Objectives. Using Jackson Heart Study data, we examined whether perceived discrimination was associated with prevalent hypertension in African Americans.
Methods. Everyday discrimination, lifetime discrimination, burden of discrimination, and stress from discrimination were examined among 4939 participants aged 35 to 84 years (women = 3123; men = 1816). We estimated prevalence ratios of hypertension by discrimination, and adjusted for age, gender, socioeconomic status, and risk factors.
Results. The prevalence of hypertension was 64.0% in women and 59.7% in men. After adjustment for age, gender, and socioeconomic status, lifetime discrimination and burden of discrimination were associated with greater hypertension prevalence (prevalence ratios for highest vs lowest quartile were 1.08 [95% confidence interval (CI) = 1.02, 1.15] and 1.09 [95% CI = 1.02,1.16] for lifetime discrimination and burden of discrimination, respectively). Associations were slightly weakened after adjustment for body mass index and behavioral factors. No associations were observed for everyday discrimination.
Conclusions. Further understanding the role of perceived discrimination in the etiology of hypertension may be beneficial in eliminating hypertension disparities.
Previous research reported an association between perceived discrimination and health outcomes.1,2 In particular, hypertension (which is higher among African Americans than Whites) was linked to discrimination,3–6 although findings were not always consistent.4,7–10 Although traditional biobehavioral risk factors might explain some of the African American–White disparity in hypertension, differences in exposure to discrimination by race might also contribute. African Americans’ exposure to discrimination could influence their risk for hypertension through various mechanisms. These include negative coping behaviors, such as unhealthy eating, sedentary lifestyles, and tobacco and alcohol intake. The experience of discrimination could also cause emotional distress,3 which can trigger physiological responses involving the hypothalamic–pituitary–adrenal axis and the sympathetic–parasympathetic systems, which play an important role in the pathophysiology of hypertension.7,11,12
Although exceptions exist, most previous studies of perceived discrimination and hypertension in African Americans employed relatively small samples13,14 and focused on a single measure of discrimination.3,5 The Jackson Heart Study (JHS), the largest prospective study of cardiovascular disease in African Americans, offered a unique opportunity to examine the association between multiple dimensions of discrimination (everyday, lifetime, burden, and stress from discrimination) and hypertension in a large sample of African Americans. The measures of various dimensions of discrimination obtained in JHS included everyday discrimination, which captured daily hassles associated with discrimination,15 lifetime discrimination, which captured acute and observable experiences similar to life events in the stress literature, and perceived burden and stress from discrimination.4 Because discrimination was conceptualized as a multifactorial construct, the inclusion of multiple measures might be necessary to accurately capture the impact of discrimination on health.2,16–18
We hypothesized that perceived discrimination would be positively associated with hypertension, and that health behaviors would partially mediate the association. Because previous work documented associations of discrimination with hypertension among women, but not men,5 and because of differences in hypertension prevalence by gender, we tested for effect modification by gender.
Between 2000 and 2004, 5301 African Americans (1941 men, 3360 women) were recruited from Hinds, Madison, and Rankin counties in the Jackson, Mississippi, metropolitan area. Participants came from 4 recruitment pools: community random sample (17%), volunteer sample (22%), the Atherosclerosis Risk in Communities Study (30%), and family members (31%). Recruitment was limited to noninstitutionalized African Americans aged 35 to 84 years, except in the family sample, in which those aged 21 to 34 years were eligible.19 Details of the design and recruitment were published elsewhere.20–22 The study was approved by the institutional review boards of University of Mississippi Medical Center, Jackson State University, and Tougaloo College. All participants provided informed consent.
The JHS discrimination instrument (JHSDIS) was administered by trained African American interviewers during the visit 1 clinic examination after blood draws and blood pressure measurement; 96% of participants completed the JHSDIS. Questions were asked to identify experiences with and reactions to perceived everyday and lifetime discrimination. Everyday discrimination was based on the scale by Williams et al.23 and had good internal consistency (α = 0.88).18 Participants were asked questions such as, “How often on a day-to-day basis do you have the following experiences? You are treated with less courtesy. You are treated with less respect. You receive poorer service than others at restaurants.” A total of 9 items were asked, and responses ranged from 1 (“never”) to 7 (“several times a day”). The mean of the 9 responses was used as the score for frequency of everyday discrimination.
Lifetime discrimination was adapted from the work of Krieger and Sidney3 and Krieger24 and had an internal consistency (α = 0.78)18 comparable with that observed for major life-event scales.13 Participants were asked about the occurrence of unfair treatment over their lifetime across 9 domains (yes/no): at school, getting a job, at work, getting housing, getting resources or money, getting medical care, on the street or public place, and getting services. The count of the domains (range = 0–9) was the lifetime discrimination score.
Burden of lifetime discrimination was measured by the following: “When you had experiences like these over your lifetime, have they been—very stressful, moderately stressful, or not stressful?” (coded 4, 2.5, and 1, respectively). “Overall, how much has discrimination interfered with you having a full and productive life,” and “Overall, how much harder has your life been because of discrimination? With response options, a lot, some, a little, or not at all?” (coded 4–1). Responses for the 3 items were reverse coded and summed to create a mean score ranging from 1 to 4, where higher scores indicated greater burden. Internal consistency of this scale was 0.63,18 suggesting a fair amount of heterogeneity among the items.13,25 Because previous research suggested that the degree of stress was a key determinant of the adverse health consequences of discrimination,2 the “stress” component of the “burden” index was examined separately (hereafter referred to as stress from discrimination). Burden and stress analyses were restricted to persons who reported at least 1 experience of lifetime discrimination.
Everyday, lifetime, and burden of discrimination were divided into quartiles to investigate whether there were threshold effects. In additional analyses, everyday, lifetime, and burden scores were examined as continuous variables after transforming them into SDs.
Participants were also asked to indicate the main reason for everyday and lifetime discrimination (age, gender, race, height or weight, or other). In secondary analyses, these responses were combined with everyday discrimination scores to create 5 categories: no discrimination, low discrimination (below the median) attributed to race, high discrimination (at or above the median) attributed to race, low discrimination attributed to nonracial factors, and high discrimination attributed to nonracial factors.
low discrimination (below the median) attributed to race,
high discrimination (at or above the median) attributed to race,
low discrimination attributed to nonracial factors, and
high discrimination attributed to nonracial factors.
Identical categories were created for lifetime and burden of discrimination.
Trained staff obtained seated blood pressure measurements with a Hawksley random-0 sphygmomanometer (Hawksley & Sons Ltd., Lansing, Sussex, UK) in the right arm of seated participants after a 5-minute rest. The average of 2 measures taken 1 minute apart was used in these analyses. Hypertension was defined as systolic blood pressure greater than or equal to 140 millimeters mercury, diastolic blood pressure greater than or equal to 90 millimeters mercury, or being on antihypertensive medications.10 Medications taken in the past 2 weeks were brought to the examination and transcribed verbatim, with subsequent medication coding by a pharmacist using the Medispan Dictionary.26
Education, income, and occupation were used to characterize socioeconomic status (SES). Education was measured as years of schooling completed, and included 4 categories: less than high school; high school graduate or general equivalency diploma; some college (1–3 years), vocational school, or associate degree; and college graduate or higher (≥ 4 years). Persons were assigned to 1 of 4 income categories (poor, lower-middle, upper-middle, and affluent) based on family size, US Census poverty levels, and year of baseline clinic visit (2000–2004). Income ranges for the categories were: poor: less than poverty level; lower-middle: 1 to 1.5 times the poverty level; upper-middle: more than 1.5 but less than 3.5 times the poverty level; and affluent: 3.5 or more times the poverty level. Occupation was coded according to the 2000 US standard occupational codes27 and classified as managerial or professional; service, sales and office; farming, fishing, and forestry; construction, extraction, and maintenance; and production, transportation, and material moving.
Additional covariates included age, gender, and possible behavioral or anthropometric mediators, including body mass index (BMI, defined as weight in kilograms divided by the square of height in meters), physical activity, cigarette smoking, alcohol consumption and diet. Physical activity was measured as the sum of the 4 index scores (active living, work or occupational, home life, and sport) from the JHS physical activity instrument.28 Cigarette smoking was classified as current, former, and never smokers. Alcohol consumption was classified as persons who abstained, had 1 to 7 drinks per week, 8 to 14 drinks per week, or more than 14 drinks per week. Diet was assessed using the Delta Nutrition Intervention Research Initiative, a short-form food frequency questionnaire.29 Selected dietary measures related to components of the Dietary Approaches to Stop Hypertension trial diet30 (percent calories from fat, sodium [mg/1000 kcal], potassium [mg/1000 kcal], calcium [mg/kcal], and fiber [g/1000 kcal]) were examined.
Of the 5301 participants who completed the examination, 362 were excluded because of missing data on hypertension (n = 59), all discrimination measures (n = 283), or education (n = 20). Unknown income and occupation data were coded as a separate category so participants could be retained in analyses, leaving 4939 participants. Regression models for each measure of discrimination were restricted to 4924, 4919, 4302, and 4337 participants for everyday, lifetime, burden, and stress from discrimination, respectively.
We examined the distribution of key variables by gender and discrimination and tested the differences using χ2 tests or t tests. We also examined age-adjusted differences in discrimination and risk factors between hypertensives and normotensives by gender. We estimated associations of discrimination with hypertension before and after adjustment for age, gender, SES, and risk factors. We also performed gender-stratified analyses because hypertension prevalence varies by gender,10 and because the impact of discrimination on hypertension was reported to vary by gender.5 However, because there were no interactions between gender and discrimination (all P > .05), only pooled results are given. Because the prevalence of hypertension was high in our cohort, Poisson regression31 was used to estimate prevalence ratios (PRs) of hypertension by discrimination before and after adjustment for covariates. Trend tests were conducted by including categories of discrimination as ordinal variables. Similar analyses were conducted using discrimination measures as continuous variables. We also tested for interactions between discrimination and age and SES. All tests were 2-tailed, and a P value < 0.05 was considered statistically significant. All statistical analyses were performed using SAS (version 9.1; SAS Institute Inc, Cary, NC).
Table 1 presents baseline characteristics of the sample. Women made up 63.4% of the sample. Hypertension prevalence was higher among women than men (64% vs 60%, respectively). Although men and women had similar levels of education, more men had affluent income, and more women were in management occupations. Men reported slightly higher levels of everyday, lifetime, and burden of discrimination, but differences were small. Women were more likely to report that lifetime discrimination made their lives very stressful. Among those reporting at least some discrimination, 43% of women and 55% of men attributed everyday discrimination primarily to race, and 57% of women and 70% of men attributed lifetime discrimination primarily to race.
Baseline Characteristics by Gender in Perceived Discrimination and Hypertension: The Jackson Heart Study, Jackson, MS, 2000-2004
|Women (n = 3123), % or Mean (SD)||Men (n = 1816), % or Mean (SD)||Pa|
|Age, y||55.1 (12.7)||53.9 (12.8)||.001|
|< high school||17.4||18.7|
|High school grad||20.2||19.3|
|≥ bachelor's degree||33.1||32.7|
|Everyday (n = 4924)c||2.1 (1.0)||2.2 (1.0)||< .001|
|Lifetime (n = 4919)d||2.9 (2.1)||3.3 (3.1)||< .001|
|Burden (n = 4328)ef||2.3 (0.8)||2.4 (0.8)||.066|
|Stress from discriminationf|
|Not stressful||20.1||24.5||< .001|
|Attribution of everyday discrimination||< .001|
|Attribution of lifetime discrimination||< .001|
|Attribution of burden||< .001|
Note. The sample size was (n = 4939).
aP values based on χ2 and t tests.
bIncome categories were based on family size, US Census poverty levels, and year of baseline clinic visit (2000–2004): poor < poverty level; lower-middle = 1–1.5 times the poverty level; upper-middle > 1.5 but < 3.5 times the poverty level; and affluent ≥ 3.5 times the poverty level.
cMedian (25th, and 75th percentiles): women: 1.8 (1.3, 2.6); men: 2.0 (2.0, 2.8).
dMedian (25th, and 75th percentiles): women: 3.0 (1.0, 4.0); men: 3.0 (2.0, 5.0).
eMedian (25th, and 75th percentiles): women: 2.2 (1.7, 2.8); men: 2.3 (1.8, 2.8).
fBurden and stress from discrimination are restricted to persons who reported at least 1 instance of lifetime discrimination.
Table 2 shows selected covariates by categories of discrimination. The average age decreased from the lowest to highest quartile for everyday and lifetime discrimination, which was consistent with previous research.7 Higher levels of lifetime discrimination and higher levels of SES were positively associated, but patterns were not as clear for the other 2 measures. Current smoking was positively associated with reports of everyday discrimination and burden from discrimination. Greater everyday discrimination was associated with higher BMI, and everyday and lifetime discrimination were associated with more physical activity. Dietary fiber and sodium intake were associated with greater everyday discrimination and burden from discrimination.
Selected Covariates by Categories of Discrimination in Perceived Discrimination and Hypertension: The Jackson Heart Study, Jackson, MS, 2000–2004
|Q1(n = 1163)||Q2(n = 1192)||Q3(n = 1305)||Q4(n = 1264)||P for Trendd||Q1(n = 1288)||Q2(n = 832)||Q3(n = 1576)||Q4(n = 1223)||P for Trendd||Q1(n = 1074)||Q2(n = 1064)||Q3(n = 1222)||Q4(n = 968)||P for Trendd|
|Demographic and SES factors|
|Age, y, mean||59.7||55.3||53.4||50.7||< .001||57.2||54.7||54.1||52.8||< .001||53.5||52.6||54.4||56.3||< .001|
|Female, %||64.0||68.5||61.1||60.0||.021||66.1||66.2||63.8||57.6||< .001||64.5||65.8||60.5||61.7||.045|
|≥ bachelor's degree, %||26.3||37.2||37.2||30.1||.153||22.1||28.4||35.3||44.0||< .001||26.9||35.6||37.9||33.6||.047|
|Income > $50 000, %||20.0||28.2||29.2||22.8||.184||18.0||22.3||27.8||31.3||< .001||20.4||28.8||28.0||25.6||.078|
|Professional, %||29.9||40.1||40.3||32.3||.419||24.6||32.9||38.4||46.0||< .001||29.4||41.2||39.6||35.8||.227|
|Current smoker, %||11.6||11.1||12.9||15.8||.006||13.0||12.1||13.5||12.6||.802||12.5||10.5||14.0||14.8||.02|
|> 14 drinks/wk, %||2.2||1.3||2.5||2.5||.205||2.3||1.9||2.5||1.7||.446||2.5||1.8||1.9||2.1||.517|
|Physical activity, mean||8.1||8.4||8.6||8.4||< .001||8.0||8.2||8.5||8.6||< .001||8.4||8.4||8.5||8.4||.467|
|Total dietary fiber, g, mean||15.9||14.9||16.3||16.8||< .001||16.1||15.4||15.9||16.5||.215||15.6||15.8||16.3||16.4||.009|
|Sodium, mg, mean||4019||3865||4167||4469||< .001||4135||4019||4147||4212||.277||4081||4097||4161||4305||.024|
Note. BMI = body mass index (defined as weight in kg divided by the square of height in m); Q = Quartile; SES = socioeconomic status. All variables except age are adjusted to the age distribution of the full sample. The sample size was (n = 4939).
aEveryday discrimination is the mean of 9 items scored 1–7. Ranges for Everyday Score Q1: 1.0–1.2; Q2: 1.3–1.7; Q3: 1.8–2.5; Q4: 2.6–7.0.
bLifetime discrimination is the count of 9 items. Ranges for Lifetime Score Q1: 0–1.0; Q2: 2.0–2.0; Q3: 3.0–4.0; Q4:5.0–9.0.
cBurden is the mean of 3 items scored 1–4. Ranges for Burden Score Q1: 1.0–1.6; Q2: 1.8–2.1; Q3: 2.3–2.8; Q4: 3.0–4.0.
dP values for trend across discrimination categories.
Table 3 displays covariates by hypertension status. Being hypertensive was associated with lower SES in both women and men, although associations were weaker and not always statistically significant in men. For women and men, each measure of discrimination was higher for hypertensives than normotensives, although the differences were small and not statistically significant. Hypertensives had a higher BMI and lower level of physical activity in both genders.
Selected Characteristics of Study Participants by Hypertension Status and Gender in Perceived Discrimination and Hypertension: The Jackson Heart Study, Jackson, MS, 2000–2004
|Hypertension (n = 2000)||No Hypertension (n = 1123)||Pa||Hypertension (n = 1084)||No Hypertension (n = 732)||Pa|
|Age, y, mean||59.3||47.8||< .001||57.7||48.3||< .001|
|≥ bachelor's degree, %||30.4||38.0||< .001||31.6||33.4||.47|
|Income > $50 000, %||19.0||25.4||.002||30.6||33.7||.2|
|Professional, %||35.5||44.0||< .001||30.0||32.5||.26|
|Stress from discriminationb||1.1||1.0||.12||0.98||0.91||.06|
|Current smoker, %||10.3||9.5||.52||18.4||17.2||.55|
|> 14 drinks/wk, %||0.38||0.72||.23||5.4||4.3||.28|
|BMI, mean||34.1||30.7||< .001||31.0||28.4||< .001|
|Physical activity, mean||8.1||8.4||.005||8.5||8.9||.009|
|Total dietary fiber, g, mean||14.7||15.1||.18||17.9||18.2||.43|
|Calcium, mg, mean||839||859||.34||916||864||.05|
|Sodium, mg, mean||3771||3882||.2||4619||4800||.16|
|Potassium, mg, mean||2588||2576||.83||3113||3096||.84|
|Calories from fat, %||34.8||35.1||.41||34.9||35.1||.51|
Note. BMI = body mass index (defined as weight in kg divided by the square of height in m). All variables except age are adjusted to the age distribution of the full sample. The sample size was (n = 4939).
aP values via χ2 tests or t tests.
bBurden and stress from discrimination are restricted to persons who reported at least 1 instance of lifetime discrimination.
Table 4 displays the prevalence ratios of hypertension by dimensions of discrimination. Everyday discrimination was not associated with hypertension, and no gradients in hypertension across categories were observed. Greater lifetime discrimination was associated with a slightly higher prevalence of hypertension in a graded fashion after adjustment for age, gender, and SES (P for trend = .006; PR for highest vs lowest quartile = 1.08; 95% confidence interval [CI] = 1.02, 1.15). The prevalence of hypertension increased by 4% for each 1 SD increase in lifetime discrimination (PR = 1.04; 95% CI = 1.01, 1.06). This pattern was slightly weakened, but persisted after adjustment for behavioral risk factors. Gender-stratified analyses (data not shown) suggested that age- and SES-adjusted associations were stronger in women (PR for highest vs lowest quartile = 1.10; 95% CI = 1.00, 1.20) than in men (PR for highest vs lowest quartile = 1.04; 95% CI = 0.94, 1.10), but tests for heterogeneity by gender were not statistically significant.
Prevalence Ratios of Hypertension by Categories of Discrimination in Perceived Discrimination and Hypertension: The Jackson Heart Study, Jackson, MS, 2000–2004
|Age, Gender Adjusted, Ratio (95% CI)||Age, Gender, SES Adjusted,a Ratio (95% CI)||Age, Gender, SES, Risk Factors Adjusted,bc Ratio (95% CI)|
|Q2||0.99 (0.93, 1.04)||0.99 (0.94, 1.05)||0.97 (0.92, 1.03)|
|Q3||0.99 (0.94, 1.04)||1.00 (0.94, 1.05)||0.98 (0.93, 1.04)|
|Q4||1.00 (0.95, 1.06)||1.00 (0.94, 1.06)||0.98 (0.92, 1.04)|
|SD units||1.01 (0.99, 1.03)||1.01 (0.99, 1.03)||1.01 (0.98, 1.03)|
|Q2||1.00 (0.94, 1.06)||1.00 (0.94, 1.07)||1.00 (0.94, 1.06)|
|Q3||1.03 (0.98, 1.08)||1.04 (0.99, 1.10)||1.02 (0.97, 1.08)|
|Q4||1.06 (1.00, 1.12)||1.08 (1.02, 1.15)||1.05 (0.99, 1.12)|
|SD units||1.03* (1.01, 1.05)||1.04** (1.01, 1.06)||1.03* (1.00, 1.05)|
|Q2||1.04 (0.98, 1.11)||1.06 (0.99, 1.13)||1.04 (0.98, 1.11)|
|Q3||1.04 (0.98, 1.11)||1.06 (0.99, 1.12)||1.04 (0.97, 1.10)|
|Q4||1.08 (1.01, 1.15)||1.09 (1.02, 1.16)||1.06 (1.00, 1.13)|
|SD units||1.02* (1.00, 1.04)||1.02* (1.00, 1.05)||1.02 (0.99, 1.04)|
|Not stressful (Ref)||1.00||1.00||1.00|
|Moderately stressful||0.99 (0.93, 1.04)||0.99 (0.94, 1.05)||0.99 (0.94, 1.05)|
|Very stressful||1.08 (1.02, 1.15)||1.08 (1.02, 1.15)||1.06 (1.00, 1.13)|
|Attribution of lifetime|
|No discrimination (Ref)||1.00||1.00||1.00|
|Race/low||1.02 (0.95, 1.09)||1.02 (0.95, 1.10)||1.01 (0.94, 1.09)|
|Race/high||1.05 (0.99, 1.12)||1.07 (1.01, 1.14)||1.04 (0.97, 1.11)|
|Nonracial/low||1.02 (0.95, 1.10)||1.02 (0.95, 1.10)||1.00 (0.92, 1.08)|
|Nonracial/high||1.07 (0.99, 1.15)||1.08 (1.00, 1.16)||1.04 (0.96, 1.12)|
|Attribution of burden|
|Race/high||1.07 (1.01, 1.13)||1.08 (1.02, 1.14)||1.06 (1.00, 1.12)|
|Nonracial/low||1.06 (0.99, 1.13)||1.04 (0.98, 1.12)||1.03 (0.97, 1.11)|
|Nonracial/high||1.04 (0.97, 1.12)||1.03 (0.96, 1.10)||1.01 (0.94, 1.08)|
Note. CI = confidence interval; Q = Quartile; SES = socioeconomic status. The sample size was (n = 4939).
aSES: education—< high school, high school graduate, some college, and ≥ bachelor's degree; income categories were based on family size, US Census poverty levels, and year of baseline clinic visit (2000-2004): poor < poverty level; lower-middle = 1-1.5 times the poverty level; upper-middle > 1.5 but < 3.5 times the poverty level; and affluent ≥ 3.5 times the poverty level; occupation—transportation/other, production, construction/extract, farming, sale/office, service.
bRisk factors: body mass index, physical activity, cigarette smoking, alcohol consumption, and diet.
cAnalysis was restricted to participants who had no missing risk factors (n = 4523, 4516, 3976, and 3985 for everyday, lifetime, burden, and stress, respectively).
dBurden and stress from discrimination are restricted to persons who reported at least 1 instance of lifetime discrimination.
*P ≤ .05; **P ≤ .01
Greater burden from discrimination (among those who reported at least 1 instance of lifetime discrimination) was also associated with greater prevalence of hypertension after adjustment for age, gender, and SES (P for trend = 0.01; PR for highest vs lowest quartile = 1.09; 95% CI = 1.02, 1.16). Each 1 SD increase in burden from discrimination was associated with a 2% greater prevalence of hypertension (PR = 1.02; 95% CI = 1.00, 1.05). This pattern was slightly weakened and was no longer statistically significant after adjustment for behavioral risk factors. There was no evidence of heterogeneity of effects by gender, although stratified analyses (data not shown) suggested slightly weaker associations in women than in men (PR for highest vs lowest quartile = 1.05; 95% CI = 0.97, 1.10 in women, and 1.13;95% CI = 1.00, 1.30 in men). Similar results were observed when stress from discrimination was examined in place of the overall burden score.
Attribution analyses showed that higher levels of lifetime discrimination were similarly associated with hypertension regardless of whether discrimination was attributed to racial or nonracial factors (age-, gender-, and SES-adjusted PR 1.07; 95% CI = 1.01, 1.14 for racial attribution and 1.08; 95% CI = 1.00, 1.16 for nonracial attribution). In the case of burden of discrimination, however, only high burden attributed to racial factors was associated with a higher prevalence of hypertension (age-, gender-, and SES-adjusted PR = 1.08; 95% CI = 1.02, 1.14 for high burden attributed to race, and 1.03; 95% CI = 0.96, 1.10 for high burden attributed to nonracial factors).
Previous studies of discrimination and hypertension have often been limited to a single dimension of discrimination.5,32 This study examined the association of hypertension with multiple dimensions of discrimination (everyday, lifetime, burden, and stress) in a large sample of African Americans. Both greater lifetime discrimination and greater burden of discrimination were associated with greater hypertension prevalence. Being in the highest quartile of lifetime discrimination or burden was associated with an 8% to 9% higher hypertension prevalence in age-, gender-, and SES-adjusted models. These associations were slightly reduced after adjustment for BMI and behavioral risk factors, suggesting that these factors could play a mediating role. In the case of lifetime discrimination, the association was similar, regardless of whether discrimination was attributed to racial or nonracial factors. However, in the case of burden, the association between high burden of discrimination and hypertension appeared to be largely restricted to those who attributed the discrimination to race.
Previous studies that investigated associations of lifetime discrimination with hypertension have not always had consistent findings. One study of discrimination and hypertension reported complex interactions between lifetime discrimination, social class, and responses to discrimination. Among professional African Americans, systolic blood pressure was 9 to 10 millimeters mercury lower among those reporting that they challenged unfair treatment and had not experienced lifetime racial discrimination.3 In contrast, among working class African Americans, systolic blood pressure was higher among those who reported that they typically accepted unfair treatment, but did not report experiencing discrimination compared with those who typically challenged unfair treatment but did report experiences of discrimination.3 Two other studies with African American samples reported no associations between lifetime discrimination and blood pressure,4,33 but sample sizes in these studies were small (< 400). A large study of African American women found no associations between lifetime discrimination (as assessed by a subset of the items we used) and incident hypertension.34 We added to the existing work by documenting associations between lifetime discrimination and hypertension prevalence in a large sample of African Americans living in the US South.
The examination of burden and stress from discrimination was a novel feature of our study. We found that higher burden from discrimination was associated with greater prevalence of hypertension. Because research suggested that the stressfulness of life events is important for predicting health outcomes,2,11,35,36 we examined the stress-related item of the burden scale separately and found similar associations. At least 1 other study also found that stress from discrimination was associated with hypertension in African Americans, even when the number of domains for lifetime discrimination was not reported.4 Our results suggested that not only the experience of lifetime discrimination, but also the burden and stress derived from the experience, might be relevant to the health consequences of discrimination.
In contrast to results for lifetime discrimination and burden, we found no associations of everyday discrimination with hypertension. Previous studies of everyday discrimination and hypertension yielded mixed results.5,37–39 Everyday discrimination was found to be associated with diastolic blood pressure reactivity, but was unrelated to baseline blood pressure in a small sample of Black women.37 A larger study found that greater everyday discrimination was associated with higher diastolic blood pressure in older African Americans (n = 2800) but not in older Whites.38 A study of 1110 African Americans in Pitt County, NC, found no association between everyday discrimination and hypertension in men or women, although everyday discrimination attributed to nonracial factors was associated with greater odds of hypertension in women.5 Analyses based on the Black Women's Health Study found no evidence that everyday discrimination was related to incident hypertension.34 Brown et al.39 also found no associations of everyday discrimination with systolic or diastolic blood pressure in a large sample from the Study of Women's Health Across the Nation. It was plausible that everyday experiences of unfair treatment, which are designed to capture long-term current levels of minor instances of discrimination, would be more related to short-term changes in blood pressure than to high blood pressure, which is likely to develop over the life course. By contrast, lifetime measures attempted to capture the extent of lifetime exposure to discriminatory experiences in multiple domains of life.
Another novel aspect of our study was the examination of whether the factors to which participants attributed their unfair treatment modified the association between discrimination and hypertension. For lifetime discrimination, associations were similar regardless of attribution. However, burden analyses suggested that high burden from discrimination might be especially unhealthy when it is attributed to racial factors. Few studies investigated differential effects depending on the attribution. One small study found that the associations of everyday discrimination with diastolic blood pressure reactivity appeared stronger when the mistreatment was attributed to racial or ethnic discrimination.37 Roberts et al.5 found that everyday discrimination attributed to nonracial factors was associated with hypertension prevalence in African American women, whereas racial discrimination was not. Our findings of possible greater effect of burden from discrimination attributed to racial factors suggested that more work is needed to understand the psychological and physiological consequences of discrimination, especially when it is attributable to racial factors.
The JHS was restricted to a single site, which limits generalizability beyond the southeastern region. Given the sampling strategy, JHS was not a representative sample of the African American population. The location of the sample in the US South might have affected the validity of the self-reported measures as well as their implications for health. Our research design was cross-sectional, which limited our ability to draw causal inferences, although it was unlikely that hypertension caused discrimination. As in all observational studies, residual confounding was a possibility for the weak associations we observed. There was some evidence that coping strategies might modify the health consequences of discrimination.1,3 This topic deserves further examination. If exposure to discrimination was ubiquitous in this study, our ability to detect associations might have been limited. As suggested by a recent review, clinic measures of hypertension might be insufficient to detect associations with discrimination,6 and ambulatory blood pressure measurements might be necessary, especially to detect the impact of recent exposures linked to everyday experiences.40,41
The JHS was the largest study of cardiovascular disease in African Americans. The JHSDIS instrument enabled us to analyze the association of hypertension with multiple dimensions of discrimination and to contrast associations when discrimination was attributed to racial or nonracial factors. The study of discrimination and health was challenging. Perceptions of discrimination might be strongly affected by contextual beliefs about fairness, resulting in differential validity of measures as proxies for true health-related discrimination exposures. The aspects of discrimination that affect health might not be fully understood, resulting in crude and imprecise measurement. Finally, there was likely to be a complex causal chain between exposure to discrimination and hypertension, with intervening factors making the detection of effects difficult in observational studies with limited measurement. Institutional discrimination, which can shape exposures to a range of socioenvironmental factors related to blood pressure, could have a greater effect than interpersonal discrimination.6
Identifying the health consequences of discrimination remains an important public health concern, as it could explain health disparities within and between race/ethnic groups. Cross-sectional studies like ours were a reasonable first approximation. However, different study designs, including longitudinal studies and ones with improved measurement of discrimination that capture the timing of exposure and the proximal and short-term behavioral and physiological consequences of experiences of discrimination as well as the moderating effects of coping strategies, might be needed to better determine if and how interpersonal discrimination affects blood pressure.
The Jackson Heart Study (JHS) was supported by the National Institutes of Health (contracts N01-HC-95170, N01-HC-95171, and N01-HC-95172, provided by the National Heart, Lung, and Blood Institute and the National Center for Minority Health and Health Disparities). This research was also supported by the National Heart, Lung, and Blood Institute (award number K01HL08468-04 to M. Sims and award number 1 K01 HL88735-03 to M. A. Bruce) and the Michigan Center for Integrative Approaches to Health Disparities (award number P60MD002249, funded by the National Center on Minority Health and Health Disparities, to A. V. Diez-Roux).
We thank the participants and staff at the JHS.
Human Participation Protection
All study procedures and protocols were approved by the institutional review boards of University of Mississippi Medical Center, Jackson State University, and Tougaloo College. All participants provided informed consent.