Notice & Comment

Two Decades Ago, The FDA and NIH Mandated the Use of Race to Categorize Subjects and Report Results in Medical and Scientific Research They Oversee. It was a Huge Mistake, by David E. Bernstein

As of this writing, the federal government is considering using race and ethnicity to allocate access to a new Coronavirus vaccine to combat Covid-19 when one becomes available. More specifically, the government is considering giving preference to African Americans and Latinos because they have been disproportionately affected by the pandemic. 

Dean Dayna Bowen Matthew of the George Washington University Law School, who is serving as a consultant on the prioritization issue, explained the emerging rationale to the New York Times,  “It’s racial inequality — inequality in housing, inequality in employment, inequality in access to health care — that produced the underlying diseases. That’s wrong. And it’s that inequality that requires us to prioritize by race and ethnicity.”

There are obvious dangers to allocating medical resources by race (and let me be clear that I am use race throughout this post for lack of a better term, but I reject the notion that scientifically-speaking there is any such thing), especially in a politically sensitive an area such as vaccines, where the public is already all-too-prone to accept various conspiracy theories and quackery that leads them to oppose vaccination. Instead of expanding the use of race in this way, science and medicine should be moving away from considering race and ethnicity at all.  

Unfortunately, the FDA and NIH have mandated the use of race and ethnicity since the late 1990s. As a result of this mandate, the use of race has become so common in the scientific and medical communities that most people in the field fail to consider whether there is any justification for doing so. As one scientist reports, “we don’t tend … to think a lot about that [race] variable, what it means, how it’s defined, how it’s being used. We just sort of use it blindly.”  

This is very unfortunate, because, in addition to other problems discussed below, the FDA and NIH mandated that the “race variable” be based on the arbitrary (but now standard in American life) racial and ethnic classifications established by the Office of Management and Budget in 1977 for civil rights enforcement purposes. At the time, the OMB warned that the “classifications should not be interpreted as being scientific or anthropological in nature.” This did not stop the FDA and NIH from institutionalizing them into medical and scientific research.

My in-progress book, The Modern American Law of Race, will have a detailed chapter on how the FDA and NIH regulations came to be and why “race” should not be used in scientific research and medical practice. I will summarize the latter case here.

Any discussion of race in science and medicine must start with the recognition that variations in DNA that may have scientific or medical implications are not specific to race, as such, but to geographical distance between different populations. Additionally, there is no known example of polymorphism that is found exclusively in any particular “racial” group.  

Moreover, even if researchers discover that a drug worked effectively in clinical studies for Group A and not the public at large, this does not provide a sound rationale for individual treatment decisions. As one scientist points out, physicians and researchers relying on a finding of a racial difference in drug treatment will be prone to make two errors: to be overconfident in how it will help members of the in group (because it’s been designated, e.g., a “drug for African-Americans”), and not to consider whether in some cases members of other groups may benefit from the drug.  

A related problem is that a study that finds a difference in medical outcomes between “racial” groups may be picking up differences in socioeconomic status or cultural norms, including differences in diet. Researchers and physicians may mistake the results of these sociological differences as being the result of genetic differences, and therefore draw incorrect conclusions from the results.

Even if at one time race may have been useful as a crude proxy for genetic heterogeneity,  as DNA testing has become more available and much less expensive, race is a poor substitute for looking at actual discernible genetic differences between people.  “Pooling people in race silos,” an editorial in Nature Biotechnology declared, “is akin to zoologists grouping racoons, tigers, and okapis on the basis that they are all stripey.” 

Even if one assumed arguendo that in general race can be a valid scientific construct that has medical salience, the OMB categories are too indeterminate to be useful.  First, there is the problem of identifying study subjects’ race or ethnicity in a consistent manner. Researchers primarily rely on self-identification. Self-identification, however, is notoriously unreliable and variable. 

Moreover, OMB racial categories used by researchers—such as White, Black/African American, and Asian—mask vast differences within each category. The White category, including everyone from Scots to Turks to Algerians, is of little value in gauging ethnicity or race in a way that would be scientifically useful.   

People of Black African descent might seem to constitute an obvious “racial” group that might benefit from the use racial data in medical studies. However, as Dorothy Roberts points out in her book Fatal Invention, Black Africans are highly genetically diverse, more so, for example, than “white” Europeans, and some Black African groups are more genetically similar to Saudis and Ashkenazic Jews than to other Africans.

With regard to African Americans in particular, the legacy of slavery means that they rarely if ever can pinpoint their exact regional or ethnic origin within Africa. As a result, if clinical studies of a pharmaceutical product show a mildly elevated risk or heightened benefit for African Americans as a group, we don’t know whether this is due to a clustering of genes within a small African ethnic group whose members have a strong reaction to the drug, or  because of  a much broader mild extra risk that is generalizable to African Americans

The OMB category of Asian, meanwhile, is absurdly non-specific and unscientific. It includes people with origins everywhere from the Philippines to the Indian subcontinent. There are vast differences among the various ethnic groups that comprise the two billion or so people who live within the Indian subcontinent, much less between South Asians and East Asians. 

Hispanic/Latino is an even more problematic category. Latinos’s origins can be any combination of African, Asian, European, and Indigenous. Nor are they culturally homogenous. There is no reason to believe that data about Dominican residents of New York City is applicable to indigenous Mexican farm workers in California.  

Multi-racial individuals present additional problems. “How white is white?,” scientists asked in the Journal of the American Medical Association. “Is a person who has only one grandparent of another race categorized the same as one who has one great-grandparent or two great-grandparents? What happens if two grandparents are white, one is black, and another is Asian?” 

Complicating matters further, some of the United States’ “racial” categories are in fact internally multiracial. Genome-wide studies of African Americans, for example, have estimated the European admixture in African American subjects at 7 to 24%. The vast majority of African Americans in the United States have some European or Native American ancestry.

An additional concern is that racial and ethnic categories are not consistent worldwide, pointing both to the unscientific nature of such categories and the problematic nature of requiring American researchers to insist on the use of American categories in worldwide research. Among the eight countries analyzed in one study, the number of categories used among researchers that correlate with American categories ranged from one to seventeen for “Asian,” from three to twenty-four for “white,” and from zero to twenty-eight for black/African-American. The authors of a recent studyconcluded that using NIH-required categories of race in pharmacogenetics research, that is, research on how people respond differently to drug therapy based upon their genetic makeup, “may be of limited utility and potentially even damaging in the clinical setting.”

Perhaps the most persuasive rationale for mandating that subjects of medical studies be categorized by race is that it allows for the study of health disparities among populations. Medical research presumes that these disparities are primarily the product of cultural and sociological forces rather than genetic variation. Using the OMB categories, however, can obscure more than they illuminate, as there are vast differences in health measures among subgroups in the various categories. Heterogeneity among minority populations can cause researchers to reach misleading conclusions.

For example, researchers found that 55% of “Asian” children in California had sufficient immunizations when they started kindergarten. But once the data was broken down further, only 21% of Southeast Asian children did. Making health policy for Southeast Asian children based on the “Asian” figure would be a mistake.

The same study reported that 6.2 percent of Native American infants in New Mexico had low birth weight. However, specific tribal rates ranged from a high of 10.4 for the Mescalero Apache to a low of 1.8 percent for the Santa Clara Pueblo. Merely knowing the “Native American” rate in New Mexico would not properly guide a public policy response.

The extent that poor health outcomes are based on minority status, as such, has also been questioned. For example, Hispanics, Native Americans, and Asians in California all have significantly lower mortality rates and higher life expectancy than do non-Hispanic whites, despite having higher poverty rates and lower levels of health insurance coverage. A researcher operating from the assumption implicit in much of the literature that being “Hispanic” is a predictor of relatively poor health would be starting his research from a skewed perspective.

Another rationale for including race in scientific studies is that it’s needed to ensure that minority populations are properly represented in those studies. Regardless of its merits, this goal also has its downsides. As Otis Brawley, former NIH Official and Director of the National Cancer Institute’s Office of Special Populations Research argued, requiring minorities to be included as research subjects might lead researchers to put undue pressure on potential minority subjects to get them into trials to ensure that they met federal funding requirements. Other critics have noted that considering race in scientific research encourages the belief that race is in fact a biological, scientific truth.

Moreover, using race in medical research inevitably encourages doctors to take their patients’ race into account in clinical practice. While some doctors may do so in only the most scientifically justified and narrow ways, others may unconsciously resort to stereotype, resulting in medical mismanagement. Particularly troubling, various treatment algorithms used by clinicians in a range of medical contexts “guide decisions in ways that may direct more attention or resources to white patients than to members of racial and ethnic minorities.” The algorithms are based on data regarding the risks and benefits of various treatment options to identifiable populations groups, particularly African Americans. Not only are the data themselves often of dubious validity, but the way they are used in algorithms assumes that the data are picking up differences in genetic response to medical interventions, even though no evidence that this is correct is generally presented.

Meanwhile, any potential gains from using race in clinical practice must be weighed against the risks of alienating patients by labeling them by race even if they object to or are deeply suspicious of such labeling. As Jonathan Kahn comments in his book Race in a Bottle, “race, if used with care and attention can produce valuable results; but used carelessly, casually, or clumsily, it can blow up in your face.”

Ironically, perhaps, even while advances in genetics has made it ever-more-feasible to study genes rather than use race as a crude proxy, the use of race in medical research has received a significant boost from politically-progressive scientists and health researchers. They believe that focusing the attention of the research community on race is essential to remedying societal inequities. Even some strong critics of the required use of race in biomedical research, such as Professor Kahn, insist that it must be used in health-disparity research, even though the latter use helps normalize the former. 

Dean Matthew, as we have seen, goes even further, arguing that a pending vaccine should be distributed by “race.” Distributing an essential medical product based on unscientific, arbitrary categories raises even more troubling questions than does the more general question of using race in medical research and clinical practice, and should be dismissed out of hand if for no other reason than the government has no scientific or other reasonable basis for determining who qualifies as African American or Hispanic/Latino. I can’t imagine anyone seriously wants the government deciding for the purpose of early vaccination whether a fair-skinned biracial woman qualifies as African-American, or whether an individual with one mixed-race Mexican, one Italian-Argentine, and two European grandparents counts as Hispanic. The result would be about as scientific and justifiable as the old race laws of the South determining how much African ancestry made someone a “Negro.” Going back to the days of quadroons and octoroons would hardly be progress.

Nevertheless, Matthew’s perspective is consistent with an emerging view among progressives that public health policy must shepherd biomedical resources into political initiatives to redress social injustice. These progressives seem to ignore the inherent and general dangers of making race a salient feature of politically-supervised medical research, much less distribution of medical goods like vaccines. 

Given both experience and what we know about how politics actually operates, their confidence that the government can competently and innocently use medical research framed by arbitrary racial categories to serve social justice goals while not encouraging racialist and racist thinking among researchers seems wildly optimistic. Their support for the idea that we should allow the government to use research based on arbitrary, scientifically ridiculous OMB racial categories to allocate medical resources to people based on those categories seems both fantastical and an unjustified triumph of unscientific racialist thinking. Unfortunately, this is what NIH’s and FDA’s imposition of the OMB categories into scientific research has wrought. 

David Bernstein is a University Professor at the Antonin Scalia Law School at George Mason University. Follow him on Twitter here.

Print Friendly, PDF & Email