Medicine in black and white: how flawed race science influences our healthcare

When I was pregnant some years ago, a friend of mine—a doctor at a busy London hospital—told me that South Asian women tend to give birth earlier than other women. That passing comment was enough for me to begin cheerfully planning around the likelihood that my baby’s arrival was more imminent than I had thought. I prepared the nursery ahead of time. I confidently told my brother-in-law, who was visiting from overseas, that his nephew would be born long before he had to return home. In the event, my son came two weeks late and my brother-in-law missed the birth.

I recalled this when I started researching the use of race in medicine. I remembered how unquestioningly I had accepted my friend’s assertion. This wasn’t just because I trusted her medical expertise, but also because it slotted neatly into a wider story we have been told: that certain ethnicities and racial groups have different propensities to illness, from diabetes to heart disease to schizophrenia. These assumptions have been creeping into diagnostics and drug evaluation for years, in different ways. But they are assumptions which, as my own pregnancy taught me, can often be misleading. Indeed, in some circumstances, fatal.

On the surface, there are of course average physical differences between populations, most obviously in skin colour and facial features. A handful of inherited diseases also show up in higher frequencies in certain communities than in others: the serious blood condition sickle cell, for example, occurs at greater rates in black Americans than white Americans. There are those who also like to point to the possibility of deeper variations, including in “build.” Reviewing Kenya’s phenomenal success in long-distance athletics a few years ago, the Atlantic’s Max Fisher speculated that the local population’s shape, lungs and muscles might give them some “genetic” edge.

But the truth is more complex than it might seem. Scientific reality almost always defies our crude categories of race, which are themselves byproducts of recent human history. Yet claims about “racial difference” reflect and reinforce the widespread belief that race matters in some deeper, visceral way. On far-right websites, I have seen supposed medical differences cited as positive proof of this. Yet most of the apparent differences are illusions, which can lead both doctors and patients down dead ends.

Type caste

For 70 years, most biologists and geneticists have been telling us that race has no basis in biology, that it is a social construct and we are all essentially similar underneath. Yet we continue to hear doctors confidently advise us that race does matter in diagnosis and treatment—that diabetes, say, is simply more common in South Asians, or that this treatment might work better in “black” patients. So which one is it? Is race biologically meaningful or not?

Race, if it sits anywhere, lies at the margins of our genomes, manifesting itself in superficial ways such as skin colour. And then only along gradients, which themselves begin to disappear in our deeply intermixed world. The genes for light skin are, for example, found in the oldest communities in Africa, as well as in Europe and Asia. Other measurable traits, such as height or “build,” have a hazier connection with race, and possibly no genetic connection at all. Some researchers have reasoned that it may well be lifestyle, training and ambition—rather than anything in the genome of entire populations—that have in recent years history pushed certain Kenyan athletes to the top of those running medal tables. Just the same explanations for why white British athletes have, at times in history, been exceptionally successful, too. As for sickle cell, it is found in all ethnic groups. While it may be rare in white Americans, it is not absent—and this is because it’s linked not to skin colour or ethnicity, but to malaria. It’s found in peoples of all complexions with roots in regions of the world in which malaria is common, because those with the sickle cell mutation enjoy some protection against that other disease.

Individual differences within the racial categories we commonly use far outweigh group differences, and especially so if the categories are scientifically arbitrary. Historically, in the United States, anyone with the smallest degree of African ancestry has been considered black, a definition since embraced by racist and anti-racist politicians alike. Whatever the politics, though, if a person can have three white European grandparents and still be black, how much genetic meaning can the label really retain?

The conflation of social and seemingly biological categories in medicine can sometimes take a turn for the bizarre, especially when money is involved. Take the American marketing claim that popular blood pressure drug Bystolic works especially well in “Hispanic patients.” This implies that there might be some distinct Hispanic ethnicity, causing drugs to act differently in this group. But “Hispanic” encompasses people with native American, black African and white European ancestry—the idea of some distinct biological “Hispanic-ness” makes no sense.

The more fundamental problem is that even when groups are defined with more precision in terms of ancestry, they still don’t explain much physical variation of any importance. Genetics does not, of course, say that every single person on the planet is different to or the same as the next. What it does say is that we are genetically more similar to people we are related to, and that this similarity tails off as the closeness of that relationship gets weaker. This means that communities of people who have lived near each other for a long time will inevitably be a little more alike, because their ancestors will have bred with each other. But humans have moved around a lot throughout history, so beyond our immediate families, these genetic commonalities tend to be weak.

"There are no discrete 'types' of people"

The result? However racial groups are defined, physical variation within them will, for nearly anything we can measure, overwhelm any average difference that might be observed between them. As far back as 1972, the Harvard biologist Richard Lewontin established that almost all difference sits within populations, and this has since only been vindicated by further genetic evidence.

There are no discrete “types” of people and no hard boundaries around any population groups. There is no gene that exists in all the members of one racial group and not another. Being of the same race doesn’t mean two people will necessarily have more in common genetically.Such is the fuzziness that, statistically speaking, it is possible for my genome to be more similar to a randomly-chosen white Briton than to a randomly-chosen Indian.

Do no harm?

Medicine is supposed to be the application of good science, including genetics. Yet the truths of that science are not always reflected in how any of us—patients and, more disturbingly, often medics too—tend to think about human difference when it comes to health. What science has taught us about human variation seems to disappear out the window in the doctor’s surgery.

In my case, I didn’t pause to ask myself whether South Asian women do tend to give birth early. The scientific jury is out on that “fact.” If it were true, it would still be rash to assume it has a genetic basis. There is a markedly raised risk of prematurity and associated problems among African Americans, too, but when Tyan Parker Dominguez of the University of Southern California reviewed the literature on this a few years ago, she noted that black immigrant women in the US had “significantly better pregnancy outcomes” than their black American counterparts. More strikingly, “the US-born daughters of black immigrant women” ran into “a significantly higher percentage” of problems than their foreign-born mothers, suggesting that those problems are “likely a function of their experiences in the American social system,” such as the stress of growing up in a racist society,“rather than an expression of genetic vulnerability.”

We all know, or should know, that we are each individuals, with our own unique physiology, and particular medical profiles. So while there may be average differences in early births between one group and another, I might have stopped to ask myself whether or not I would be typical of my group. And yet, such is the collective preoccupation with “stock” that I didn’t do that. So where does this obsession come from?

The largest factors that make people within races seem more similar than they are biologically are language, culture and—in racially-divided societies—also class. Both that and culture certainly impact on diet, lifestyle and health. But it is this trick of the eye, conflating superficial physical difference with deeper cultural or social differences, which makes us believe that race has biological significance. In any society where race has importance, all of us will be susceptible to this illusion, and doctors are not immune.

The modern-day use of racial categories in medicine has been fuelled in part by the habit of collecting racial demographic data when conducting clinical research. Since 2000 the NHS has recorded patients’ ethnicities, ostensibly to measure discrimination. But in practice doctors also use this kind of information in diagnosis and treatment.

Very few scientists would pretend that race is an infallible way of making sense of human variation, but some might defend it as a bit of a fudge—a way of lumping together people who share some possible statistical similarity, an average gene frequency for certain diseases. In an imperfect world race can feel at least a little useful. If we knew, for example, that a certain socially identifiable group of people responds to a drug slightly differently on average than another group, then might that not be useful in assigning treatment?

In theory, maybe. In practice, we need to analyse the statistics with care. Suppose that at a population level, scientists spot some marginal difference in a health outcome between two groups—let’s say that British Chinese people had been shown to have on average somewhat higher rates of asthma—that doesn’t mean you can then diagnose or treat on the assumption that every British Chinese patient will be predisposed to asthma. Each person is too different from the next.

This is true with categories that have much more underlying medical meaning than race—such as age ranges, for example. Overlaps and fuzzy boundaries inevitably come into play. There is always debate, for instance, about which ages should be covered by a cancer screening programme.

Sticking too rigidly to certain categories can blind doctors to thinking beyond them. This is also true when it comes to sex which, unlike race, is a category with some obvious biological basis. But it does not extend to making all women different to all men in every way. While women are more likely to get breast cancer, it does—albeit more rarely—strike men, too. A failure to imagine men with breast cancer is one reason why it goes under-diagnosed in them. The British Heart Foundation co-funded a 2016 study which found that women were 50 per cent more likely than men to be given the wrong diagnosis after a heart attack. While women do sometimes present with different symptoms from men, one of the authors, Chris Gale, a professor at Leeds University, concluded that there was a need to “shift the perception that heart attacks only affect a certain type of person… a middle-aged man who is overweight, has diabetes and smokes.”

Coming back to race, the dangers are all the greater because—with very rare exceptions, such as sickle cell—innate differences tend to be both small and statistical. And when it comes to those rare -exceptions, relying on race as a diagnostic short-cut can have devastating consequences. American paediatrician Richard Garcia describes the case of a friend who as a child repeatedly failed to receive a correct diagnosis for the serious genetic condition cystic fibrosis because it is known to be more common among white Americans, 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—shockingly late for a cystic fibrosis diagnosis—and her colour had to be invisible, before she was correctly diagnosed.

Retrofitting race

Where race does matter for health is in the social consequences of discrimination, and in cultural habits. Diabetes has become epidemic in South Asia, for instance, not just because there may be some underlying susceptibility, but also because economic growth has thrust millions of Indian families into both greater prosperity but also more unhealthy diets and sedentary lifestyles.

Another important example is raised blood pressure, something which the World Health Organisation estimates contributes to nearly 13 per cent of deaths globally. Hypertension, persistently high blood pressure, is strongly associated with being black—or at least, it is in the US and the UK: American studies have suggested that it is twice as common in black Americans. A huge difference, no doubt. But is it all down to some innate racial difference in physiology? No. Being “black” is certainly not a reliable indicator of the risks globally if the extremely low rates of hypertension in Africa are anything to go by.

Something else is going on. Living in an urban environment is associated with higher blood pressure, as is having fewer years in education. But the most significant risk factor is salt. “Diet is the underlying cause of hypertension,” says Richard Cooper, of Loyola University Medical School in Chicago. It is why he has seen some of the highest adjusted global rates of hypertension in Finland. Most of the population is ethnically Finnish and “white,” but their diets have traditionally been lower in fruit and vegetables and higher in fatty meat and salt than those elsewhere in Europe.

Black Americans tend to be poorer than white Americans, which makes it likely that, on average, they will have poorer diets featuring cheaper processed foods with more added salt. So here perhaps we begin to solve the riddle of black Americans and higher blood pressure. Diet, in both Finns and black Americans, is the key—not ancestry or skin colour.

The danger for researchers—and regulators—is that when they approach real gaps between social groups with some notion of race at the front of mind, this distracts from stronger contributing factors. In 2005 the US Food and Drug Administration went so far as to approve Bidil, a drug used to treat heart failure that can develop from hypertension, as the world’s first drug to be marketed solely to black patients. In the UK, guidelines issued by the National Institute for Health and Care Excellence recommend that “black people of African or Caribbean family origin” under the age of 55 be given a different drug for hypertension to patients belonging to other groups. From what we know to date about the underlying genetics of human variation, these groups are highly unlikely to be different enough to warrant this kind of racialised healthcare. Yet here it is being used on both sides of the Atlantic.

The use of race might still be defended as a practical proxy—a way of getting at important variation between groups, even if it isn’t rooted in any genetic difference. To understand whether it may be warranted in this case, Richard Cooper, along with Canadian epidemiologist Jay Kaufman, at McGill University, crunched the numbers. They discovered that assigning a drug by race for hypertension actually makes very little difference. Combing through published studies on common hypertension drugs, they did find some small statistical gaps, but none of them profound.

In the UK, ACE-inhibitors are recommended as a drug of first choice to white patients under the age of 55 but not recommended for black patients under the age of 55. Kaufman and Cooper’s data showed that for 100 white people given this drug, only 52 would respond as hoped. Meanwhile, if 100 black people were given this drug they are usually denied, 41 of them would benefit from it. The gap might be real, but it is modest. According to Cooper, it would be more fruitful for researchers to look at other factors, such as class, education and diet. Giving race more weight creates the illusion that hypertension is down to some innate difference.

So why do we persist with race? Maybe because all of us—medics included—have become so deeply attached to the racial -categories we see on census forms that we cannot believe they don’t have more biological meaning. We latch on to any scrap of evidence that might give them significance. Cooper, who has looked at hypertension all over the world, describes the science around it as being “retrofitted to accommodate race.”

Walls and bridges

Medicine, I have come to realise, is a bit like engineering, the subject I studied at university. Although doctors draw heavily from the sciences to guide them in their diagnoses, and although pharmaceuticals are designed using cutting-edge molecular methods, when it comes to any one patient, there is much art and guesswork involved. It’s about getting that person fixed, by hook or by crook, whether that means prescribing a painkiller or a sugar pill. There are thousands of ways to build a bridge and there are almost as many to treat a patient for a -common illness.

Lumping people into categories, especially those that might appear to be “natural,” is a tempting way to narrow down the options and reach quicker conclusions. Not every patient will fit the group profile, but it’s assumed that the proxy is good enough for most. Race becomes a shorthand in a cash-strapped, time-poor real world in which medics must diagnose and treat.

What is forgotten is that building medical walls between races inadvertently biologises race. It is dangerous if not handled with care because it leads patients—let’s face it, all of us at one time or another—to believe that there is more to race than there really is. I have seen racists invoke supposed differences in health to argue that there could well be cognitive differences between races as well. Their logic is that if one set of physical differences exists, why not another?

The lingering presumption that racial gaps in health must be innate also conveniently allows society to wash its hands of the real causes of the disparity. The US is a perfect case in point. Life expectancy at birth of a black American is today three and a half years lower than for white Americans. Black Americans are hit by almost every cause of death at higher rates than white Americans. When medical researchers look for genetic explanations for these gaps, they neglect how unlikely it would be for an entire racial group to be so deeply biologically disadvantaged. They neglect that, in countries like the UK, gaps like these are seen between deprived and wealthier regions.

According to Dorothy Roberts of the University of Pennsylvania, there’s zero logic in expecting black Americans to be so medically unusual: to be more susceptible to everything. “How could it possibly be that a group called black people,” an arbitrary group which includes individuals with significant Asian and Native American ancestry, could “for an innate biological reason… have a particular health outcome? That just doesn’t make sense,” she says. “The most plausible, to me the only possible explanation could be because of inferior social conditions.” As the social conditions of black Americans have improved, unsurprisingly, so too, and markedly, has health inequality on race lines—although there remains a long way to go.

But race divides in health, even if overwhelmingly not rooted in genes, remain a social reality. So what alternative is there for busy and under-resourced doctors to using race as a medical shortcut?

It may lie in speeding the move towards personalised medicine, where each person’s unique health profile, determined by genetic factors as well as non-genetic ones, will guide doctors to the best possible treatment. This would require collecting information on every patient’s lifestyle and diet, perhaps also on socioeconomic and environmental factors such as air quality, alongside biology. Under this system—if it can be created ethically and effectively—there would be fewer drugs that don’t work, fewer adverse reactions, and hopefully no need for inaccurate proxies. Until then, clinical researchers could improve things by ceasing to search so compulsively for innate racial differences, and instead collecting more data on the other social, cultural and environmental factors that we know globally have a strong bearing on health.

But race has a power to make us believe things that lie beyond the facts. Take schizophrenia. In recent years it has been described by some as a “black disease,” given that people of black Caribbean ancestry living in the United Kingdom receive proportionately more diagnoses than other groups. Although it is understood to have some heritable component, environmental risk factors, such as living in an urban environment and being an immigrant, have also been shown to be at least as important as genetics. One study published in Schizophrenia Bulletin in 2012 found that patients with psychosis were almost three times as likely to have been exposed to adversity as children.

Framing schizophrenia as a “black disease” contrasts, then, with an observation made by Nazi scientist Otmar von Verschuer a year before the outbreak of the Second World War. Looking at patient data, he wrote, “Schizophrenia is strikingly more frequent among Jews. According to statistics from Polish insane asylums, among insane Jews schizophrenia is twice as common as among insane Poles.” He concluded that it must be viewed in Jews as “a racial characteristic.” At that time and in that place, then, schizophrenia wasn’t a “black disease,” but a Jewish one.