Hertz RP, Unger AN, Cornell JA, Saunders E. Racial Disparities in Hypertension Prevalence, Awareness, and Management. Arch Intern Med. 2005;165(18):2098-2104. doi:10.1001/archinte.165.18.2098
Copyright 2005 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2005
Effectively reducing cardiovascular disease disparities requires identifying and reducing disparities in risk factors. Improved understanding of hypertension disparities is critical.
Cross-sectional analysis of nationally representative samples of black and white adults 20 years and older who participated in the National Health and Nutrition Examination Survey (NHANES) 1999-2002 (white, n = 4624; black, n = 1837) and NHANES III conducted in 1988-1994 (white, n = 7121; black, n = 4709). We examined differences in hypertension prevalence, awareness, treatment, and blood pressure (BP) control among both treated and prevalent cases across the 2 periods.
Hypertension prevalence increased significantly from 35.8% to 41.4% among blacks and from 24.3% to 28.1% among whites and remains significantly higher among blacks. Awareness is higher among blacks (77.7% vs 70.4%; P<.001), as is treatment (68.2% vs 60.4%; P<.001). These results are driven by higher rates in black women. Blood pressure control rates among those treated have increased in both races, primarily as a result of increased BP control in black and white men (27.3% and 44.7%, respectively; P≤.03). Despite the improved BP control rates, disparity in BP control among treated cases increased, with 59.7% of treated whites and 48.9% of treated blacks now reaching BP goal (P<.001). Racial differences in BP control rates among those treated cannot be explained by nonpharmacologic management or health insurance, but educational attainment is associated with BP control.
The higher prevalence of hypertension in blacks and the growing disparity in BP control among those treated pharmacologically are causes for concern.
It is well established that health status, access to care, and quality of care are not equal among all populations within the United States.1,2 Although cardiovascular mortality is on the decline overall, rates have declined less among blacks than among whites and continue to be higher among blacks.3 Explanatory factors for these racial differences include differences in treatment of heart disease, for example, unequal access to cardiovascular procedures such as coronary arteriography and coronary artery bypass graft surgery.4 These treatment disparities do not totally explain outcome differences, however. It is also necessary to understand and address differences in the prevalence and management of risk factors for heart disease. A focus on racial differences in hypertension, a known major risk factor for cardiovascular disease,5 is warranted.
Hypertension is on the rise in the United States, among the population as a whole, and among black and white adults specifically.6,7 Based on data from the National Health and Nutrition Examination Survey (NHANES) 1999-2000 and NHANES III conducted in 1988-1994, age-adjusted hypertension prevalence increased 3.7 percentage points among US adults overall. Among whites and blacks, increases of 3.1 percentage points and 4.6 percentage points, respectively, were observed.7 Based on NHANES 1999-2000, the age-adjusted prevalence of hypertension is also higher among blacks vs whites (38.8% vs 27.2%).6 Although these studies provide race-specific data on hypertension prevalence, and one also provides race-specific data on hypertension awareness and control,7 the studies did not specifically focus on racial disparities. Our study expands on earlier work by analyzing differences between black and white adults and studying explanatory factors for these differences. Also, our study updates the current literature by incorporating recently released NHANES 2001-2002 data into a larger, more robust data set, NHANES 1999-2002, and by exploring changes from NHANES III to the present among black and white adults. Given that cardiovascular disease is the leading cause of mortality in the United States among both blacks and whites and that the Centers for Disease Control and Prevention’s Healthy People 2010 calls for eliminating health disparities and specifically eliminating disparities in cardiovascular disease,8 increasing our understanding of disparities in hypertension prevalence, awareness, and management is warranted.
We analyzed the subsample of black non-Hispanic and white non-Hispanic adults (20 years and older) participating in NHANES, an ongoing population-based statistical survey to estimate the health status of the noninstitutionalized civilian US population, based on interview, examination, and laboratory information from representative samples of US households. Beginning in 1999, NHANES was converted from a static survey in which data collected over a period of years were not released until survey completion, to a continuous annual survey providing for periodic data releases. Data collected over successive 2-year intervals are released to the public as they become available, together with survey weights for estimating population averages over each 2-year period separately or over the whole period based on cumulative data. We aggregated NHANES data releases for 1999 through 2000 and 2001 through 2002 into a combined data set (NHANES 1999-2002) to increase sample size for greater estimator reliability (see “NHANES Analytic Guidelines,” June 2004, http://www.cdc.gov/nchs/data/nhanes/nhanes_general_guidelines_june_04.pdf). We assessed changes over time in hypertension by comparing data from NHANES 1999-2002 with corresponding data from NHANES III.
Hypertension was defined in accordance with “The Seventh Report of the Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure” (JNC 7).9 A person is deemed to have hypertension if the NHANES examination indicates that systolic blood pressure (BP) is 140 mm Hg or higher or diastolic BP is 90 mm Hg or higher or if he or she reports current use of antihypertensive medication. A hypertensive person is considered “aware” if he or she gives a positive response to the question, “Have you ever been told by a doctor or other health professional that you had hypertension, also called high blood pressure?” A hypertensive person is classified as “treated” if he or she reports currently taking antihypertensive medication. A person is considered “controlled” if systolic BP is lower than 140 mm Hg and diastolic BP is lower than 90 mm Hg.
We applied survey weights to NHANES data to estimate hypertension prevalence, as well as awareness, treatment, and BP control rates among non-Hispanic black and non-Hispanic white adults 20 years and older. Estimates are derived by race for the total population, for men, for women, and for the 3 age groups of 20 to 39 years, 40 to 59 years, and 60 years and older. To allow for comparisons between races and across time, hypertension prevalence rates for the total populations and for men and women separately are age-sex adjusted to the Census 2000 adult population distribution. Total and sex-specific rates of hypertension awareness, treatment, and BP control are age-sex adjusted to the adult hypertensive population as estimated from NHANES 1999-2002. Calculations of these rates as well as mean systolic BP and diastolic BP were performed using SAS version 9.1 (SAS Institute Inc, Cary, NC) statistical software, and estimated standard errors were computed using the procedures SURVEYFREQ and SURVEYMEANS, which take into account the effect on estimator variance attributable to the complex NHANES multistage stratified cluster sample design. The normal approximation to the binomial distribution was used for carrying out significance tests of racial differences and changes over time and to construct 95% confidence intervals around estimates of change. To adjust for confounders in analyzing the effect of race on hypertension control among treated cases, we estimated a multivariable logistic regression model incorporating covariates for age, sex, health insurance status, education level, body mass index category, and self-reported behavioral change to lower BP (controlling diet, exercising more, and reducing salt and alcohol consumption). Pairwise interactions between race and each of the other independent variables were tested for significance. The model was estimated using the SAS SURVEYLOGISTIC procedure, which takes the NHANES survey design into account in the estimation of model coefficients and their variances.
Table 1 provides the demographic characteristics of black and white adults with hypertension in the United States. Hypertension occurs earlier among blacks. Based on NHANES 1999-2002, 62.8% of black adults and 44.7% of white adults with hypertension are younger than 60 years. Blacks are less likely to have health insurance and less likely to have graduated from high school.
From NHANES III (1988-1994) to NHANES 1999-2002, the prevalence of hypertension has significantly increased in both black and white adults in the United States and currently affects 41.4% of blacks and 28.1% of whites (Table 2). These current rates reflect increases of 15.7% within each group (5.6 percentage points among blacks and 3.8 percentage points among whites) since NHANES III. The greatest increase in prevalence is observed among white women (22.2%), with statistically significant increases also occurring in black women and black men. Prevalence increased significantly among middle-aged blacks, and both blacks and whites 60 years and older. Based on the most recent NHANES survey, hypertension prevalence in blacks in each age group exceeds prevalence in whites, and each comparison is statistically significant. Prevalence ranges from 13.3% to 81.0% among black adults by age group and from 7.3% to 65.4% among white adults by age group.
Table 3 indicates that 77.7% of blacks and 70.4% of whites with hypertension are aware of their diagnosis (P<.001). The difference in awareness between black and white men is not statistically significant (70.0% vs 66.6%, respectively); however, the 84.0% awareness rate among black women significantly exceeds the 73.5% awareness rate among white women. Over time, awareness rates have increased significantly among blacks 40 years and older (P = .05), and among whites 60 years and older (P = .02).
Treatment rates have increased significantly among both blacks and whites overall, black and white men, and black women (Table 4). Currently, treatment rates are higher among blacks than among whites (68.2% and 60.4%, respectively; P<.001), a difference that is driven by the higher treatment rate among black women compared with white women (75.8% vs 64.0%; P<.001). Treatment rates increase by age among blacks, from 40.3% to 66.3% to 77.4%. Among whites, treatment rates are 29.6%, 62.5%, and 65.3%, respectively, across the 3 age groups (20-39 years, 40-59 years, and ≥60 years). Over time, increases in treatment rates are observed in all but the youngest black and white adults.
Since NHANES III, BP control rates among those treated pharmacologically have increased significantly in both black and white adults, although progress in goal attainment among whites exceeds that among blacks (Table 5). The rate of BP control among the treated population increased 17.6% in blacks, from 41.6% to 48.9% (P = .02), compared with 24.2% in whites, from 48.0% to 59.7% (P<.001). The improvement in both racial groups has been driven by progress in BP goal attainment in men; however, white men are still more likely than black men to reach BP goal, and white women are more likely than black women (64.8% vs 51.6% [P = .005] and 55.4% vs 46.7% [P = .01], respectively). The disparity in BP control among treated cases is particularly evident in the 40- to 59-year age group, in which only 50.1% of blacks compared with 70.5% of whites achieve BP goal (P<.001).
Viewed with respect to the total hypertensive population, the rate of BP control (Table 5, “BP control among prevalent cases”) has also increased significantly for both races since NHANES III, by 37.9% for blacks and 36.3% for whites (P<.001). This results from the effect of compounding significant improvements over time in the treatment rate and the rate of BP control among treated cases. There is no significant overall racial disparity in this measure of BP control; however, significant disparities exist for black vs white men (29.9% vs 35.8%; P = .04) and among the middle aged (32.4% vs 43.4%; P = .002)
When treatment is successful, there is no racial difference in mean systolic BP or diastolic BP, as shown in Table 6. Among those treated but not controlled, mean BPs are higher among blacks than among whites in each age group.
Table 7 presents a multivariate analysis of predictors of failure to reach BP goal among those treated with pharmacotherapy. After adjustment for other variables, blacks treated for hypertension were more likely than whites not to reach BP goal (odds ratio [OR], 1.43; 95% confidence interval [CI], 1.16-1.77); women were more likely than men to fail treatment (OR, 1.47; 95% CI, 1.13-1.90); and persons with less than a high school education were more likely to be uncontrolled than those with at least a high school education (OR, 1.45; 95% CI, 1.04-2.04). Behavioral modification in response to physician’s advice (ie, control diet, exercise more, reduce salt, and reduce alcohol consumption) did not have a statistically significant association with BP control rate. Although lack of health insurance is associated with lack of BP control, it does not reach statistical significance (OR, 1.8; P = .10). However, the estimated effect may be attenuated in a model across all ages because most older persons (who account for a large fraction of the hypertensive population and hence the sample) have insurance coverage through Medicare. To investigate this hypothesis, we estimated another model restricted to ages 20 to 59 years and found the effect of “no insurance” to be strengthened and statistically significant (OR, 2.4; P = .006). Nevertheless, a significant racial disparity in BP control rate among persons aged 20 to 59 years persists after controlling for differences in health insurance coverage as well as the other covariates, with blacks being over twice as likely as whites to fail to reach BP goal (OR, 2.2; P<.001).
Compared with whites, blacks are more likely to have hypertension, more likely to be aware of it, and more likely to be pharmacologically treated, but less likely to achieve BP control while receiving treatment. Although treatment rates for both races have increased significantly from 1988-1994 to 1999-2002, and the effectiveness of treatment as measured by BP control rates has improved for both races, blacks continue to lag behind whites in achieving BP goal, and the disparity has grown wider over time. This finding calls into question the 2003 National Healthcare Disparities Report, which concluded that black patients do not experience any disparity in BP management.2
While awareness and treatment are necessary first steps in BP management, successful management is measured by the extent to which BP control is achieved. Our analysis of BP control rates among the total hypertensive population indicates no statistically significant BP control disparity overall, but significantly lower BP control rates among black men and blacks aged 40 to 59 years. Clearly, pharmacologic success is possible but may be more difficult among blacks: blacks controlled on pharmacotherapy have on average the same systolic and diastolic BP levels as whites, whereas those uncontrolled have significantly higher levels compared with whites.
It is noteworthy that improvements in BP control among those treated in both races are due to improvements in BP control among men. Women who are treated for hypertension are no more likely now to reach goal than they were in 1988-1994, and they are less likely to reach goal compared with men. Given that, compared with men, there are more hypertensive women in the United States, women are more likely to be aware of their diagnosis, and women are more likely to be treated, the failure of women to achieve BP control deserves attention.
Although lack of insurance has been suggested as a reason for racial disparity in hypertension control,10- 12 our study and others have not found this to be explanatory.13- 15 We found, however, that lack of insurance is an independent risk factor for BP control in both races among young and middle-aged persons. Other possible explanatory factors for racial disparity are patient behaviors, some of which can be studied using NHANES data. Inclusion of behavioral modification related to diet and exercise in the multivariate analysis did not explain BP control; however, failure to find an association may be related to the way in which these questions were asked. Subjects were asked if they were “told by a doctor” to control weight, reduce salt, exercise more, or reduce alcohol consumption. Only those who answered affirmatively were asked whether they were currently following this advice. Thus, there may have been misclassification of persons engaged in behavioral modification.
Health literacy may also influence patient behavior.16 Patients may have difficulty reading labels on pill bottles, reading and understanding educational brochures, or comprehending appointment slips. In addition to difficulty understanding written materials, they also may have difficulty understanding oral communication and conceptualizing risk. In our analysis, we found significant differences in educational attainment among black and white adults with hypertension, and in our multivariate analysis of persons treated for hypertension, we found that not being a high school graduate was an independent risk factor for poor BP control. Although we were unable to assess health literacy, low educational attainment may be a proxy for literacy.
Another limitation of this study was the inability to assess physician variables, including practice patterns, physician race, and the role that physician/patient racial discordance may have on patient outcomes. It is also likely that we overestimated the BP control rates in both racial groups, and underestimated the disparity in goal attainment between the 2 groups. By defining uncontrolled hypertension as a BP of 140/90 mm Hg or higher, we applied the same goal to those with and without diabetes, an approach taken by previous studies.6,7,14 Applying a hypertension control rate of a BP lower than 130/85 mm Hg to adults with diabetes would yield a slightly lower estimate of BP control in the United States.7 Given the higher prevalence of diabetes in blacks,17 using the standard threshold of a BP of 140/90 or higher is likely to result in a disproportionate overestimate of BP control in blacks.
The importance of eliminating racial disparities in cardiovascular disease morbidity and mortality through secondary prevention including hypertension control is unambiguous. Primary prevention of risk factors for cardiovascular disease is equally important.9,11,18,19 Our analysis updates the literature by adding NHANES 2001-2002 data and comparing NHANES III with NHANES 1999-2002 data. The disproportionate prevalence of hypertension among blacks has been well established,6,7 and this analysis, using the latest nationally representative data, supports the conclusion that prevalence is growing in both races and continues to be higher in blacks. A public health strategy that includes weight reduction, dietary changes including reduction in sodium consumption, increased physical activity, and decreased alcohol consumption is warranted.9,19 Eliminating disparities through primary prevention of cardiovascular risk factors, and secondary prevention of cardiovascular disease, should be the ideal.
Correspondence: Robin P. Hertz, PhD, US Outcomes Research–Population Studies, Pfizer Global Pharmaceuticals, 235 E 42nd St, New York, NY 10017 (firstname.lastname@example.org).
Accepted for Publication: May 23, 2005.
Financial Disclosure: None.
Funding/Support: This study was funded by Pfizer Inc, New York, NY.
Disclaimer: Drs Hertz and Unger had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.
Acknowledgment: We thank Laura Charleville for help in preparing this manuscript and Michael B. Lustik, MS, for his review and comments.