|
|
| |||||||||||||||||||||||||||||||||
Background Although the general relations between race, socioeconomic status, and mortality in the United States are well known, specific patterns of excess mortality are not well understood.
Methods Using standard demographic techniques, we analyzed death certificates and census data and made sex-specific population-level estimates of the 1990 death rates for people 15 to 64 years of age. We studied mortality among blacks in selected areas of New York City, Detroit, Los Angeles, and Alabama (in one area of persistent poverty and one higher-income area each) and among whites in areas of New York City, metropolitan Detroit, Kentucky, and Alabama (one area of poverty and one higher-income area each). Sixteen areas were studied in all.
Results When they were compared with the nationwide age-standardized annual death rate for whites, the death rates for both sexes in each of the poverty areas were excessive, especially among blacks (standardized mortality ratios for men and women in Harlem, 4.11 and 3.38; in Watts, 2.92 and 2.60; in central Detroit, 2.79 and 2.58; and in the Black Belt area of Alabama, 1.81 and 1.89). Boys in Harlem who reached the age of 15 had a 37 percent chance of surviving to the age of 65; for girls, the likelihood was 65 percent. Of the higher-income black areas studied, Queens–Bronx had the income level most similar to that of whites and the lowest standardized mortality ratios (men, 1.18; women, 1.08). Of the areas where poor whites were studied, Detroit had the highest standardized mortality ratios (men, 2.01; women, 1.90). On the Lower East Side of Manhattan, in Appalachia, and in Northeast Alabama, the ratios for whites were below the national average for blacks (men, 1.90; women, 1.95).
Conclusions Although differences in mortality rates before the age of 65 between advantaged and disadvantaged groups in the United States are sometimes vast, there are important differences among impoverished communities in patterns of excess mortality.
We estimated the 1990 death rates of people 15 to 64 years of age in eight persistently poor geographic areas and eight geographically proximate, but more advantaged areas, matched for race. We studied the extent to which disadvantaged people in the prime of life have shorter life expectancy than people in more advantaged groups and than the national averages. We also compared mortality rates among different types of communities with persistent poverty.
Methods
Study Groups
Summary data on the 16 areas studied are shown in Table 1. The geographic areas where the groups resided are described in greater detail in the Appendix. Our analyses of black populations include only blacks in those areas, and the analyses of white populations include only whites. In each area, two geographically proximate groups of people of the same race — one poor and one of higher income — were selected for study. In keeping with the distribution of income nationally, the black groups (both poor and higher-income) tended to be less well off than the white groups, and the disparities between the more disadvantaged and less disadvantaged groups were greater in the urban North than in the rural South.
|
Statistical Analysis
For each group, we combined information from death certificates for the three years from 1989 to 1991 with age-stratified counts of persons of each sex from the 1990 U.S. Census to calculate age- and sex-specific rates of death, both overall and from specific causes. To reduce bias due to undercounting in the census, we adjusted the population counts. Specific estimates of undercounting in local areas could not be obtained for every group. We replicated the analyses for Harlem with adjustment for both local and national rates of undercounting. Because the conclusions were the same regardless of the adjustment made, estimates for all the study populations are based on the adjustment for undercounting nationally.
We then computed age- and cause-specific standardized mortality ratios and annualized excess-death rates.9,10 To ensure that the reported measures were comparable among the study groups, we standardized our calculations with reference to the age distribution of the white population nationwide, according to sex.11 For each group, we calculated the number of deaths that would be expected among whites nationwide if their age- and cause-specific death rates were the same as those of the study group. To compute the standardized mortality ratios, we divided the number of expected deaths by the number of observed deaths in the white population nationwide. We calculated excess-death rates as 100,000 times the difference between expected deaths and observed deaths, divided by the white population nationwide.
We used Greville's method10,11 to derive the likelihood that a 15-year-old living in a given area would survive to a specified age. We used standard life-table methods to calculate the average number of years of life lost between the ages of 15 and 65 in each group.12 Using standard multidecrement life-table techniques,12,13 we also estimated how many of the years of life lost between the ages of 15 and 65 in that group could be attributed to a particular cause of death, when competing mortality risks were taken into account.
We classified deaths according to their underlying causes, using the diagnostic categories of the International Classification of Diseases, 9th Revision (ICD-9). After examining many causes, we report on those found to be especially important in explaining the disparities in death rates: diseases of the circulatory system (ICD-9 codes 390 to 459); cancer (codes 140 to 208); accidents (codes E800 to E949); human immunodeficiency virus (HIV) infection (codes 042 to 044); homicide (codes E960 to E969); infectious disease, pneumonia, and influenza (codes 001 to 041, 045 to 139, and 480 to 487); and a separate category that included all the remaining causes.
Using standard methods, we estimated the sampling variability associated with each measure of mortality.10,12,13 For every possible pair of groups, we conducted two-tailed statistical tests of the hypothesis that the measures of mortality would be the same. For the poor populations, the estimated 95 percent confidence intervals never exceeded 11 percent above or below the standardized mortality ratio, 84 deaths more or less than the excess-death rate, 0.03 more or less than the probability of survival, or 0.88 year more or less than the number of years of life lost.
Results
Summary Measures
Table 2 shows annualized excess-death rates and standardized mortality ratios according to sex for people 15 to 64 years old in each study group, relative to the national averages among whites, and the probability of survival to the age of 65 among persons who had survived to the age of 15. The death rates for both sexes in every group of poor blacks were excessive. The disparity was not as great in the Black Belt area of Alabama as in central Detroit, Watts, or Harlem, and it was greatest in Harlem. Mortality rates among blacks in Black Belt Alabama were similar to the national average for blacks and to the rates in northern Alabama, the comparison group, but all the other groups of poor blacks had higher mortality rates than their comparison groups. The disparity between Harlem and its comparison group (Queens–Bronx) was particularly large, with standardized mortality ratios of 4.11 among men and 3.38 among women in Harlem, as compared with ratios of 1.18 and 1.08, respectively, in Queens–Bronx. Black residents in the latter area had mortality rates similar to the national average for whites.
|
Probability of Survival
Typical white Americans who were 15 years old had a higher likelihood of surviving to the age of 65 than blacks nationwide or poor 15-year-olds in any of the poor study groups. With the exception of white residents of Detroit, the poor whites had higher probabilities of survival than the blacks (P<0.001). Among the poor black groups, people in Harlem, central Detroit, and Watts had a lower probability of surviving to the age of 65 than people in Black Belt Alabama (P<0.001). The lowest probability of survival to the age of 65 was in Harlem, among blacks: 0.37 in men and 0.65 in women. For men, this represents less than half the probability of survival of whites nationwide. For poor whites, the lowest probability of survival was in Detroit (men, 0.60; women, 0.77). The probability that a 15-year-old girl in Harlem would survive to the age of 45 was the same as the probability that a typical white girl anywhere in the United States would survive to the age of 65; for boys in Harlem, this probability was lower (P<0.001) (data not shown).
There was marked variation among the groups studied in the average number of years of life lost between the ages of 15 and 65 (Figure 1 and Figure 2). Among men from Harlem, an average of 14 years were lost; among men from central Detroit, 11 years; among men from Watts, 12 years; and among men from Black Belt Alabama, 7 years. Poor whites generally lost more years of life than whites either nationally or in the four comparison groups of higher-income whites. Except in Detroit, poor whites lost fewer years than poor blacks. Queens–Bronx was the only black area that compared favorably with some advantaged white areas.
|
|
Specific causes of excess death in the poor groups are shown in Table 3. Diseases of the circulatory system were important contributors to excess mortality in every group, and in most instances they were the leading cause. HIV was the principal cause of excess death among men in Harlem and the Lower East Side, and the second most common cause among women. Accidents were the leading cause of excess mortality among men in Appalachia, and the second most common cause among men in Black Belt Alabama. Homicide was the leading cause of excess death among men in Watts, the second most common cause among men of either race in Detroit, and the third most common cause among men in Harlem. Cancer was an important cause of excess mortality among both sexes in many groups, notably residents of Harlem, Watts, and both the poor black and poor white areas of Detroit, and infections were a prominent cause among both men and women in Harlem.
|
Discussion
Our findings paint a stark portrait of social inequalities in mortality, but also a varied and complicated one. The differences between the advantaged and disadvantaged groups were sometimes vast. The situation in Harlem in 1990 was particularly dire. Comparison of the estimates by McCord and Freeman7 with ours shows that in Harlem mortality among women relative to that nationwide has not improved since 1980, whereas mortality among men has deteriorated. On the other hand, groups that might have been expected to have excess-mortality rates equivalent to or higher than the rates in Harlem did not.
Our results show that there are important differences among impoverished communities in patterns of mortality. The data suggest that reducing excess mortality may require interventions that address local conditions and mortality risks. For example, our findings underscore the need for interventions to reduce the incidence of diseases of the circulatory system in all the locations examined, but they suggest that HIV, homicide, and accident prevention need to be emphasized differently in different places.
Our findings are generally consistent with the association between race and excess mortality in the United States that is often reported. However, the poverty rate and the location of a group (urban and northern vs. rural and southern) are also important. White residents of Detroit fared as poorly as residents of some black areas that we studied. One black comparison group (that in Queens–Bronx) had a mortality rate only slightly higher than the national average for whites.
Thirty percent of the people in the Queens–Bronx area were black immigrants from the West Indies, but analyses that do not include this foreign-born population indicate that this factor does not explain the favorable mortality profile. An alternative explanation is that when a black population has the same degree of economic advantage as a white population, it also has a favorable mortality rate. In the groups we studied, the number of years of life lost generally increased with the percentage of people in the group who were living in poverty, with the poverty rate accounting for more than half the racial differences in mortality (data not shown).
Although we could not study the extent to which individual economic well-being was itself associated with the likelihood of premature death, our findings were consistent with recent analyses showing that controlling for indicators of individual socioeconomic status accounts for some, but not all, the racial disparity in mortality among people of comparable age.5,14 Other factors that may matter include population density, household crowding, and correlates of residential segregation, such as residence in an area that is medically or socially underserved, one with dilapidated housing stock or a high crime rate, or one with excessive exposure to environmental hazards. Relevant psychosocial factors may be associated with chronic uncertainty, racial stress, or ongoing problems with social injustice and community disruption.15,16,17,18
The data we obtained in Harlem and Black Belt Alabama highlight the importance of accounting for social factors that are not represented in typical measures of socioeconomic status. Black Belt Alabama had the lowest excess mortality of the poor black groups, although it had the highest rate of poverty, whereas Harlem had the highest excess mortality but the lowest poverty rate. These findings remained valid after adjustment for cost-of-living differences between the rural South and the urban North.19,20
More generally, our findings underscore the importance of conducting repeated analyses at the population level in several areas as complements to analyses of individuals or single areas. Population analysis can take into account a more complete and longer-term set of social factors that affect mortality,21 and the study of multiple populations yields more evidence about the generalizability of data and permits extreme or anomalous populations to be identified. The consideration of current or historical features peculiar to such populations can improve an understanding of the social epidemiology of excess mortality.
That young people in some areas of the United States cannot expect to survive through middle adulthood is a disturbing manifestation of social inequality. Living with high levels of risk and uncertainty may affect social behavior and health habits and exacerbate the risk of disease.4 In addition, young people and middle-aged adults provide financial support and care for their families. As public policy shifts away from supporting poor families, we should consider the implications of excess mortality for the ability of adults in persistently poor populations to shoulder more economic responsibilities than they already have.
Supported by a grant (U83/CCU51249-01) from the Centers for Disease Control and Prevention, by the W.T. Grant Foundation, and by the Russell Sage Foundation.
We are indebted to Sherman James for comments, to Al Hermalin and Barbara Anderson for statistical consultation, to Lisa Neidert for data consultation, to Sandra Crump for editorial assistance, and to Kimberly Giamportone and Rachel Stevens for help in the preparation of the manuscript.
Source Information
From the Department of Health Behavior and Health Education, School of Public Health (A.T.G., P.B.B.), the Population Studies Center (A.T.G., J.B., M.M.H.), and the Departments of Economics (J.B.) and Sociology (M.M.H.), University of Michigan, Ann Arbor; the National Bureau for Economic Research, Cambridge, Mass. (J.B.); and the Urban Institute Health Policy Center, Washington, D.C. (T.A.W.).
Address reprint requests to Dr. Geronimus at the Department of Health Behavior and Health Education, University of Michigan School of Public Health, 1420 Washington Heights, Ann Arbor, MI 48109-2029.
References
The 16 population groups (defined according to race and socioeconomic status) for which data are presented in this article included residents of the following neighborhoods, counties, and health center districts.
Poor blacks — Harlem: the Central Harlem Health Center District. Central Detroit: the Central, University, Central Business District, Foch, Jefferson–Mack, Airport, St. Jean, Chene, and Jeffries neighborhoods. Watts: the Watts area of South Central Los Angeles and adjacent areas to the south and west. Black Belt Alabama: Dallas, Fayette, Greene, Bibb, Sumter, Hale, Lamar, Marengo, Marion, Perry, and Pickens counties.
Poor whites — Lower East Side (and Brooklyn): areas in the southern part of the Lower East Side Health Center District (Manhattan) and the Williamsburg–Greenpoint Health Center District, directly across the East River in Brooklyn. Detroit: areas on the northeastern and southern edges of Detroit, including Delray, Clark Park, Chadsey, Condon, Springwells, Jeffries, State Fair, Burbank, Denby, Finney, Mt. Olivet, Grant, Davison, Pershing, and Nolan. Appalachia: Clay, Jackson, Knox, Lee, Leslie, Owsley, Rockcastle, and Wolfe counties (Kentucky). Northeast Alabama: De Kalb, Jackson, and Marshall counties.
More-affluent blacks — Queens–Bronx: sections of the Jamaica East and Flushing Health Center Districts (eastern Queens) and the Pelham Bay Health Center District (north Bronx), including Woodlawn, Wakefield, Williamsbridge, Baychester, Edenwald, Eastchester, Glen Oaks, New Hyde Park, Floral Park, Bellerose, Bellaire, and Queens Village. Northwest Detroit: Palmer Park, Pembroke, Bagley, Redford, Rosedale Park, Evergreen, and Greenfield. Crenshaw–Baldwin Hills: areas of suburban Los Angeles, including Crenshaw, Ladera Heights, Leimert Park, Baldwin Hills, and Windsor Hills. Northern Alabama: Blount, Winston, Calhoun, Walker, Cherokee, Clay, Talladega, Cleburne, Colbert, Cullman, St. Clair, De Kalb, Etowah, Franklin, Jackson, Shelby, Lauderdale, Lawrence, Marshall, and Morgan counties.
More-affluent whites — Queens: sections of the Maspeth–Forest Hills Health Center District, including Rego Park, Forest Hills, Middle Village, and Maspeth. Sterling Heights, Mich.: a northern suburb of Detroit. Western Kentucky: Ballard, Calloway, Fulton, Graves, Hickman, Marshall, and Carlisle counties. Southwest Alabama: Monroe, Clarke, Choctaw, Conecuh, Dallas, Washington, Perry, Sumter, Marengo, and Wilcox counties.
| |||||||||||||||||||||||||||||||||
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
