Race-Based Care and Beliefs Regarding the Etiology of Racial Differences in Health Outcomes

Abstract

Introduction:

Physicians’ perspectives regarding the etiology of racial health differences may be associated with their use of race in clinical practice (race-based practice). This study evaluates whether attributing racial differences in health to genetics, culture, or social conditions is associated with race-based practice.

Methods:

This is a cross-sectional analysis, conducted in 2022, of the Council of Academic Family Medicine Education Research Alliance 2021 general membership survey. Only actively-practicing U.S. physicians were included. The survey included demographic questions; the Racial Attributes in Clinical Evaluation (RACE) scale (higher scores imply greater race-based practice); and 3 questions regarding beliefs that racial differences in genetics, culture (e.g., health beliefs), or social conditions (e.g., education) explained racial differences in health. Three multivariable linear regressions were used to evaluate the relationship between the RACE score and beliefs regarding the etiology of racial differences in health.

Results:

Of the 4,314 survey recipients, 949 (22%) responded, of which 689 were actively-practicing U.S. physicians. In multivariable regressions controlling for age, gender, race, ethnicity, and practice characteristics, a higher RACE score was associated with greater belief that differences in genetics (β=3.57; 95% CI=3.19, 3.95) and culture (β=1.57; 95% CI=0.99, 2.16)—but not social conditions—explained differences in health.

Conclusions:

Physicians who believed that genetic or cultural differences between racial groups explained racial differences in health outcomes were more likely to use race in clinical care. Further research is needed to determine how race is differentially applied in clinical care based on the belief in its genetic or cultural significance.

INTRODUCTION

Racially and ethnically minoritized peoples are at greater risk of poor health outcomes than their White American counterparts.¹ This reality has been used to justify the inclusion of race and ethnicity in medical recommendations, guidelines, and algorithms driving treatment thresholds and interventions.² Often, when race and ethnicity are used, it is without mention of the mechanisms through which these identities result in poor health.³ Using race without recognizing it primarily as a sociopolitical⁴ construct can stigmatize the racially minoritized as biologically inferior and normalize their poor health. It can also worsen health disparities by codifying them as normal.⁵ The extent to which physicians rely on race-based algorithms and guidelines is likely influenced by their beliefs regarding the drivers of racial differences in health. More research is needed to understand how physicians operationalize race and respond to their beliefs about the etiology of racial differences in health outcomes.

Little is known regarding the relationship between physicians’ understandings of racial health differences and their use of race in clinical care. In an evaluation of a 2005 survey of family medicine physicians,⁶ Warshauer-Baker and colleagues found no association between the belief in genetic differences as the source of racial differences in health and the value physicians assigned to race in clinical care. However, in a study⁷ of general internists, conducted with 2010 survey data, Sellers et al. found a positive association between belief in race as a biological construct (“biological race”) and race-based practice, suggesting that perceptions of race as biological are related to how physicians approach patient care. Nevertheless, physicians’ use of race in clinical care has likely changed since these studies were conducted due to increased scrutiny of race-based practice^5,8–10 in medical education,¹¹ professional societies,^12,13 and the U.S. Congress.¹⁴ Thus, a more current evaluation is needed.

In addition to beliefs regarding the biological significance of race, physicians may also hold ideas about its cultural and social significance. In Hunt et al.’s qualitative study of primary care clinicians,¹⁵ the authors found that participants believed attitudes and behaviors (“racial culture”) unique to ethnically and racially minoritized patients contributed to their poor health. Moreover, these clinicians expressed that considering their patients’ race allowed them to mitigate unhealthy aspects of their patients’ culture. Likewise, it is possible that clinicians who believe that racial differences in health are caused by racial differences in social conditions (e.g., poverty, education) respond by using race to identify patients who need the most support.¹⁶ However, clinicians may also see race-based care as essentializing race and minimizing the role of structural racism,¹⁷ and therefore eschew guidelines including race.¹⁸ No studies have quantitatively assessed the relationship between beliefs that racial differences in culture and social conditions contribute to health differences and the use of race in clinical care.

To address this gap, this study evaluated the relationship between physician’s self-reported use of race in medical decision-making and their belief in genetic, cultural, and social differences between racial groups. The authors hypothesized that physicians who attribute racial differences to cultural and genetic factors will be more likely to use race to guide clinical decision-making.

METHODS

The data were derived from the 2021 general membership survey by the Council of Academic Family Medicine (CAFM) Education Research Alliance (CERA).¹⁹ Potential respondents were sent an email invitation that included a link to the CERA survey, which was conducted through the online program SurveyMonkey^®. The study was open between September 29, 2021 and October 29, 2021, and, during that time, members received up to 5 requests to complete the survey. No incentives were given. The study was approved by the American Academy of Family Physicians (AAFP) IRB.

Study Population

CAFM is a joint initiative of 4 major academic family medicine organizations: Society of Teachers of Family Medicine (STFM), North American Primary Care Research Group (NAPCRG), Association of Departments of Family Medicine (ADFM), and Association of Family Medicine Residency Directors (AFMRD). Participants were members of CAFM organizations who did not identify as program directors, clerkship directors, or department chairs. The survey was delivered to 4,314 individuals (4,167 U.S. and 147 Canada). The sample was limited to actively practicing U.S. clinicians who identified their highest degree as MD, DO or MD/PhD. Participants who identified as practicing in Canada, did not see patients in a clinical setting, and who reported their highest degree as a non-medical degree were excluded.

Measures

The use of race in medical decision-making was determined by the RACE (Racial Attributes in Clinical Evaluation) scale. The RACE scale is a validated 7-item measure²⁰ used in prior studies to evaluate physician’s use of race in clinical care.^7,21–25 RACE scale items—such as “I consider my patients race when making decisions about which medications to prescribe”—use a 5-point Likert scale (with 0 indicating “none of the time,” and 4 indicating “all of the time”). Item responses were summed (range 0 to 28) with higher scores indicating greater use of race in clinical care.

Self-reported age (continuous), gender (female, male, other/non-binary, choose not to disclose), race (American Indian or Alaska Native, Asian, Black or African American, Native Hawaiian or Pacific Islander, White, choose not to disclose), and ethnicity (Hispanic/Latine v Non-Hispanic/Latine) were collected. Race was recategorized as Asian, Black or African American, White, and Other (American Indian, Alaska Native, Native Hawaiian, Pacific Islander) due to low cell counts in “other” racial categories. Choosing not to disclose was considered missing.

Medical school graduation year and current practice affiliation (allopathic, osteopathic, non-medical school) were reported. Duration of practice was defined as number of years since completion of medical school. Current practice region was defined as Northeast (New England and Middle Atlantic), Midwest (West North Central and East North Central), South (West South Central, East South Central, South Atlantic) and West (Mountain and Pacific). Participants described their practice location as urban, rural, suburban. Participants reported whether they were practicing in an underserved area (yes, no, unsure); responses of “no” or “unsure” were grouped together.

The study team developed and piloted 3 items to categorize respondents’ beliefs regarding the etiology of racial health differences. The developed items included: (1) “To what extent do genetic differences contribute to differences in health outcomes (e.g., differences in disease prevalence and prognosis) between racial groups?;” (2) “To what extent do differences in social conditions (e.g., income, education, environment) contribute to differences in health outcomes between racial groups?;” and (3) “To what extent do differences in culture (e.g., beliefs about health, value of diet or exercise) contribute to differences in health outcomes between racial groups?” Questions were scored on a 5-point scale Likert scale (with 1 indicating “does not contribute” and 5 indicating “it is the sole contributor”).

Statistical Analysis

Descriptive statistics summarized participant characteristics. Univariate linear regressions were used to analyze the relationship between RACE and items assessing the etiology of racial differences in health outcomes. Separate multivariable linear regressions evaluated the relationship between RACE and each racial difference item, controlling for participant age, gender (ref=female), race (ref=White), ethnicity (ref=non-Hispanic/Latine), practice region (ref=Midwest), practice location type (ref=urban), institutional affiliation (ref=allopathic), and underserved community (ref=no/unsure). All variables were selected a-priori. Since years in practice was correlated with age, it was excluded from the model. Observations with missing data were not included in the multivariable analysis. Analyses were conducted using Stata SE 16²⁶ with significance determined by p-values ≤0.05.

RESULTS

Of the 4,314, who received the survey package, 949 (22%) responded, of which 689 identified as practicing U.S. physicians. Of these 689 practicing physicians, 654 (95%) completed the RACE scale items and the 3 questions regarding the etiology of racial health differences in health. The average respondent identified as White (84%), female (61%), middle-aged (mean age 46.2 years; SD 10.6), affiliated with an allopathic medical school (55%), and employed in an underserved area (58%) (Table 1). The median RACE score was 11 (IQR 8–14), and the median scores for beliefs that social, cultural, and genetic differences were the cause of racial differences in health outcomes were 4 (IQR 4–4), 3 (IQR 3–4), and 2 (IQR 1–3), respectively. Characteristics were unchanged after limiting the sample to observations without missing data (N=567).

Table 1.

Characteristics of Survey Participants

Characteristic	Full sample	Reduced samplea
	N=654	N=567
	n (%)	n (%)
Age, mean (SD)b	46.2 (10.6)	46.2 (10.7)
Gender
Female	392 (61.1)	348 (61.4)
Male	249 (38.8)	218 (38.4)
Non-binary/Other	1 (0.2)	1 (0.2)
Missing	12
Race
White	514 (83.6)	473 (83.4)
Asian	65 (10.6)	61 (10.8)
Black	31 (5.0)	28 (4.9)
Other	5 (0.8)	5 (0.9)
Missing	39
Ethnicity
Hispanic/Latine	47 (7.4)	34 (6.0)
Non-Hispanic/Latine	590 (92.6)	533 (94.0)
Missing	17
Practice affiliated with medical school
Yes, allopathic	361 (55.4)	307 (54.1)
Yes, osteopathic	19 (2.9)	17 (3.0)
No	272 (41.7)	243 (42.9)
Missing	2
Underserved practice area
Yes	378 (58.0)	323 (57.0)
No/ Unsure	274 (42.0)	244 (43.0)
Missing	2
Urban practice location
Yes	324 (49.5)	279 (49.2)
No, Suburban	224 (34.3)	197 (34.7)
No, Rural	106 (16.2)	91 (16.0)
Practice region
Midwest	201 (30.7)	182 (32.1)
Northeast	116 (17.7)	103 (18.2)
South	167 (25.5)	140 (24.7)
West	170 (26.0)	142 (25.0)
RACE scale, n; median (IQR)	11 (8,14)	11 (8,14)
Etiology of racial differences in health outcomes
Social conditions, n; median (IQR)	4 (4,4)	4 (4,4)
Culture, n; median (IQR)	3 (3,4)	3 (3,4)
Genetics, n; median (IQR)	2 (1,3)	2 (1,3)

^aExcludes observations with any missing data (i.e., sample use in multivariable analysis).

^b36 observations were missing data on age.

RACE, Racial Attributes in Clinical Evaluation.

In unadjusted models (Table 2), RACE scores were positively associated with genetics (β=3.62; 95% CI=3.25, 3.99) and culture (β=1.58; 95% CI=1.00, 2.16), but not associated with social conditions (β= −1.06; 95% CI= −2.12, 0.00). In separate multivariable analyses controlling for age, gender, race, ethnicity, practice region, location type, institutional affiliation, and practice in an underserved area, RACE scores remained positively associated with genetics (β=3.57; 95% CI=3.19, 3.95) and culture (β=1.57; 95% CI=0.99, 2.16), and not associated with social conditions.

Table 2.

Association (β, 95% CI) Between Beliefs Regarding the Etiology of Racial Differences in Health Outcomes and RACE

Beliefs	Unadjusted (N=567)	Adjusteda (N=567)
Social conditions	−1.06 (−2.12, 0.00)	−0.72 (−1.80, 0.35)
Culture	1.58 (1.00, 2.16)	1.57 (0.99, 2.16)
Genetics	3.62 (3.25, 3.99)	3.57 (3.19, 3.95)

^aAdjusted for age, gender, race, ethnicity, institutional affiliation, underserved area, practice location type, and practice region.

RACE, Racial Attributes in Clinical Evaluation

DISCUSSION

This study found that physicians who believed that genetic or cultural differences between racial groups explained racial differences in health were more likely to self-report using race in clinical care. The belief that differences in social conditions explained racial differences in health was not associated with self-reported race-based practice. Physicians ranked differences in social conditions followed by differences in culture and genetics in order of greater to lesser importance in their contribution to racial differences in health.

The finding of a positive association between race-based practice and the belief that genetic differences between racial groups explain racial differences in health outcomes varies from the findings of Warshauer-Baker and colleagues.⁶ There are several explanations for this discrepancy. First, the outcome measures were slightly different. The Warshauer-Baker study asked physicians to rate the importance of race/ethnicity in clinical care, whereas this study used the 7-item RACE scale to measure the extent to which physicians engaged in race-based practice. While values are often related to behavior, they are not the same. Second, how physicians interpret race and value it in clinical care has likely changed over the last 15 years. Institutional efforts^12,13 to move away from race-based care are recent and likely increase practice heterogeneity through challenging current practice norms. Finally, changes in race-based practice may be advancing within academic and community settings at different rates. The Warshauer-Baker study sample included all types of family physicians while this study was limited to academic family physicians.

This study found that race-based practice was positively associated with the belief that cultural differences between racial groups contribute to racial differences in health. To the authors’ knowledge, no other study has quantified this relationship. A few qualitative studies have shown that clinicians associate race with patient’s health values and behaviors,^15,27–29 which can impact their approach to patient care.^15,29 For example, a recent study found that views about patients’ end of life preferences resulted in clinicians avoiding advanced care planning with racially and ethnically minoritized patients.²⁹ While the RACE scale has 1 item related to the general consideration of race, most of the scale addresses genetic risk explicitly or implicitly, through questions about the relevance of race in decisions regarding the titration and selection of medications. Therefore, the finding that belief in genetic difference, compared to cultural difference, was more strongly associated with RACE, and thus race-based care, is not surprising. It is possible that beliefs about patients’ values and health behaviors may be better associated with how physicians approach patient counseling and their expectations regarding their patients’ health.

In this study, physicians believed that differences in social conditions were the most important driver of racial differences in health, but this belief was not associated with race-based practice. Qualitative studies suggest that clinicians associate race with SES,^15,28 though there is no literature as to how these associations affect patient care. The lack of association between race-based practice and the belief that racial differences in health result from differences in social conditions suggests that physicians who hold this belief practice in varied ways. Some physicians may view race as identifying patients in need of supportive services and may rely on race-based algorithms and recommendations to reduce health disparities. These physicians may also believe that exposure to discrimination results in epigenetic changes that warrant race-based tools and guidelines that suggest biological differences between racial groups.¹⁶ On the other hand, physicians may perceive race-based tools and algorithms as minimizing, and therefore upholding, structural racism by attributing race, instead of racism, to poor health outcomes.¹⁸ They may also believe that race-based algorithms worsen health disparities.³⁰ For example, until 2021, the vaginal birth after cesarean (VBAC) calculator³¹—used to estimate the probability of a successful vaginal delivery after a cesarean—used Black race and Hispanic ethnicity to predict an increased risk of failure, reducing the likelihood of these patients being offered a vaginal delivery, a preferable birthing option.³²

There are several strengths to this study. This is the first study to evaluate, and find an association between, belief in cultural race and race-based practice. Second, there are only a few quantitative studies^6,25 evaluating family physicians use of race in clinical care, and this study adds to that literature. Finally, this study utilized the validated RACE scale to assess how physicians employ race in care.

Limitations

There are also several limitations to this study. First, the RACE scale measures physician’s self-reported use of race and not their observed use of race. As such, participants may have underreported the extent to which they relied on race to guide care, resulting in measurement bias and limiting the association between beliefs regarding the etiology of racial differences in health and RACE. The RACE scale was also limited in its assessment of the varied ways in which physicians incorporate race into care, with a greater focus on behaviors that imply belief in genetic differences between racial groups, thereby reducing the ability to find an association between belief in differences in culture and social conditions and race-based practice. Second, given the sensitivity surrounding conversations of race, the study may be limited by selection bias. However, study questions were a component of a larger survey measure addressing other topics and most individuals who returned the survey package, and met the study inclusion criteria, completed the RACE questionnaire along with the 3 questions regarding the etiology of racial differences. Third, this sample was majority female, but, as of 2015, women were slightly less than half of academic family physicians in the U.S.³³ As such, it is possible that this sample is not sufficiently representative of academic family medicine physicians. Additionally, the study population was limited to those belonging to academic family medicine institutions, limiting generalizability of findings to the greater family medicine physician community or academic physicians of other specialties.

CONCLUSIONS

Race was differentially employed in clinical care based on the belief that racial differences in genetics and culture contributed to racial health differences. Physicians also believed that differences in social conditions best explained racial differences in health. The belief that race provides insight into patients’ genetic disease risk is contentious and has been disavowed by some medical societies.^12,13,34 Moreover, the focus on biological race and racial culture may detract from the social factors driving racial disparities, providing a false sense of comfort in current efforts to reduce these disparities.

Additional research is needed to understand how physicians use assumptions about racial culture to tailor their care and whether these beliefs result from medical training emphasizing cultural competency or simply reflect beliefs prevalent in broader society. More work may be needed to shift providers from the cultural competency mindset, which can lead to stereotyping of racial groups, to one of cultural humility,³⁵ where, in the words of Tervalon and Murray-Garcia, a physician is “flexible and humble enough to assess anew the cultural dimensions and experiences of each patient.”³⁶ Moreover, education that explicitly endorses viewing patients as individuals with their own unique attitudes and behaviors is needed.