Summarize the attached article

RIVERS OF DISPARITIES By Thomas LaVeist, Keshia Pollack, Roland Thorpe Jr., Ruth Fesahazion, and Darrell Gaskin DOI: 10.1377/hithaff.2011.0640 HEALTH AFFAIRS 30, NO. 10 (2011): 1880-1887 @2011 Project HOPE- Place, Not Race: Disparities The People-to-People Health Foundation, Inc. Dissipate In Southwest Baltimore When Blacks And Whites Live Under Similar Conditions Thomas LaVeist (tlaveist@ jhsph.edu) is director of the ABSTRACT Much of the current health disparities literature fails to Center for Health Disparities Solutions and the William C. account for the fact that the nation is largely segregated, leaving racial and Nancy F. Richardson groups exposed to different health risks and with variable access to Professor in Health Policy at the Johns Hopkins Bloomberg health services based on where they live. We sought to determine if racial School of Public Health, in health disparities typically reported in national studies remain the same Baltimore, Maryland. when black and white Americans live in integrated settings. Focusing on Keshia Pollack is an assistant a racially integrated, low-income neighborhood of Southwest Baltimore, professor in the Department of Health Policy and Maryland, we found that nationally reported disparities in hypertension, Management, Johns Hopkins diabetes, obesity among women, and use of health services either Bloomberg School. vanished or substantially narrowed. The sole exception was smoking: We Roland Thorpe Jr. is a found that white residents were more likely than black residents to research scientist in the Department of Health Policy smoke, underscoring the higher rates of ill health in whites in the and Management, Johns Hopkins Bloomberg School, Baltimore sample than seen in national data. As a result, we concluded and a faculty associate at the that racial differences in social environments explain a meaningful Center for Health Disparities Solutions. portion of disparities typically found in national data. We further concluded that when social factors are equalized, racial disparities are Ruth Fesahazion is a doctoral candidate in the Department minimized. Policies aimed solely at health behavior change, biological of Health Policy and differences among racial groups, or increased access to health care are Management, Johns Hopkins Bloomberg School. limited in their ability to close racial disparities in health. Such policies must address the differing resources of neighborhoods and must aim to Darrell Gaskin is an associate professor of health economics improve the underlying conditions of health for all. in the Department of Health Policy and Management, Johns Hopkins Bloomberg School. here is extensive documentation of Americans than among white Americans. Con- T persistent health disparities among sequently, the overlap between these two factors the racial and ethnic groups that complicates efforts to determine whether it is make up the United States. In re- "race and class" or "race or class" that produces cent years, researchers have at- disparities in health status. tempted to understand the reasons for these dis- The second challenge resulting from using na- parities and to find ways to eliminate them. tional data is complex and, we suspect, even Much of this research relies on data from na- more powerful. The United States is segregated tional studies. These data are suboptimal for along racial lines. Racial segregation creates dif- the study of racial health disparities for two ferent exposures to economic opportunity and to reasons. other community resources that enhance health. First is the confounding of race with socio- Likewise, segregation produces differential ex- economic status.' Health status varies by both posure to health risks.'- Thus, racial disparities race and socioeconomic status, and socioeco- may be confounded by disparities based on nomic status tends to be lower among black place. As a result, estimates of racial disparities 1880 HEALTH AFFAIRS OCTOBER 2011 30:10 Downloaded from HealthAffairs.org on May 22, 2020. Copyright Project HOPE-The People-to-People Health Foundation, Inc. For personal use only. All ri euse permissions at HealthAffairs.orgfrom national samples may be biased because they fail to consider the differing opportunity and risk profiles of communities where racial and ethnic groups live. Confounding race with socioeconomic status and segregation can have profound implications for the development of policies to address racial and ethnic health disparities. For instance, some researchers have focused on biological factors, assuming that disparities result from biological differences among racial groups. Others have focused on health behaviors thought to be con- centrated among different cultural groups. But it may be that racial and ethnic disparities observed in national data reect disparities based on features of the communities where peo- ple live. Racial and ethnic health disparities may be driven mostly by placethat is, the segregated communities where the preponderance of these minority populations may live. To shed further light on this issue, we sought to design an analysis of racial disparities in health-related outcomes that also accounts for social determinants based on place.We designed the Exploring Health Disparities in Integrated Communities study to compare health outcomes among black and white Americans who live in the same socioenvironmental conditions and have similar financial resources.6 In this article we briey describe the study, summarize several key findings that have been previously reported in scientific journals, and outline the relevant policy implications. Study Data And Methods SELECTING THE commuurrv We identified com munities in the United States that contained at least 35 percent African American and at least 35 percent white residents; had a ratio of black towhite median income between 0.85 and 1.15', and had a ratio of blacktowhite high school graduation rates (among people age twentyfive and above) between 0.85 and 1.15. Of 66,438 census tracts identified in the 2000 US census, 425 tracts met our inclusion criteria.We selected two contiguous tracts in Southwest Baltimore, Maryland, as our rst study site. The study design is novel in that it compares black and white Americans who are exposed to the same set of socioeconomic, social, and envi ronmental conditions. Thus, we can discern whether or not racial differences in health related outcomes stem from something endemic to the people or whether they are more properly attributable to differences in conditions within communities. sunvsv METHODS We conducted in-person in terviews with adult residents (age eighteen and older) of the Southwest Baltimore study site. Trained interviewers administered a structured questionnaire and measured blood pressure for approximately 42 percent (71 : 1, 489) of the 3,555 eligible residents of the study area. The study methods are described in greater detail elsewhere.6 The study questionnaire incorporated ques- tions from three national surveys: the National Health Interview Survey, 2003; the National Health and Nutrition Examination Survey, 19992004; and the Medical Expenditure Panel Survey, 2002. Replication of questions enabled us to compare results from analyses of the na- tional samples, which do not account for segre- gation, with results from Southwest Baltimore, which is racially integrated. We used data from the National Health Inter- view Survey to compare national and study-site data for obesity,7 smoking,\" and diabetes.9 This survey is conducted annually by the National Center for Health Statistics and includes the civilian, noninstitutionalized population of the United States.7 The interviews are conducted in the homes of adults age seventeen and older.We used the Sample Adult Core section of the survey for our comparison because of similarities with the time period and age range for our Southwest Baltimore study. Detailed information regarding the National Health Interview Survey can be found elsewhere.\" We used data from the National Health and Nutrition Examination Survey to compare na- tional and study-site data for hypertension.ll This is an annual survey from which public-use data les are released every two years. The survey is a nationally representative sample of the US civilian, noninstitutionalized population, with an oversample of lowincome people, partici- pants ages 1219, adults age 50 and older, black Americans, and Mexican Americans.l2 We used data from the Medical Expenditure Panel Survey to compare use of health services13 nationally and in the study area. This survey, conducted by the Agency for Healthcare Re- search and Quality, includes the US civilian, non- institutionalized population. It is an authorita- tive source ofinformation on the nation's health care use and is used by the agency to monitor progress on eliminating health care disparities.\" ANALYTI: s-rlm'rsav This article is a synthesis of previously reported information in which we focused separately on the health outcomes re- viewed here. As such, our research offers a com- prehensive look at disparitiesif anybetween black and white people living in similar circum- stances.\"'\"'\" Using the Southwest Baltimore data, we analyzed health-related outcomes for hypertension, diabetes, obesity among women, OCTOBER 2011 30110 HEALTH AFFAIRS Duwlllamlm I'm-n HmltlIAl'l'airs mg on May 21 mu Cowugm Plulcn HOPb l1): Pcupic-Io-Pcupic Hcallh roumanun. [m- m wrmmi m unly All ngllli men-m kwsc pannimuns a! callllAllhusMrg 1881 DRIVERS OF DISPARITIES smoking, and use of health services-categories Exhibit 2 summarizes key findings from pre- for which well-documented racial disparities ex- viously published studies based on the South- ist on a national basis. We then conducted the west Baltimore sample. same analysis using the national data. Our goal HYPERTENSION A study published by Roland was to determine if the racial disparities found in Thorpe and colleagues" sought to examine hy- the national surveys (which do not account for pertension disparities between blacks and racially segregated living environments) dif- whites in Southwest Baltimore compared to fered from the results of our analysis. Using this those shown in the National Health and Nutri- approach, we could estimate whether disparities tion Examination Survey. That study defined found in national studies existed when we ac- hypertension as systolic blood pressure greater counted for differences in place. than or equal to 140 mm Hg, diastolic blood pressure greater than or equal to 90 mm Hg, respondent's self-report of taking antihyperten- Study Results sive medications, or some combination. We used We compared selected characteristics of the logistic regression analysis to adjust for age, sex, Southwest Baltimore study site with each of marital status, household income, education, the national surveys (Exhibit 1). Because the na- health insurance, self-rated health, weight, ex- tional surveys are representative of the US pop- ercise, diabetes, drinking, and smoking. We re- ulation, there are only minor variations among plicated the same model in both samples. them. However, in some cases there are substan- We found a racial disparity in hypertension in tial differences between the Baltimore sample Southwest Baltimore, where blacks suffered and the national samples. For example, respon from hypertension at a higher rate than whites. dents in the Baltimore sample tended to be youn- However, the race odds ratio was 29 percent ger and were less likely to have received formal smaller in the racially integrated community education beyond high school. Also, the Balti- compared to the national survey. We concluded more respondents were more likely to rely on that racial differences in social environments public health insurance or to be uninsured. Con- explained about one-third of the racial difference sequently, all analyses adjusted for these dif- in hypertension typically found in national data. ferences. More than two-thirds of the disparity in hyper- EXHIBIT 1 Variables Related To Characteristics Of Black And White Residents, Southwest Baltimore Study Compared With National Studies, Selected Years 1999-2004 EHDIC-SWB NHANES' (2003) NHIS (2003) (1999-2004) MEPS (2002) Variable Black White Black White Black White Black White Age, years (mean) 38 4 43.9 44.8 47.3 42.3 46.9 437 45.1 SEX (/%) Male 43.1 45.6 38.0 44.4 46.4 49.0 40.4 47.2 Female 56.9 54.4 62.0 55.6 53.6 51.0 59.6 52.8 EDUCATION (/) Less than high school 35.4 47.5 22.93 19.2 50.4 28.7 28.9 27.9 High school/GEL 36.4 23.6 30.7 29.0 18.7 23.5 37.5 31.6 More than high school 185 183 46.4 51.8 30.7 47.7 33.2 39.9 INCOME (/0) Less than $10,000 25.2 22.1 40.4 33.5 14.9 5.9 37.8 29.2 $10,000-19,999 34.6 33.1 12.7 8.1 30.5 21.7 23.3 21.9 $20,000-34,999 19.7 22.6 15.5 14.4 14.1 11.3 21.8 22.7 $35,000-54,999 11.2 11.6 15.3 16.1 16.9 19.8 11.5 14.7 $55,000 or more 9.1 103 158 27.7 23.6 41.3 5.6 11.5 HEALTH INSURANCE (0/) Private 42.3 42.2 51.9 67.1 66.9 84.6 57.8 42.9 Public 60.4 69.8 32.3 27.1 33.1 15.4 28.1 42.3 Uninsured 34.9 40.2 20.1 17.2 15.0 10.0 14.1 14.8 SOURCES See below. NOTES EHDIC-SWB is Exploring Health Disparities In Integrated Communities-Southwest Baltimore. NHIS is National Health Interview Survey. NHANES is National Health and Nutrition Examination Survey. MEPS is Medical Expenditure Panel Survey. GED is general equivalency diploma. Note 6 in text. Note 10 in text. "Note 12 in text. Note 14 in text. 1882 HEALTH AFFAIRS OCTOBER 2011 30:10 Downloaded from HealthAffairs.org on May 22, 2020. Copyright Project HOPE-The People-to-People Health Foundation, Inc. For personal use only. All rights res . All rights reserved. Reuse permissions at HealthAffairs.org.EXHIBIT 2 Summary Of Previously Published Findings From The Exploring Health Disparities In Integrated Communities-Southwest Baltimore Study EHDIC-SWB National survey Health-related outcome Odds ratio 95% CI Odds ratio 95% CI Major finding Hypertension 1.42 1.09-1.86 2.01 1.63-2.48 Racial disparity related to hypertension prevalence smaller in EHDIC-SWB than in NHANES, but still significant Diabetes 1.07 0.71-1.58 1.61 1.26-2.04 Racial disparity related to diabetes found in NHIS but not in EHDIC-SWB Obesity 1.25 0.90-1.75 1.99 1.71-2.32 Racial disparity related to obesity found in NHIS but not in EHDIC-SWB Smoking Blacks have lower odds of being a current smoker Lifetime smoker 0.55 0.41-0.72 0.62 0.49-0.79 and smoke fewer cigarettes per day compared to Current smoker 0.71 0.56-0.90 0.93 0.72-1.21 whites in EHDIC-SWB, but no racial disparity in Cigarettes smoked per day 0.68 0.61-0.75 0.86 0.74-1.01 smoking in NHIS Use of health services 1.44 1.00-1.87 0.74 0.51-1.07 No disparity in having a medical care visit in MEPS, but more likely to have a visit in EHDIC-SWB SOURCES See below. NOTES All odds ratios are black:white. EHDIC-SWB is the Exploring Health Disparities in Integrated Communities-Southwest Baltimore Study. Cl is confidence interval. NHANES is National Health and Nutrition Examination Survey. NHIS is National Health Interview Survey. MEPS is Medical Expenditure Panel Survey. Note 11 in text. Note 9 in text. 'Note 7 in text. "Note 8 in text. Note 13 in text. tension, however, appears to result from some- SMOKING Laveist and colleagues examined thing other than place. racial disparities in tobacco use, comparing DIABETES A study published by Thomas our Baltimore study data with data from the Na- LaVeist and colleagues' used research methods tional Health Interview Survey. In Southwest similar to those used in the hypertension study to Baltimore, whites had greater odds of being a determine if the racial disparity in diabetes re- current smoker and reported smoking more cig- ported in national data is similar when black and arettes per day compared to blacks. However, the white Americans live under similar social condi- national survey showed no significant racial dif- tions. In the National Health Interview Survey ference in smoking status or in the number of data, black Americans had greater odds of having cigarettes smoked per day. For both black and diabetes compared to whites. In Southwest Bal- white residents, the prevalence rates for both timore, however, white and black Americans had lifetime and current smoking were much greater similar odds of having diabetes. Moreover, dia- in the local sample. However, when comparing betes prevalence for black Americans was similar blacks and whites across samples, we found that in both samples (10.4 percent and 10.5 percent, the magnitude of difference between the samples respectively), whereas the rate for whites was was greatest for whites. We concluded that differ- much higher in Southwest Baltimore (10.1 per- ences in social and environmental exposures re- cent compared to 6.6 percent nationally). We sulting from segregation partially account for concluded that racial disparities in diabetes racial differences in smoking patterns normally may stem from differences in the social and envi- found in national data. ronmental health risks in communities where USE OF HEALTH SERVICES Darrell Gaskin and black and white Americans typically live. colleagues used the Southwest Baltimore sam- OBESITY Sara Bleich and colleagues' investi- ple to study racial disparities in the use of health gated whether there were racial disparities in care services. In the Medical Expenditure Panel obesity among women. In the National Health Survey data, blacks were less likely to have had a Interview Survey data, black women exhibited health care visit within the past year when com- nearly twice the odds of being obese when com- pared with whites. However, in the Southwest pared with white women, after covariates were Baltimore sample, blacks were more likely to controlled for. However, in Southwest Balti- have had a visit. Moreover, in the Southwest more, black and white women had similar odds Baltimore data, among those who had at least of being obese. We concluded that there were no one health care visit, there was no disparity in the racial disparities in obesity among black and number of follow-up visits. But in the national white women exposed to similar living con- sample, blacks had fewer follow-up visits. This ditions. indicates that disparities in initiating care may OCTOBER 2011 30:10 HEALTH AFFAIRS 1883 Downloaded from HealthAffairs.org on May 22, 2020. Copyright Project HOPE-The People-to-People Health Foundation, Inc. For personal use only. All rights reserved. Reuse permissions at HealthAffairs.org.DRIVERS OF DISPARITIES 1884 be inuenced by community-level factors such as the availability of providers and transportation. In contrast, disparities in continued use of care may be related to factors associated with inter- actions between patients and physicians. Discussion In this article we report on a set of previously published findings generated from a novel study design that allowed us to compare black and white Americans who live under similar social and economic conditions and receive health care in the same marketplaceWe found that the racial disparity we normally see in national samples was attenuated or completely erased when white and black Americans live under similar condi- tions. The sole exception was smoking, where we found that white residents in the Baltimore sample were more likely to smoke than black residents. When social factors and medical care are equalized, racial disparities are minimized. It is important to point out that the smaller racial disparity in our studies resulted from dra- matically higher rates of ill health among whites in Southwest Baltimore compared with the na- tional samples. Conversely, blacks' rates of ill health from Southwest Baltimore tended to be similar to rates found in national samples.When whites are exposed to the health risks of a chal- lenging urban environment,15 their health status is compromised similarly to that of blacks, who more commonly live in such communities. We acknowledge that the Southwest Baltimore community is urban and low income, and thus not representative of all communities. Results may differ in other types of communities higher income, suburban, rural0r in other re- gions of the country. We are in the planning phase of a similar study to be conducted in a high-income community. At this point in our research, we have generated a set of findings that strongly suggests that at least some portion of the disparities normally attributed to race should more appropriately be attributed to placecom- munities where people live, work, play, and pray. POLICY CHALLENGES The challenge in develop ing policies that focus on place is that race also determines place. That is, members of racial minorities have fewer options in the housing market.16 A racially segmented housing market affects health through several routes: limited appreciation in home values leading to restricted opportunities for building wealth;"'\" increased exposure to health risks;15"9'21 decreased avail- ability of resources necessary to live a healthy lifestylefz'23 less access to quality health care?"28 and limited access to social capital, such as networks of friends.29'5 HEALTH AFFAlRS OCTOBER 2011 30:10 The challenge in developing policies that focus on place is that race also determines place. The impact on health disparities may be great- est if policy makers address the systemic struc- tures that produce inequities in opportunity. Sol- utions for the health outcomes discussed in this articleincluding obesity and tobacco usehave often emphasized personal choice. We certainly agree that people have the responsibility to maxi- mize their health outcomes. However, a large body of research has documented that these pub- lic health problems are complex and multi faceted, inuenced by myriad factors interacting at the individual, family, community, and soci etal levels. Strategies to eliminate health dispar ities re quire solutions that address multiple lev els, not just individual responsibility. Even ifwe assume that the eradication of racial segregation is beyond the reach of policy pre scriptions, we can pursue strategies that lessen the impact of place in producing race disparities. This can be done through the adoption of poli cies that redress the inequitable distribution of power and resources across communities. HEALTH IN ALL POLICIES Adopting a "health in all policies\" approach is one such strategy. This approach recognizes that health is affected by policies that do not explicitly address health, including those in the arenas of housing, agri culture, and the environment. Thus, to improve population health, policy makers need to con sider sectors outside the strict confines of \"health.\"1 Internationally, the healthinall policies approach has been described as having the potential to contribute to population health and reduce health inequalities}2 One tool to use in achieving health in all pol- icies is the health impact assessment. It is a proc- ess whereby the wide-ranging health impacts of a proposed policy, especially those outside the realm of health, are evaluated in order to inform decision making}3 For example, the City of Ba]- timore conducted a health impact assessment of its proposed zoning code, which would certainly inuence the character of its neighborhoods.\" As a result ofthe health impact assessment, city leaders proposed several recommendations that Duwnlaaum 1mm HealtlIAh'ain mg on May 22. 2020 Copyright swan HOPtil'be J'ouplermrl'euplo Hmtlh Ionn'tanon, m. inrpemml use only All ngl: reserves Ruse penmsnuns at Healumn'mmug sought not only to mitigate disparities in land Health reform increases Medicaid payments to use, but also to create healthy environments for primary care physicians, but these increases are all residents." This is an example of how, in temporary. Federal and state policy makers practice, a health-in-all-policies approach can should make efforts to further eliminate dispar- be used to create equitable environments for ities in reimbursement rates, thus encouraging all racial and ethnic groups and to create healthy more physicians to practice in minority com- and safe communities. munities. Another possible way to address social deter- Finally, some existing policies appear promis- minants of health associated with place is to ing, including those that foster access to afford- improve health care resources in disadvantaged able and nutritious food, enforce bans on smok- communities. The goal of interventions should ing, and limit the location and number of outlets be to lessen risk, increase resilience among indi- selling alcohol. Similar place-based policies viduals and communities, and ensure access to should be formulated and disseminated so that medical care. These steps would address health their uptake is wide and far-reaching. The health- needs associated with living in a high-risk com- in-all-policies approach recognizes that individ- munity. uals exist in the context of their interpersonal The Affordable Care Act of 2010 attempts to networks; overarching community organiza- address problems associated with health care tions; and public policies at the local, state, financing and delivery systems that are height- and federal levels that affect the various tiers ened by residential segregation. The organiza- of social organization. 3536 Scholars have pro- tion and financing structure of US health care posed that the ecological model be applied to puts people of low socioeconomic status and public policies for some of the health conditions their communities at a disadvantage. Disparities described in this article, which would yield op- in reimbursement rates between Medicaid and portunities for people to make healthy choices other third-party payers create "medical deserts" easily. 37-40 in some minority communities. This research was supported by Grant No. P60MD000214-01 from the National Center on Minority Health and Health Disparities of the National Institutes of Health, and a grant from Pfizer, Inc. NOTES 1 LaVeist TA. Disentangleng race and 7 Bleich SN, Thorpe RJ Jr., Sharif- Disparities in Integrated Commun socioeconomic status: a key to Harris H, Fesahazion R, LaVeist TA ities (EHDIC) Study. Soc Sci Med. understanding health inequalities. J Social context explains race dispar- 2008;67 (10):1604-11. Urban Health. 2005;82(2 Supply): ities in obesity among women. J 12 Centers for Disease Control and iii, 26-34. Epidemiol Community Health. 2010; Prevention. National Health and 2 Navarro V. Race or class versus race 64(5):465-9. Nutrition Examination Survey [In- and class: mortality differentials in B LaVeist TA, Thorpe RJ, Mance GA, ternet]. Atlanta (GA): National the United States. Lancet. 1990; Jackson J. Overcoming confounding Center for Health Statistics; [cited 336 (9725):1238-40. of race with socioeconomic status 2011 Aug 6]. Available from: http:// 3 LaVeist TA. Racial segregation and and segregation to explore race dis- www.cdc.govchs/ hanes.htm longevity among African Americans: parities in smoking, Addiction. 13 Gaskin DJ, Price A, Brandon DT, an individual-level analysis. Health 2007;102(Supply):65-70. LaVeist DT. Segregation and dispar- Serv Res. 2003;38(6 Pt 2):1719-33. 9 LaVeist TA, Thorpe RJ, Galarraga JE, ities in health services use. Med Care La Veist TA, Wallace JM Jr. Health Bower KM, Gary-Webb TL. Environ- Res Rev. 2009;66(5):578-89. risk and inequitable distribution of mental and socio-economic factors 14 Agency for Healthcare Research and liquor stores in African American as contributors to racial disparities Quality. Medical Expenditure Panel neighborhood. Soc Sci Med. 2000; in diabetes prevalence. J Gen Intern Survey [Internet]. Rockville (MD): 51(4):613-7. Med. 2009;24(10):1144-8. AHRQ; [cited 2011 Aug 6]. Available Williams DR, Collins C. Racial resi- 10 Centers for Disease Control and from: http://www.meps.ahrq.gov/ dential segregation: a fundamental Prevention. About the National mepsweb/ cause of racial disparities in health. Health Interview Survey [Internet] 15 Morello-Frosch R, Jesdale BM. Sep- Public Health Rep. 2001;116(5): Atlanta (GA): National Center for arate and unequal: residential seg- 404-16. Health Statistics; [cited 2011 Aug 6]. regation and estimated cancer risks LaVeist T, Bowen-Reid T, Jackson J, Available from: http://www.cdc.gov/ associated with ambient air toxics in Gary T, Thorpe R, Gaskin D, et al. nchshis/about_nhis.htm US metropolitan areas. Environ Exploring health disparities in inte- 1 Thorpe RJ, Brandon DT, LaVeist TA. Health Perspecti grated communities: overview of the Social context as an explanation for 386-93. EHDIC Study. J Urban Health. 2008; race disparities in hypertension: 16 Massey DS, Denton NA. American 85 (1):11-21. findings from the Exploring Health apartheid. Boston (MA): Harvard