown v. Board and the Mystification of Race,” Payne, now at Rutgers University, argued that the Brown vs Board of Education decision didn’t ultimately get to the heart of why racism persists, and would continue to persist in the U.S. He writes:
“Clearly, part of the miscalculation involved a widespread tendency to overestimate the power of the law to make change and to underestimate the degree of racial intransigence outside of the South. Those miscalculations, though, may reflect a larger pattern. What the initial misreadings of Brown tell us is that by mid-century, national discourse about race has been thoroughly confused, the nature of racial oppression had been effectively mystified. A part of that mystification was a process by which the systematic character of white supremacy had been reduced to something called “segregation…” John Cell points out that the term is ‘profoundly ambiguous and self-contradictory” and contents “that this state of ambiguity and contradiction was skillfully and very deliberately created. Confusion has been one of segregations greatest strengths and achievements.”
What Payne is getting at is that focusing on one narrow policy shift – even if the shift is the right one to make – can mistakenly leave us apathetic to the larger forces at work. He understood that desegregation didn’t fix the larger issues of racism and oppression in this country, and it may in fact create the impression that all was well, leading to a more underground form of oppression.
To be sure, Jones and his colleagues were correct in saying that race may be useful in descriptive statistics but not in prescriptive guidelines, and that before using a guideline that involves race, physicians should pose these questions: Is the need for race correction based on robust evidence and statistical analyses (e.g., with consideration of internal and external validity, potential confounders, and bias)? Is there a plausible causal mechanism for the racial difference that justifies the race correction? And would implementing this race correction relieve or exacerbate health inequities?
These are principles that others, like Kowalsky, Jones, and Wright, have supported.
But removing race as a variable without acknowledging the other, deeper, systemic factors at play
runs the risk of winning a battle and losing – or even ignoring – the larger ‘war’ that is racism in medicine. The very outcomes that Jefferson, Sims, and Cartwright fueled are still at play, even if clinical practice guidelines no longer explicitly use race. Unbiased healthcare provision is not attained by removing these “corrections,” but a goal that can only be achieved through the harder path of shifting systemic barriers that consistently lead to worse health outcomes in Black patients.
VI: Representation without “Correction”
Beginning in 2020, several hospitals began to halt the use of race as a variable in their clinical decision making. The University of Washington stop using eGFR in their clinical decision making, and a multidisciplinary, multi-institution, group created an alternative “race-free” equation to assess kidney failure. In 2021, physicians from the University of Pennsylvania recommended that the American Thoracic Society revise their guidelines for assessing lung function/spirometry in 2021 (which was then backed by the New York City Health Commissioner). This was followed by a broader push by the University, even as a law professor (with a medical degree) at that same institution still perpetuates racist theories around intelligence.
Last year, in the Lancet Digital Health journal, Jones and his group published a paper recommending a revision to the atherosclerotic cardiovascular disease calculator without race. A few months later, a group in the University of Pittsburgh looked at Kowalsky’s claim and in their study in JAMA Pediatrics, using the data from sixteen previous studies (almost 180,000 children) found that replacing race with previous history of UTI and fever duration was found to be similarly accurate at predicting risk of UTI in statistical models. The American Academy of Pediatrics has offered another management guideline.
In mid-March, the National Academy of Sciences Committee on the Use of Race, Ethnicity, and Ancestry Population Descriptors in Genomics Research, of which Wojcik participated in, and which includes researchers, sociologists, and physicians, released a report entitled “Using Population Descriptors in Genetics and Genomics Research.” The report made clear the ongoing concerns around race, writing:
“Researchers have frequently used population descriptors as a shorthand for capturing the continuous and complex patterns of human genetic variation resulting from history, migration, and evolution. Of particular concern is the long-standing use of race, and more recently, ethnicity, as this shorthand. In humans, race is a socially constructed designation, a misleading and harmful surrogate for population genetic differences, and has long history of being incorrectly identified as the major genetic reason for phenotypic differences between groups. Rather, human genetic variation is the result of many forces – historical, social, biological – and no single variable fully represents this complexity…Race and racism have recently gained renewed attention from the U.S. scientific community. Recognition by the U.S. biomedical research community of the need to address the complex issue of population descriptors in genetics research has never been greater…this is a crucial moment to offer concrete guidance to the research community.”
The committee made thirteen recommendations for future research, with the following seven being directly tied to using race as a variable:
Recommendation 1: Researchers should not use race as a proxy for human genetic variation. In particular, researchers should not assign genetic ancestry group labels to individuals or sets of individuals based on their race, whether self-identified or not
Recommendation 2: When grouping people in studies of human genetic variation, researchers should avoid typological thinking, including the assumption and implication of hierarchy, homogeneity, distinct categories, or stability over time of the groups.
Recommendation 3: Researchers, as well as those who draw on their findings, should be attentive to the connotations and impacts of the terminology they use to label groups.
Recommendation 4: Researchers conducing human genetics studies should directly evaluate the environmental factors or exposures that are of potential relevance to their studies, rather than rely on population descriptors as proxies…Genetics and genomics researchers should collaborate with experts in the social sciences, epidemiology, environmental sciences, or other relevant disciplines to aid in these studies whenever possible.
Recommendation 6: Researchers should tailor their use of population descriptors to the type and purpose of the study, in alignment with the guiding principles, and explain how and why they used those descriptors. Where appropriate for the study objectives, researchers should consider using multiple descriptors for each study participants to improve clarity.
Recommendation 7: For each descriptor selected, labels should be applied consistently to all participants. For example, if ethnicity is the descriptor, all participants should be assigned ethnicity rather than labelling some by race, others by geography, and yet others by ethnicity or nationality.
Recommendation 8: Researchers should disclose the process by which they selected and assigned group labels and the rationale for any grouping of samples. Where new labels are developed for legacy samples, researchers should provide descriptions of new labels relative to old labels.
To “win the war” against racist medical care involves a longer, likely treacherous, road; one that must begin by meaningfully including the voices of those impacted. In the 18th and 19th century Sims, Cartwright, and Jefferson dictated how Black Americans were to be treated medically. Four centuries later the power dynamics aren’t entirely different. In other words, it’s not simply about race corrections in clinical practice guidelines – it’s about power: who has the power to make decisions for others – in this case whole groups of people. How are the power structures organized in hospitals healthcare centers that do research? Who are the decision makers who set the policies for how patients are treated, and how does this inherently disadvantage some groups over others even if it isn’t explicitly written?
“We need to collect much more comprehensive data about patients’ social and economic exposures, and do so longitudinally, so that we generate the data researchers need in order to understand what specific factors are generating the pervasive health inequities seen throughout US society,” Jones told me. This too will require a ton of work—but so did the human genome project. If doctors agree that something is important, then the needed research will get done.”
In Wright’s paper, he and his colleagues made several of their own recommendations as a step towards acknowledging that an anti-racist approach to healthcare would need much more than simply dropping race from clinical decision-making tools. For instance, they suggested physicians abide by the American Academy of Pediatrics’ Words Matter guideline, to reduce bias in how care is provided, while also acknowledging inequities as directly impacting health and access to healthcare.
“Even in the face of insurmountable evidence that race is not an independent biologic proxy, there is definitely a role for what race represents in terms of capturing differential lived experiences and exposures many of which are rooted in racism,” Wright told me. “It will be necessary to rigorously incorporate socially determined factors that frame health status into the development of clinical care guidelines to address inequities, structural or otherwise, eliminate disparities and achieve more equitable outcomes.
In his essay, Charles Payne wrote that desegregation doesn’t get rid of racism, instead it should shed light on all the deeply rooted ways that segregation persists. Removing affirmative action in college admissions shouldn’t lead to a race neutral approach to education – it should instead consider how social inequities, many fueled by the social construct of race, often begin years or decades (and generations) earlier, and thus, so should interventions, to level the playing field. It remains to be seen if the decision on this policy, as expected in June, will view equity through this lens.
“If we can help people address their racial biases rather than suggesting that they don’t really exist or that race doesn’t matter, we have a better chance of moving the needle,” Hall told me. “The only way to achieve progress in African American health is to name it and address it head-on. I also firmly believe that this debate is a critical, and inclusive, component for improving African American health and no medical decision should be purely race-based.”
Indeed, to move from “race-based” to race conscious medicine requires bravely challenging the very structures that were responsible for creating the healing profession in the first place. As Wojick pointed out ‘there are no boundaries between people.’ Will the profession, and those within it, be courageous enough help dissolve those boundaries, even the ones the system perpetuates in order to remain vital?
“To be ‘race-conscious’ of something is to consider it as part of the decision-making process. As a start, when making far-reaching decisions that impact the African American community, having an expert in African American health in the room would be more valuable than not having one contribute,” Hall told me. “Most of the race-based mistakes were driven by bias, not true science, and there was no one Black in the room to question the decision. Now that we have an opportunity to contribute (in some settings), I’m sure better, more solution-focused decisions can be developed.”