HEALTH, BEHAVIOR, AND HEALTH CARE DISPARITIES: DISENTANGLING THE EFFECTS OF INCOME AND RACE IN THE UNITED STATES

Transcription

1 Inequalities in Health in the U.S. HEALTH, BEHAVIOR, AND HEALTH CARE DISPARITIES: DISENTANGLING THE EFFECTS OF INCOME AND RACE IN THE UNITED STATES Lisa C. Dubay and Lydie A. Lebrun The literature on health disparities in the United States typically focuses on race/ethnicity or on socioeconomic status (SES) separately, but not often together. The purpose of the study was to assess the separate effects of race/ethnicity and SES on health status, health behaviors, and health care utilization. Cross-sectional analyses were conducted using the 2008 National Health Interview Survey (n = 17,337 non-elderly adults). SES disparities within specific racial groups were examined, as were race disparities within high and low SES groups. Within each racial/ethnic group, a greater proportion of low versus high SES individuals were in poor health, a lower proportion had healthy behaviors, and a lower proportion had access to care. In both SES groups, blacks and s had poorer health outcomes than whites. While whites were more likely to exercise than blacks and s, they are more likely to be smokers and less likely to have no or moderate alcohol consumption. Blacks had similar or better health care use than whites, especially for cancer screening; s had lower use within each SES group. Race/ethnicity disparities among adults of similar incomes, while important, were dwarfed by the disparities identified between high- and low-income populations within each racial/ethnic group. Health and health care disparities research has received a great deal of attention in the United States in recent years, with a particular focus on black and populations. Conclusions from this body of literature indicate that minority individuals have poorer health outcomes, fewer protective health International Journal of Health Services, Volume 42, Number 4, Pages , , Baywood Publishing Co., Inc. doi:

2 608 / Dubay and Lebrun behaviors, and worse access to and quality of health care, compared to white individuals (1 7). Consequently, health promotion programs have prioritized the improvement of health behaviors and health care among minority populations (8 10). Disparities based on socioeconomic status (SES) have also been widely documented. Similarly, conclusions from these efforts indicate that people with lower income and lower levels of education have worse health behaviors, outcomes, and health care than people with higher income and higher education (11 14). Most of the literature focuses separately on race-based disparities or on SES-based disparities, but not both together, making it difficult to disentangle the distinct effects of race/ethnicity and SES on health and health care (15). In this article, we attempt to fill this gap in the literature on disparities in the United States. National health policy in the United States tends to focus on reducing racial and ethnic disparities, with less attention paid to disparities based on SES, despite clear evidence that SES and race each have separate effects on health (16 19). This policy focus occurs for both political and empirical reasons. On the political side, there is willingness as a nation to attempt to repair the historic repercussions of racism and discrimination against minority populations. In contrast, disparities caused by SES are often perceived to result from lack of individual agency in a market-based society and to be unworthy of policy interventions. On the empirical side, there is a much greater prevalence of information on race relative to SES and social class in national health data, making it difficult to comprehensively examine these disparities (16, 20 22). Nonetheless, when we do not consider SES and race disparities together, we risk making inaccurate conclusions about the nature of health disparities and creating ineffective policies to address them (23, 24). Having a thorough understanding of these interrelationships is critical for informing successful policies to address health disparities in the United States (25). The primary objective of this study was to disentangle the effects of race/ ethnicity and SES (i.e., income) across a broad range of health and health care outcomes. To do this, we stratified our analysis by income levels (i.e., lowvs. high-income) and examined the extent to which race/ethnicity disparities exist within income groups; we also examined income disparities within racial/ ethnic groups. Only by understanding these relationships can appropriate policies to reduce disparities be developed. To the extent that race/ethnicity disparities are small or non-existent within income groups, but income disparities are large within racial/ethnic groups, the current policy focus is likely to be misguided. Rather, an additional focus on eliminating SES disparities for all racial/ethnic groups would be warranted. Such a focus would not only reduce disparities for racial/ethnic minorities who are disproportionately represented in low SES groups, but also improve health outcomes for low-income whites.

3 METHODS Data Sources Data from the 2008 National Health Interview Survey (NHIS) were analyzed. The NHIS is conducted annually by the National Center for Health Statistics and contains nationally representative information about demographic characteristics, health status, health-related behaviors, and health care access and utilization for the U.S. civilian, non-institutionalized population. The response rate for the NHIS was 62.6% (26). Analyses were limited to non-elderly adults, ages 18 to 64 years (n = 17,337). Study Variables Income, Race, and Disparities in U.S. / 609 Outcomes were analyzed in the following areas: health status (self-reported general health, overweight/obesity), health-related behaviors (physical activity, tobacco use, alcohol consumption), general health care-seeking behaviors (regular source of care, physician consultation, dentist visit), and screening procedures (colorectal screening, Pap test, mammogram, blood pressure test, cholesterol check). For general health status, survey participants were asked to rate their health as excellent, very good, good, fair, or poor; we recoded responses into two categories (poor/fair vs. excellent/very good/good). We also created a dichotomous variable to indicate respondents who were overweight or obese (Body Mass Index 25). Healthy People 2010 and the U.S. Centers for Disease Control and Prevention were consulted to determine guidelines for recommended weekly physical activity (27, 28). Individuals who reported engaging in vigorous physical activity at least three times per week for at least 20 minutes each time, as well as individuals who reported engaging in moderate physical activity at least five times per week for at least 30 minutes each time, were coded as meeting exercise recommendations. A dichotomous variable was created to indicate whether individuals had never smoked (as opposed to current or former smokers). Survey participants also reported on alcohol consumption habits, and their responses were categorized into lifetime alcohol abstainer, or current moderate/infrequent/light drinker versus current heavier drinker or former drinker. For health care utilization, individuals reported whether they had seen or talked to any general doctor in the past 12 months, whether they had a regular source of care when they got sick, and whether they had any dentist visits in the past 12 months. Whether respondents received timely cancer screenings for colorectal cancer, cervical cancer, and breast cancer, according to national recommended cancer screening guidelines, was also examined (27, 29). For colorectal cancer, a dichotomous variable to indicate whether individuals age 50 to 64 had received either a colonoscopy, sigmoidoscopy, or proctoscopy in the past two years or a

4 610 / Dubay and Lebrun fecal occult home blood stool test in the past two years was created. For cervical cancer, a dichotomous variable to indicate whether women age 18 to 64 had received a Pap test in the past three years was created. For breast cancer, we created a variable to indicate whether women age 40 to 64 had received a mammogram in the past two years. Variables were also constructed to indicate whether individuals reported receiving a blood pressure test in the past two years and a cholesterol check in the past five years, according to national goals and recommendations (27). The primary independent variables of interest were income and race/ethnicity. Participants reported household income, and responses were recoded into low-income (< 200% of the federal poverty level) or high-income ( 200% of the federal poverty level), while taking into account family size in calculating the ratios of income to poverty threshold. Participants self-reported racial and ethnic identity and were classified in accordance with the federal standards released by the Office of Management and Budget (30). The racial/ethnic groups consisted of whites (not or Latino), blacks or African-Americans (not or Latino), s or Latinos, and other racial/ethnic groups (composed primarily of Asians, Native Hawaiians, or other Pacific Islanders, but also including American Indians or Alaska Natives). Results for the other racial group are not presented here because of limited sample sizes. Other independent variables included respondents sex, age, marital status, geographic region of residence, insurance status, and education. Survey details and specific questionnaire wording can be obtained from the website of the National Center for Health Statistics (31). Statistical Analysis Crude proportions for all health status, health behavior, and health care utilization indicators were calculated for each racial/ethnic and income group. Bivariate analysis was used to test for differences between low- and highincome groups by race/ethnicity and for differences between race/ethnicity groups within both low- and high-income populations. Finally, multiple logistic regression was used to determine the independent effect of race/ethnicity on each outcome, after adjusting for education, insurance status, sex, age, marital status, and geographic region of residence. Multivariate models were estimated separately for low-income populations and high-income populations. Odds ratios and 95% confidence intervals for logistic models were calculated using Stata software (version 10) (32). To obtain appropriate standard errors and estimates that were nationally representative, all analyses accounted for the complex sampling design of the survey and sampling weights.

5 RESULTS Table 1 presents results for the analysis of health status and health behaviors. The first two columns display descriptive statistics for low-income and highincome adults within each racial/ethnic group. The third column presents the corresponding unadjusted odds ratios for income disparities in health within racial/ethnic groups. The final two columns present unadjusted odds ratios for these outcomes across race/ethnicity groups, for low-income populations and high-income populations, respectively. Table 2 provides adjusted odds ratios indicating the associations between race/ethnicity, health status, and health behaviors, within low-income populations and high-income populations. Table 3 presents descriptive and unadjusted results for health care-seeking and health screening behaviors, and Table 4 presents adjusted analyses for associations between race/ethnicity and health care-seeking and screening by income level. Health Status Within each race/ethnicity group, low-income adults had significantly higher odds of reporting that they were in fair or poor health relative to those with higher incomes (p ). Differences in the rates of poor health between highand low-income groups were large (e.g., 23% of low-income whites vs. 6.5% of high-income whites). Among low-income populations, 23% of whites reported being in fair or poor health compared with 23.1% of blacks and 16.9% of s. Relative to whites, s had lower odds of reporting fair or poor health (p 0.001). These results remained statistically significant after multivariate adjustment. High-income blacks had higher odds of reporting being in fair or poor health than whites, but these differences disappeared after adjustment for other factors. Low-income whites were slightly less likely to be overweight or obese than high-income whites (p 0.05); however, no other income-based differences were found for this outcome among the other race/ethnicity groups. Among low-income adults, 60.2% of whites, 67.8% of blacks, and 72.1% of s were overweight or obese; rates were similar for high-income adults. For both low- and high-income populations, unadjusted and adjusted odds ratios indicated that relative to whites, blacks and s had greater odds of being overweight or obese (p ). Health Behaviors Income, Race, and Disparities in U.S. / 611 For each racial/ethnic group, low-income adults had lower odds of engaging in physical exercise that met recommendations, relative to their high-income

6 612 / Dubay and Lebrun Table 1 Health status and behaviors by income level and race/ethnicity among adults under age 65, 2008 Unadjusted OR % Lowincome % Highincome Income disparities within race/ethnicity a Race disparities within low-income (< 200% FPL) HEALTH STATUS Poor/fair health status Non- White Non- Black Overweight or obesity (BMI 25) Non- White Non- Black **** 3.23**** 2.49**** 0.80* *** 1.39**** 1.71**** Race disparities within high-income ( 200% FPL) 1.34* **** 1.50****

7 Income, Race, and Disparities in U.S. / 613 HEALTH BEHAVIORS Exercise meets recommendations Non- White Non- Black Never smoked Non- White Non- Black Lifetime alcohol abstainer or moderate/infrequent/light drinker Non- White Non- Black Source: 2008 National Health Interview Survey a Reference group is high-income. OR: Odds Ratio, FPL: Federal Poverty Level *p 0.05, **p 0.01, ***p 0.001, ****p **** 0.56**** 0.58**** 0.61**** 0.66**** **** **** 0.67**** 1.88**** 3.07**** 1.44** 2.04**** 0.69**** 0.66**** 1.74**** 1.70**** **

8 614 / Dubay and Lebrun Table 2 Adjusted associations between race/ethnicity and health status and behaviors, by income level, 2008 a Adjusted OR Low-income (< 200% FPL) High-income ( 200% FPL) HEALTH STATUS Poor/fair health status Non- White Non- Black Overweight or obesity (BMI 25) Non- White Non- Black **** 1.49**** 1.75**** **** 1.48**** HEALTH BEHAVIORS Exercise meets recommendations Non- White Non- Black Never smoked Non- White Non- Black Lifetime alcohol abstainer or moderate/infrequent/light drinker Non- White Non- Black 0.73** **** 3.30**** 1.60**** 1.87**** 0.76** 0.74*** 1.98**** 1.93**** ** Source: 2008 National Health Interview Survey a Adjusted for sex, age, marital status, geographic region of residence, education, and insurance status. OR: Odds Ratio, FPL: Federal Poverty Level *p 0.05, **p 0.01, ***p 0.001, ****p

9 counterparts (p ). Low-income whites and blacks also had lower odds of never being smokers compared with high-income whites and blacks (p ), and low-income whites had lower odds of having limited alcohol consumption than high-income whites (p ). These differences were particularly large for exercise and smoking outcomes: High-income individuals were about 10 percentage points more likely to meet exercise recommendations and to never have smoked relative to low-income adults. Within each income group, health behavior patterns varied across race/ ethnicity, with whites showing more evidence of health-promoting behaviors for exercise and minorities showing more health-promoting behaviors for smoking and alcohol outcomes. Only 28.5% of low-income whites met the recommended exercise levels, but even fewer blacks and s met the recommendations (21% for both). Among high-income whites, 40.7% met the exercise recommendations compared with 32.2% of blacks and 31.3% of s. Corresponding odds ratios favoring whites at both income levels were significant in both the unadjusted and adjusted models (p 0.01), except that disparities between low-income s and whites became nonsignificant in the adjusted model. With respect to both smoking and drinking, there were differences by minority status within each income group as well as differences across different behaviors. About 45% of low-income whites had never smoked compared to 60.6% of blacks and 71.5% of s. Similar patterns of never smoking were found for high-income groups: 57.1% for whites, 69.9% for blacks, and 69.4% for s. Corresponding odds ratios were significant in both the unadjusted and adjusted models (p ). Rates of relatively low alcohol consumption within the low-income population were 74.2% for whites, 80.5% for blacks, and 85.5% for s. Compared with low-income whites, low-income blacks and s had higher odds of consuming relatively low amounts of alcohol in both the adjusted and unadjusted models. Among high-income adults, only s had significantly higher adjusted odds of limited alcohol consumption relative to whites. General Health Care-Seeking Income, Race, and Disparities in U.S. / 615 Within each race/ethnicity group, low-income adults had lower odds of having a regular source of care, a consultation, or a visit with a doctor or a dentist in the past 12 months, compared with their high-income counterparts (p ). These differences were greatest for dental care, where high-income adults were at least 20 percentage points more likely to have had a dental visit than low-income adults. The vast majority of low-income whites and blacks, about 75%, had a usual source of care, but fewer s, 59.4%, had one. In unadjusted and adjusted

10 616 / Dubay and Lebrun Table 3 General health care-seeking and screening practices by income level and race/ethnicity among adults under age 65, 2008 Unadjusted OR % Lowincome % Highincome Income disparities within race/ethnicity a Race disparities within low-income (< 200% FPL) Race disparities within high-income ( 200% FPL) GENERAL HEALTH CARE-SEEKING Regular source of care when sick Non- White Non- Black Doctor consult, past 12 months Non- White Non- Black Dentist visit, past 12 months Non- White Non- Black **** 0.50**** 0.44**** 0.66**** 0.70**** 0.55**** 0.32**** 0.43**** 0.43**** **** **** **** 0.78* 0.44**** **** 0.65**** 0.51****

11 Income, Race, and Disparities in U.S. / 617 SCREENING PRACTICES Colorectal screening or home blood stool test, past 2 years Non- White Non- Black Pap test, past 3 years Non- White Non- Black Mammogram, past 2 years Non- White Non- Black Blood pressure test, past 2 years Non- White Non- Black Cholesterol check, past 5 years Non- White Non- Black Source: 2008 National Health Interview Survey a Reference group is high-income. OR: Odds Ratio, FPL: Federal Poverty Level *p 0.05, **p 0.01, ***p 0.001, ****p **** * 0.44**** 0.57** 0.57**** 0.35**** 0.46**** 0.47**** 0.56**** 0.49**** 0.44**** 0.35**** 0.48**** 0.46**** * 1.50** * **** 1.53**** * **** **

12 618 / Dubay and Lebrun Table 4 Adjusted associations between race/ethnicity and general health care-seeking and screening practices, by income level, 2008 a Adjusted OR Low-income (< 200% FPL) High-income ( 200% FPL) GENERAL HEALTH CARE-SEEKING Regular source of care when sick Non- White Non- Black Doctor consult, past 12 months Non- White Non- Black Dentist visit, past 12 months Non- White Non- Black SCREENING PRACTICES *** ** ** 0.77** Colorectal screening or home blood stool test, past 2 years Non- White Non- Black Pap test, past 3 years Non- White Non- Black Mammogram, past 2 years Non- White Non- Black Blood pressure test, past 2 years Non- White Non- Black **** ** **** ** **

13 Income, Race, and Disparities in U.S. / 619 Table 4 (Cont d.) Adjusted OR Low-income (< 200% FPL) High-income ( 200% FPL) SCREENING PRACTICES (Cont d.) Cholesterol check, past 5 years Non- White Non- Black 1.94**** 1.75**** 1.56**** 1.46**** Source: 2008 National Health Interview Survey a Adjusted for sex, age, marital status, geographic region of residence, education, and insurance status. OR: Odds Ratio, FPL: Federal Poverty Level *p 0.05, **p 0.01, ***p 0.001, ****p models, s had lower odds of having a usual source of care relative to whites (p 0.001). About 85% of high-income whites and blacks had a usual source of care, compared with only 77.1% of s. High-income blacks and s had lower odds of having a usual source of care relative to whites in unadjusted analyses, but not in adjusted analyses. Within low-income populations, blacks and whites showed similar rates of having at least one doctor s visit or consultation in the past 12 months, 60.4% and 58.1%, respectively, while adults had rates below 50%. These differences were significant in both the adjusted and unadjusted models (p 0.01). Similar patterns existed for higher-income adults, except that differences were not as large and there were no racial/ethnic differences in the adjusted models. Among low-income adults, whites were significantly more likely to have had a dental visit in the past year, 44%, compared with s, 34.9%. While these differences were significant in the unadjusted models, they were not significantly different in the adjusted models. For high-income adults, similar patterns were found, with more whites reporting a dental visit in the past year compared to their minority counterparts. In this population, however, the corresponding odds ratios were statistically significant in the adjusted models. Screening Practices Within each race/ethnicity group, low-income adults had consistently lower rates of screening for various types of cancers and other health problems,

14 620 / Dubay and Lebrun compared to their high-income counterparts. With the exception of blood pressure screening, these income-based differences were large, typically ranging from 10 to 20 percentage points depending on the screening practice and race/ethnicity. White and black low-income adults over 50 years of age had relatively low rates of colorectal screening in the past two years, 23.3% and 28.6%, respectively, but s had even lower rates of 15.8%. The differences between lowincome whites and s were statistically significant in the unadjusted models, but there were no racial/ethnic differences in the adjusted models. About one-third of white and black high-income adults received colorectal screening, which was significantly higher than rates for s, 25.3%, in unadjusted but not adjusted models. In both the high- and low-income groups, black women were significantly more likely than white women to have had a Pap test in the past three years. While 70.3% of low-income and 84.2% of high-income white women received such a screening, comparable rates for black women were 78.1% and 86.3%. For low-income groups, both unadjusted and adjusted models indicated that the odds of receiving a mammogram in the past two years were higher for blacks relative to whites. Among high-income women, there were no significant racial/ethnic differences. About 90% of low-income whites and blacks, and 95% of high-income whites and blacks, had a blood pressure test in the past two years. Low-income s had significantly lower odds of having a blood pressure test compared with low-income whites, in both the unadjusted and adjusted models, with only 74.2% of s having had such a test (p ). High-income s also had lower odds of receiving blood pressure screening than whites in both the unadjusted and adjusted models (p ). About 55% of low-income whites had their cholesterol checked in the past five years, compared with 65.1% of low-income blacks and 55.4% of low-income s. Rates were about 15 to 25 percentage points higher for each race/ ethnicity group in the high-income population. In the adjusted models, both low- and high-income blacks and s had significantly greater odds of receiving a cholesterol test relative to their white counterparts (p ). DISCUSSION Clear evidence of race/ethnicity disparities was found in this analysis across a wide range of health status, health behavior, health care use, and health screening outcomes, within both low- and high-income groups. In general, whites had better

15 Income, Race, and Disparities in U.S. / 621 outcomes than racial/ethnic minorities with two key exceptions. Minority groups reported better behaviors with respect to smoking and drinking. In addition, blacks appeared to have comparable or better outcomes than whites with respect to use of most health care screening services. Over the past two decades, race disparities have been on the national health policy agenda, beginning with the goal of reducing race disparities in health under Healthy People 2000 and continuing with former President Bill Clinton s call for the elimination of such disparities under Healthy People In addition, during this period, both the Office of Minority Health at the Department of Health and Human Services and the National Center for Minority Health and Health Disparities at the National Institutes of Health were established. The targeted research, policy, and education initiatives that grew out of this focus likely can be credited for the relatively similar access to care currently enjoyed by blacks and whites within comparable income groups. The research presented here also indicates that more targeted policies are needed to understand and eliminate disparities between whites, s, and other races. While important, the race/ethnicity disparities among adults of similar incomes identified in this article were dwarfed by the magnitude of the disparities identified between high- and low-income groups for those of the same race. For each outcome examined, except overweight and obesity, high-income adults were found to be in better health, to engage in healthier behaviors, to have greater use of general health and dental services, and to receive more timely screening for cancer and other health conditions, compared to low-income adults of the same race/ethnicity. Results from this analysis are consistent with the small body of evidence that examines race disparities within the context of income and class disparities (15 19, 33, 34). These studies have identified race disparities in health status, activities of daily living, and mortality when individuals are compared within the same socioeconomic group measured by their income, education, or class. These studies also find disparities between socioeconomic groups among adults of the same race/ethnicity that are considerably larger than those based on race, providing consistent evidence of the independent effects of both race/ethnicity and SES or social class on health and access to health care. The focus in the United States on eliminating race/ethnicity disparities in health and health care is both long-standing and laudable. Racism, discrimination, and residential segregation have well-documented links to poor health outcomes for blacks and are considered by many to be the fundamental cause of race disparities in health (15, 16, 22, 35, 36). These forces also limit educational attainment and job opportunities for blacks and, more recently,

16 622 / Dubay and Lebrun other minorities, placing them at greater risk of having low incomes (22, 33, 35, 37 39). In this article, we present descriptive evidence regarding the extent and magnitude of race and income disparities rather than attempt to identify the factors that cause such differences. To the extent that racial minorities are disproportionately low-income, the elimination of race disparities will not be possible unless income disparities are also addressed. It is hoped that the stark differences we present will focus policy attention on this issue. Low-income adults are at greater risk of being uninsured and exposed to environments that are detrimental to their health than their higher-income counterparts. The Patient Protection and Affordable Care Act recently enacted will certainly reduce income disparities in insurance coverage and access to care, but it will not eliminate them nor will it necessarily change the patterns of care associated with low income. Numerous other factors place low-income adults at greater risk for poor health outcomes, beyond access to health care. Addressing income disparities in the nature of care provided, availability of healthy foods, safe places to exercise, toxin-free environments to live and work, and less stressful work and living conditions will also be necessary to produce a more equitable distribution of health. Current federal efforts targeted at reducing disparities are short-sighted from the perspective of population health. As long as societal factors result in minorities disproportionately having low incomes, an exclusive focus on reducing race disparities will be ineffective. The economic, environmental, and social factors that put low-income populations at much greater risk for poor health outcomes than their higher-income counterparts, regardless of race, must be addressed to eliminate both race-based and income-based disparities in population health. Acknowledgments Both authors contributed to study design, data analyses and interpretation, and manuscript preparation. The data used in this study are publicly available from the U.S. Centers for Disease Control and Prevention, National Center for Health Statistics; we take responsibility for the accuracy of the data analysis. We have no conflicts of interest, financial or otherwise, to disclose. Some of the results included in this article were previously presented at the 2008 annual meeting of the American Public Health Association and the 2009 annual meeting of Academy Health. We thank Thomas A. LaVeist, Ph.D., professor in health policy and Director of the Center for Health Disparities Solutions at the Johns Hopkins University Bloomberg School of Public Health, for providing guidance in data analysis and thoughtful comments on draft versions of this article.

17 Income, Race, and Disparities in U.S. / 623 REFERENCES 1. Smedley, B., Stith, A., and Nelson, A. (eds.). Unequal Treatment: Confronting Racial and Ethnic Disparities in Health Care. Institute of Medicine, Washington, DC, Kurian, A., and Cardarelli, K. Racial and ethnic differences in cardiovascular disease risk factors: A systematic review. Ethn. Dis. 17: , Agency for Healthcare Research and Quality National Healthcare Disparities Report. Department of Health and Human Services, Rockville, MD, Long, J., et al. Update on the health disparities literature. Ann. Intern. Med. 141: , Mayberry, R. M., Mili, F., and Ofili, E. Racial and ethnic differences in access to medical care. Med. Care Res. Rev. 57: , Manton, K., Patrick, C., and Johnson, K. Health differentials between blacks and whites: Recent trends in mortality and morbidity. Milbank Q. 65(Suppl 1): , Wong, M., et al. Contributions of major diseases to disparities in mortality. N. Engl. J. Med. 347: , Adams, E., Breen, N., and Joski, P. Impact of the National Breast and Cervical Cancer Early Detection Program on mammography and Pap test utilization among white,, and African American women: Cancer 109: , Giles, H., et al. Racial and ethnic approaches to community health (REACH) An overview. Ethn. Dis. 14:S15 S18, Fitzpatrick, L., Sutton, M., and Greenberg, A. Toward eliminating health disparities in HIV/AIDS: The importance of the minority investigator in addressing scientific gaps in Black and Latino communities. J. Natl. Med. Assoc. 98: , Fiscella, K., and Williams, D. Health disparities based on socioeconomic inequities: Implications for urban health care. Acad. Med. 79(12): , Adler, N., and Newman, K. Socioeconomic disparities in health: Pathways and policies. Health Aff. 21:60 76, Lantz, P., et al. Socioeconomic disparities in health change in a longitudinal study of U.S. adults: The role of health-risk behaviors. Soc. Sci. Med. 53:29 40, Feinstein, J. The relationship between socioeconomic status and health: A review of the literature. Milbank Q. 71: , LaVeist, T. A. Disentangling race and socioeconomic status: A key to understanding health inequalities. J. Urban Health 82(2 Suppl 3):iii26 34, Kawachi, I., Daniels, N., and Robinson, D. E. Health disparities by race and class: Why both matter. Health Aff. 24(2): , Muntaner, C., Hadden, W. C., and Kravets, N. Social class, race/ethnicity and allcause mortality in the U.S.: Longitudinal results from the National Health Interview Survey. Eur. J. Epidemiol. 19(8): , Navarro, V. Race or class versus race and class: Mortality differentials in the United States. The Lancet 336(8725): , 1990.

18 624 / Dubay and Lebrun 19. Navarro, V. Race or class or race and class: Growing mortality differentials in the United States. Int. J. Health Serv. 21(2): , National Center for Education Statistics. Digest of Education Statistics, U.S. Government Printing Office, Washington, DC, National Center for Health Statistics. Health, United States, U.S. Government Printing Office, Hyattsville, MD, Isaacs, S. L., and Schroeder, S. A. Class: The ignored determinant of the nation s health. N. Engl. J. Med. 351(11): , Schulman, K., et al. The roles of race and socioeconomic factors in health services research. Health Serv. Res. 30: , Smith, G. Learning to live with complexity: Ethnicity, socioeconomic position, and health in Britain and the United States. Am. J. Public Health 91: , Nickens, H. The role of race/ethnicity and social class in minority health status. Health Serv. Res. 30: , Division of Health Interview Statistics National Health Interview Survey (NHIS) Description. ftp://ftp.cdc.gov/pub/health_statistics/nchs/dataset_documentation/nhis/2008/srvydesc.pdf (accessed April 19, 2010). 27. U.S. Department of Health and Human Services. Healthy People U.S. Government Printing Office, Washington, DC, U.S. Centers for Disease Control and Prevention. Physical Activity for Everyone. (accessed January 14, 2009). 29. U.S. Preventive Services Task Force. Guide to Clinical Preventive Services. Agency for Healthcare Research and Quality, Rockville, MD, Office of Budget and Management. Directive No. 15: Race and Ethnic Standards for Federal Statistics and Administrative Reporting. Washington, DC, National Center for Health Statistics. National Health Interview Survey (NHIS): Questionnaires, datasets, and related documentation nchs/about/major/nhis/quest_data_related_1997_forward.htm (accessed January 14, 2009). 32. Stata statistical software: Release 10.0 [computer program]. StataCorp, College Station, TX, Williams, D. R. Race, socioeconomic status, and health. The added effects of racism and discrimination. Ann. N.Y. Acad. Sci. 896: , Williams, D. R., and Jackson, P. B. Social sources of racial disparities in health. Health Aff. 24(2): , Williams, D. R., and Collins, C. Racial residential segregation: A fundamental cause of racial disparities in health. Public Health Rep. 116(5): , LaVeist, T. A. Segregation, poverty, and empowerment: Health consequences for African Americans. Milbank Q. 71(1):41 64, LaVeist T. A., Nickerson, K. J., and Bowie, J. V. Attitudes about racism, medical mistrust, and satisfaction with care among African American and white cardiac patients. Med. Care Res. Rev. 57(Suppl 1): , LaVeist, T. A., Rolley, N. C., and Diala, C. Prevalence and patterns of discrimination among U.S. health care consumers. Int. J. Health Serv. 33(2): , 2003.

19 Income, Race, and Disparities in U.S. / LaVeist, T. A., Sellers, R., and Neighbors, H. W. Perceived racism and self and system blame attribution: Consequences for longevity. Ethn. Dis. 11(4): , Direct reprint requests to: Lisa C. Dubay, Ph.D. Senior Fellow The Urban Institute 2100 M Street NW Washington, DC ldubay@jhsph.edu