A Scientific Approach to the Use of Racial and Ethnic Identities in Medicine and Kidney Health

The concept of race emerged from 18th century European scientists, most notably Carl Linnaeus, who formalized a pseudoscientific hierarchical classification of human subspecies based on physical appearance and continental origin. Linnaeus embedded presumed innate personal attributes into racial categories, building the pillars of racism.1,2 Since the 1950s, United Nations Educational, Scientific and Cultural Organization and social scientists have recognized race as a social construct. Scientifically, race is commonly used as a misnomer—there is only one race, the human race.1,2 La Veist defined race/racial identity in America as an inconsistently measured and socially assigned latent/proxy construct used to create heterogeneous population groups with no intrinsic differences outside of appearance and no direct relation to any medical or physiologic condition.3 Simply put, racial and ethnic identities are socially assigned group constructs circulated within society based on cultural history, political contexts, and how people see and think about themselves and others. Social systems that foster structured, racially mediated discrimination result in inequitable distribution of and access to health- and life-affirming resources and opportunities, creating health disparities. Despite this scientific grounding, the appropriate use of racial and ethnic identities in medicine and kidney health continues to challenge many researchers and clinicians who are attached to a belief that race and ethnicity are biological and causal to poor health.4,5 For over many decades, nephrology has been emblematic of racial and ethnic disparities including incident ESKD, and kidney transplantation waitlist duration and transplant receipt.6 Although the causes of these multidecade disparities are multifactorial and largely structurally mediated, findings linking higher serum creatinine values among Black individuals with more muscle mass contributed to incorporating a Black race coefficient into several kidney-specific clinical algorithms including eGFR equations.7 The recent removal of race as an individual-level modifier from the eGFR formula raised concerns that this decision was driven by a social/racial justice movement at the expense of science.4,5 In this perspective, we examine the methodologically appropriate use of racial and ethnic identities in kidney-related research and clinical care. Understanding the Use of Racial and Ethnic Identities as Variables in Kidney Research Multiple variables including sociodemographic, clinical, and laboratory values are commonly used in health research and clinical care. These variables differ widely in their characteristics, some quantitative (continuous or categorical) and others qualitative (ordinal—ordered or nominal—unordered). Sociodemographic variables are critical for understanding group-level differences/prediction models because of their strong association with many medical conditions.2,8 Carefully including racial and ethnic identities as covariates and examining stratified associations in health-related research can help adjust for and capture the influence of many measured/unmeasured factors that may affect health. This approach includes many factors for which race and ethnicity often serve as imprecise proxies such as social oppression, discrimination, and intersectionality with other identity-based biases. Moreover, the distribution of many health-related variables, such as insurance status, employment, education, income, and wealth, has also been racialized, reinforcing the need to carefully use race or ethnicity in research.2 Sociodemographic variables are also important for public health and health system research/quality improvement. Public health agencies often create community/population-level messages regarding CKD and ESKD risk factor education, outreach, interventions, and policies. Health systems research and quality improvement projects may use sociodemographic group-level differences in CKD quality of care, pharmaco-equity, and/or clinical outcomes to inform system-level interventions to improve kidney health for all groups. However, to equally assign group-level differences based on a given variable to each group member with reasonable certainty, that variable must (1) be consistently measured, (2) have a natural order or ranking relative to other variables, and (3) be directly related to the physiological/health condition or outcome of interest.8 Nominal variables (e.g., racial and ethnic identities, marriage status, religion, political party) have no intrinsic order and no direct relation to physiological/health conditions or outcomes, although they may have strong associations. Thus, group-level differences in racial and ethnic identities cannot be assigned equally to each group member. Their inappropriate use includes instituting race-specific medication use or individual-level racial/ethnic modifiers in a clinical algorithm (e.g., prior race-based eGFR estimation).8 Thus, health-related research findings of racial and ethnic group differences are best used to inform population-level messaging, prioritizing of screenings, monitoring of progress and modifying systems to address group-level disparities, and creating health and social policies. They can also be used to inform hypothesis-driven interventions such as clinical trials. Within all of these efforts, acknowledging that race and ethnicity are broad categories representing people of incredibly diverse cultures, languages, and experiences is critical. Use of Racial and Ethnic Identities in Clinical Trials There is a common but mistaken belief that broad representation ensures clinical studies capture genetic differences across specific racial/ethnic or ancestry groups. This mistaken belief has led to inappropriate racialization of gene variants and contradicts well-established evidence of greater genetic variation within versus between racial groups.9 For instance, APOL1 risk variants are highly prevalent in persons of West African descent, but infrequent in East and South Africa. APOL1 testing based on Black race or continental ancestry alone may overestimate group risk, while underestimating risk for other groups who may not self-identify with Black race (e.g., Afro-Latino individuals). In sum, the presence or absence of a gene variant does not define a racial/ethnic group. Health relevant gene variants are (1) uncommon, (2) not directly related to ancestry or racial or ethnic identities, and (3) their prevalence varies widely across and within groups. Therefore, group differences in outcomes because of these uncommon gene variants cannot be captured in most clinical trials and when found are due to the gene variant and not racial or ethnic identities. Racial/ethnic group differences in clinical trial outcomes often occur, not because of innate group differences, but rather because of the downstream effect of discrimination and marginalization emanating from maldistributed structural drivers/social determinants of health (SDOH). This negatively affects access to health-affirming resources/opportunities and increases stress, affecting neurocognitive systems/pathways, which may manifest as adaptive/maladaptive dietary and health behaviors (distinct from indigenous cultural dietary and health behaviors) and biological systems/pathways leading to differential expression of neurohormonal, inflammatory, immune, oxidative stress pathways, and epigenetic changes that affect drug/device response and side effect profiles, even after controlling for comorbidities. Broad representation in clinical trials improves their scientific quality because of enhanced transparency, confidence among underrepresented groups that the trial has been safely conducted in people like me, and the capture of a wide range of sociopolitically induced biological variation across populations. Accordingly, National Institutes of Health clinical research inclusion policies might be improved by ensuring researchers enhance their framing of race and ethnicity-specific study questions. Precision Medicine The recommendation to use ancestry-informative markers (AIMs) in lieu of racial or ethnic identities to bring greater scientific rigor to the use of race in medicine or precision medicine warrants scrutiny.2,9,10 The use of continental level AIM has been conflated with racial biology without taking into account the vast heterogeneity and changing ancestries/genetics within and across continents.10 However, AIMs merely estimate the proportion of an individual's ancestry from different regions of the world, often mirroring continental categorizations of populations.10 Like racial and ethnic identities, AIMs represent nominal variables and are best described as genetically measured social constructs reflecting lived experience, culture, anthropology, epidemiology, sociology, history, and possibly discrimination in racialized societies.10 Although AIMs provide insight regarding health-related genomic variations within and across large populations, they cannot inform individual-level precision care.8 By contrast, APOL1, BRCA2, cystic fibrosis, and other gene variants can be helpful in assessing medical risk and informing both a diagnostic and treatment plan thereby advancing individual-level precision medicine.8 Thus, we may use well-defined gene variants, but not AIMs, for individual-level diagnostic and treatment plans. Clinical Care Clinical findings, laboratory values, and evidenced-based guidelines should inform individual patients' therapies. Social histories and lived experiences, where racial and ethnic identities reside along with social networks, SDOH and family history (which may elucidate hereditary risk), help inform care plans. This may include the order of differential diagnosis and testing, frequency of follow-up, culturally and structurally tailored behavioral and lifestyle recommendations, referrals for social support, and other relevant health-enhancing interventions. Group-level racial/ethnic differences in the prevalence of CKD and common risk factors (e.g., diabetes, hypertension, family history) are well recognized. Although we cannot assign their group-level risk equally to each group member, we can note an individual has membership in a group with higher/lower risk of a given condition and explore in more depth factors for which racial and ethnic identities may be an imprecise proxy (e.g., access to care, neighborhood disadvantage, marginalization and discrimination, SDOH) to identify supportive strategies. Thus, we should not use racial or ethnic group-level stereotypes or associations in disease prevalence or clinical outcomes to inform direct individual-level clinical care, including to make or exclude a diagnosis, initiate a specific drug therapy, or create individual-level racial/ethnic algorithm modifiers.8 Summary The appropriate use of racial or ethnic identities in medicine requires, first, recognition that they refer to socially created groups and not to nonexistent biological human subspecies.1,2 Recent scientific and social justice efforts have reified the recognition of race as a social construct and the appropriate use of racial and ethnic identities as socially assigned unordered nominal proxy variables (not ordered biological variables) in nephrology.7 Thus, the removal of racial and ethnic identities as individual-level modifiers in clinical formulas and the cessation of race-based medical therapy is not grounded in social/racial justice, although a desirable outcome, but in methodologic rigor.8 Direct patient care plans should be formulated using assessments of clinical and lab data and evidence-based guidelines.2,8 This approach can then be individualized by integrating shared decision making with social and family history that together may uncover and activate ancillary interventions needed to help patients achieve their optimal health.2,8