Use of race and ethnicity in medicine

INTRODUCTION — Race and ethnicity are constructs commonly used in medicine, but their direct relevance to medical care is limited [1]. Current use is often based on the inaccurate assumption of a biologic basis for race. The historical assumption that race is related to innate biologic and/or personal attributes can reinforce harmful stereotypes and assumptions and contribute to stigmatization, inappropriate race-based medical care, and inequitable allocation of health-affirming resources and opportunities. This topic will outline the common fallacies associated with the use of race and ethnicity in medical care and decision-making and provide guidance on the appropriate "conscious" use of race and ethnicity in medical care.

Other discussions of the role of race, ethnicity, and culture in clinical medicine are provided separately. (See "Racial and ethnic inequities in obstetric and gynecologic care and role of implicit biases" and "The patient’s culture and effective communication".)

TERMINOLOGY — Key terms and conceptual definitions related to race and ethnicity in medicine are outlined in the table (table 1).

CONTEXT — Despite its official status in government, research, and health professions, the term race is a misnomer and there is only one human race (Homo sapiens) [2]. In practice, race is used as the sociopolitical label to differentiate groups of humans based on how they look [3,4]. Racial discrimination has evolved into a system of power to control groups of people based on how they look (racism) and has expanded to exclude and/or marginalize groups based on other attributes (eg, ethnicity as defined by culture/language/religion) [3,4].

Historical fallacies and racism — Historically, pseudoscientific publications reinforced racism by describing innate group differences in physiology, biology, intellect, and other personal attributes as originally defined by Carl Linnaeus in the 1700s. These false narratives have been perpetuated in large part by the medical sciences, in both overt and insidious ways [5,6]. Because race was thought to be associated with innate biologic differences, it was often offered as a causal explanation for observed health conditions and outcomes, leading to inappropriate systematic differences in clinical care.

Instead, race and ethnicity should be understood as social constructs, with no causal biologic basis. This concept is widely recognized in the social sciences [7]. Only more recently has the medical community started gaining awareness of this, slowly leading to a change in how race is described in the medical literature and used in medical care and research.

A common example of misuse of race is in the treatment of hypertension. Many clinicians have been taught that Black patients with hypertension are more likely to respond to diuretics than White patients and use race as the basis for choosing therapy. However, this is due to misinterpretation of the use of race in clinical studies. For example, in one report, diuretic therapy achieved blood pressure <140/90 in 49 percent of Black participants compared with 42 percent of White participants [8,9]. Although this 7 percent difference was statistically significant, it has minimal clinical significance regarding selection of antihypertensive agents for individual patients. Nonetheless, many clinicians make race-based decisions about treatment of hypertension, without recognizing that such decisions are derived from a relatively small intergroup difference in a single study. Instead, we should use the antihypertensive regimen best suited for the overall medical profile of each patient. This example highlights our conditioned response to notice even slight racial differences and apply those observations to all people of each group. (See "Choice of drug therapy in primary (essential) hypertension", section on 'Role of patient race in selection of initial monotherapy'.)

Limitations and pitfalls of using race and ethnicity for clinical care

●Race and ethnicity as proxies for other factors – The observed differences in health conditions and outcomes between race- or ethnicity-defined groups are usually driven by a myriad of factors, including social determinants (or "drivers") of health, cultural risk factors, effects of racism, or in some instances associated but not directly related biologic factors (eg, population-level differences in gene frequency for certain variants). Medical science and clinical care should focus on those direct causes of the health outcomes rather than race or ethnicity. (See 'Race as a proxy for other risk factors' below.)

●Wide intragroup variation – Within any race- or ethnicity-defined group, there are wide variations in physical, social, and genetic characteristics. This wide intragroup variance (imprecision) precludes any clinically relevant assumptions about individual members of a group. Clinical care should always be based on individual patient characteristics rather than assumptions based on their membership in a group.

●Reinforcement of implicit or explicit bias – When a clinician incorporates race or ethnicity into a clinical narrative or care decision, they may knowingly or unknowingly reinforce assumptions about the individual's characteristics or experiences. This common practice of assuming and documenting a patient's race or ethnicity reinforces the misconception of an innate biologic basis for race and is subject to implicit bias or stereotypes [10].

●Race-based medicine may exacerbate health disparities – Given the above pitfalls, medical practices that focus on race or ethnicity to guide clinical care can exacerbate health disparities. For example, by focusing on race, a clinician may ignore the underlying proximate drivers of health (eg, social, environmental, or genetic). Similarly, using algorithms that incorporate race or ethnicity as a factor often lead to unjustified variations in diagnosis or treatment. (See 'Use of race in medical algorithms and decision aids' below.)

RACE AS A PROXY FOR OTHER RISK FACTORS — Historically, the medical profession has used race explicitly (eg, choosing to use a race-based algorithm), or implicitly (eg, making assumptions about a patient's innate risk factors and motivations rather than seeking more specific information). We should not do this. Instead, lifestyle risks should be described as specific behaviors (eg, smoking, diet, betel nut consumption, breastfeeding initiation) rather than as racial characteristics or associations. Race and ethnicity should only be considered in the context of a complete evaluation, including social and family history to explore established causative risk factors underlying group level associations, with an individual medical history and physical examination. In select clinical scenarios, race or ethnicity may influence the order of targeted laboratory testing.

Social and cultural drivers of health — In the United States and many other parts of the world, health inequalities exist between populations, and groups that have been socially and economically marginalized often experience an unbalanced distribution of the social drivers of health (SDOH), also known as the social determinants of health, as well as other negative effects of racism/discrimination [11]. In these situations, if a clinician or guideline focuses on race as cause of the resultant health outcomes, they are less likely to recognize and address the true underlying drivers. These false assumptions can lead to patient stigmatization, inappropriate or ineffective interventions, and inaccurate assessment of the true risks that can affect anyone in similar circumstances.

Structured discriminatory practices force communities that have been marginalized to suffer from the maldistribution of resources. One example is a greater burden of environmental exposures and related health consequences. These burdens may include lack of access to healthy food options, exposure to heavy metals from lead in water, proximity to industrial waste dumps, direction of commercial traffic through urban minority low-income communities leading to increased small particulate matter and noise pollution, lack of trees or parks resulting in excess heat, and more [12,13]. These burdens contribute to higher risk of chronic illness such as obesity, diabetes, kidney and pulmonary disease, premature cardiovascular disease, cancer, mental health disorders, and other conditions. Moreover, the deliberate creation of low-income communities with limited resources has predictably promoted a lower sense of self-worth, higher rates of crime, violence, suicide, and premature death from firearms and other weapons.

Specific examples of environmental injustice include lead poisoning in the water supply in Flint, Michigan, discussed separately (see "Lead exposure, toxicity, and poisoning in adults", section on 'Sources of exposure'), and residential "redlining" (the discriminatory practice of categorizing neighborhoods based on perceived mortgage investment risk [14]), which influences neighborhood environments and health risks. (See "Asthma in adolescents and adults: Evaluation and diagnosis", section on 'Social determinants of health and asthma'.)

Examples of health disparities that are sometimes attributed to race but are often actually mediated by SDOH that result from historic and current social and economic marginalization include:

●Diabetes – Primary drivers of increased risk include socioeconomic status, access to healthcare, quality food, and safe physical activity [15].

●Unhealthy diet – Primary drivers of increased risk include limited access to healthy foods, cultural exposure (shared traditions or habits and/or maladaptive responses to marginalization), predatory marketing, and limited access to quality medical care [16].

●Hypertension – Primary drivers of increased risk include factors related to socioeconomic status, in addition to the stress and physiologic changes related to constant exposure to discriminatory experiences and social adversity [17].

●Lower rates of breastfeeding initiation – Primary drivers of increased risk include cultural exposure (experience of family and peers), lower maternal age and maternal educational level, reduced support systems, reduced access to medical care (eg, counseling during prenatal and postnatal care), and economic drivers (eg, need to return to work). As an example, birthing facilities that serve majority-Black populations are less likely to help with breastfeeding initiation or offer lactation support during the birth hospitalization [18,19].

●Smoking – Primary drivers of increased risk include cultural exposure, predatory marketing, maladaptive behavior to cope with stress, and economics (less expensive than vaping) [20].

●Asthma – Primary drivers of increased risk include exposure to pollutants and allergens, which are associated with living in poor quality housing with high density of dust; insects and rodents; and close proximity to highways or industries that generate small particulate matter [21].

●Lead exposure – Primary drivers of increased risk include poor quality housing and/or proximity to industries with lead pollution [22].

●Maternal morbidity – In the United States there are large differences in rates of maternal mortality, with highest rates occurring among non-Hispanic Black individuals [23]. Causes are multifactorial but likely include direct and indirect effects of racism. (See "Severe maternal morbidity", section on 'Racial and ethnic minorities'.)

Infectious disease exposures — Many infectious diseases exhibit substantial geographic variation. In most cases (eg, hepatitis A, tuberculosis, malaria), these are due to differences in SDOH, including sanitation, housing conditions, vectors, access to medical care (eg, immunization programs), or geographic distribution of the infectious agent (eg, malaria). These proximate causes should not be confounded with race or ethnicity.

Infectious diseases that can be vertically transmitted from mother to child are more complex because they reflect exposures of previous generations. As examples, hepatitis B and hepatitis C, which are more common in the Western Pacific and African regions, are also more common in the children of individuals who migrate to a lower-prevalence region, giving the appearance of an innate racial difference.

Physical characteristics — Certain physical characteristics that are associated with race or ethnicity are relevant to medical decision-making. In these cases, the focus should be on the proximate cause, which is the physical characteristic itself, rather than race or ethnicity groupings. This approach recognizes the wide variation in physical characteristics within any self-identified race or ethnic group, and the substantial overlap in characteristics between different groups.

As examples:

●Skin pigmentation

•Risk for vitamin D deficiency – Conversion of previtamin D to vitamin D by sunlight is controlled by skin pigmentation and temperature. Persons in an equatorial environment tend to have darker skin pigmentation which reduces the rate of conversion of previtamin D to vitamin D by sunlight. People with more skin pigmentation who migrate from equatorial latitude to less sunny and colder environments may be at an increased risk for low vitamin D levels [24]. Vitamin D status is also affected by sun exposure (eg, clothing) and dietary vitamin D intake. (See "Vitamin D insufficiency and deficiency in children and adolescents".)

•Transcutaneous bilirubin measurements in infants – Transcutaneous bilirubin measurements tend to overestimate serum bilirubin in infants with darker skin pigmentation and underestimate serum bilirubin in those with lighter skin pigmentation. Clinical algorithms should recognize this systematic error and incorporate other methods (eg, appropriate thresholds for direct measurement of serum bilirubin) to ensure appropriate care. (See "Unconjugated hyperbilirubinemia in term and late preterm newborns: Screening", section on 'Transcutaneous bilirubin (TcB)'.)

•Pulse oximetry measurements – Pulse oximetry can provide falsely elevated readings in patients with darker skin. This concern for "hidden hypoxemia" is discussed separately. (See "Pulse oximetry", section on 'Skin pigmentation'.)

•Appearance of skin lesions – The appearance of skin lesions and dermatologic conditions varies with skin pigmentation, and clinicians often have lower diagnostic accuracy when assessing darker skin [25]. Clinicians must learn to recognize these differences, and medical information resources should support clinicians by including images and descriptions of dermatologic conditions in the full range of skin phototypes [26-28].

●Hair type and distribution – Similarly, hair type and distribution are associated with, but not directly related to, race or ethnicity. These characteristics may be relevant for the evaluation of hirsutism or timing of pubertal milestones, but should be interpreted in the context of familial patterns rather than race or ethnicity per se. (See "Evaluation of premenopausal women with hirsutism" and "Definition, etiology, and evaluation of precocious puberty", section on 'Epidemiology'.)

Complex or undetermined factors — For some other health-related issues that vary between race- or ethnicity-defined groups, the proximate cause is unclear:

●Pubertal onset – In the United States, the timing of pubertal onset is somewhat earlier in Black girls compared with Hispanic or White girls, based on several population studies. However, race is probably a proxy for other factors, which may include elevated body mass index, family history (mother's pubertal onset), and possibly stress or other SDOH [29]. Moreover, there is much more variance within each group than between the groups. (See "Definition, etiology, and evaluation of precocious puberty", section on 'Threshold for evaluation'.)

●Gastric cancer – The prevalence of gastric cancer is much higher in East Asian populations (Mongolia, China, South Korea, and Japan) [30]. True risk factors are probably Helicobacter pylori infection and diet (salt and salt-preserved foods, and nitroso compounds), with some familial predisposition. (See "Risk factors for gastric cancer".)

EFFECTS OF RACISM — Racism and its many manifestations is a powerful risk factor for poor health, morbidity and mortality [31]. Members of affected groups suffer the direct and indirect effects of racism, which may be inaccurately ascribed to innate biologic characteristics. Mediators of racism may include:

●Disparities in social drivers (SDOH) – Differences in resources, opportunities, environmental exposures, and health care, which were generated by and often perpetuated by systemic racism. (See 'Social and cultural drivers of health' above.)

●Mistrust of medical care – The history of racism in health care and a patient's personal experiences with racism may lead to mistrust or disconnection from medical care [32,33].

●Effects of chronic stress – Persistent discrimination, marginalization, and limited access to resources leads to increased stress for oppressed groups. This excess stress leads to what has been termed allostatic load or "weathering", which refers to the health disadvantage from a cumulative lifetime exposure to adverse socioeconomic conditions and discrimination [34]. Proposed mediators of this effect are hyperactivation of neurohormonal and sympathetic nervous systems, increase oxidative stress and inflammatory gene expression, suppressed immune function, and epigenetic changes which can be transmitted intergenerationally [35].

Allostatic load can be measured by physiologic parameters that usually include systolic and diastolic blood pressures, body mass index, glycated hemoglobin, albumin, creatinine clearance, triglycerides, C-reactive protein, homocysteine, and total cholesterol, and is strongly associated with premature aging and mortality risk [35-38]. Appropriate race- and ethnicity-conscious health care recognizes the possibility that allostatic load may contribute to a patient's condition, and may include a more extensive medical evaluation to uncover specific drivers that should be addressed to optimize health [34]. (See 'Race- and ethnicity-conscious care' below.)

These biologic changes are induced directly or indirectly by racism and discrimination and are not innate to one's race or ethnicity.

GENETIC RISK FACTORS ASSOCIATED WITH RACE OR ETHNICITY — There are some group-level associations between race or ethnicity and the prevalence of certain medical conditions. The vast majority of these associations are due to nonbiologic factors. Only a small proportion (1 to 5 percent) of these associations are due to group-level differences in the prevalence of different gene variants. These are due to geo-evolutionary factors and consolidation of genetic traits within an isolated population or "founder effect", and not due to race or ethnicity. We must use the actual clinical, lab, and/or gene variant data to make direct clinical and treatment decisions, and not race or ethnicity. We can use race or ethnicity information in prioritizing our differential diagnosis, and perhaps for prioritizing testing for specific genetic risk factors, but not in making treatment decisions. We should never exclude a potential diagnosis in a patient because of their race or ethnicity.

A few examples are:

●Sickle cell disease – Sickle cell disease and many other hemoglobinopathies are highly prevalent in populations where malaria is present, but they are not directly related to race or ancestry-informed markers. Because most of the slave trade from Africa to the United States emanated from sub-Saharan Africa where malaria is highly prevalent, these hemoglobinopathies are more prevalent among Black Americans from sub-Saharan Africa or among the descendants of formerly enslaved Africans in the Americas and in persons from the Mediterranean and southeast Asia [39]. By contrast, these hemoglobinopathies are nearly absent in several other parts of Africa. (See "Pathophysiology of sickle cell disease".)

●Chronic kidney disease – Similar to sickle cell disease, APOL1 variants associated with risk for chronic kidney disease arose in response to infectious agents (trypanosomiasis) and are highly prevalent in some parts of sub-Saharan Africa. Consequently, the APOL1 variants are also more common in Black descendants of formerly enslaved Africans in the Americas than people with predominantly European ancestry. Patients with two high risk APOL1 alleles are at higher risk for kidney failure, but appear to require another activating event or condition such as systemic lupus erythematosus or HIV. (See "Gene test interpretation: APOL1 (chronic kidney disease gene)".)

●Breast cancer – Pathogenic variants in BRCA1 and BRCA2 are strongly associated with breast and ovarian cancer and are more prevalent in certain groups defined by ethnicity or geographic area, including persons of Ashkenazi or Eastern European Jewish ancestry. This association is related to consolidation of these variants in an isolated population but is not caused by ethnicity. (See "Genetic testing and management of individuals at risk of hereditary breast and ovarian cancer syndromes".)

●Cystic fibrosis – Cystic fibrosis is relatively common in people with European ancestry (less so in Scandinavian groups). It is also reasonably common in Black and Hispanic populations but is often underdiagnosed because of the historical fallacy that the disease is uncommon in these groups. Also, the pathogenic CFTR polymorphisms vary between populations and the gene panels used for screening are biased towards detecting the most common variants in people with European ancestry. Black or Hispanic people with cystic fibrosis are less likely to have genotypes that respond to the new highly effective CFTR modulator drugs [40]. These observations reinforce the importance of treating based on gene polymorphisms/variants and not race or ethnicity, and the need to expand research, screening tools, and drug development to optimize treatment for all. (See "Cystic fibrosis: Treatment with CFTR modulators".)

Finally, a lot has been made of using ancestry-informative markers as an approach to understand race and ethnicity and as a possible tool for precision (or "personalized") medicine. Ancestry-informative markers are single-nucleotide polymorphisms used to estimate the proportion of ancestry of an individual derived from each population. However, these genetic markers do not measure any biologic properties, and serve only as markers of lineage, which is linked to a combination of diverse factors such as culture, anthropology, demography, epidemiology, history, and sociology [41]. Thus, like race and ethnicity, ancestry-informative markers should not be used in directly treating an individual patient. A more informative clinical approach for a given patient is to inquire about their individual family history of disease.

USE OF RACE IN MEDICAL ALGORITHMS AND DECISION AIDS — Race has been included as a factor in developing medical algorithms or decision aids, which are generally based on group-level differences observed in population studies. Using race as one of many factors to ascertain overall risk is appropriate. However, extrapolating group-level differences to determine an individual's risk has important pitfalls. In some cases, this practice reflects the false assumption that race has a biologic basis (ie, that innate physiologic differences between racial groups cause or exacerbate a medical or physiologic condition). In other cases, race is included as a factor in an attempt to address or account for racial or ethnic disparities in a clinical condition or outcome but is really just a proxy for other determinants (eg, social drivers of health [SDOH]).

The use of assigning group-level racial differences equally to each group member as an individual-level modifier is not methodologically appropriate because race nor ethnicity meet the criteria required to include an individual-level variable in a clinical algorithm:

●Criteria for individual-level variables in clinical algorithms – To be used as a predictor of an individual's level of risk, a clinical variable must be consistently measured, be ordered, and be directly related to the outcome of interest. For example, HbA1c meets these criteria for diabetes and its complications. It is consistently measured, it is ordered (increasing HbA1c concentrations predict increasing risk), and it is directly related to the diagnosis of diabetes and risk of complications.

●Why race does not meet minimum criteria – Race and ethnicity do not meet any of the three major criteria. As social constructs (like political party, religion, or favorite color), race and ethnicity are not consistently measured, are not ordered (ie, these are nominal variables, with no inherent ranking), and have no direct relation to any medical condition. The methodologic use of assigning social group-level differences in a condition or outcome equally to each member of the group is known as an "ecologic or population fallacy" [42]. Thus, similar to political party or religion, race should not be used as individual-level predictor in medical algorithms even if group-level associations exist.

Thus, both scientific and social justice call for efforts to stop using race and ethnicity as a modifier in medical algorithms designed to predict individual-level risks.

Despite increasing recognition of this concern, several algorithms and formulas with individual-level racial or ethnic modifiers are still being used for clinical care (eg, heart failure prognosis, cardiac surgery mortality, kidney transplant failure, risk of ureteral stone, rectal cancer survival, breast cancer risk, osteoporosis, and fracture risk) [43,44]. From a methodologic perspective, all individual-level racial or ethnic modifiers in algorithms and formulas should be removed, but a process for their removal needs to include thoughtful steps for updating the algorithms and formulas and ensuring equity in care during the transition from race-based to race-conscious medicine [44,45].

Encouragingly, many algorithms and formulas are being re-examined with the intent to deconstruct (ie, identify the true underlying risk factors rather than race or ethnicity), revise, and retire the race or ethnicity modifier. Notable examples include vaginal birth after cesarean, estimation of glomerular filtration rate (GFR), and management of urinary tract infection for infants and children [44].

Below are a few examples, which are discussed in more detail separately:

●Vaginal birth after cesarean birth – Historically, calculators and equations to estimate the likelihood of successful vaginal birth after cesarean birth included a factor for race. The race factor was removed in recognition that risk is driven by social and physical characteristics rather than race itself [46]. (See "Choosing the route of delivery after cesarean birth", section on 'MFMU Network calculator for use at entry to prenatal care'.)

●Estimation of glomerular filtration rate (GFR) – Older equations or calculators for estimating GFR incorporated race, which resulted in a higher eGFR for self-identified Black individuals, and thus reduced the sensitivity for detecting kidney function impairment. The preferred modern equation removes the race factor with minimal effect on accuracy. (See "Assessment of kidney function", section on 'Estimation of GFR'.)

●Pulmonary function tests – Interpretation of pulmonary function tests that include race and ethnicity factors may underestimate the severity of pulmonary disease in Black individuals. The data and ongoing controversies are discussed separately. (See "Selecting reference values for pulmonary function tests", section on 'Effect of race/ethnicity'.)

BEST PRACTICES

Race- and ethnicity-conscious care — There is a role for race in medicine since we are in a race-conscious society with racialized structures and systems. We would be naïve to never consider race (race-blind), but we don't want to perpetuate the misuse of race (race-based). Instead, we should use a race-conscious approach [47]. This means we are aware that a person's race or ethnicity causes them to be treated differently, experience life differently, and have different levels of access to life- and health-affirming resources and opportunities. We cannot know how much each of these has impacted each individual patient, and there may be group level overlap in the downstream effects of the differential treatment by other identities such as sex, gender identity, class, neurodiversity, and more.

Best practices for race-conscious medicine include (table 2):

●Appropriate documentation of race or ethnicity – In clinical practice, we should not use race as an identifier in the chief complaint or history of present illness as it has no direct relation to most medical conditions or physiologic processes [10]. These descriptive terms can be provided in the social history along with other factors that may contribute to a health condition, such as social drivers of health (SDOH) or environmental exposures. In this context, race and ethnicity provide clinically relevant information because they are risk factors for racism and related adverse effects on health, which may influence access to care or inform the treatment plan. It is best to ask the patient to describe their own race or ethnicity, because self-identified group membership is an important lens through which one views and lives in a racialized society [10].

●Recognize and correct historical and contemporary race-based myths and misconceptions – Historical fallacies about innate group differences in physiology and biology have been perpetuated in medical literature and medical education, and some persist to this day [48]. Awareness of these fallacies is important to providing appropriate care and to building trust with patients. (See 'Historical fallacies and racism' above.)

●Promote race- and ethnicity-conscious screening guidelines – Recognition of group-level associations can be appropriately used to sensitize a clinician to specific disease risks. Where group-level differences in disease prevalence, severity, and onset exist, different group-level recommendations for education and screening might also be appropriate. This population-level use of race in the United States is appropriate because race represents populations of people who have different outcomes, not based on their biology, but their treatment in society or, in rare instances, group-level differences in gene polymorphism/variant prevalence.

As examples, we agree with recommendations to utilize familial risk assessment tools in patients with ancestry associated with BRCA1/2 gene mutations in order to determine need for genetic counseling and/or genetic testing [49]. (See "Genetic testing and management of individuals at risk of hereditary breast and ovarian cancer syndromes", section on 'Criteria for genetic risk evaluation'.)

Similarly, awareness of increased risk for prostate cancer should inform shared decision making regarding early prostate screening in Black/African American men [50]. (See "Screening for prostate cancer", section on 'Approach to screening'.)

●Prioritize the differential diagnosis for social or genetic risk factors – Race and ethnicity should not be used as a proxy for social or genetic risk factors, as discussed above (see 'Race as a proxy for other risk factors' above). However, because some group-level associations exist, these associations may be used to prioritize the differential diagnosis and guide further evaluation, including a complete social and family history, medical history, and physical examination. In some cases, the differential diagnosis may be informed by knowledge of higher or lower group-level prevalence of some conditions, such as cystic fibrosis, BRAC1/BRAC2, sickle cell disease, or APOL1 alleles associated with risk for chronic kidney disease. A patient's race or ethnicity may inform the sequence of an evaluation, but should never be used to make or exclude a diagnosis. It is confirmed or excluded by testing for the condition. Direct medical treatment should never be based on race or ethnicity, but only on clinical and laboratory data.

●Use our understanding to guide evaluation and patient health education – People experiencing social and economic disadvantages warrant extra attention to appropriate screening and counseling for a wide range of conditions, including cancers (colorectal, human papilloma virus, lung), cardiovascular disease, diabetes, infectious diseases (HIV, hepatitis C, tuberculosis), substance use (tobacco, alcohol, other), mental health, oral health, and social risks (poverty and intimate partner violence) [51]. These goals call for measures to enhance access to primary care and to prioritize outreach and support for at-risk populations.

For an individual patient, it is reasonable to recognize that their membership in a racial or ethnic group is associated with a higher health risk as an early step in an evaluation. Then, it is critical to explore the true risk factors, such as exposure to adverse SDOH including racism/discrimination or in uncommon cases, the likelihood of a pathogenic gene variant. As examples:

•Diabetes has a higher incidence in many racial and ethnic minority groups. A patient's individual risk should be based on actual laboratory measures (eg, hemoglobin A1C), which should not be interpreted with a race-based algorithm. However, in determining when to screen a patient, it is reasonable to consider a lower threshold for screening (ie, lower body mass index cutoff) for certain racial and ethnic populations given what we know of increased trends in prevalence across difference groups [52]. Knowledge of the increased risk may also help to sensitize the clinician and patient to the need for regular screening and follow-up. In this way we are recognizing a patient’s racial group membership and how it may indirectly impact their health, typically through the exposure to adverse social risk factors (race-conscious approach).

•Average age-adjusted blood pressure is higher among members of the Black/African American community compared with the White community. We can use that information to justify more targeted blood pressure education and ensure regular screening.

●Be cautious in using race-based algorithms – When you encounter a race-based algorithm in medicine (one that assigns a modifier to an individual based on their race or ethnicity), be aware that the use of race may reflect the false assumption that race has a biologic basis, and thus interpret results with caution. Consider using an alternative algorithm or risk calculator that does not include a factor for race.

Public health/allocation of resources — Collection of data on disease prevalence or complications by race, ethnicity, and other identities, is often performed in medical research and by public health departments and individual health systems. These data are crucial to understanding health disparities and their mediators. If group-level outcomes differ, public health/health services research should seek to determine whether this is due to group-level differences in SDOH or systematic differences in health care, such as the use of evidenced-based therapeutics or testing. Systematic differences in health care may be mediated by differences in health insurance status, provider recommendation, patient health belief/behavior (eg, medical mistrust or cultural practices), or family/community level factors such as social support or environmental exposures.

When properly analyzed, data on group-level differences can be used to drive group-targeted interventions to eliminate disparities, such as [53]:

●Community level messaging and education

●Screening of at-risk populations

●Policies and programs to eliminate health care disparities

●Monitoring the results of group level interventions

●Advocacy to reduce inequities in the distribution of the SDOH (housing, poverty, opportunities)

SOCIETY GUIDELINE LINKS — Links to society and government-sponsored guidelines from selected countries and regions around the world are provided separately. (See "Society guideline links: Diversity, equity, and inclusivity in medicine".)

SUMMARY AND RECOMMENDATIONS

●Context – The use of race and ethnicity in medicine is often based on the inaccurate assumption of a biologic basis. Race and ethnicity are social constructs, and do not provide a causal explanation for observed health conditions and outcomes. (See 'Context' above.)

●Terminology – Key terminology is described in the table (table 1).

●Race as a proxy for other risk factors – Differences in health conditions and outcomes between race- or ethnicity-defined groups have been inaccurately ascribed to innate biologic characteristics, but are usually driven by other factors, including social drivers of health (SDOH), cultural/lifestyle risk factors, differential community-level exposures, effects of racism, or (rarely) differences in the prevalence of select genetic risk factors. (See 'Race as a proxy for other risk factors' above.)

●Effects of racism – Racism itself is a powerful risk factor for poor health, morbidity, and mortality. The effects of racism on health may be mediated by inequities in resources or opportunities, or exposure to environmental hazards and chronic stress. (See 'Effects of racism' above.)

●Genetic factors – There are some group-level associations between race or ethnicity and the prevalence of certain medical conditions, but these conditions are not equally distributed within any racial or ethnic group and do not define a group. Only a small proportion of the group-level differences are related to differences in the prevalence of gene variants. (See 'Genetic risk factors associated with race or ethnicity' above.)

●Use of race in medical algorithms – Race has been included as an individual-level modifier in many medical algorithms or decision aids, reflecting the false assumption that there are innate physiologic differences between racial groups that cause or exacerbate a medical or physiologic condition. Many algorithms are being re-examined with the intent to deconstruct, revise, and retire race modifiers. Notable examples include estimation of glomerular filtration rate (GFR), prediction of successful vaginal birth after cesarean birth, and management of urinary tract infection for infants and children. (See 'Use of race in medical algorithms and decision aids' above.)

●Best practices – A race-conscious approach to medical care means we are aware of the fact that based on their race or ethnicity, groups of people are treated differently, experience life differently, and have different levels of access to life- and health-affirming resources and opportunities. Best practices include efforts to (table 2):

•Avoid including race and ethnicity in the chief complaint or medical history narrative. Instead, include self-identified race or ethnicity as part of the social history.

•Recognize and correct historical race-based myths and misconceptions.

•Promote race- and ethnicity-conscious screening guidelines.

•Use self-identified race and ethnicity as part of the complete history that will be used to prioritize the differential diagnosis for social or genetic risk factors, but do not use race or ethnicity to assume or exclude a diagnosis.

•Use our understanding of race and ethnicity to guide evaluation and individualized patient health education.

•Be cautious in using race-based algorithms. (See 'Race- and ethnicity-conscious care' above.)

●Public health considerations – Awareness of group-level differences can be used to drive targeted interventions to eliminate health disparities through education, screening, monitoring, and advocacy/action to directly reduce inequities. (See 'Public health/allocation of resources' above.)