Why racialized medicine doesn’t work
“An Englishman Tastes the Sweat of an African” reads the caption of the black and white engraving.
Dated 1725, it documents everyday life in the slave trading post of Calabar, West Africa. Part of a work that later became a kind of instructional guide for European seafarers, it shows a man being examined before he’s sold into slavery and shipped to the New World. There is an obvious inhumanity to it, the flavor of a cattle market. But it’s made unnerving also by showing a strange act that is the engraving’s central focus. The tall, well-built African slave dressed only in a loincloth crouches on his knees so that a skinny-legged, fully clothed Englishman with a sword hanging by his side can reach his face with an outstretched tongue to lick his chin.
In the following engraving, a ship sets off with its cargo of slaves, as families left behind mourn their loss, heads in their hands. The fatal transaction is complete. But there’s still the memory of that odd lick. People have wondered just what the Englishman was doing. The accompanying description says he was confirming the slave’s age and checking that he wasn’t sick, although it’s not immediately clear how a lick would achieve this. We can only assume there was some method to the technique.
Centuries later a young economist found the picture and came up with an alternative explanation. The licking, Roland Fryer at Harvard University has suggested, may have been to gauge the saltiness of the African man’s body, because being a little brinier might better equip him to handle the long sea voyage to the New World. It’s an idea inspired by a scientific theory that black Americans, mostly the descendants of slaves, process salt differently from white Americans. The slave ships that transported people from places like Calabar to the New World would have seen immense loss of life along the way as a result of fluid loss caused by dehydration, vomiting and diarrhea, the theory goes. The proportion who naturally retained more salt would have fared better, producing a genetic bottleneck among the slaves who survived to the end.
This process of human evolution on a rapid scale, acting only on black slaves, left them fundamentally different at the other end. Those who reached America were the saltiest.
Fast-forward to the present day. In the twenty-first century, this historical theory has been commandeered to explain why black Americans today suffer persistently high blood pressure, commonly known as hypertension, at higher rates than other groups in the country. Hypertension is made worse by eating too much salt, but if black Americans naturally happen to retain more salt because of the legacy of the slave trade, then the theory goes that this might be the reason they suffer hypertension, rather than because of their diets. Their bodies, some scientists and doctors argue, simply aren’t the same.
Hypertension is a simple thing to measure and it’s a widespread problem all over the world. My mother happens to have it. In and of itself, it doesn’t cause her any problems, but as the doctor reminds her, she needs to watch her sodium intake to bring her blood pressure down because high blood pressure increases her risk of heart disease and stroke. When I visit her for lunch I can’t help but notice how liberal she is with the salt cellar. She doesn’t need to remind me that it’s a source of shame to serve Indian food underseasoned.
“Hypertension is probably the oldest chronic disease we know about,” says Richard Cooper, a seventy-three-year-old public health researcher at Loyola University Medical School in Chicago. Cooper has spent decades investigating blood pressure, stirred by his days as a medical student in eastern Arkansas when he saw patients dying of strokes while still only in middle age. Part of the problem, he has found, is that it’s a strangely nebulous disease. “There are some people who argue that hypertension is not a disease, it’s a condition, that it’s a state, like anxiety,” he says. “We don’t really know initiating cause.” It isn’t rooted in any one organ of the body; you can’t run a scan or a biopsy for it. “It just emerges out of the mists.”
Yet it’s a major killer. The World Health Organization estimates that raised blood pressure accounts for nearly 13 percent of all deaths worldwide.
Despite how common hypertension is, only in around the 1940s or 1950s did American physicians begin to notice more black patients than usual coming to them with hypertension. In the United States, studies suggest that hypertension is almost twice as common in black Americans as in other groups. If you search the UK National Health Service website for the factors associated with high blood pressure, alongside salt you’ll see lack of exercise, too much alcohol, smoking, older age, and one more: being of African or Caribbean descent. Hypertension is thought to be so powerfully correlated to blackness that UK clinical guidelines even recommend different drugs for black people and white people under the age of fifty-five.
Preventable heart disease and stroke resulting from high blood pressure are two to three times as likely to kill a black American as a white American. What is stranger is that this mirrors death rates from other causes, too. The life expectancy of a black person born in the United States today is three and a half years lower than that of a white person. Almost every major cause of death and disability hits blacks harder if they live in the United States; even infant mortality among the black population shows the same elevated rates. So in a country where black people die at disproportionate rates anyway, the mysterious quantity known as hypertension has long been a vehicle for speculation about racial differences.
Cooper saw it for himself growing up in Little Rock, Arkansas, and when he was at medical school in the 1960s. “There were patients who were getting a transfusion and would say, “You’re not going to give me black blood are you?” You were constantly confronted with that,” he recalls. At that time, wards weren’t integrated and black Americans had access only to lower standards of healthcare. Growing up in Arkansas, he says, was like living under apartheid.
The notion of black exceptionalism runs right through the history of American medicine, Cooper adds. “There were very raw and basic racist ideas that were the norm in the medical school.” One common belief was that some diseases, such as tuberculosis and syphilis, manifested differently in blacks. Another was that even if black and white bodies were similar physiologically, black bodies had less value. In the Tuskegee experiment of 1932, the United States Public Health Service teamed up with researchers at the Tuskegee Institute in Alabama (at the time a college for black people) and deliberately denied black patients antibiotics they knew could cure them so as to track the effects of syphilis. The men were observed until they died, their internal organs slowly ravaged. Only after ethical concerns were aired in the press did the study end, forty years later.
Hypertension is one of the conditions to have survived this era with its racial stereotypes intact. American doctors have for decades wondered whether the gaps they see could be due to some intrinsic difference between races. They used to ask whether hypertension might even be a uniquely different illness in black people, connected somehow to skin pigmentation or testosterone levels, or to the heat and humidity of Africa.
But as time passed and population studies were done, it turned out that people living in Africa, especially rural Africans, have the lowest levels of hypertension in the world. “The people are skinny, they don’t eat much salt, they’re very active. There is no way that blood pressures cannot be low. It’s just not possible. They don’t have diabetes and they don’t have hypertension,” says Cooper, who has carried out blood pressure studies on tens of thousands of people across Africa, Europe, and the Americas. People in Nigeria and Ghana in West Africa, from where most black Americans can trace their ancestry, are known to have far lower blood pressures than in other countries.
In fact, Cooper adds, topping the hypertension charts are Finland, Germany, and Russia. “They have terrific hypertension.” White North Americans and Canadians, meanwhile, tend to have lower levels than Europeans, including those in England, Spain, and Italy. Hypertension, then, isn’t a global problem for those with black skin; it’s a local one. We know that black Americans certainly do have higher rates of hypertension on average than white Americans, and the same appears to be the case in Britain with black and white Britons.
For US physicians, the question has been, “Why do black Americans have higher rates of hypertension?” In the 1980s, attempting to neatly square the circle of low hypertension in black Africans with high hypertension in black Americans, a doctor named Clarence Grim came up with what became known as the “slavery hypertension hypothesis”—the theory, later championed by the Harvard economist Roland Fryer, that black Americans were naturally predisposed to retain more salt because of a rapid process of natural selection on the slave ships that brought their ancestors to the New World.
It was an evocative story, giving the tragic brutality of slavery an extra poignancy. Sensitivity to salt, which had helped some through the brutal journey across the Atlantic, landed their unfortunate descendants in the twentieth century with the fatal scourge of hypertension. Western diets had damned them, and there was nothing they could do. The media loved it. Fans of the theory included Oprah Winfrey and the resident health expert on her talk show, Doctor Oz. With Fryer’s apparent historical evidence supporting Grim’s hypothesis, combined with the picture of the slave being licked—it all seemed to be tied up in a bow.
Not everyone was taken in by the tale, though. Biologists raised eyebrows at the suggestion that evolutionary change could happen over such a short time scale. The historian Philip Curtin, an expert on the African slave trade, argued that dehydration and salt depletion were not significant causes of death on slave ships. Neither, he noted, had there ever been a shortage of salt in West Africa to make people there particularly more salt retaining. If anything, Curtin concluded, the historical evidence ran counter to the hypothesis rather than supporting it. Richard Cooper adds that salt sensitivity, if it is seen to be higher in black Americans, is likely to be a product of being primed over a lifetime for all the factors that give them hypertension in the first place, particularly poor diet. This is why other demographic groups who have higher hypertension, including men and the elderly, are also more sensitive to salt.
Even so, there was a widespread expectation that harder proof would one day be found, most likely in our genes. If anything could settle the debate once and for all, it would be the glittering new science of genomics.
In 2009 researchers thought they had finally discovered the evidence they were looking for. A team led by scientists at the National Human Genome Research Institute in Bethesda, Maryland, took DNA samples from around a thousand people and discovered five genetic variants linked to blood pressure in black Americans. The effects were admittedly modest, but they seemed promising. In that moment it felt as though there really might be a whole host of tangible genetic differences between races that would help science get to the root of disproportionate ill health in black Americans.
The search for “black genes” already had a precedent in the research surrounding sickle cell anemia, a serious blood condition more prevalent in people who have ancestry in malaria-afflicted regions such as West Africa. The sickle cell gene is known to provide some resistance to malaria, which can otherwise be fatal. This is a watertight evolutionary explanation for why such a debilitating illness has persisted. But sickle cell disease also exists outside Africa, including in Saudi Arabia and India, which means it can be found in people of all different skin shades. As the UK’s National Institute for Health and Care Excellence states, the sickle cell gene is actually found in all ethnic groups. Within the African continent, it isn’t seen in high rates everywhere; for example, in South Africa malaria is less of a problem and the incidence of the sickle cell is therefore lower.
In the United States these nuances were lost purely for demographic reasons. Many white Americans tend to be of European extraction, where sickle cell anemia is rare, and black Americans tend to have West African roots, where it’s more common, so it came to be seen as a “black disease.” Once viewed this way, it reinforced existing assumptions about essential differences between blacks and whites. Two independent facts began to align in people’s minds. First, that there may be different genes determining health according to race. Second, if black people suffer illness and death at higher rates than white people, could this then be genetic?
Hypertension, however, turned out to be not so straightforward. In 2012 another team of researchers, Clarence Grim among them, tried to replicate the 2009 study, this time with more than twice as many people. They failed. They just couldn’t see the same correlations.
As scientists struggled to find the genetic evidence they thought must be out there, one team led by a researcher at the Harvard School of Public Health decided to look at factors other than race that might correlate with high blood pressure. They discovered that level of education, which often correlates with income level and social class, was a far better predictor of someone’s having hypertension than the person’s percentage of African ancestry. Each year of education was associated with an additional .5 millimeter decrease of mercury in blood pressure readings. A year later, a study in Cuba showed that being black or white there made no difference to average blood pressure or hypertension. Others pointed out that living in an urban environment was strongly associated with blood pressure rises, as was being an immigrant or adopting a westernized lifestyle. Richard Cooper suggests, too, that the effects of chronic exposure to discrimination could well account for some of the black-white differences in blood pressure in the few countries where they’re seen.
Concerns were raised that the nongenetic influences on blood pressure were being oddly neglected by most medical researchers, even though this looked like an obvious medical gold mine of explanations for disparities in health.
“It has been very hard to find genetic factors that affect salt sensitivity. What has been found is not more common in blacks than whites,” concludes Cooper. “There is no evidence of any significant selection across the African diaspora.” Despite heavy resources poured into finding a gene, researchers have still found no association or mechanism that can fully account for higher hypertension in black Americans. The weight of evidence so far points elsewhere. If hypertension is more common among certain groups in certain countries, it probably isn’t because of something in the genes. Most of it is likely to be down to the same reasons that my mother has hypertension—diet, stress, and lifestyle.
“Diet is the underlying cause of hypertension,” Cooper insists. In Finland, for example, with among the highest rates of hypertension in the world, diets have traditionally been low in fruits and vegetables and high in fatty meat and salt. Food in the American South, traditionally associated with black Americans, is similarly rich in salt and fats. Cheap processed foods also have more added salt. Studies have shown that reducing salt intake by just 6 grams a day could prevent 2.5 million deaths annually from stroke and coronary heart disease worldwide. For scientists seeking an easy explanation for differences in hypertension rates across different groups, there’s one available in every kitchen: the food people eat. Part of the reason that diet is understudied is that it isn’t easy to measure with precision. People don’t always know what’s in the food they’re eating, and they rarely report accurately what they do eat. It’s easier to measure simple variables such as race, age, and sex.
But there’s more to the lure of the slavery hypertension hypothesis than data collection, says Cooper. “It’s a very useful window into people’s thinking. Below the surface, there is this very strong prior bias to believe in a mechanism like that. Racialized thinking is such a deep part, just like gendered thinking is such a deep part, of our psychology that we can’t just by conscious effort free ourselves from it completely. It keeps popping up in ways when we’re unprepared or not vigilant.”
People wanted so much for the story to be true, to be able to link the trauma of slavery to the trauma of black American deaths today, that they couldn’t look past it to more mundane explanations. Hypertension, Cooper says, is a case of science being retrofitted to accommodate race. The data, the theories, the facts themselves, are rotated and warped until they fit into a racial framework we can relate to. This is the power of race. It is the power to twist science to its own ends.
“Typically in my field, we don’t just collect data and then show the data that we collected,” I’m told by Jay Kaufman, an epidemiologist and statistician based at McGill University in Canada. Data in its most raw form is rarely useful. Almost always, it has to be packaged, interpreted in some way to make sense.
He gives me a simple example. In 1970 the death rate in the city of Miami, Florida, was 8.92 per thousand people. In Alaska at the northwest tip of the United States, that same year it was only 2.67. This is the unvarnished truth. If you happen to live in Miami, you are more likely to die than if you live in Alaska. But then, if you happen to live in Miami, you’re also more likely to be retired. In 1970 the city, with a total population around 60 percent of the whole of Alaska, had more than 92,000 people over the age of 65, while Alaska had only around 2,000. If you’re between 15 and 24 in either of these places, the death rates are around the same. So statisticians adjust death rates by age to give more useful comparisons. Miami and Alaska then turn out to be about equally lethal as each other.
Adjustments like these are an everyday part of data analysis. Very little of what the public is fed in terms of statistics is unadjusted, and that includes health and medical data. “The logic of these statistical adjustments is based on ideas from trials,” explains Kaufman. In experiments for new treatments, the gold standard is the double-blind randomized clinical trial, in which patients who are roughly the same in every other respect are randomly selected to receive either the treatment or a placebo but neither they nor the researchers know who has got what. Researchers can then be sure that the effects they see are because of the drug and not influenced by expectations.
In real life, though, it’s not always possible or ethical to carry out randomized trials, which is why adjustments are made after the fact, artificially removing the effects of variables such as age and weight. When they do this, says Kaufman, researchers create what are in essence imaginary worlds where these things no longer matter. They are worlds made of manipulated data, sending a clear signal through the noise of reality. “This is sort of my niche within this field, the use of statistics and comparisons.” He has years of experience in this, the darkest art of statistics. Kaufman is a traveler through imaginary worlds.
“Once we start adjusting and describing some imaginary world, then the question is: What is that for? Why are we making an imaginary world? Which imaginary world are we going to make? What are we trying to learn from this imaginary world that we can’t learn from the real world?” When it comes to data about racial difference, the logic of adjustment is that if all the social and environmental aspects of racial difference can be removed and the data still shows a gap between groups, the cause of the gap must be biological.
In one study published in the Journal of Allergy and Clinical Immunology in July 2017 Kaufman spotted an article by a large team of American medical doctors who claimed that the airways of black people become more inflamed than those of white people when they have asthma. In the United States, it’s well known that black Americans carry the heavier burden of asthma. They’re almost three times as likely to die from asthma-related causes than non-Hispanic white Americans. Black children are four times as likely as white to be admitted to hospital for asthma. It’s also well known that asthma is affected by the environment, including smoking, air pollution from busy roads and factories, and living among cockroaches, dust mites, and mold. Yet this study claimed that there was also something intrinsic to their bodies that made black asthma patients suffer more severely. The authors stated that black people might even need their own therapies.
Kaufman decided to pick through their figures.
The original data the researchers had on airway inflammation showed no significant differences between white and black patients. So they decided to control for factors including lung function and degree of control of the disease, as well as body mass index, age, and gender, until finally they came up with adjusted data that showed that there was a small but significant difference between black and white patients, that black people’s airways responded uniquely differently to asthma.
The lead author of the paper herself tells me that her study can’t explain higher asthma rates among black people overall, only the severity of their condition. But as Kaufman explains, their adjustment can’t even necessarily explain this. “In a trial we would never adjust something that itself is affected by the exposure.” It’s like testing a drug with a known side effect of weight gain. You can’t judge the effects of that drug by adjusting for weight, because the drug itself causes patients to gain weight. In the asthma study, “they’re saying let’s imagine a world in which black people don’t have a more severe disease. Then what would their inflammatory response be? What use is that?” In Richard Cooper’s words, the science is being retrofitted to accommodate race.
“People get trained in schools to build models and make adjustments. This is the way we do things. And then they just apply it to race as though it’s the same as a pill you would take. It’s completely bizarre,” warns Kaufman. “Most practitioners of medical research with medical degrees and basic science degrees don’t really have much background in statistics. Many people with perfectly good intentions end up committing a lot of statistical errors because of lack of training and something we call “wish bias,” which is this idea that you want to find something interesting so you keep sifting through the data and fishing around until you find something interesting. That’s a practice that generates many incorrect findings.”
As useful as imaginary worlds can be, Kaufman tells me he cannot understand why medical researchers persist in applying statistical methods to race when it’s obvious that the methods cannot work the way they want them to, that they produce imaginary worlds of little or no use. “It’s epidemic in our literature that these adjustments are kind of nonsensical. In economics, this approach that epidemiologists and biomedical people use doesn’t fly at all.” He often sees scientists picking out a handful of variables to adjust for racial differences, without explaining why those ones were chosen and not others. At other times, he sees evidence of residual confounding, where the variables are measured poorly in the first place, making the final statistics even less reliable. This is especially true when it comes to adjusting for complex factors such as socioeconomic status.
The logic of statistical adjustment also holds true only if adjustment is actually possible. And racial difference is not a simple, measurable quantity like age or weight. The effects of racial discrimination, especially in a society as historically divided as the United States, run incalculably deep. It’s not just about income or educational disparity. Black people in poorer neighborhoods live with worse levels of transportation, waste disposal, and policing, and they are more likely to have environmental hazards located nearby, such as bus garages, sewage treatment plants, and highways. These areas are also targets for cigarette and fast-food marketing.
Discrimination is a problem that extends all the way to the doctor’s office. A hefty 432-page report published by the National Academy of Sciences in 2003 confirmed that evidence of racial and ethnic gaps in healthcare is “remarkably consistent across a range of illnesses and healthcare services,” despite surveys showing that most Americans believe blacks and whites receive the same quality of care. More recently, Roosa Tikkanen, a researcher at the Commonwealth Fund, a private healthcare foundation based in New York, found that black and minority patients in New York and Boston are disproportionately treated at poorer, less well-equipped public hospitals rather than wealthier, high-quality teaching hospitals. She tells me that in some cases a public hospital may serve three times as many black patients as a private hospital located just a few blocks away.
Tikkanen suggests that institutional segregation and structural racism may be at play. Interviewing ambulance and emergency medical staff, she discovered that if someone was picked up from a low-income neighborhood, in the Bronx or a certain part of Brooklyn, and that person happened to be from a minority background, ambulance personnel would by default tend to take them to a public hospital or a handful of so-called “safety net” hospitals that have a history of welcoming poor and nonwhite patients. “Even after you account for the fact that they have worse insurance, minority patients are disproportionately seen in the public system,” she adds.
In The Social Life of DNA, Columbia University sociologist Alondra Nelson explains that in eighteenth-century New York, the mortality rate of black infants was twice that of white. More than half the black population died in childhood, and those who lived had their bodies driven to the breaking point by the physical stresses of slavery. American medical experts claimed at the time that black people were naturally more robust, more resistant to diseases that killed others, including gallstones, tuberculosis, pneumonia, and syphilis. It was a narrative that served slavery, allowing slaveholders to subject people to harsher labor and living conditions on the assumption that this couldn’t harm them. Nelson argues that this legacy of seeing people as intrinsically different, today manifested in lower living standards, has a direct impact on illness and death.
One 2018 study even found a possible relationship between racism in a geographical area and the health of newborns in that area. Researchers saw, astonishingly, a direct correlation between the proportion of Google searches for the N-word in an area and the prevalence of black babies being born prematurely or with a low birth weight. They also noted a similar heightened birth risk among women with Arab surnames in the six months after the 9/11 attacks. What is clear is that researchers are nowhere near understanding all the health impacts of the social factors around race.
In a paper published in the medical journal Lancet in 2017, a raft of public health researchers, including Mary Bassett, the New York City commissioner for health, warned that scientists were too often turning to biology to answer questions that could so clearly be better explained by social inequality. We know, for example, that 38 percent of non-Hispanic black children in the United States live below the poverty line, compared with less than 15 percent of children overall. Black Americans are clearly disadvantaged in employment and the legal system. In one study, in which an identical set of resumes was sent out to employers, white applicants with criminal records were called back more often than black applicants without criminal records.
“There is a rich social science literature conceptualizing structural racism, but this research has not been adequately integrated into medical and scientific literature,” the authors of the study wrote. Structural racism is the elephant in the room. Of almost 48,000 articles they found on race and health, only 2,000 mentioned the word “racism” even once.
The way data is collected and organized is also a problem.
Since 1993 the National Institutes of Health, the largest funder of medical research in the world, has had a general policy of requiring the clinical trials it supports to include women and minorities and also to collect data by race, across at least six categories. The purpose of this was never to look for differences between groups, but to ensure that medical studies include a broader spread of the population. As a demographic exercise, then, it made no difference that all black people, whether born in Africa or the United States, are today lumped by the National Institutes of Health into one general category, despite the prevalence of hypertension in their birth countries being completely different. It also shouldn’t matter that the “white” category includes those from the Middle East and North Africa, as well as Northern Europe. Or that immigrants from Russia, which has generally very high levels of hypertension, are in the same category as white Americans, who tend to have far lower hypertension.
These were never considered to be genetically similar groups, just social and demographic ones. Collecting racial data like this was meant to be an exercise of ticking off boxes.
“All this is a record-keeping function that comports with other federal categories and guidelines, which are social categories,” explains Dorothy Roberts, a law and sociology professor at the University of Pennsylvania. “They are the same categories as are used in the census to keep track of who is recruited into scientific studies. It’s not a requirement for researchers to design their studies in any particular way.” But that’s not the way the data always ends up being used.
While it’s important to know where and why racial disparities exist in health, the American habit of collecting data by race has the unintended consequence of driving researchers to use it, hunting for gaps and trying any means possible to explain those gaps. “The US government provides millions and millions of dollars of grant money targeted to this question. If the government is giving you money, saying we want you to answer this question, and this is all people know how to do, then they’re going to answer that question whichever way they know how. So that provides some incentive,” explains Jay Kaufman. It would be the same if funding agencies suddenly began collecting data by hair color as well as by race and sex. You would be almost certain to see studies suggesting biological differences between people with brown hair and blond hair. Just having the data invites comparisons.
But pooling people into groups is always imperfect, and the larger the group, the more imperfect it becomes. Environmental and cultural factors can play out in ways that outside observers may not anticipate.
For example, in Britain it’s common for health researchers to lump all South Asian communities into one convenient category. The National Health Service website lists being South Asian as associated with a higher risk of cardiovascular disease. But this conclusion ignores the cultural and socioeconomic differences between Indians, Pakistanis, and Bangladeshis, even just within London. Rates of smoking tend to be high among Bangladeshis but very low among Indians, and smoking is a high-risk factor for cardiovascular disease. Vegetarianism is common among many Indians, but rare among Pakistanis, and diet is also a crucial component of cardiovascular disease. Hypertension is another condition also considered to be slightly higher among South Asians living in the UK. Yet despite the fact that India, Pakistan, and Bangladesh were one country until 1947, those of Indian origin tend to have higher blood pressure than those of Pakistani origin. Those of Bangladeshi ancestry have lower levels than Britain’s white population.
With data as fuzzy and meaningless as this, adjusting for race or ethnicity can be a minefield, yet researchers routinely do it anyway. And sometimes with perplexing results, as Kaufman has found. One study he saw modeled hypertension and blood pressure for people in every country in the world in 1990 and in 2015. “So powerful was this model,” he wrote, “that the authors even specified the mean blood pressure in 1990 for countries that did not even exist at that time.”
The list of drugs available to treat hypertension reads like a Willy Wonka Chocolate Factory product catalogue: there are beta-blockers to slow down your heart rate, alpha-blockers to relax blood vessels, diuretics to shift salt out of your body, ACE-inhibitors that work on the hormones, calcium-channel blockers, vasodilators, and more. And within each of these categories are different brands, each competing fiercely for a piece of the lucrative blood pressure pie. They represent a global market worth about $80 billion. So it’s small surprise that every pharmaceutical company wants to establish its unique selling point, to mark out its drug from the competition.
So it was with NitroMed, the marketers of BiDil, a pill that combined two different existing generic drugs, one that relaxed blood vessels and one that helped to combat heart failure, known to develop from hypertension. Around thirty years ago it became clear that mixing treatments in this way might make patients live longer, and BiDil was one of the earliest pills to do it. But there was a problem. The Food and Drug Administration refused to approve it because clinical trials hadn’t been carried out fully, so the company was unable to show just how reliably it worked. Its patent was running out, leaving the company in a jam. How could they get their pill approved as quickly and cheaply as possible?
The solution they came up with was unprecedented: they carried out a clinical trial on black patients only.
Studies already suggested that black patients tended to respond a little less effectively to ACE-inhibitors, for reasons that weren’t yet fully understood. It’s a fact so widely accepted that hypertension pills in the United States often specify different usage guidelines for different races. In the United Kingdom, clinical advice goes so far as to state that ACE-inhibitors should be given to white patients under the age of fifty-five to treat hypertension, but not to black patients. BiDil wasn’t an ACE-inhibitor, which meant it might make a promising new first drug of choice for black patients, I’m told by Jay Cohn, the pill’s developer, who is a cardiologist based at the University of Minnesota.
Cohn’s original tests on BiDil had shown that the small number of black American patients who were included in the trial (just forty-nine people) seemed to respond better to the pill than other groups did. By 2004 the results of a new trial, this time carried out on around a thousand patients, were published in the New England Journal of Medicine; they confirmed that BiDil, when taken in addition to existing medication, reduced mortality rates by 43 percent. Given that the drug had already been shown to be effective in 1987, this outcome couldn’t have come as a complete surprise. But what was certainly different this time was that every single one of the patients tested was black.
Cohn tells me that this was a practical decision. “We did not have adequate support to do a trial in the full population,” he admits. The cost of full-scale clinical trials can easily run into many millions of dollars. “So we determined that maybe the best way to go would be to study the most responsive population, which was a self-designated black population.” This didn’t mean that BiDil didn’t work in white patients, only that they didn’t have the funds needed to do larger trials that included everyone. “We would have needed a larger sample size to study a general population than we could get away with, with a black population.” On the basis of this one group-specific trial, in June the following year the Food and Drug Administration approved it as a medical treatment to be marketed solely to African Americans.
It was the world’s first black pill.
As soon as the decision was made, it divided people. Health campaigners and some high-profile groups such as the Association of Black Cardiologists welcomed it as a positive move, signaling recognition, finally, of the historically neglected medical needs of black Americans. Others, including many doctors, saw it as little more than a cynical marketing ploy to leach more profit out of a drug on which the patent was about to expire. NitroMed, the company that owned the marketing rights to BiDil at the time, gained thirteen years of patent protection, and with this it could sell the combination drug for as much as six times the price it could charge for the individual pills separately. The pharmaceutical goldmine had been plumbed a little deeper.
That said, the going wasn’t easy. NitroMed struggled to sell BiDil, in part because of skepticism among doctors, but also because of its eye-watering price. Since then, marketing rights have been sold to another pharmaceutical firm, Arbor Pharmaceuticals in Atlanta, Georgia. Go to its website and you’ll see black models on nearly every page, smiling and reassuringly healthy-looking. “Although heart failure is on the rise all across America, it hits the African American community hardest,” it states. There’s no doubt that BiDil is still being marketed as a black pill.
Yet, as I’m reminded by Jonathan Kahn at the Mitchell Hamline School of Law in Minnesota, who has tracked the case from the beginning, “Race became relevant in the creation of this drug not for medical reasons, but for legal and commercial reasons.” What’s more, he explains, BiDil set a precedent. Seeing its success with the Food and Drug Administration, pharmaceutical firms began to file patent applications for other treatments that had been shown to work better in certain racial and ethnic groups. Looking at US patent applications filed between 2001 and 2005, the years running up to BiDil’s approval, Kahn found that 65 mentioned race or ethnicity. Between 2006 and 2016, there were 384 that did.
Although no more explicitly race-specific drugs have been approved since BiDil, Kahn tells me that he has noticed an increased use of racial categories in drug labeling, the section that informs you about usage and dosage. “You know, Asians respond differently from Caucasians, that kind of language.” For example, the website for the beta-blocker Bystolic highlights how well it works in black and Hispanic patients. In the first half of 2008, Bystolic was the most heavily advertised drug in the United States. Drug labeling might seem of little relevance, but, he explains, it is a “very powerful force for how medical professionals, and by extension anyone who’s on these drugs, is being taught to think about the relationships of race to biology.” When you see a drug specified for use in a certain group, it implies that this group of people is biologically different from others.
In reality those working in medical research know that race is hard to define; it is a poor proxy for how human variation really works. But when there are few easy ways to distinguish people, it can feel as good as any. The ultimate aim for many in medicine is not racialized medicine but personalized medicine, to be able to sequence an individual’s genome and then tailor therapies to suit that individual. With personalized medicine, in principle nobody will ever need to take a drug that doesn’t work on them, or that gives them a bad reaction. But sequencing everyone’s individual genome is expensive and ethically fraught, and we don’t yet have all the data we need to analyze them anyway. Given these limitations, grouping people by race is seen as an imperfect but practical approximation. Most doctors and medical researchers will admit that it’s a fudge, but they use it anyway. A proxy can save money and time.
But how useful a fudge is race to medicine really? This is a question of statistics and demographics. Given all the medical data we have, how likely is it that an individual placed in a given racial group will benefit from a drug? And how likely is it that someone left out of that group could have benefited from it?
Take sickle cell anemia. It was once suggested that only black infants in the United States be screened for the sickle cell gene, because screening all infants would pose an unnecessary cost. Jay Kaufman and Richard Cooper worked together to dissect the statistics around sickle cell and found that, indeed, sickle cell trait prevalence in self-identified white Americans is only 250 per 100,000 members of the population, whereas in people who self-identify as black it is between 6,500 and 7,000. On this basis, it seems sensible to screen only black infants. On the other hand, there are many more white Americans than black Americans. The odds of a black newborn having the sickle trait may be 6.7 percent, but the odds of any newborn having it are in the same order of magnitude, around 1.5 percent. This is why US states today screen newborns universally, regardless of ethnicity or race.
For BiDil, this kind of number-crunching isn’t possible because the 2004 clinical trials included only self-identified black Americans and no white Americans. Statistical comparison just couldn’t be done. What we have instead is plenty of studies done over the decades on racial differences in response to common hypertension treatments. On the basis of these, the National Institute for Health and Care Excellence in the United Kingdom recommends treating black people with calcium-channel blockers as the drug of first choice, rather than ACE-inhibitors or other alternatives.
Kaufman and Cooper looked through all the published papers that studied responses to blood pressure medication. Their aim was to figure out just how many individuals actually benefit from this racial distinction. They discovered that the perceived racial differences in drug response are in fact relatively small compared with differences within racial groups—exactly to be expected, given everything scientists know about the genetics of human variation. So although there might be statistically significant differences at a population level, this isn’t always useful when it comes to designing a treatment for any one individual patient. For example, they found that for ACE-inhibitors, which are given to white patients under the age of fifty-five in the United Kingdom but not to black patients, data suggest that for a hundred white people given the drug, forty-eight of them would fail to respond as hoped. Meanwhile if a hundred black people were given this drug they are usually denied, forty-one of them would benefit from it.
In this case, they conclude, assigning treatment by race is about as useful as flipping a coin.
I ask Jay Cohn, the eminent cardiologist who invented BiDil, whether his drug, the world’s first black pill, works well in patients who aren’t black.
“Oh, of course it does! I use it all the time in white patients,” he replies. “Everyone responds.”
Cohn has known this all along, of course, and he has always been honest about it. His intentions were never to be racist, he tells me with a laugh. And I believe him. His goal was simply to get the drug approved any way he could. Labeling it as a “black pill” was only ever driven by a commercial imperative. And in the end, this is business. Indeed, pharmaceuticals are big business.
But patients who are prescribed different drugs or who read the labels telling them that race matters can find it difficult to parse just how complicated these facts really are. When it comes to big business, there are also those who may want to conceal facts. In January 2017 the independent news organization ProPublica revealed that Tom Price, a Republican congressman nominated by Donald Trump to become head of the Department of Health and Human Services, had persistently lobbied on behalf of Arbor Pharmaceuticals, which owns the marketing rights to BiDil, to remove a certain study from a government website. This study, carried out in 2009 by heart researchers at the University of Colorado on more than 76,000 people, showed that across all the racial and ethnic groups they studied, the combination of drugs that made up BiDil was not associated with significant reductions in mortality or hospitalization. The very first study on BiDil had looked at only 49 black Americans, and the second at just over a thousand. This study was far larger and therefore likely to be more significant.
The United States has a federal agency, the Agency for Healthcare Research and Quality, to help patients and doctors make informed choices about medical treatments. Yet according to the news report, one of Tom Price’s aides had emailed the agency “at least half a dozen times” to have the University of Colorado study removed. It turned out that Arbor Pharmaceuticals had previously donated to Price’s campaign fund. The month after the ProPublica news report came out, Price was confirmed as secretary of health and human services, although he resigned before the year was out, having been criticized for his use of expensive charter flights.
For now BiDil is still on the market, but the expiration date on its patent is coming up. In the meantime, race has become a firmly and widely accepted variable in medicine, according to the law professor Jonathan Kahn. He uses three clichés to describe what has happened in the last couple of decades. “The road to hell is paved with good intentions,” is the first. “I don’t think the use of race in these contexts was nefariously plotted,” he says. Almost everyone in the medical research community believed they were doing the right thing, curing as many people as possible in the most efficient way.
The second is “the law of unintended consequences, resulting from that. And finally, that it’s creating an accident waiting to happen.” For Kahn, well-intentioned people have reintroduced race to medical science without fully understanding either their reasons or the consequences.
“It’s not that the use of BiDil in black patients is immediately going to lead to the resubstantiation of slavery or scientific racism, but it’s a step down the road of re-biologizing race in a way that feeds deep strains of racism,” he tells me. “What we’re seeing now in the US and again on the rise in Europe is a sort of ethno-nationalism. And any sort of indirect or direct approval or imprimatur of using race as a biological category becomes, no matter how well intended, dangerous in those contexts.”
If the idea of race did no harm to anyone, perhaps there would be nothing wrong in making use of it sometimes, especially when it might at least provide some small kernel of usefulness. In the cash-strapped world of healthcare, why not?
But Evelynn Hammonds, the historian of science at Harvard University, warns me just how dangerous it is to wander down this road. The idea of race is not harmless. It brings with it centuries of political baggage, the blood of millions. And it has never been a neutral variable. “When Jesse Owens won the medals in the 1936 Olympics, some people argued that he was not a full-blooded Negro, that he was actually mixed, and so they measured his whole body. That was 1936. They measured his whole body and made arguments like this: the thighbone is within the normal range of what the normal Negroid thighbone should be,” she says. “That’s always the hotbutton question at the end of the day. These people are different from each other, they’re fundamentally different from each other, in disease, in athletic capacity, and, fundamentally in intelligence. That’s the narrative. It acts as a kind of animating force under the surface.”
Even when we know that there are population-level differences in disease frequency, blindly sticking to racial generalizations can be life-threatening. An American pediatrician, Richard Garcia, once described the case of a friend who repeatedly failed to receive a correct diagnosis for cystic fibrosis because it was thought to be a white disease, and she was black. Only when a passing radiologist happened to spot her chest x-ray, without knowing to whom it belonged, was her condition instantly spotted. She had to wait until she was eight years old, and her skin color had to be invisible, before she could be diagnosed.
In 2003, Duana Fullwiley, an anthropologist then at the Harvard School of Public Health, spent six months watching medical researchers in laboratories in California. Their job was to find genetic differences in how people responded to drugs. It was a fairly young, diverse, international team, not at all stuffy or old-fashioned. And she noticed that all the scientists were routinely using racial categories not only to select their subjects but also to confidently pick out statistical differences between these racial groups.
So as Fullwiley observed them, she asked each scientist she interviewed one simple question: “How would you define race?”
Not one of them could answer her question confidently or clearly. The interviews were punctuated by long, awkward pauses and shy, embarrassed laughs. When pushed, some admitted that the concept of race made little sense, that the hard and fixed census categories actually didn’t mean very much. One said, “You can only judge race to a certain degree of confidence.” Another hesitated before admitting, “I need to think more about it.”
Fullwiley concluded that most of the researchers “were unsure of the meanings of the race categories that they used, yet they continue to assert that there is a biological basis to them, which they will soon corroborate.” Race was the entire premise upon which they were doing their research, but they were unable to tell her what it was. Their work instead seemed to rest upon a hope that if they just persisted, they would eventually come to find meaning in these categories. What they couldn’t yet define would then be defined.
Somehow it would become real.
For all the studies that point to innate racial differences in health, the genetic evidence so far rarely tallies. Hypertension was just one case in point. Even enormous experiments looking at the genomes of thousands of people have turned up little. Although hundreds of gene variants linked to blood pressure have been found, collectively they explain just a percent or so of the variation we see, says Jay Kaufman. “We’ve had a decade of genome-wide association studies now, we’ve spent billions and billions of dollars, and we still are at the position that it looks like 97 percent of the mortality disparity between blacks and whites in the United States has nothing to do with genes.”
And this makes sense, he adds. It would be bizarre to imagine that black Americans are somehow so uniquely biologically disadvantaged that they would naturally die of almost everything at higher rates than everyone else.
Dorothy Roberts agrees that there’s no logic in expecting black Americans to be so medically unusual. “How could it possibly be that a group called black people, which first of all is defined differently around the world—it’s been defined differently even within the United States, but the current definition in the United States is anyone with any discernible African ancestry—how could that hugely varied group, which could include someone with mostly European ancestry, someone with mostly Asian ancestry, someone with mostly Native American ancestry, how could it possibly be that that group could have a particular health outcome for an innate biological reason? That just doesn’t make sense,” she says and laughs. “The most plausible, to me the only possible, explanation could be because of inferior social conditions.”
Kaufman suggests that part of the reason the United States clings to the idea of black exceptionalism when it comes to health may be because, in some way, it lets society off the hook. It lays the blame for inequality at the feet of biology. If poor health today is intrinsic to black bodies and has nothing to do with racism, it’s not anyone’s fault. “It says it’s not about our organization of society that’s somehow unfair or unjust or discriminatory. It’s not that we treat people badly. It’s not that we give people worse life chances. It’s not that we are unfair or unjust as a society. It’s just that these people have some genetic defect, and it’s just the way they are.”
During the era of slavery, recall, doctors proclaimed that blacks were uniquely tough, that they were resistant to pain and common diseases. “So you have at one time a literature saying that black people are especially robust, and at another time you have a literature saying that black people are especially predisposed to illness,” he says. “It’s a contradiction, but each one serves its own purpose.”
Richard Cooper thinks that a psychological sea change is needed. What we think of as fact today will have to be replaced by a better understanding of race, just as what was accepted as fact yesterday has started to be replaced. Race is not a universal constant, a biological rule. “Race is a story we tell ourselves,” he tells me. “Everybody has a general belief in race and then they have stories about it, either something they’ve seen or experienced. And the two reinforce each other. History, psychology, politics, we all have our belief systems and myths and things which a hundred years from now are not going to be true.”
Yet we keep looking back to race because of its familiarity. We can’t help it. For so long, it has been the backdrop to our lives, the running narrative. We automatically translate the information our eyes and ears receive into the language of race, forgetting where this language came from. “I think that scientists, they are trapped by the categories they use. They will either have to jettison it or find different ways of talking about this,” says Hammonds. “They’ll have to come to terms with that it has a social meaning.” This doesn’t mean that racial categories shouldn’t be used in medicine or in science more generally. But it does mean that those who use them should fully understand what they mean, be able to define them, and to know their history.
They should at least know what race is.