Successfully reported this slideshow.

From Politics to Parity: Using a Health Disparitiies Index to Guide Legislative Efforts for Health Equity


Published on

Published in: Health & Medicine
  • Be the first to comment

  • Be the first to like this

From Politics to Parity: Using a Health Disparitiies Index to Guide Legislative Efforts for Health Equity

  1. 1. RESEARCH AND PRACTICEFrom Politics to Parity: Using a Health Disparities Index toGuide Legislative Efforts for Health Equity Bryant Cameron Webb, BA, Sean L. Simpson, PhD, and Kristen G. Hairston, MD, MPHThe US Congress took its first stand against Objectives. We created an index quantifying the longitudinal burden of racialminority health disparities with the passage of health disparities by state and compared this index to variables to guide thethe Minority Health and Health Disparities construction of, and validate support for, legislative efforts aimed at eliminatingResearch and Education Act in October 2000. health disparities.This measure commissioned the creation of Methods. We evaluated 5 focus areas of greatest racial disparities in healththe National Center for Minority Health and from 1999 to 2005 and compiled state health disparities index (HDI) scores. WeHealth Disparities and mandated health dis- compared these scores with variables representing the purported social de-parities research and reporting.1 The goal was to terminants of health.conscientiously move the discourse from identi- Results. Massachusetts (0.35), Oklahoma (0.35), and Washington (0.39) aver-fying health disparities to eliminating them. aged the fewest disparities. Michigan (1.22), Wisconsin (1.32), and Illinois (1.50) The years that followed brought progress in averaged the greatest disparities. The statistical reference point for nationwide average racial disparities was 1.00. The longitudinal mixed model procedureresearching causation, positing solutions, and yielded statistically significant correlations between HDI scores and Black statesupporting promising models for the elimination population percentage as well as with the racial gap in uninsured percentages.of these disparities. The congressionally man- We noted a trend for HDI correlations with median household income ratios.dated Institutes of Medicine report Unequal Conclusions. On the basis of the HDI-established trends in the extent andTreatment grouped factors of causation into 3 distribution of racial health disparities, and their correlated social determinantsbasic areas: health system–level factors, care of health, policymakers should consider incorporating this tool to advise futureprocess variables, and patient-level variables.2 efforts in minority health legislation. (Am J Public Health. 2011;101:554–560. doi:In addition, its authors proffered strategies 10.2105/AJPH.2009.171157)for achieving health equity with recommen-dations that included legal, regulatory, andpolicy interventions.2 When considered in sum, these state and METHODS With increasing research and dialogue, legis- federal legislative efforts have ushered in a newlators began considering the next appropriate age of health disparities discourse. Champions Six of the most glaring indicators of racialintervention. Congress has continued to debate of comprehensive health equity legislation and ethnic health disparities, as referenced bysolutions to the health care issues facing minority argue that it is a necessary adjunct to existing the Department of Health and Human Services,communities, introducing 6 comprehensive mi- efforts to eliminate racial and ethnic health were the foundation for this HDI. These focusnority health equity bills since 2000. In 2007, disparities. Skeptics assert that the disparities areas are: cancer screening and management,the 110th Congress saw the introduction of an can be largely addressed through generally cardiovascular disease (CVD), diabetes, HIV/unprecedented 16 health disparities–focused improving access to quality care for all Amer- AIDS, immunizations, and infant mortality.6bills.3 Despite the increasing legislative interest in icans. Most, however, appreciate the utility in Mortality was selected as the primary outcomeimproving minority health, no comprehensive bringing the discussion into the legislative for each condition as a proxy for disease severity,minority health measure has become public law arena as this issue continues to gain momen- because of its value as a more precise endpointsince the Minority Health and Health Disparities tum. than morbidity and its compelling significanceResearch and Education Act of 2000. In this context, we recognized a need to to legislators. We omitted immunization statistics, The proposition of legislative remedies for create tools to guide legislative efforts for as immunization is unique among the 6 focushealth disparities has not been unique to achieving minority health equity. The goals areas as it is a preventive measure rather thanCongress. Since 1999, a majority of state of our analysis were 3-fold: (1) to establish an a distinct cause of mortality.legislatures have passed minority health legis- index depicting variations in US racial health We obtained mortality statistics from thelation.4 These laws have addressed cultural disparities; (2) to evaluate the association be- National Center for Health Statistics Com-competency, health professional recruitment and tween this health disparities index (HDI) and pressed Mortality File from 1999 to 20057 forretention, and disease burden or risk factor known social determinants of health; and (3) to each of the 5 disease processes. We specificallymanagement. Furthermore, 40 states created use statistical correlations to help guide mi- calculated crude mortality rates per 100 000 foroffices of minority health, and 4 states designated nority health legislative interventions at the Black and White populations aged 20 to 64official minority health contacts.5 state and federal levels. years in all 50 states. For each disease process,554 | Research and Practice | Peer Reviewed | Webb et al. American Journal of Public Health | March 2011, Vol 101, No. 3
  2. 2. RESEARCH AND PRACTICEwe calculated a state disparity value (SDV). social determinants of health.8 These inclu- because of missing values. Additionally, 10For example, we calculated the SDV for CVD ded income and social status, education and observations were removed from this latterin Maryland during 2003 by using the for- literacy, health services, culture, and social analysis because of missing values for the gapmula: environments. in uninsured percentages. We omitted from the To evaluate income and social status, we analysis states with unreliable mortality data, asð1Þ SDVCVD in 2003 ¼ calculated income disparities as a ratio of defined in the Compressed Mortality File as ðBlack mortalityfrom CVD in MD during 2003 White-to-Black median household incomes values fewer than 20 deaths per annum in any À White mortalityfrom CVD in MD during 2003 Þ by using the 2000–2005 Annual Social and HDI subcategory.14 =White mortalityfrom CVD in MD during 2003 Economic Supplements of the US Census Bureau’s Current Population Survey.9 Next, RESULTS as our variable for education and literacy, we This value is representative of the ratio of calculated the ratio of Black-to-White high Thirty-three states had sufficient data forexcess Black mortality to the White mortality school dropout rates as reported by the National the calculation of HDI scores from 1999 toper 100 000 individuals for a single disease Center for Educational Statistics between 2000 2005, with 17 states omitted.1 Among theprocess in a specific state. Additionally, we and 2005.10 As a proxy for culture and social states for which scores were compiled, Massa-calculated a cumulative US disparity value environment, we used 2 indicators: geographical chusetts (0.35), Oklahoma (0.35), Washington(USDV) for each disease process during a par- region, as divided by the US Census Bureau,11 (0.39), Nevada (0.53), and Kentucky (0.57)ticular year. We calculated the HDI value per and the percentage of each state’s population that had the lowest 7-year average HDI scores, in-year, per disease process, per state with the is Black. These variables were meant to loosely dicative of the least racial health disparity.formula: model the cultural and social environment in California (1.17), North Carolina (1.20), Michigan each state. (1.22), Wisconsin (1.32), and Illinois (1.50) hadð2Þ HDICVD in ðyearÞ ¼ We depicted health services through 3 the highest 7-year average HDI scores, making SDVCVD in ðyearÞ =USDVCVD in ðyearÞ separate variables. First, differences in access them the states averaging the most racial health to health services were represented by the disparity between 1999 and 2005 (Table 1). We calculated a comprehensive HDICOMP difference between Black and White unin- When we used Pearson’s correlation coeffi-score for each year by using the following sured rates as reported in Behavior Risk cients, the HDI was positively correlated toformula, for example: Factor Surveillance System questionnaire racial disparities in median household income responses. In addition, state Medicaid pro- (P < .001), state Black population (P < .001),ð3Þ HDI2003 ¼ gram eligibility was selected as a proxy for and Medicaid eligibility scores (P < .01). We ½HDICancer in 2003 1 HDICVD the size of the state’s safety net for its most found a negative correlation between HDI 1 HDIDiabetes 1 HDIHIV vulnerable populations. This value was scores and state health spending (P < .001). 1 HDIinfant mortility Š=5 extracted from Public Citizen Health Re- The longitudinal mixed model analysis in- search Group’s analysis of state Medicaid cluded 5 years of data (2000–2004) for each We calculated these HDI scores for the 7 programs, which evaluated each state Med- state rather than the entire 7-year data set.years in the analysis for all 50 states as suffi- icaid program for its coverage of optional The 1999 and 2005 data were eliminatedcient data allowed. eligibility groups.12 Finally, to capture the in- from the analysis because of the missing A state HDI score of 1.00 represents racial creased spending that would be associated with values for median household income andhealth disparities equivalent to those existing health disparities, we evaluated state health state health spending for those years. In theamong the entire US population. An HDI score spending, calculated as a percentage of the gross final analysis, 32 states were included (Ten-of 0.00 would represent racial health parity, state product.13 nessee was ultimately omitted because ofand states with HDI scores of 0.01 to 0.99 a lack of a state Medicaid eligibility score),experience less racial health disparity in mor- Statistical Analysis and a total of 150 observations. The longi-tality than the inequalities among the US We first employed Pearson’s correlation co- tudinal mixed model analysis yielded a cor-population at large. Lastly, states with HDI efficients to assess the association between relation between HDI scores and bothscores exceeding 1.00 exhibit more racial state HDI scores and variables representing the racial disparities in uninsured percentageshealth disparity than US averages, indicating social determinants of health. We completed (P < .01) and state Black population (P < .01).larger differences in Black and White mortality a second, more sophisticated longitudinal We found a trend toward significance be-for the selected conditions. mixed model analysis to reassess correlations tween HDI scores and racial disparities in in the data from 2000 to 2004. In the mixed median household income (P = .074). TableSelection of Variables model we used yearly data for disparities in 2 details the outcomes of the statistical To test the validity of the HDI as a tool uninsured percentages, state health spending, analysis both for Pearson’s correlation co-to measure health disparities, we first evalu- and disparities in median household income, efficients and the longitudinal mixed modelated its correlation to several of the known but averaged data for high school dropout rate analysis.March 2011, Vol 101, No. 3 | American Journal of Public Health Webb et al. | Peer Reviewed | Research and Practice | 555
  3. 3. RESEARCH AND PRACTICE Tracking Progress in Reducing TABLE 1—Health Disparities Index Scores by Rank, 1999–2005 Disparities Statea 1999 2000 2001 2002 2003 2004 2005 Average Rankb According to HDI scores, most states failed to significantly reduce the extent of health Massachusetts 0.27 0.44 0.32 0.33 0.52 0.22 0.31 0.35 1 disparities between 1999 and 2005, an ob- Oklahoma 0.46 0.35 0.28 0.36 0.26 0.29 0.44 0.35 2 servation consistent with the findings of the Washington 0.63 0.18 0.15 0.36 0.48 0.34 0.61 0.39 3 National Healthcare Disparities Reports issued Nevada 0.44 0.38 0.57 0.51 0.51 0.67 0.64 0.53 4 during the study period.15–17 The states with the Kentucky 0.65 0.52 0.42 0.69 0.57 0.53 0.59 0.57 5 fewest disparities in 1999 generally maintained New York 0.57 0.60 0.60 0.53 0.58 0.60 0.69 0.60 6 their status among the most equitable states by Florida 0.68 0.66 0.59 0.66 0.56 0.59 0.55 0.61 7 2005, whereas states with the greatest disparities Minnesota 0.85 0.65 0.59 0.29 0.36 0.64 0.91 0.61 8 largely remained among the states with the Colorado 0.62 0.82 0.66 0.93 0.78 0.36 0.60 0.68 9 highest HDI scores throughout the duration of Connecticut 0.63 0.84 1.02 0.73 0.69 0.58 0.81 0.75 10 the study. The profiles of the states at either end Arkansas 0.90 0.89 0.73 0.73 0.73 0.65 0.76 0.77 11 of the rankings were noticeably divergent with Mississippi 0.76 0.72 0.81 0.88 0.75 0.69 0.85 0.78 12 regard to the Black population percentage and Alabama 0.94 0.76 0.92 0.74 0.83 0.63 0.75 0.79 13 racial disparities in uninsured percentages, Georgia 0.86 0.82 0.78 0.78 0.78 0.82 0.78 0.80 14 whereas disparities in median household income Indiana 0.96 0.72 0.76 0.91 0.86 0.83 0.81 0.84 15 were not significantly different among states at Kansas 0.80 0.87 0.78 1.08 0.69 0.72 1.15 0.87 16 either end of the HDI rankings (Table 3). Delaware 1.21 0.86 0.97 0.73 0.92 0.65 0.93 0.90 17 Although few states exhibited any meaningful Ohio 0.82 0.86 0.85 0.95 0.95 1.02 0.94 0.91 18 diminution of health disparities, some displayed Tennessee 0.96 1.03 0.93 0.95 0.85 0.85 0.84 0.91 19 steady improvement in their HDI scores over Missouri 1.15 0.83 0.95 1.14 0.89 0.88 0.92 0.97 20 the study duration. Maryland, for example, Texas 0.92 0.94 0.98 1.04 1.06 1.00 1.07 1.00 21 began the analysis as the state exhibiting the Nebraska 1.12 0.91 0.92 0.88 1.21 1.01 0.98 1.01 22 third-most racial health care disparity in 1999, Louisiana 1.07 1.02 1.03 1.11 1.12 1.06 0.96 1.05 23 with an HDI score of 1.25. By 2005, however, New Jersey 1.10 1.09 1.12 1.08 1.11 1.04 1.05 1.08 24 this score had steadily decreased to a value of South Carolina 1.11 1.13 1.13 1.05 1.18 1.11 0.90 1.09 25 0.93, a value that was below the average for Maryland 1.25 1.24 1.23 1.02 1.11 0.96 0.93 1.11 26 racial health care disparity across the nation. Pennsylvania 1.33 1.07 1.09 1.03 1.21 1.08 1.02 1.12 27 The greatest improvement was made in the Virginia 1.03 1.12 1.20 1.32 0.96 1.20 1.02 1.12 28 interval between 2001 and 2002, as the HDI California 1.09 1.16 1.16 1.18 1.18 1.23 1.22 1.17 29 score decreased from 1.23 to 1.02. Of note, North Carolina 1.12 1.28 1.25 1.24 1.13 1.25 1.10 1.20 30 Maryland passed a total of 4 state laws ad- Michigan 1.24 1.05 1.18 1.29 1.21 1.24 1.35 1.22 31 dressing health disparities in 2002 and 2003.18 Wisconsin 1.11 1.41 1.03 1.19 1.55 1.65 1.34 1.32 32 We are not, however, attributing Maryland’s Illinois 1.47 1.44 1.42 1.52 1.57 1.58 1.51 1.50 33 decreasing HDI value solely to the passage of Note. Health disparities index is an index depicting state variations in US racial health disparities. Values < 0.00 represent that state legislation. Rather, the legislative a Black mortality rate that is better than the White mortality rate. Values of 0.00 represent parity in Black–White mortality. commitment to the reduction of disparities in Values > 0.00 but < 1.00 represent state mortality disparities that are below the national average for mortality disparities. Maryland reflects a statewide emphasis on im- Values of 1.00 represent state mortality disparities that equal the national average for mortality disparities. Values > 1.00 represent state mortality disparities that exceed the national average for mortality disparities. proving the health of minority communities. a Alaska, Arizona, Hawaii, Idaho, Iowa, Maine, Montana, New Hampshire, New Mexico, North Dakota, Oregon, Rhode Island, Furthermore, this commitment was statistically South Dakota, Utah, Vermont, West Virginia, and Wyoming had insufficient data for the calculation of a health disparities reflected in the reduction of health care dispar- index score. b State rank with respect to 33 states with sufficient data to be included in the analysis. ities as depicted by the HDI values. Tracking Health Disparities in MinorityDISCUSSION created to depict variations in the extent and Health Focus Areas distribution of the racial disparities that are The HDI offers states the opportunity to The HDI has several strengths, including its known to persist. The HDI builds upon the track disparities in mortality among cancer,evaluation of health disparities. Rather than corpus of health disparities literature to depict CVD, diabetes, HIV/AIDS, and infant mortal-focusing the analysis on well described gaps the racial health inequalities for each state ity. Figure 1 illustrates the HDI disease-specificbetween Black and White health statuses within the larger, national context of health scores for these 5 focus areas in the 2 states ataccording to various indicators, the HDI was disparities discourse. polar ends of the HDI rankings. This index556 | Research and Practice | Peer Reviewed | Webb et al. American Journal of Public Health | March 2011, Vol 101, No. 3
  4. 4. RESEARCH AND PRACTICE Application of the HDI to the TABLE 2—Statistical Correlations Between Health Disparities Index Scores Development of Minority Health and Selected Variables Legislation Health Disparities Index Scores Through statistical correlations, the HDI becomes a valuable tool to evaluate health- Pearson’s Correlation Coefficients inequalities in relation to variations in the Coefficient P Longitudinal Mixed Analysis, P social determinants of health. As such, corre- Geographical region NA .243 lations found between the HDI and the se- State Black population (n = 238) 0.65 <.001 .001 lected variables have implications for both Disparities in uninsured percentages (n = 218) 0.1 .124 .008 understanding racial and ethnic health dis- Disparities in high-school drop-out rates (n = 224) –0.09 .203 .466 parities and crafting future minority health Disparities in median household income (n = 204) 0.78 <.001 .074 legislation. State health spending (n = 204) –0.38 <.001 .958 Blacks are 3 times more likely to live in State Medicaid eligibility score (n = 238) 0.18 .006 .779 poverty than are Whites, a remarkable fact when one considers how income signifi- cantly influences health status, access toconstitutes another mechanism by which states management. By allowing states to identify health care, and health insurance coverage.19can track disparities in certain disease pro- disease processes that contribute significantly The positive correlation between the HDIcesses to identify state priorities in addressing to the health disparities burden in their state, and racial disparities in median familydisparities. Importantly, each of these condi- specific interventions can be crafted to assist income corroborates studies touting the im-tions can be addressed through better pre- in addressing these inequalities in minority pact of disparities in income on racial healthvention, surveillance, and disease communities. inequalities. Additionally, we found a correla- tion between states with the largest Black populations and those having the most racial TABLE 3—Changes Over Study Duration (1999–2005) in Health Disparities health disparity. Notably, the income dispar- Index (HDI) Scores and Selected Variables Among States With the Lowest and ities and Black state population variables, Highest HDI Scores in 1999 themselves, were positively correlated, sup- porting claims of their interrelatedness in State HDI Score HDI Rank Black Population, % Health Insurance Gap, % Income Gap contributing to poor health outcomes. Least disparities We employed the longitudinal mixed Massachusetts model analysis because of the potential for 1999 0.27 1 5.34 6.80 1.76 confounding variables. Not only did it ex- 2005 0.31 1 6.51 8.60 1.81 plicitly model state change in HDI scores Nevada over time, but it also simultaneously and 1999 0.44 2 6.50 NA 1.75 explicitly modeled between- and within-state 2005 0.64 7 8.02 NA 1.53 variation. In discerning the relationship of Oklahoma variables to the HDI over the duration of the 1999 0.46 3 6.38 –0.40 1.82 study, this analysis afforded flexible model- 2005 0.44 2 8.11 9.00 1.55 ing of covariance structure of the repeated Most disparities measures, allowing us to determine whether Maryland variables such as Black percentage of the 1999 1.25 31 27.72 5.50 1.47 population and disparities in income were 2005 0.93 20 29.52 10.60 1.5 confounders or independently associated Pennsylvania with the HDI. 1999 1.33 32 9.60 4.30 1.66 Using the longitudinal mixed model analysis, 2005 1.02 25 10.68 10.80 1.79 we identified correlations between the HDI Illinois and both racial disparities in uninsured per- 1999 1.47 33 14.50 13.40 1.67 centages as well as the state Black population. 2005 1.51 33 15.13 15.40 1.82 Additionally, we noted a statistically insignifi- Notes. NA = not available. Of the 17 states without HDI scores, 16 have Black populations of less than 4%, and represent the cant trend correlating the HDI and racial 16 states with the smallest Black populations. disparities in median household income. The presence of both the correlation for stateMarch 2011, Vol 101, No. 3 | American Journal of Public Health Webb et al. | Peer Reviewed | Research and Practice | 557
  5. 5. RESEARCH AND PRACTICE We found no correlation between the HDI and state inclusion of optional eligibility groups in their Medicaid programs. Greater access should not be confused with greater accessibility. For a state’s most vulnerable populations, increasing access without the infrastructural support to increase actual ac- cessibility to quality care will not sufficiently reduce racial health disparities. In a review of comprehensive minority health equity legislation that has been in- troduced into Congress, several trends in policy-based interventions recur. Each bill has called for health disparities research and improving health care quality, and most included provisions to increase access to care, health workforce diversity, and training for cultural competency. Several bills proposed grants for community programming, public health education, and promoting healthy lifestyles. Each intervention is supported by extensive health disparities research and evidence-based recommendations for health disparities solutions. The challenge, how- ever, comes in the prioritization of these interventions and the specification of efforts that constitute best practices in achieving Note: CVD = cardiovascular disease; HDI = health disparities index. Value of 0.00 represents racial/ethnic parity in mortality health equity. rates, while a value of 1.00 represents the average Black–White mortality disparities across the United States. Advisory tools are needed for identifying FIGURE 1—Component disease processes of HDI for a) Massachusetts and b) Illinois. Y-axis specific community program grants that values are HDI scores for each disease process or for state disparities as a whole. legislation should support, as well as the ap- propriate improvements to create quality, ac- cessible care. Using the HDI as a guide,Black population and the trend for disparities our analysis, however, we found no statisti- minority health legislation should focusin income support claims that these variables cally significant correlation between dropout primarily on decreasing the number ofare independently related to health disparities rates and the HDI, indicating that this factor uninsured, as well as closing the significantand not merely a function of each other. As alone is not related to health disparities across income gaps that persist between Blackssuch, we believe that legislation to eliminate states. and Whites. Furthermore, the preponder-racial health disparities must be accompanied by As a primary reason for escalating health ance of health disparities in states withefforts to address the income disparities that care costs, chronic diseases, such as those larger Black populations indicates that theimpact health. composing the HDI, account for more than states with the greatest disparities are also Higher education levels have been linked 75 cents of every dollar spent on health care the states with the largest segment of theto the utilization of preventive services and in the United States.21 Additionally, minority population that would benefit from suchlonger life. Conversely, the lower rates of health disparities cost $229.4 billion in direct targeted legislative initiatives. The percent-educational attainment documented among medical care expenses between 2003 and age of the population affected in the statesBlacks compared with Whites and Asians are 2006.11 In our analysis, state health spending with the most racial health disparity shouldconsidered a factor in poor health status.20 was inversely related to racial health disparities encourage legislators to seek to improveWith the impact of increased dropout rates in mortality. Although increased spending may minority health on behalf of these individ-unlikely to manifest for many years, these be associated with less mortality, this is not uals, their constituents.values function as an educational snapshot of a sustainable mechanism to combat disparities in For state legislators, the HDI would bethe community in comparison with other health care in light of our nation’s economic an effective tool to identify the need forstates, allowing us to reasonably extrapolate climate, and current health care trends in disease-specific minority health legislation, asthose values to the community as a whole. In spending. well as a mechanism to track the progress in558 | Research and Practice | Peer Reviewed | Webb et al. American Journal of Public Health | March 2011, Vol 101, No. 3
  6. 6. RESEARCH AND PRACTICEdecreasing mortality. Federally, the reality national racial health disparities. We chose our Contributorsthat the Black–White mortality gap has not point of comparison with focus on the unique B. C. Webb originated the study, completed the analysis, led the writing, and supervised all aspects of its imple-improved significantly over the past 4 decades disparities within each state. Even still, we mentation. K. G. Hairston assisted with study designwarrants the utilization of the HDI to better recognize that this could understate particu- and oversaw study implementation. S. L. Simpson pro-define this egregious disparity and reliably larly concerning Black mortality rates in states vided statistical support. All authors helped to concep- tualize ideas, interpret findings, and review drafts of thetrack current and future efforts to close that with higher mortality among Whites as well. Although this analysis focused on evalu- We were unable to calculate HDI scoresating trends in the social determinants of health, for 17 states because of unreliable mortality Acknowledgmentsfuture variables could be increasingly specific to data. Notably, of the states without HDI scores, This research was supported financially by the Studentpredict the impact that diversifying the physician 16 have Black populations of less than 4% National Medical Association/Pfizer David M. Satcher Research Fellowship.workforce, improving patient lifestyle factors, and are the states with the smallest Black Ronny A. Bell, PhD, Capri G. Foy, PhD, Jaimie Hunter,and increasing funding for community grants populations in the nation. To better elucidate MPH, David L. Mount, PsyD, John H. Stewart IV, MD,would have on the elimination of health dispar- health disparities causation and the most Ashley C. Augustus, MPH, and Orita McCorkle all contributed to creative, analytical, and structural guid-ities. With both state and federal legislation, promising policy solutions, reporting for some ance and support on this understanding of the correlation between primary outcome—be it mortality, incidence, orsocial determinants and the HDI can effectively prevalence—must be sufficient to reliably evaluate Human Participant Protectionserve as a sound basis for legislative initiatives the health status of minority populations in No human participants were involved.using pertinent positive and negative correlations each identify targets for policy intervention up- Finally, we realize that the index proves Referencesstream of the differences in health status among useful in assessing correlation between the 1. Thomson GE, Mitchell F, Williams M. Examining theracial and ethnic minorities. extent of health disparities and the selected Health Disparities Research Plan of the National Institutes of Health: Unfinished Business. Washington, DC: National variables, although it does not assert causation. Academies Press; 2006.Limitations Instead, this analysis can serve 2 other major 2. Smedley BD, Stith AY, Nelson AR, eds. Unequal We used mortality rather than incidence purposes: (1) to track progress in eliminating Treatment: Confronting Racial and Ethnic Disparities indata to construct the HDI. Although health disparities in each state and (2) to Health Care. Washington, DC: National Academies Press; 2002.mortality was selected as a surrogate for dis- identify areas of interest among the socialease severity, many of the interventions to determinants of health and subsequently guide 3. Key health disparities-focused legislation introduced in the 110th Congress [Kaiser Family Foundation Webeliminate health disparities would be meant to future legislation for reducing racial health site]. 2007. Available at: disease and might be better modeled disparities. health/7724.cfm. Accessed October 8, using incidence if sufficient data for these In conclusion, the HDI serves as a novel and 4. Disparities in health 2010 introduced legislationmeasures existed in all states. Additionally, useful mechanism to guide legislative efforts for [National Conference of State Legislatures Web site].a likely time lag exists between variations in the improving minority health. By tracking the 2009. Available at: tabid=14298. Accessed October 8, determinants of health and resultant extent and distribution of racial health dispar- 5. State profiles: minority health and health equitychanges in mortality. Through this analysis, ities, identifying correlated social factors, and offices [National Conference of State Legislatures Webhowever, we hope to discern how long it would positing potential initiatives to reduce dispar- site]. 2009. Available at: for changes in policy to manifest in ities, the HDI displays its great potential utility aspx?TabId= 14299. Accessed October 8, 2009.improvements in HDI scores. in such analyses. Policymakers should strongly 6. Eliminating racial and ethnic health disparities Another limitation was our use of crude consider incorporating this tool to evaluate [Centers for Disease Control and Prevention Web site]. 2009. Available at: rates as opposed to age-adjusted health disparities and strategically plan future disparities.htm. Accessed October 8, 2009.rates. As we were comparing 2 populations legislative initiatives. j 7. Data file documentation, compressed mortality file,with different age structures, it would follow 1999-2005. Hyattsville, MD: National Center for Healththat age-adjusted rates could be beneficial. Statistics; 2009. About the AuthorsBecause Whites have an older age distribution Bryant Cameron Webb is an MD candidate at Wake 8. Williams DR, Jackson PB. Social sources of racialthan do Blacks, age confounds the relationship Forest University School of Medicine, Winston-Salem, NC, disparities in health. Health Aff (Millwood). 2005;24(2): and a JD candidate at Loyola University Chicago School of 325–334.between race and mortality. Future studies Law, Chicago, IL. Kristen G. Hairston, MD, MPH, is withshould consider adjusting the mortality rates 9. US Census Bureau. Current population survey, the Section on Endocrinology and Metabolism at Wake 2008 detailed data tables. Household income 2007.accordingly. Forest University School of Medicine. Sean L. Simpson, Available at: Accessed Octo- Next, our index compares the mortality for PhD, is with the Department of Biostatistical Sciences at ber 17, 2008. Wake Forest University Health Science’s Division of PublicBlacks in each state against the mortality of Health Sciences. 10. US Department of Education, Institute of EducationWhites within the same state. As such, it does Correspondence should be sent to Bryant Cameron Sciences, National Center for Education Statistics. Totalnot compare within-state Black mortality to Webb, 973 E 61st St, Apt 1, Chicago, IL 60637 (e-mail: dropout rate by state and race 2000–2001 to 2004– Reprints can be ordered at http:// 2005 [statistics from Common Core of Data database].a national standard for mortality rates, but by clicking the ‘‘Reprints/Eprints’’ link. Available at: Accessed October 1,rather state racial health disparities against This article was accepted June 21, 2010. 2008.March 2011, Vol 101, No. 3 | American Journal of Public Health Webb et al. | Peer Reviewed | Research and Practice | 559
  7. 7. RESEARCH AND PRACTICE11. US Census Bureau. Census regions and divisions ofthe United States [US Census Bureau Web Site]. 2009.Available at: Accessed February 27, 2009.12. Ramirez de Arellano AB, Wolfe SM, Public CitizenHealth Research Group. Unsettling scores: a ranking ofstate Medicaid programs, 2007(Public Citizen Website). 2007. Available at: AccessedOctober 10, 2009.13. LaVeist TA, Gaskin DJ, Richard P. The economicburden of health inequalities in the United States [TheJoint Center for Political and Economic Studies Web site].2009. Available at: November 6, 2009.14. Compressed mortality file 1979-2007 [Centers forDisease Control and Prevention Web site]. Available at: AccessedDecember 9, 2010.15. National Healthcare Disparities Report, 2003. FullReport [Agency for Healthcare Research and QualityWeb site]. 2003. Available at: Accessed November9, 2009.16. National Healthcare Disparities Report, 2004. FullReport [Agency for Healthcare Research and QualityWeb site]. 2004. Available at Accessed November9, 2009.17. National Healthcare Disparities Report, 2005. FullReport [Agency for Healthcare Research and QualityWeb site]. 2005. Available at: Accessed November9, 2009.18. Nathan-Pulliam S, Hussein CA. Addressing healthcare disparities: promising practices in Maryland [Mary-land Department of Health and Mental Hygeine Website]. 2004. Available at: Accessed November 9,2009.19. Income, poverty, and health insurance coverage inthe United States: 2005 [US Census Bureau Web site].2005. Available at: Accessed November 9, 2009.20. Stoops NA. Half-century of learning: historicalCensus statistics on educational attainment in the UnitedStates, 1940 to 2000 [US Census Bureau Web site].2007. Available at: AccessedOctober 10, 2009.21. Majestic E. Public health’s inconvenient truth: theneed to create partnerships with the business sector[Centers for Disease Control and Prevention Web site].2009. Available at: Accessed November 6, 2009.22. Satcher D, Fryer GE Jr, McCann J, Troutman A,Woolf SH, Rust G. What if we were equal? A comparisonof the Black-White mortality gap in 1960 and 2000.Health Aff (Millwood). 2005;24(2):459–464.560 | Research and Practice | Peer Reviewed | Webb et al. American Journal of Public Health | March 2011, Vol 101, No. 3