Perceived interpersonal racism in relation to incident coronary heart disease among Black women

Abstract

Background:

Racism is highly prevalent in the United States. Little data exists about whether perceived interpersonal racism is associated with risk of coronary heart disease (CHD).

Methods:

We followed 48,305 participants in the Black Women’s Health Study through biennial mailed and web health questionnaires from 1997, when they provided information on perceived interpersonal racism and were free of cardiovascular disease and cancer, until the end of 2019. We averaged participant’s responses to five validated questions about perceived interpersonal racism in everyday activities, such as “people act as if they think you are dishonest”. We summed the positive responses to three questions about perceived racism in interactions that involved jobs, housing, and police; score ranged from 0 (No to all) to 3 (Yes to all). CHD cases were defined as non-fatal myocardial infarctions confirmed through medical records, fatal cases identified through the National Death index, and self-reported revascularization events. We used Cox proportional hazard models adjusting for major confounders to estimate hazard ratios.

Results:

During 22 years of follow-up, we identified 1,947 incident CHD cases. For women who reported experiences of racism in employment, housing, and involving the police relative to women who reported no such experiences, the age-adjusted hazard ratio for CHD was 1.35 (95% CI: 1.13–1.61, p trend=0.006) and multivariable HR for CHD was 1.26 (95%CI: 1.05–1.51, p trend=0.05). For women in the highest quartile of perceived interpersonal racism in daily life relative to women in the lowest quartile, the age-adjusted hazard ratio for CHD was 1.25 (95% confidence interval (CI): 1.07–1.46, p trend=0.006). After multivariable adjustment, the HR was attenuated and no longer statistically significant.

Conclusions:

Perceived experiences of interpersonal racism in employment, housing and with the police were associated with higher incidence of CHD among Black women, while perceived racism in everyday life was not associated with higher risk.

Introduction

Cardiovascular disease (CVD) is the leading cause of death in the United States. Black women have disproportionally higher burden of coronary heart disease (CHD) and CHD-related mortality and earlier onset than women of other racial and ethnic groups^1,2. Myocardial infarctions (MI) also occur at younger ages in Black women, and these events are more likely to be fatal than those of White women^1,2. Many of these CHD-related deaths are preventable^3,4. Underlying reasons for these racial disparities are poorly understood.

The American Heart Association (AHA) stated that the most significant opportunity to reduce cardiovascular disease lies in addressing the social determinants of health, and studies are needed to assess racism as “a fundamental driver of health disparities”^3,4. Structural racism^5–7, refers to the ways societies, based on race, limit access to the resources (e.g., housing, health care, education) needed for living a healthy life^8,9.

Individuals may be unaware of the existence of structural racism. On the other hand, perceived interpersonal racism refers to self-reported experiences of unfair treatment perceived by the recipient as being due to her/his race. Currently, little is known regarding the longitudinal association between perceived racism with incident CHD.

In the present study, we examined whether perceived interpersonal racism, in everyday circumstance and in settings involving job, housing, and interactions with the police, is associated with incident CHD, using data from the largest follow-up cohort study of Black women in the United States, the Black Women’s Health Study (BWHS)¹⁰.

Methods

Data, Materials, and Code Disclosure Statement:

Data are available upon reasonable request to the senior author. Information on the procedures to obtain and access data from the Black Women’s Health Study is described at http://www.bu.edu/bwhs under “information for researchers.”

Study population:

The BWHS is an ongoing, prospective study of Black women from across the continental United States^10,11. In 1995, 59,000 women aged 21–69 years (median age=38 years) enrolled by completing mailed questionnaires. They were primarily subscribers of Essence magazine (94% of BWHS participants) and residing in 17 states across the mainland U.S. At baseline, respondents provided data on demographic characteristics, socioeconomic factors, medical conditions, and lifestyle factors. Every two years, participants update health information on follow-up mailed and web questionnaires. BWHS participants came from neighborhoods with a wide range of socioeconomic status. Follow-up through 2019 is complete for 86.2% of person-years. All but 5% of BWHS participants were U.S. born. The Boston University Medical Campus Institutional Review Board approved the study.

Incident coronary heart disease (CHD):

We defined incident CHD as the first occurrence of either fatal or nonfatal myocardial infarction (MI), coronary artery bypass grafting (CABG) or percutaneous coronary intervention (PCI). Fatal CHD cases were ascertained through the National Death Index. The other outcomes were reported by participants on their biennial heath questionnaires. Women were classified as having had fatal CHD if it was listed as the underlying cause of death (the International Classification of Diseases, 9th Revision, 410–414, ICD10 code I20-I25)¹². Non-fatal cases were confirmed through medical records review. Participants who reported occurrences of physician-diagnosed MI and CABG/PCI were asked for permission to obtain medical records from their relevant hospitalization. Adjudication was carried out by cardiologists and trained researchers blinded to participant exposure status. Agreement between medical records adjudicators was 100%.

For MI adjudication, we used Framingham Heart Study criteria for determination of MI status^13,14. A participant was considered to have had a definite MI if at least two of the following three criteria were documented in the records: 1) symptoms indicative of ischemia; 2) changes in biomarkers of myocardial necrosis; 3) serial changes in electrocardiograms indicating the evolution of an infarction, including but not limited to ST-elevation MI and the loss of initial QRS potentials (development of pathologic Q-waves)”. Other clinical factors including angiographic data (if available) and imaging were also considered by adjudicators. Among participants whose appropriate records were obtained for review, 56% (251 out of 450) were confirmed for MI. For CABG and PCI, 272 of 305 self-reported cases (90%) were confirmed through medical record review.

For the current analysis, the primary outcome was incident definite CHD, which included non-fatal MI confirmed through medical records review, fatal CHD identified through the National Death index, and self-reported revascularization events. We also conducted separate analyses of MI only. In secondary analyses, we included probable MIs (n=70).

Perceived interpersonal racism:

BWHS questions on racism/discrimination were adapted from previously validated instruments (Cronbach’s alpha =0.74, and test-re-test reliability coefficients=0.70^15,16 , Table S1). The questions were designed to measure perceived interpersonal racism as experienced by the individual in everyday life and when dealing with situations that involved employment, housing, or police. For BWHS, the weighted k was 0.73 for the summary score based on five questions on everyday racism and 0.67 for the three questions on experiencing racism on the job, in housing, and involving police¹⁷. The information used on experiences of perceived interpersonal racism was collected in the 1997 questionnaire¹⁵. Five questions asked about perceived interpersonal racism in everyday activities, such as how often “people act as if they think you are dishonest”. Three questions were about perceived interpersonal racism in interactions that involved jobs, housing, and police^18–20. We calculated a score for perceived interpersonal racism in everyday life by averaging participants’ responses to the five questions and dividing into quartiles (quartile 1=lowest; quartile 4=highest). We calculated a perceived interpersonal racism score for interactions that involved jobs, housing, and police by summing the positive responses to the three questions; the score ranged from 0 (No to all three questions) to 3 (Yes to all three questions).

Prior BWHS analyses have found associations of higher racism scores with higher risk of stroke²¹, type 2 diabetes²², obesity¹⁹, clinical insomnia²³, and with lower subjective cognitive function²⁰.

Covariate assessment:

Covariates were chosen a priori based on the literature. We included age (continuous), body mass index (<23, 23–24.9, 25–29.9, 30–34.9, ≥35 kg/m²), years of education (≤12, 13–15, 16, ≥17 years), geographic region (Northeast, South, Midwest, West), neighborhood socioeconomic status (SES, quintiles)²⁴, vigorous physical activity (<1hr, 1–4 hrs, ≥5 hrs per week), cigarette smoking (never smoked, current smoker<15 cigs/day, current smoker ≥15 cigs per day, former smoker quit <10 years ago, former smoker quit ≥10 years ago), history of diabetes (yes, no), history of hypertension (yes, no).

Hypertension was defined as a report on a biennial questionnaire of use of an antihypertensive drug, or a report of hypertension together with use of an antihypertensive medication or diuretic. In a validation study, 138 out of 139 (99%) of self-reported cases of hypertension were confirmed by medical records^25,26. Type 2 diabetes was defined as self-report of diabetes at age 30 or older. In a validation study, 217 (94%) of 229 women who reported diabetes were confirmed by their physicians to have type 2 diabetes²⁷.

Statistical analyses:

Participants were follow-up continued until the occurrence of incident CHD, death, loss to follow-up, or the end of follow-up (December 31, 2019), whichever came first (Figure S1). We performed a time to event analysis using Cox proportional hazard models to assess the relation of perceived interpersonal racism to risk of incident CHD. For all analyses, we stratified by age, questionnaire cycle, education, and neighborhood SES. Values for covariates were taken from the 1997 questionnaire, the same time point as the racism questions. In subgroup analyses, we used the Wald test to assess potential effect modifications by age, education, hypertension, BMI, type 2 diabetes, neighborhood SES, and geographic region. All analyses were performed using SAS software, version 9.4 (Cary, North Carolina).

Results

For the present study, the analytic sample consisted of 48,305 BWHS participants who completed the 1997 questions on perceived interpersonal racism and who were free of cardiovascular disease (N excluded=1,196) and cancer (N excluded=1,539) in 1997 (Figure S1). During 22 years of follow-up from 1997 through the end of 2019, we identified 1,947 incident CHD cases among the 48,305 women who answered perceived interpersonal racism questions and were free of cardiovascular disease and cancer at baseline. In 1997, their mean age (standard deviation) was 40.5 (±10.5) years. Women with higher levels of perceived interpersonal racism in everyday life were more likely to smoke, have hypertension, and be more educated, than those with lower levels of perceived racism. Women who reported being treated unfairly due to their race at their job, in housing, and interactions with the police, were more likely to live in a neighborhood with high SES, reside in the West, smoke, and have a higher educational level than those who reported no such occurrences (Table 1).

Table 1.

Age-standardized baseline characteristics in Black Women’s Health Study participants by categories of perceived interpersonal racism in 1997

	Quartiles of perceived interpersonal racism in everyday life
	Q1 (lowest)	Q2	Q3	Q4(highest)
No.	8947	13651	13751	11956
Age (mean, std dev)	40.6 (10.5)	40.6(10.5)	40.5(10.5)	40.5(10.5)
Neighborhood Socioeconomic Status, %
Q1, Lowest quintile	28.2	25.3	25.8	28.0
Q2	18.5	18.7	18.1	17.5
Q3	17.8	18.9	18.9	17.7
Q4	17.8	18.3	18.9	18.2
Q5, highest quintiles	17.7	18.8	18.4	18.6
Geographic region, %
Northeast	27.9	26.8	25.9	26.1
South	32.4	32.5	31.2	30.3
Midwest	20.9	22.3	23.5	24.0
West	18.5	17.9	19.0	19.3
Body Mass Index, kg/cm2 (mean, std dev)	27.4(6.4)	27.6(6.5)	27.9(6.6)	28.6(6.9)
Pack years of smoking (mean, std dev)	3.4(8.3)	3.8(8.6)	4.0(8.9)	4.6(9.9)
Vigorous activity, ≥5 hours/week, %	13.0	13.3	13.3	13.8
Alcohol: Current drinker, %	26.6	27.6	29.0	31.3
Hypertension, %	17.0	17.8	18.3	20.1
Diabetes, %	4.6	5.0	4.7	5.4
Education, %
≤12 years	22.1	17.5	16.3	17.5
13–15 years	36.6	35.4	34.4	36.3
16 years	23.1	25.5	25.3	23.5
≥17 years	18.2	21.7	24.0	22.7
	Categories of perceived interpersonal racism in job, housing, and interactions with the police
	No to all	Yes to 1	Yes to 2	Yes to 3
No. of positive responses	13929	15463	11419	5149
Age (mean, std dev)	40.2(10.4)	40.3(10.4)	40.2(10.4)	40.3(10.3)
Neighborhood Socioeconomic Status, %
Q1, Lowest quintile	28.3	26.2	24.6	26.0
Q2	19.4	18.8	16.9	16.7
Q3	18.0	18.6	18.8	17.9
Q4	18.0	18.2	19.6	19.1
Q5, highest quintile	16.3	18.2	20.1	20.4
Geographic region, %
Northeast	26.0	25.9	27.4	27.5
South	34.2	33.7	29.0	25.0
Midwest	22.3	22.5	23.2	24.4
West	17.2	17.7	20.3	23.1
Body Mass Index, kg/cm2 (mean, std dev)	27.8(6.6)	27.8(6.6)	27.9(6.6)	28.0(6.6)
Pack years (mean, std dev)	3.6(8.4)	3.9(8.9)	4.0(8.9)	4.3(9.4)
Vigorous activity, ≥5 hours/week, %	12.5	13.6	14.0	14.6
Alcohol: Current drinker, %	26.6	28.6	30.7	31.8
Hypertension, %	18.1	18.1	17.6	17.6
Diabetes, %	4.9	5.1	4.5	4.5
Education, %
≤12 years	22.9	16.5	13.7	13.8
13–15 years	35.3	36.2	36.3	36.0
16 years	22.8	25.8	25.8	23.4
≥17 years	19.0	21.5	24.2	26.8

Compared with women in the lowest quartile of perceived interpersonal racism, those in the highest quartile had a higher risk of incident CHD. In age-adjusted models, the hazard ratio (HR) was 1.25 (95%CI: 1.07–1.46, p trend=0.006, Table 2). Multivariable adjustment attenuated the association to 1.10 (95%CI: 0.94–1.29, p trend=0.20). The age-adjusted HR for women who perceived being treated unfairly at their job, in housing, and by the police was 1.35 (95%CI: 1.13–1.61, p trend=0.006), compared to women who reported no such experiences. In the multivariable model, the HR became 1.26 (95%CI: 1.05–1.51, p trend=0.05).

Table 2:

Associations of perceived interpersonal racism in 1997 with incident coronary heart disease

Incident Coronary Heart Disease
Perceived interpersonal racism	N case/Person-years	Model 1 HR (95% CI)	Model 2 HR (95% CI)
Everyday life	N cases=1947
Q1	444/166778	Ref	Ref
Q2	551/262404	1.00 (0.86,1.15)	0.94 (0.81,1.09)
Q3	491/267121	1.02 (0.88, 1.19)	0.97 (0.83, 1.13)
Q4	461/231367	1.25 (1.07, 1.46)	1.10 (0.94, 1.29)
p trend		0.006	0.20
In job, housing, and interactions with the police	N cases=1804
No to all	521/266435	Ref	Ref
Yes to 1	578/298114	1.13 (0.99,1.30)	1.10 (0.96, 1.27)
Yes to 2	458/220970	1.09 (0.94, 1.27)	1.05 (0.90, 1.22)
Yes to 3	247/99314	1.35 (1.13, 1.61)	1.26 (1.05, 1.51)
p trend		0.006	0.05
Incident Myocardial Infarction
	N case/Person-years	Model 1 HR (95% CI)	Model 2 HR (95% CI)
Everyday life	N cases=628
Q1	135/170380	Ref	Ref
Q2	187/266715	1.20 (0.93, 1.56)	1.15 (0.88, 1.49)
Q3	172/271200	1.22 (0.94, 1.59)	1.14 (0.87, 1.48)
Q4	134/235448	1.27 (0.96, 1.69)	1.09 (0.82, 1.46)
p trend		0.11	0.59
In job, housing, and interactions with the police	N cases=583
No to all	157/270415	Ref	Ref
Yes to 1	186/302608	1.09 (0.86, 1.40)	1.06 (0.83, 1.36)
Yes to 2	154/225112	1.13 (0.87, 1.46)	1.07 (0.82, 1.39)
Yes to 3	86/101483	1.43 (1.06, 1.95)	1.34 (0.98, 1.83)
p trend		0.04	0.12

Model 1: stratified for age (continuous), questionnaire cycle (continuous), neighborhood SES (quintiles), and educational level (≤12, 13–15, 16, ≥17 years).

Model 2: further adjusted for body mass index (<25, 25–29.9, 30–34.9, ≥35 kg/m²), geographic region (Northeast, South, Midwest, West), vigorous physical activity (<1hr, 1–4 hrs, ≥5 hrs per week), cigarette smoking (never smoked, current smoker<15 cigs/day, current smoker ≥15 cigs per day, former smoker, quit <10 years ago, former smoker, quit ≥10 years ago), history of diabetes (yes, no), and history of hypertension (yes, no).

Test for proportional hazard assumption.

Incident CHD: P=0.86 for perceived racism in everyday life, p=0.42 for perceived racism in job, housing, and interactions with the police.

Incident MI: P=0.31 for perceived racism in everyday life, p=0.28 for perceived racism in job, housing, and interactions with the police.

Associations were similar but not statistically significant when we assessed MI only (Table 2). Based on a total of 628 incident MIs, the multivariable adjusted HRs were 1.09 (95%CI: 0.82–1.46, p trend=0.59) for the highest quartile vs. the lowest quartile for perceived interpersonal racism in everyday life, and 1.34 (95%CI: 0.98–1.83, p trend=0.12) for perceived interpersonal racism in experiences involving the job, housing, and the police relative to no such experiences

Associations were similar across subgroups of age, BMI, education, neighborhood SES, hypertension, diabetes, smoking, and geographic region (Tables S2 and S3). Analyses that included probable MIs yielded results similar to those based on confirmed MIs. P value for test for proportional hazard assumption for incident CHD was 0.86 for perceived racism in everyday life, and 0.42 for perceived racism in job, housing, and by the police.

Discussion

In this prospective cohort study of 48,305 Black women followed from 1997 until 2019, Black women who reported perceived racism in experiences involving job, housing, and the police had an estimated 26% higher CHD incidence, relative to those who reported none of these experiences. For perceived interpersonal racism in everyday life, there was a small nonsignificant increase in risk of incident CHD for Black women who were in the highest quartile of perceived racism compared to those in the lowest quartile.

Experiences of racism can result in acute⁷ and chronic stress^8,9 among Black individuals. Perceived interpersonal racism, even in the absence of overt discrimination, is a significant stressor⁸. Higher perceived interpersonal racism has been associated with unhealthy behavior and lifestyles^28,29, higher stress^30,31, higher inflammatory markers³², and hormone dysregulation^33,34. In previous studies of hypertension, perceived interpersonal racism has been associated with higher risk of hypertension^17,35–39. With regard to CHD, perceived racism was not associated with coronary artery calcification⁴⁰, carotid intima-media thickness⁴¹, and the Framingham CVD 10-year risk score⁴², but was marginally associated with more carotid plaque⁴³. It was also associated with higher prevalence of self-reported CVD among Black Americans with a history of mood disorder⁴⁴.

Previous studies have largely focused on hypertension as the endpoint, with a number of reports of perceived racism associated with higher risk of hypertension^17,35–39. Several studies have examined the relation between perceived racism with subclinical CVD^40,41,43, prevalent CHD^44–46, and incident stroke²¹, but little is known regarding the longitudinal association between perceived racism with incident CHD. Further, literature suggests that the associations with health outcomes from different types of racism might differ. Participants may interpret experiences of racism as major lifetime discrimination and therefore may discount and underreport their exposure to racism in day to day life^47,48. Black Americans at different levels of socioeconomic status (SES) may experience different types of racist events (e.g. people with low SES might report higher levels of threat/harassment, while individuals with higher SES might report more racism in the work place⁴⁸).

In our study, there were differences in associations with CHD between perceived racism in daily life, and perceived racism at the job, in housing and by the police. The types of perceived racism may be reported with different accuracy by participants. Participants might have a more accurate recall of experiences of racism in job, housing and with the police than experiences of racism in daily life⁴⁸. Being denied at job promotion, mortgage application, or being treated unfairly by the police might be harder to forget and might have greater association with Black women’s social-emotional-physical wellbeing, compared with experiences of racism in daily life such as receiving poor services at a restaurant. In addition, different forms of discrimination may operate through different pathways. Perceived racism in daily life may be associated with a Black women’s CHD risk through influencing her to smoke or her eating habits. Major racism events, such as being denied promotion at work due to perceived racism, may operate through a different mechanism, such as through effects of trauma on the sympathetic nervous system. Perceptions of racism may change over time. Individuals who experiences racism may report these sensitive events differently at different times in adulthood or change their interpretations and perceptions of these experiences over time. Future studies are needed to examine perceived racism through a life course approach with repeated measurements of racism at different time periods of life.

Our study provides longitudinal evidence on the relation between perceived interpersonal racism with incident cardiovascular endpoints, suggesting that Black women with higher levels of perceived interpersonal racism may subsequently have a higher risk of incident CHD.

Strengths of our study include the prospective cohort study design, long-term follow up, and large sample size of Black women from across the U.S. The incident CHD endpoints included were all validated, with the exception of self-reported revasculization events. Because the confirmation rate for reported revasculization events for which we obtained records was high, 90%, we accepted self-reports for which we were unable to obtain medical records as cases. Because the data on perceived racism were collected before the CHD events occurred, biased reporting of racism is unlikely to explain the study results. A validated instrument was used for questions on perceived experiences of interpersonal racism, both in everyday life and in experiences including the job, housing, and the police^15,18–20.

Our study has several limitations. Racism is complex and takes on many forms. Our investigation was limited to perceived interpersonal racism, which is subjective by definition and may reflect different perceptions of levels of actual racism for different individuals. Future studies are needed to examine the associations of structural racism with cardiovascular health endpoints, as well as evaluate the joint associations of perceived racism and structural racism. In our study, we examined perceived racism with only a one-time measurement. Further studies are needed to examine this association with repeated measurements of racism over time. Despite our efforts to control for a comprehensive list of confounders, the study is observational in nature and may still have some unmeasured or residual confounding factors that were not accounted for. Finally, our study consisted of Black women with sufficient literacy to complete health questionnaires every two years. Almost all BWHS participants were recruited from among subscribers to Essence magazine who may have differed from Black women nationally in their health status and other important ways, affecting generalizability of the results. Future studies among Black women with lower levels of education (i.e., <12 years) as well as in a national sample with a wider range of SES are needed. In our analytical sample, most of the women had completed ≥12 years of education. On the other hand, the range of neighborhood SES was wide and there was an association of perceived racism with CHD risk across strata of neighborhood SES.

In conclusion, perceived experiences of racism in job, housing, and by the police were associated with a higher incidence of CHD among Black women, while as higher risk estimated for perceived experiences of racism in daily life was smaller and not statistically significant. Our study suggests that experiences of racism may contribute to the disproportionately high incidence of CHD in Black women.

Clinical perspective.

1. What is new?

   • Perceived interpersonal racism in employment, housing and in interactions with the police was associated with increased incidence of CHD among Black women, while perceived racism in everyday life was not associated with increased risk.
2. What are the clinical implications?

   • Experiences of racism may contribute to the disproportionately high incidence of CHD in Black women.

Abbreviations

BWHS

Black Women’s Health Study

BMI

Body mass index

MI

Myocardial infarction

CHD

Coronary heart disease

CVD

Cardiovascular disease

HR

Hazard ratio

CI

Confidence interval