What exactly is a “racial health disparity”? This article explores five lenses that have been used to answer that question. It contends that racial health disparities have been presented—by researchers both within academia and outside of it—as problems of five varieties: biology, behavior, place, stress, and policy. It also argues that a sixth tradition exploring class—and its connection to race, racism, and health—has been underdeveloped. The author examines each of these conceptions of racial disparities in turn. Baked into each interpretive prism is a set of assumptions about the mechanisms that produce disparities—a story, in other words, about where racial health disparities come from. Discursive boundaries set the parameters for policy debate, determining what is and is not included in proposed solutions. How one sees racial health disparities, then, influences the strategies a society advocates—or ignores—for their elimination. The author ends by briefly discussing problems in the larger research ecosystem that dictate how racial health disparities are studied.
More than 35 years ago, the US Department of Health and Human Services (HHS) published an eight-volume report on “Black and Minority Health” (HHS 1985). Dubbed the “Heckler Report” (after Secretary of Health and Human Services Margaret Heckler), it was a comprehensive survey of racial disparities, offering data on overall mortality, health care access, the racial composition of health personnel, maternal and child health, violence, and chronic disease. The report was an analogue to An American Dilemma, the Swedish economist Gunnar Myrdal's exhaustive 1944 study on racism in the United States (Myrdal 1944). The chair of the task force that assembled the Heckler Report declared that it would be “a generating force for an accelerated national assault on . . . persistent health disparities.”
Whatever its shortcomings, the Heckler Report signaled federal recognition of racial health disparities as a core public health priority. It led to the creation of the Office of Minority Health in HHS, and it sparked research that culminated in the National Institute on Minority Health and Health Disparities. A 2019 special supplement published by that agency showcased new ways of measuring racial health disparities, evaluating interventions to ameliorate them, and establishing causation and understanding their impacts over a person's life course (Jones et al. 2019), all refined over the decades. As David R. Williams, Jourdyn Lawrence, and Brigette Davis (2019) put it in an Annual Review of Public Health field survey published that same year, this “body of research illustrates the myriad ways in which the larger social environment can get under the skin to drive health and inequities in health.” Nearly four decades after Heckler, it is worth taking stock of where the field has come and where it might go.
Rather than cover every crevice of the field, I instead focus on a simple question: “What exactly is a racial health disparity?” I take a cue from writers who have probed the nature of “poverty” (Gans 1972; Katz 2015; O'Connor 2001). Like a racial health disparity, poverty on its surface is simple: a blunt metric of deprivation. Yet this simplicity masks contestation about its causes, ways to eliminate it, and above all, its fundamental essence. Is poverty fundamentally a problem of an individual's cumulative life choices? Is it a by-product of the geographic location where one is born or lives? Or is it rooted in the larger economic system in which one must survive? In short, what kind of a problem, as the late social critic Michael B. Katz asked, is poverty?
I contend that racial health disparities have been interpreted and explained—by researchers both within academia and outside of it—as problems of five kinds: biology, behavior, place, stress, and policy. The politics of some of these five analytic traditions have not always been straightforward and, indeed, were sometimes embraced by partisans of divergent ideological persuasions. They were also shaped by the vicissitudes of history. For most of the 20th century, explanations for racial health disparities were deeply reactionary and outright racist, but after the civil rights revolution of the 1960s, alternatives to reigning accounts had more space to flourish. These new explanatory narratives, however, did not fully situate racial health disparities in the larger political-economic order. Class, I conclude, is a sixth way of problematizing racial health disparities that remains underdeveloped in American analyses.
I proceed with three caveats. First, although the most common population comparison made in the examples that follow is between Black and white people, that fact reflects emphases in the discourse, particularly academic literature. But as some other examples show, the five analytic traditions are still marshaled when the health of other non-Black racialized populations has been under discussion. Second, I do not focus on nomenclature and debate over using the term “disparities” as opposed to others such as “inequalities” or “inequities.” I use “disparities” simply because it remains commonplace, even as writers have increasingly identified flaws with this phraseology, particularly with regard to what they see as lack of a “fairness judgment” (Braveman 2014; Hammonds and Reverby 2019; Lynch and Perera 2017). Last, I present these traditions as distinct. My purpose in doing so is not to depict a false mutual exclusivity. Rather, I am conveying the core kernel of each tradition, illustrated with real-world examples in which one overriding constitutive factor shapes explanations of racial health disparities more than others. From this list of five analytic traditions, readers will find other examples where the traditions operate in tandem and in more varied ways.
This is not a typological exercise for its own sake. As with poverty, a freestanding and seemingly objective quantitative indicator of a racial health disparity obscures the multiple optics one can bring to explaining the world—and the racism within it—that produced the figure (Espeland and Stevens 2008). Baked into each interpretive prism is a set of assumptions about the mechanisms that produce disparities—a story, in other words, about where racial health disparities come from. As Julie Lynch and Isabel Perera note (2017), such discursive boundaries set the parameters for policy debate, determining what is and is not included in proposed solutions. How one sees racial health disparities, then, influences the strategies a society advocates—or ignores—for their elimination.
I identified the five traditions by reviewing three types of publications: public health and social science research on racial disparities; primary source writings published in various time periods examining the relationship between race and health; and historical scholarship on race and American medicine, science, and public health. I constructed a corpus through a combination of means: using various keyword permutations and combinations (e.g., “race,” “disparities,” “racial,” “ethnic,” “health,” “inequalities,” “inequities”) and entering them into the PubMed database; following references in highly cited articles; and drawing on my own knowledge of the historical scholarship. For analysis, I mostly read inductively, identifying notions of what caused racial health disparities. If they recurred in the corpus itself and were invoked by contemporaries for at least two decades, I declared the notion an analytic tradition. I identified the underdeveloped sixth tradition—class—by considering adjacent literatures on social class and health, general racial stratification, and theoretical social science explicitly examining the race-class nexus.
Racial Health Disparities as a Problem of Biology
For most of American history, differential health outcomes were interpreted as by-products of a supposed natural hierarchy of biologically discrete races. Even when successfully challenged at various junctures, these explanations have endured.
In the antebellum period, physicians characterized enslaved people as prone to various behaviors, such as eating dirt; immune to certain diseases; or having weaker organs (Heller 1995; Hogarth 2017; Willoughby 2017). By the late nineteenth century, these discourses persisted in a more elaborated body of scientific knowledge, in which various indices of biological difference—measures of brains, lungs, skulls, and others—were cited to explain gradients in health. This practice was bolstered by new statistical techniques for capturing characteristics of population groups, plus growing knowledge of genetics, each of which allowed racist ideology to assume an objective and scientific gloss (Gould  1996; Zuberi 2003). The idea of multiple discrete human origins and evolutionary paths that mapped onto biologically rooted racial hierarchy circulated in mainstream scientific thinking as well (Stocking 1968). Known as polygenism, this idea provided further fuel for explaining racial health disparities as immutable natural facts.
The new racial science served unique American political ends, especially with the consolidation of Jim Crow in 1896. It not only justified state-sanctioned discrimination in a variety of spheres, from public accommodations to labor markets to housing options, but also turned explanations for poor Black health away from resource scarcity and toward supposedly innate traits of the formerly enslaved themselves. Assessing a widely circulated publication about “race traits” that argued respiratory capacity was lower in Black people than in white people, W. E. B. Du Bois captured this orientation away from society toward individual biology, stating that the tract's author “finally concludes that the cause of the failure of so many peoples in the struggle of life is the lack of those race characteristics for which the Aryan is pre-eminent” (Du Bois 1897, Wolff 2006).
A parallel discourse emerged around immigrants. It gained traction as record immigration to the United States in the 1890s raised anxieties about labor competition and societal change brought on by the incursion of newcomers from eastern and southern Europe and East Asia. Often, expressions of public health threat simply pointed to outsiderness. But increasingly, this discourse became explicitly hereditarian, especially when espoused by eugenicists, with some groups deemed more fit and others more “feeble-minded” and weak (Kevles  1995). Political discourse fixated on the “fecundity” of immigrants and how they might retard the national birth rate. Such concerns about the biological constitution of newcomers arose amid nativism that culminated in the Immigration Act of 1924, which imposed stringent country-specific quotas and banned Asian immigration altogether (Ngai 1999).
But the biological race concept and attendant notions of inferiority came under severe challenge after World War I. From anthropology came studies by Franz Boas that partially undermined prevailing biological classifications (Baker 1998; Yudell 2014). The so-called “first” UNESCO Statement on Race declared in 1950 that “race is less a biological fact than a social myth and as a myth it has in recent years taken a heavy toll in human lives and suffering” (UNESCO 1969), and its chief author was Ashley Montagu, who had authored Man's Most Dangerous Myth ( 1974) a decade earlier. A seminal paper by the population geneticist Richard Lewontin (1972) found that 85% of genetic variation occurred within so-called “races,” not across them, undercutting the case for biological races and, by extension, the health disparities that could be explained away by them.
Yet racial biology never died a clean death. The 1950 UNESCO statement was attacked by many in the scientific establishment, and a second revision qualified its core claims (Brattain 2007). Many geneticists, like Theodore Dobzhansky, accepted most of the criticism leveled at the biological race concept but still had trouble relinquishing the notion entirely (Yudell 2014). The notion was similarly alive in public health research, most notoriously in the Tuskegee syphilis experiments, which were predicated on the belief that the disease developed differently in Black people (Reverby 2009). At century's end, racial biology anchored openly racist books such as Charles Murray and Richard Herrnstein's The Bell Curve (1994), which argued for the existence of congenital racial differences in intelligence.
In the late 1990s and early 2000s, novel genome sequencing techniques and completion of the Human Genome Project injected new life into claims that races were biological. Geneticist Neil Risch proclaimed that new data from a study he led showed that commonsensical racial categories indeed had a genetic basis (Risch, Burchard, Ziv, and Tang 2002; Tang et al. 2005). Some scientists offered terms like “admixture” to move beyond traditional and discrete racial categories, but they ultimately still anchored observed group differences in biology (Fullwiley 2008; Rajagopalan and Fujimura 2012).
In medicine, scholars demonstrated that biological race concepts—old and new—crept up in routine literature and were regularly used to explain racial health disparities in breast cancer, respiratory disease, and other conditions. They expressed concern that in clinical practice, race might be used as a blunt—and erroneous—proxy for assessing a patient and prescribing differential treatment (Braun et al. 2007). Race correction, the practice by which a different “normal” reference range is used for diagnosing Black patients, was premised on biological race. It attracted increasing criticism by scholars (Braun et al. 2020) but remained widely used and taught with no consensus as yet on whether the practice should cease. Meanwhile, the billion-dollar market for ancestry detection kits reinforced notions of biological race in the popular imagination (Bolnick et al. 2007).
One notable aspect of racial biology's twenty-first-century resurrection was that it proceeded not under the guise of racial animus, as it had a century prior, but instead racial egalitarianism. The marketing of BiDil—a heart drug whose manufacturer claimed it would have particular effectiveness in Black patients—was defended on the grounds that it would close the gap in cardiac racial health disparities (Kahn 2014). And Esteban Burchard, an expert on racial health disparities in asthma, has argued that some of these disparities are rooted in pharmacological differences between racial groups' response to medications (Aldhouse 2016). For him, eradicating racial health disparities meant searching for genetic roots. If it once looked like racial health disparities as a problem of a biology had become a thing of the past, the present moment has left that conclusion in sharp doubt.
Racial Health Disparities as a Problem of Behavior
An equally durable explanation for racial health disparities was behavior. Here again, Jim Crow and nativism were turning points. By the late nineteenth century, newspapers and public health inspection reports regularly asserted that certain lifestyles were endemic to particular racial groups. These dispatches came with an exoticizing gaze that focused on supposedly elevated levels of substance abuse, unhygienic living habits, and inability to follow health dictates. One inspector in San Francisco's Chinatown, for example, wrote in 1885 that “the mode of life among the Chinese here are [sic] not much above those of rats on the waterfront” (Shah 2001). African Americans in Baltimore were frequently characterized as “incorrigible consumptives” by physicians and nurses overseeing tuberculosis campaigns, with similarly lurid descriptions of their homes and purportedly race-specific habits (Roberts 2009). As with biology, behavioral narratives made individuals the central unit of analysis, not social subordination and its role in poor health outcomes.
Behavioral analyses were not always the provenance of racists. But they could still be shot through with middle-class moralism. Du Bois's 1899,Philadelphia Negro, for example, did not escape fixations, prevalent in the settlement house movement, on nonnormative family structure and what Du Bois called the “laxity in morals” allegedly arising from Black migrants' adjustment to city life in the north (Du Bois 1899; Hartman 2020; Reed 1997). Some civil rights organizations, such as the Urban League (Reed 2008), similarly focused on the perceived inability of constituents to embrace proper social mores, part of what Kevin Gaines (1996) has called an “uplift ideology” common among middle-class African American civic leaders of the time. Many immigrants rallied behind a strategy of outward decorum and respectability as well. In the 1930s, Nayan Shah (2001) notes, Chinese American advocates for public housing argued that it would facilitate stable nuclear families and shield them from the lifestyles of vice-ridden Chinese bachelors who lived unhealthily and elsewhere.
In the 1960s, behavior assumed center stage in larger social policy debate. Daniel Patrick Moynihan's The Negro Family: The Case for National Action (1965) centered on Black households without male heads, which led, in the author's view, to a “tangle of pathology” that mired African Americans in poverty and unemployment. In the 1980s, some analysts and public officials claimed that public assistance incentivized dependency (Katz 2013). Both the Moynihan report and subsequent welfare reform rhetoric led to criticism from those who pointed out what they perceived to be racist and sexist overtones and assumptions.
These criticisms also informed contemporary public health approaches. Increasingly few practitioners of health promotion completely divorced “health behaviors” from larger social networks, living conditions, and life circumstances in which people lived. Yet there was no doubt, as a look through a highly used health promotion textbook indicated, that there existed an uneasy tension between frameworks that focused on how an individual cycled through various cognitive states and frameworks that took a more social-ecological view of why some people were more likely—and had the means—to behave in health-reinforcing ways (Glanz, Rimer, and Viswanath 2015). This was especially the case in the field of immigrant health, which continued to center heavily around “acculturation” and whether it contributed to helpful or harmful health behaviors, such as the adoption of certain diets or substance use (Lara et al. 2005).
The behavioral tradition held up in the public sphere as well. In 2010, Annice Kim and colleagues studied racial health disparities discourse in the broader public sphere (Kim et al. 2010). Analyzing a decade of newspaper coverage, they found that African Americans were most frequently centered and that the leading explanation offered for disparities was behavioral (versus “societal”). Furthermore, behavioral solutions accounted for “nearly half” of those foregrounded by the articles examined.
Nevertheless, awareness of stigma—and its frequently racialized underpinnings—grew enormously among both scholars and practitioners in the first two decades of the 21st century. Behavioral explanations lacked the flagrantly pathologizing quality that they had had a century prior. But narrow conceptions of racial health disparities as a problem of behavior still persisted, even as the larger field had long pursued more holistic explanations. It is to those we now turn.
Racial Health Disparities as a Problem of Space
If biology and behavior were influential but incomplete—and sometimes erroneous—explanations for racial health disparities, what other arenas did the field explore? One was space: the exploration of differential location by race and its health effects. In the United States, Du Bois's Philadelphia Negro (1899) was an exemplar of this analysis. On health and other metrics of wellness, it documented white–Black mortality and morbidity differentials not just in the aggregate but ward by ward. Du Bois advanced beyond European studies of spatial difference and health by also incorporating the role of ascriptive difference (Coleman 1982; Jones-Eversley and Dean 2018). Indeed, Du Bois might well have been the first entry in the spatial tradition of racial health disparities work.
Du Bois wrote in 1899. Unfortunately, for much of the ensuing century, space proved to be a rather Janus-faced optic. In fact, for the first two thirds of the 20th century, it was more often deployed for destructive ends. The racial segregation of immigrants and African Americans led to what one scholar has labeled “territorial stigmatization” (Wacquant 2007). At the turn of the century, public officials frequently depicted immigrant and Black neighborhoods as cauldrons of contagious disease and a threat to cities at large, providing rationales for racial segregation and even destruction of unhealthy areas (Molina 2006; Roberts 2009; Shah 2001).
These local-level actions scaled up in the middle of the century with the advent of the federal urban renewal program. In 1948, the American Public Health Association's Committee on Hygiene and Health developed rubrics to determine whether neighborhoods ought to be razed if they were deemed enough of a public health hazard. In Russ Lopez's (2009) words, these scales “supplied a scientific and seemingly impartial justification for declaring a neighborhood blighted.” They also gave cities across the country “a procedure for targeting neighborhoods for destruction,” one backed by the public health profession, that forcibly relocated selected inhabitants away from more affluent residents to whom their living quarters had posed an alleged risk.
The civil rights era, however, catalyzed new spatial thinking. Whereas previous incarnations examined threats to those outside segregated areas, subsequent analysis focused on the harms done to those within such neighborhoods. The sociologist Pierre De Vise, at the time head of the Chicago Regional Hospital Study, systematically documented links between residential segregation and health services scarcity. He argued that Chicago's reputation as a medical powerhouse was belied by the way it treated Black patients. In an analysis of what he called the city's “apartheid health system,” De Vise noted that four public hospitals were all concentrated in one section of town, necessitating onerous travel for those who depended on them. Of Cook County Hospital's patients during the previous year, “half had no health insurance of any kind, and nine-tenths were black. The average waiting time was two hours.” By contrast, in other city hospitals, only “one-fifth had no insurance, and half were black.”
At an even more granular level, De Vise examined neighborhoods with a sharp decline in white population and an increase in Black population. He found a precipitous drop in physician presence—by between 70% and 90%—as a neighborhood became more Black (De Vise 1971). Entire pockets of Chicago, then, were becoming medical deserts. De Vise's bleak portrait of Chicago mirrored that of cities like Los Angeles. There, the 1964 Watts uprising had spotlighted severe resource shortfalls in south central Los Angeles, leading a commission to call for the construction of a full-service hospital there (Viseltear 1967). If midcentury medicine was a story of boom and technological accomplishment, these reports were a reminder that it had advanced along geographically and racially segregated lines.
This research was limited in one fundamental way: “health” meant health care. Subsequent academic research on residential segregation evolved in capacious ways. One branch, pioneered by figures such as Thomas LaViest (1989), Chiquita Collins (with David Williams, 1999), and David Williams (with Chiquita Collins, 2001), explored the association between residential segregation and health in a broader sense, captured by aggregate indices such as low birth weights and infant and all-cause mortality. Later, studies appeared about specific illnesses, including cancer, diabetes, and cardiovascular disease. In general, the more segregated an area was, the worse its health outcomes were.
It was harder, however, to move from strong ecological association to a more specific articulation of the pathway between racial composition of a neighborhood and the health of individuals within it. But other work on neighborhood effects offered some answers, exploring the health consequences of geographically concentrated advertising for harmful products, limited walkability, nonavailability of certain foods, and lack of social service supports (Roux 2001). These were not solely problems in racially segregated neighborhoods, but often they were most pronounced in such areas, especially when tied to simultaneous economic hardship.
Other scholars turned to place-specific analysis. One of the most ambitious attempts came from those who examined Detroit, most of them based at the University of Michigan. Their studies walked a taut line: coming up with generalizable propositions on the one hand, while anchoring phenomena in local characteristics unique to a place on the other. The work stood out for the way it frequently situated its findings in Detroit's history of deindustrialization, white exodus, declining tax bases, and pervasive racial segregation (Schulz et al. 2002). The Detroit work did not depend exclusively on large national secondary datasets; it also drew on surveys conducted by the venerable Detroit Area Study, and occasionally, more granular data obtained by partnering with community-based groups. Besides neighborhood effects, the Detroit studies also explored the impact of decentralized city government, the city's labor market, and pollution.
Most spatially oriented research did not identify a single overriding mechanism tying segregation and health. Rather, it offered multiple, sometimes more hypothetical, pathways threading economics, race, and geography. One exception was environmental justice research. Here, the argument was forthright and direct: a major cause of racial health disparities, especially in respiratory and cardiovascular illnesses, came from toxicological hazards—bus depots, garbage dumps, highways, factories—disproportionately located in low-income minority neighborhoods.
Activists had long noted these patterns in cities large and small (Hurley 1995). But documentation of these patterns had not quite coalesced into a more systematic formal body of knowledge. That changed in the 1980s, with a series of lawsuits arguing that unequally distributed environmental burdens amounted to civil rights violations. A growing body of spatial research, much of it synthesized in Robert Bullard's Dumping in Dixie (1990), showed strong associations between adverse environmental health outcomes and racial composition of an area. Much of the work came from unusually fruitful academic partnerships with community groups, with the former supplying technical assistance and the latter providing on-the-ground knowledge about where to even look for potential risks. One early collaboration—still in existence—that drew attention from both environmental health advocates and researchers was between Columbia University's Center for Environmental Health in northern Manhattan and the West Harlem Environmental Action Network (Minkler, Vásquez, and Shepard 2006; Sze 2007). It focused on air quality issues connected to diesel fuel and the high concentration of bus depots in the area. On the other side of the country, the Southern California Environmental Justice Collaborative connected local organizations and foundations with three academic institutions—the University of California at Santa Cruz, Occidental College, and Brown. The academic partners worked to analyze publicly available data, which community organizations subsequently used to lobby successfully for better regional emissions standards. These collaborations did not just exist in large cities. In Ottawa County, Oklahoma, Native American advocates worked with researchers from Emory University, the University of Oklahoma, and the University of New Mexico to document children's elevated lead levels stemming from hazardous construction projects. Dubbed Tribal Efforts Against Lead, the collaboration gave lay residents input into the study design and provided them with evidence of serious health risks in areas where Native Americans resided (Minkler et al. 2006).
The elegance of the environmental justice narrative did conceal some ambiguities behind the patterns its champions had documented. Did the phenomena result from planners' racial antipathy toward residents themselves? Did the problems arise from policy makers' perception that political reaction from the most powerless would be muted when it came to locational choices? What role did economics—and the tendency of polluters to cluster hazards on cheap land—play alongside residential segregation? The answers to these questions could be complex. But to residents affected and their advocates, they were often scholastic and beside the point. In many regions, from New York City to Anniston, Alabama, racial minorities bore the greatest brunt of pollution, whatever the cause, and the goal was to establish this outcome's existence, then mitigate its effects. Efforts in the courtroom and at the grassroots level paid off in 1994, when President Bill Clinton signed an executive order mandating that federal agencies “identify and address” actions of theirs that might result in the unequal environmental health burdens. It affirmed a core claim of not just environmental justice but the spatial tradition: that geographic arrangements produced racial health disparities.
Racial Health Disparities as a Problem of Stress
If spatial analyses were highly ecological, another school of thinking focused on societal-to-physiological pathways. Racial health disparities here were conceptualized as a problem of accumulated stress. The most agenda-setting work was conducted by Sherman James in the early 1980s. James created a scale to measure “John Henryism,” named after a retired African American sharecropper in North Carolina whom James had met and who had worked himself to near exhaustion under trying lifelong circumstances (1993). John Henryism measured one's commitment to “hard work” and perseverance, especially in the face of adversity.
The general pattern James uncovered in his work was as follows: high levels of John Henryism in African American men were associated with alarming cardiovascular consequences, namely blood pressure levels and hypertension (James et al. 1984; James et al. 1987; James, Hartnett, and Kalsbeek 1983). Moreover, the effect was far more pronounced in those of lower socioeconomic status and did not typically hold for white men. Heavy exertion in the face of substantial social impediments created a psychosocial reaction. Or, as James put it in a reflection on his work, “the combination of high stress (now significantly correlated with low socioeconomic status) and prolonged, high-effort coping with such stress is probably responsible for this strong elevation in risk” (1994). James's work explored larger settings in which people were embedded, particularly the workplace. The general association, he found, was strongest when one focused on lower-prestige occupations (James et al. 1984). It suggested that foreclosed opportunity and racial discrimination on the job were critical forces at play.
Arline Geronimus, a colleague of James's, elaborated on his questions. In studies on maternal and child health, she found that African American mothers who gave birth at younger ages had children with better health outcomes (measured by birth weight and mortality) than their older African American counterparts (Geronimus 1992, 1996). Geronimus hypothesized a temporal element in African American health, suggesting earlier onset of aging and physiological deterioration in Black versus white mothers. As did James, Geronimus suspected this acceleration was rooted in “psychosocial” stress and dubbed the phenomenon “weathering,” posing it as a possible explanation for why younger Black mothers and their children in Geronimus's studies experienced fewer problems during childbirth.
Geronimus's work on racism and weathering grew more methodologically sophisticated. Her later studies used allostatic load (AL) scores: an aggregate measure of “chronic dysregulation (i.e., over-activity or inactivity) of physiological systems” as a result of repeated stressful events (McEwen and Gianaros 2010). Work on AL had uncovered strong associations between high scores and a host of undesirable health outcomes. Now, Geronimus applied its insights to racial health disparities. Using the AL measure with a number of large panel datasets, she suggested that typical physiological deterioration might occur sooner in African Americans than in whites because of the larger amount of external stressors African Americans confronted throughout their lives. The results could be dramatic—dispiritingly so. Models showed higher predicted Black AL scores (versus predicted white AL scores) in all age groups, especially in the adult years (Geronimus et al. 2006). On average, these models suggested that the onset of aging processes would take place an alarming 10 years earlier for African Americans than for whites. If racial health disparities were a problem of stress, they did not just come as ephemeral shocks at single points of time; rather, their effects accumulated, wearing and tearing one down over the years.
The exact external stressors at work were sometimes more implied than explicit. Everyday acts of racial discrimination and John Henryism were obvious ones. Elsewhere, Geronimus argued that her findings were not just the results of distressing person-to-person encounters. They stemmed from larger societal arrangements, including increasingly punitive social welfare policy and larger cultural stereotypes about African American single mothers, resulting in a public pathologization of this population—a practice that Geronimus outspokenly criticized (Geronimus 2000). Like James, Geronimus also explored the attenuating and amplifying effects of other axes of inequality. One was socioeconomic status (SES), where she found protective effects against Black mortality but less for morbidity (Geronimus 1992). Another found “nonpoor Blacks have a greater probability of high [allostatic load] scores than do poor Whites,” disturbing evidence for how powerful stress effects could be (Geronimus et al. 2006).
A third set of important contributions to the stress tradition came from the Coronary Artery Risk Development in Young Adults (CARDIA) study, a longitudinal dataset started in 1985. One of its most creative users was Nancy Krieger, who deployed CARDIA to explore racial disparities in blood pressure. But whereas James had used John Henryism as a key input, Krieger used self-reported discrimination. Her findings could be complicated. Although moderate amounts of discrimination led to higher levels of blood pressure, one striking discovery was that those who reported no discrimination exhibited the highest levels of blood pressure. For Krieger, this was potential evidence of a stress effect resulting from a situation where “unfair treatment is perceived by members of stigmatized groups as ‘deserved’ and nondiscriminatory.” Self-blame and denial, rooted in larger social stigmatization, thus manifested in stress and physiological consequences (Krieger and Sidney 1996). Another CARDIA study examined perinatal risk and odds of low birth weights, as Geronimus had done. It found that racial discrimination was associated with more than half of preterm deliveries and low birth weights among Black women compared to only 5% of preterm deliveries and 0% of low birth weights among white women (Mustillo et al. 2004). Here again, “racial discrimination as a psychosocial stressor” was the key factor at play.
Whatever input was used—John Henryism, AL scores, self-reported racial discrimination—the stress tradition found generally consistent results when it came to disparity, notwithstanding the occasional deviation from expected patterns. As generative as the findings themselves were, though, they also represented something else: a repudiation of the biological tradition. Stress research explored biological consequences—there was a surface similarity—but it started from the opposite premise of the field's forebears, who saw racial health disparities stemming from race as biological fact. The stress tradition, by contrast, held that racial health disparities were socially produced, the result of bodily coping with racism in its many manifestations. Geronimus and Krieger's attention to gender—and James's attention to the workplace setting—was strikingly prescient, too. Not only did it situate individual physiological deterioration in larger social contexts, its multifactorial nature also presaged what is now commonly dubbed “intersectional” analysis, long before the term became scholarly vogue.
Racial Health Disparities as a Problem of Policy
State-sponsored action—that is, public policy—was a final lens for analyzing racial health disparities. Policy has lurked under all the previous sections above. Racial biology and eugenics gave way to immigration restriction and residential segregation, as did widely held beliefs about the unhealthy group-specific behaviors of immigrants and African Americans. Contemporary spatial organization resulted from past land-use policies, leading later generations to debate what might reverse, at least partially, their deleterious effects on health. The stress research, though more implicitly than explicitly, pointed to lax enforcement of antidiscrimination laws and racially inflected welfare policy as sources of stressors.
These other traditions, however, did not explicitly center policy. For those who did, the most obvious site was health care access. Like most public and private accommodations in the United States, health care in the 20th century was racially segregated. This was entrenched by the 1946 Hill-Burton Act, which appropriated $75 million for building hospitals, 93% of which went to former Confederate states (McBride 1991; Thomas 2011). While it brought Southern health care services into closer line with the rest of the nation, it did so under Jim Crow logic. Many facilities were either all white or all Black, and those that were not remained racially segregated by ward.
It was not just hospitals, either. Most Black medical students attended Howard University College of Medicine and Meharry Medical College and were rarely admitted to other institutions, blocking wider African American entry into the profession. And after graduation, African American physicians were typically shut out of the state and local medical societies that were critical to gaining hospital privileges, leaving them professionally marginalized. Despite these constraints, Black hospitals and medical professionals—especially nurses and midwives—made significant strides in public health education and programming (Smith 1995).
In the 1950s, growing initiatives across the country documented local medical exclusion and linked it to racial health disparities. They coalesced into a broad civil rights coalition focused on desegregation of health care. In the words of David McBride, “an assertive black medical leadership believed that only confrontational civil-rights politics would in the end win black America's fight against the social diseases of poverty and racism, as well as the bodily diseases caused by infections and physical and mental debility” (1991). For them, racial health disparities were a problem of policy, stemming from legally sanctioned segregation.
Desegregation policy took many forms. The first was legal victories. In Simkins v. Cone (1963), the US Supreme Court declared segregated hospitals unconstitutional. But big-ticket legal precedents were only as good as enforcement. It would take two pieces of federal legislation, passed within a year of each other, to really affect circumstances on the ground. One was perhaps the most consequential law passed in the United States after World War II: the Civil Rights Act of 1964, which banned discrimination in both private and public spheres. The second was the passage of legislation creating Medicare and Medicaid. These new funds promised millions of dollars to health care facilities, which could now bill the government for services rendered to elderly and low-income patients. But Medicare and Medicaid also provided a big stick to break Jim Crow medicine. The federal government, civil rights activists reasoned, could threaten to withhold funds from facilities unless they complied with both high court rulings and new laws. Soon, the Department of Health, Education, and Welfare under President Lyndon Johnson implemented a sweeping inspections program to ensure facilities were in compliance. They did so under activist pressure from organizations like the National Medical Association, which represented the nation's Black physicians and patients, and the Medical Committee for Human Rights, an ad hoc organization created to provide medical assistance to civil rights workers in the South. The fight for desegregation of facilities had spillover consequences, too. It ended medical society segregation along with overtly exclusionary medical school admissions policies (Burrows and Berney 2019).
The consequences of desegregation, subsequent research has demonstrated, could be enormous. One analysis, by economist Douglas Almond and colleagues (2006), showed that after the onset of desegregation, mortality rates declined for Black infants much more sharply than for white infants, while racial health disparities narrowed. These trends were most pronounced in the rural South, where Jim Crow consolidation often blockaded access to health care services. There, the disparity was halved, with a decline of 11.4 deaths per 1,000 births, an astonishing fall that saved 4,000 lives in the decade after the new civil rights regime began. Explaining their findings, the authors pointed out that rapid treatments for pneumonia and gastroenteritis, two leading causes of infant death, were now no longer denied. Medicaid may also have contributed by financing African American access to health care institutions. A similar analysis performed by Nancy Krieger and colleagues (2013) found a striking convergence in Black infant mortality rates between “Jim Crow polities” and the rest of the country, one that began in the early 1970s and did not reverse until the end of the 20th century, though it never returned to pre-1970s levels. For the authors, this was evidence that “abolition of Jim Crow laws has contributed to shaping US infant death rates,” in this case for the better.
Access to medical care nonetheless had a two-sided quality. A literature on physician bias, for example, uncovered discriminatory beliefs among white physicians toward nonwhite patients. One disturbing set of studies found that physicians sometimes assumed Black patients could naturally bear higher amounts of pain (Hoffman et al. 2016). Conceiving of racial health disparities as a problem of biology continued in formal medical knowledge (Braun et al. 2007). At worst, this could mean differential assessments of otherwise similar patients, skewed by beliefs about the role of a patient's race.
The demographic makeup of the medical profession was also slow to change, with historically underrepresented minorities still making up low percentages of entering classes. Even the entrance of women—a major positive reversal—masked underlying racial disparities: in 2015 only 7% of graduating women physicians were Black, and only 5% were Hispanic (Chowkwanyun and Howell 2019). The larger consequences of these phenomena for racial health disparities were less clear. But it underscored that removal of formal obstacles to medical institutions did not eliminate unequal dynamics within them. Likewise, beyond personnel, access could sometimes be theoretical, as shown by persistently higher rates of noninsurance in Black and Latino populations compared to whites, even as they were dramatically reduced by the Affordable Care Act, with the Black–white and Hispanic–white disparities falling by 4% and 8%, respectively (Buchmueller and Levy 2020).
For all the success of civil rights policy in the medical sphere, it never eliminated racial health disparities for a number of outcomes. To take just one example, from the field of maternal child health, the striking disparity in Black–white maternal deaths persisted and gained striking attention in the 2020 campaign cycle. It raised a question: was health care policy alone the chief domain to concentrate on when it came to racial health disparities? This concern was anchored in a larger classic question in public health about how much medical care actually contributed to population health improvement and group-to-group health disparities. More conservative estimates put the figure at around only 10%; more generous ones put it at around 50% (Cutler, Rosen, and Vijan 2006; Williams and Collins 1995). Either way, health care itself was clearly not the full story. Neither was the realm of the nonbiomedical, including factors identified by those in the spatial and stress traditions, such as racial prejudice, residential segregation, and environmental health hazards.
The importance of the social has led many concerned with racial health disparities to move outside not just the health care domain but health entirely for policy solutions. This was the vision of Health in All Policies (HiAP), an increasingly popular framework that many policy makers embraced to combat health disparities (Hall and Jacobson 2018). Its premise was simple: to better consider the health ramifications of domains not strictly labeled “health,” such as housing, transportation, criminal justice, agriculture, and others. For instance, rather than solely designing bus routes to reduce transport time, a planner ought to also consider how rerouting buses might reduce or increase pollution in certain areas, and furthermore, whether any burdens might be disproportionately absorbed by racial minorities. Those designing new housing developments, to take another example, might incorporate health-abetting features—green spaces, walkways, convenient outlets for better foods—often absent in underresourced and segregated areas. Some HiAP proponents have advocated using formal tools, such as health impact assessments (HIAs), to translate the notion into real-world practice. Like environmental impact assessments, HIAs provide a method to project the health ramifications of potential policies, including differential racial impacts (Collins and Koplan 2009). When used by those who work outside the formal health sector, they have the effect of rendering visible the health consequences of nonhealth policies.
Racial Health Disparities as a Problem of Class
One final dimension shaping disparities was class. Racial health disparities scholarship sat alongside a parallel literature on “social class” gradients in health, some of it produced by the same researchers (Krieger, Williams, and Moss 1997; Link and Phelan 1995). It drew from a sociological literature that sought to operationalize one's economic position—how to, for example, characterize white-collar professionals as a class. And it probed many important questions: whether it was best to measure SES as a single whole or break it out into income, education, and occupational components; the limits of the SES notion itself; how to incorporate occupational prestige; and the role of income inequality in health.
Yet when this research was brought to bear on racial health disparities, class's role could sometimes be simplified. Analysts took one of three courses. The first was to treat class—measured by a quantitative measure of SES—as a variable to be “controlled” against others. More often than not, class had some effect but did not eradicate racial health disparities entirely. A second was to stratify the study population by class and race. The result generally showed that racial disparities endured, regardless of what level of class was examined. A third approach was to use interaction terms. This frequently demonstrated that lower economic class exacerbated racial health disparities.
These analyses were conducted in the language of econometrics and epidemiology. Here, class became a numeric metric conveying economic standing, while race was an ascriptive characteristic. One concluded which variable was more or less at work when it came to a particular racial health disparity. More sophisticated analyses factored in multiple levels of geography and other axes of difference such as gender. But for all their quantitative sophistication, much like “risk factor” analysis (Aronowitz 1998) that dominated much of the health sciences, they flattened dynamic social processes into tables and coefficients, mechanically isolating variables and their respective roles. And most of the time, they treated race and class as separate variables to be squared off against each other, rather than interstitial ones that shaped each other. This was true, too, of analysts who sought to explain racial health disparities by claiming they were primarily epiphenomenal effects of economic gulfs. What resulted were reams of data pointing to one or another variable's relative importance but not to the dynamic social contexts and institutional configurations that made it so.
A richer conception of class could be found in a longstanding literature on the race/class dynamic, rooted in the contributions of African American social scientists (almost all based in historically Black colleges) during the 1930s and 1940s (Cox 1948; Holloway 2002). Like the social class work, this school saw class not as a simple numeric indicator but as an individual's position relative to a labor system: that is, what sector a person worked in and whether the person received a wage or not, owned an enterprise (and if so, a small firm or a large one), operated inside or outside of formal labor markets, and so forth. This school also saw class as embedded in a dynamic system of institutions and social relations reconfigured from epoch to epoch.
Race, in this conception, came into being as an ideological construct used to rationalize New World societies' increasing dependence on enslaved people of African descent. With slavery's abolition came a reconfiguration of class relations; race continued to play a profound role, sorting the population into occupational hierarchies, and in the case of the Jim Crow United States, denying access to a host of accommodations, public programs, and occupational and residential options (Fields 1990; Morgan 1975; Wacquant 2002). For these scholars, the chief analytic goal was understanding how race operated differently as surrounding social relations changed.
The transformation of the race and class dynamic after Jim Crow and in the 21st century holds many potential implications for racial health disparities. Our approaches to understanding them must take into account many new developments. Consider just a handful. One is a massive upward distribution of income and wealth, datable to the mid-1970s and documented most vividly by economist Thomas Piketty (2013). A second is the breakdown in union membership, an overall retrenchment of a more robust welfare state, and the rise of austerity politics and budgetary stringency (Phillips-Fein 2017). This has particular ramifications for racial inequality, given the critical role of the public sector in facilitating stable and remunerative jobs for racial minorities after the civil rights movement (Katz, Stern, and Fader 2005; Pitts 2011).
Yet another is change in the very makeup of the labor force and the population at large. A Black–white binary has become increasingly untenable, with immigrants and their children making up larger demographic percentages in the United States and displacing white majorities altogether in certain states and municipalities. With demographic change has also come much more frequent use of nonwhite, non-Black racial identifiers and subidentifiers (e.g., foreign-born, third-generation, mixed-race, and so forth). This will challenge, in turn, the very definition and measurement of racial health disparities, which typically still rely on a default white reference group and assumptions about its average standing. A fourth and related development is the emergence of not interracial but intraracial economic inequality. Increasing stratification within racial categories undercuts how much race alone can signify about life chances, health and otherwise. (Katz and Stern 2006). While the rise of a Black professional class has received significant scholarly and policy attention (Boyd 2008; Lacy 2007; Pattillo 1999; Wilson 1987) for decades, a downwardly mobile—and drastically so—section of the white population has only recently generated significant interest (Case and Deaton 2015; Montez and Zajacova 2013). Its health ramifications are currently the subject of considerable debate, sparked by high-profile studies documenting declining white life expectancy.
A fifth, and final, development is capital movement (Cowie 1999; Self 2003; Souther 2017; Sugrue 1996), which has proceeded in waves since World War II and completely remade certain localities, hollowing out industrial cores and shrinking occupational opportunities in their wake, much of it felt most strongly by Black residents in cities like Camden, Detroit, Oakland, and Cleveland. The latest phase is technology that has upended entire sectors and consigned workers, many of them immigrants, to transient, low-wage service labor.
Each of these transformations comes with many potential adverse effects, many of which will be disproportionately felt by immigrants and African Americans, and will, by extension, adversely affect their health, especially those in the most marginal and subordinated of class positions. Figuring out the exact ways this will happen will be the task of future racial health disparities researchers.
For most of the 20th century, two ways of explaining racial health disparities—biology and behavior—were predominant, often at the expense of pathologized minorities. Three alternatives—what I have called the spatial, stress, and policy traditions—arose but came with their own internal tensions and limitations. Class analysis was present but often worked from a truncated and static notion of the term. Yet if the Great Recession of 2008 and the COVID-19 pandemic of 2020–21 have shown us anything, it is that the world of work and labor and the larger system of social relations is changing rapidly and undergoing continued upheaval. In addition to the robust spatial, stress, and policy traditions, class should better be incorporated in future racial health disparities discourse as a major analytic tradition. How might this occur?
Analyzing epoch-defining changes has mostly been left to qualitative social scientists and historians. These scholars' strength has been to identify precise agents, policies, and on-the-ground details that name the pieces of the system and how they operate. Their weakness, however, is limited familiarity and facility with the large-scale quantitative data of public health researchers. There is a real opportunity for cross-disciplinary collaboration and methodological complementarity here. Yet institutional structures and traditional funding agencies continue to privilege certain types of data and methods. They favor quantitative assessments of racial health disparities and interventions to reduce them, thereby devaluing studies that do not approach the problem from evaluative perspectives or statistical methods.
Thus, beyond adding or subtracting analytic traditions, it may be time to aim for more holism and methodological diversity in how we study health and why some attain well-being much more than others, especially along racial lines. For if our “long decade” of crises from 2008 to 2020 carries a single message, it is that now is no longer the time for scholarly business as usual.
My analytic debt to poverty scholars Herbert J. Gans, Alice O'Connor, Adolph L. Reed Jr., and Michael B. Katz should be obvious here. I also want to thank Alana LeBrón, Jamila Michener, Jon Oberlander, and the anonymous reviewers for their helpful comments. For an extremely probing and careful reading of an earlier draft, I thank Mo Torres.