Skip Navigation


American Journal of Epidemiology Advance Access originally published online on October 28, 2008
American Journal of Epidemiology 2008 168(11):1247-1254; doi:10.1093/aje/kwn291
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
168/11/1247    most recent
kwn291v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Related articles in Am. J. Epidemiol.
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Hearst, M. O.
Right arrow Articles by Johnson, P. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hearst, M. O.
Right arrow Articles by Johnson, P. J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology © The Author 2008. Published by the Johns Hopkins Bloomberg School of Public Health. All rights reserved. For permissions, please e-mail: journals.permissions@oxfordjournals.org.

ORIGINAL CONTRIBUTIONS

The Effect of Racial Residential Segregation on Black Infant Mortality

Mary O. Hearst, J. Michael Oakes and Pamela Jo Johnson

Correspondence to Dr. Mary O. Hearst, Division of Epidemiology and Community Health, School of Public Health, University of Minnesota, 1300 South Second Street, Suite 300, Minneapolis, MN 55454 (e-mail: hearst{at}umn.edu).

Received for publication October 9, 2007. Accepted for publication April 1, 2008.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Economic differences and proximal risk factors do not fully explain the persistent high infant mortality rates of African Americans (blacks). The authors hypothesized that racial residential segregation plays an independent role in high black infant mortality rates. Segregation restricts social and economic advantage and imposes negative environmental exposures that black women and infants experience. The study sample was obtained from the 2000–2002 US Linked Birth/Infant Death records and included 677,777 black infants residing in 64 cities with 250,000 or more residents. Outcomes were rates of all-cause infant mortality, postneonatal mortality, and external causes of death. Segregation was measured by using the isolation index (dichotomized at 0.60) from the 2000 US Census Housing Patterns. Propensity score matching methods were used. After matching on propensity scores, no independent effect of segregation on black infant mortality rates was found. Results show little statistical evidence that segregation plays an independent role in black infant mortality. However, a key finding is that it is difficult to disentangle contextual effects from the characteristics of individuals.

African Americans; infant mortality; prejudice; residence characteristics; risk factors


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
While medical and societal advances have reduced infant mortality rates over the past 100 years, the disparity between African-American (herein referred to as black) and non-Hispanic white (herein referred to as white) infant mortality rates remains (1). Considerable research has been conducted trying to understand the etiology of poor black infant survival and the persistent disparity (2–20). Proximal risk factors contribute to the high black infant mortality rate but do not fully account for the racial differences.

One important risk factor for infant death is low birth weight (<2,500 g), which reflects both the infant's and mother's health. Blacks have twice the rate of low birth weight as whites (21–23). Yet, even among normal birth weight infants, black infant mortality rates are still twice as high as those for white infants (21–23). Proposed explanations for the persistent differences in infant mortality rate include income and education of the mother (24). However, infants born to college-educated blacks continue to have more than twice the infant mortality rate as those born to college-educated whites (24, 25). It seems that other contextual factors must exist.

External causes of death, particularly during the postneonatal period, show the clearest causal path between harmful social and physical environmental exposures, such as racial residential segregation (herein referred to as segregation), and infant mortality rate. Although neonatal mortality, death in the first 27 days of life, is primarily attributed to biologic causes and is most clearly dependent upon the mother's prenatal exposures, postneonatal mortality, death at 28 days to 1 year, is historically attributed to environmental causes and infectious disease (26). Three specific causes of death during the postneonatal period are classified as "external" because of the etiology—sudden infant death syndrome, accident, and assault. Sudden infant death syndrome (27–29) is the third leading cause of infant mortality and the most common cause of postneonatal mortality. Blacks have rates of sudden infant death syndrome twice those of whites (21–23, 27, 29, 30). Accidental injury death (unintentional) is among the 10 most common causes of all infant mortality and is one of the top 3 leading causes of postneonatal mortality. Finally, compared with other racial groups, blacks have the highest rates of infant death due to homicide and certain unintentional injury (31–33). Black assault (homicide) rates, the fifth leading cause of postneonatal mortality, are more than 3 times those for whites (34–37).

Risk factors for these external causes of death include demographics (race, marital status, socioeconomic status) (32, 33), low birth weight, prematurity, high parity, stomach (prone) sleeping position, soft bedding, and bed sharing (27, 29, 32, 38, 39). Several studies have focused on the social and physical environment to which black infants and mothers are exposed. Research suggests residual confounding by unmeasured social and cultural factors (33) or mediation by such factors as housing conditions (40). One hypothesis for the increased incidence of sudden infant death syndrome in the black community, especially the segregated community, is that community members are not receiving important health messages regarding sleep position because they are isolated from mainstream society (41). It was these hypotheses of differential social and environmental exposures (perhaps due to segregation) leading to infant death that we proposed to explore.

Recent literature suggests that segregation contributes to health disparities between blacks and whites because of the social, political, and economic disadvantage brought about by segregated communities (4, 41–44). Segregation is defined as the extent to which groups of individuals living in distinct areas within a region differ by race (45). The majority of blacks in the United States reside in residentially segregated communities, regardless of individual economic status (46). In fact, 80% of middle-class black households are located in predominantly black communities and within 4 blocks of impoverished neighborhoods. This proximity, regardless of individual economic status, may contribute to the higher rates of infant death for blacks living in areas with few contextual risk factors (4).

Our conceptual model of segregation's effect on infant mortality suggests that blacks, compared with whites, are differentially exposed to negative macrosocial conditions (neighborhood deterioration, lack of employment and educational opportunities, loss of tax base), social and material conditions (housing and air quality, limited social contacts, loss of community infrastructure), and individual-level factors (less physical activity, inadequate diet, lack of social support, stress, reduced access to medical care), placing infants at risk of dying because of the isolating effects of segregation (4–6, 10, 12, 13, 18, 20, 47, 48). Segregation reflects both the people who make up segregated communities and a theoretically larger toxic environment.

Although infant mortality has been studied extensively, we know of just 10 published articles on segregation and infant mortality (9–13, 16, 19, 20, 49, 50). All 10 concluded that segregation has a harmful association with infant mortality, varying from the Metropolitan Statistical Area to the borough level. The studies were all ecologic, with no ability to take into account individual characteristics. As our conceptual model above portrays, segregation reflects the individual characteristics and a larger contextual experience; thus, it is important to account for individual-level characteristics.

In this study, we asked whether there is a harmful, independent effect of segregation due to isolation on infant survival once individual-level characteristics are taken into account. We used propensity score matching techniques to compare "exchangeable" infants born into segregated cities with those not born into segregated cities, thereby minimizing the effect of individual-level covariates. We then compared infant death outcomes between different levels of exposure to segregation.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Our data were derived from 2 sources: 1) Linked Birth/Infant Death records from the National Center for Health Statistics (outcome) (51) and 2) Housing Patterns from the US Census Bureau (exposure = segregation) (52). Linked Birth/Infant Death data are coded according to uniform coding specifications and are universally considered an accurate count of all infant births in the United States (53). The Linked Birth/Infant Death records are created by linking infant death records to birth records, with successful linkages made for nearly 99% of all infant deaths in 2000–2002 (21–23). Geographic identifiers are available in the Linked Birth/Infant Death data for cities with populations greater than 250,000 residents (n = 64 (largely central) cities). Our segregation measure is at the level of place, or cities. Place is classified in both sources according to the Federal Information Processing Standards from the US Geological Survey, Geographic Names Information System (54). Linked Birth/Infant Death records were linked to city-level segregation data by Federal Information Processing Standards code, with 100% matching. Refer to Figure 1 for the location of study places by segregation status.


Figure 1
View larger version (17K):
[in this window]
[in a new window]
[Download PowerPoint slide]
  		Figure 1. Study places by segregated (•) and nonsegregated ({blacktriangleup}) status of residents, 2000 US Census Housing Patterns and Linked Birth/Infant Death records.

 
The target population was defined as infants born to black mothers residing in US places with a total population of more than 250,000 residents in 2000–2002 (n = 677,777 livebirths). The exposure of segregation (isolation index) used for this study was calculated by the US Census Bureau using blacks compared with whites as the referent population (52). To calculate segregation measures, the US Census Bureau defines white as those who reported their ethnicity as non-Hispanic and race as white only, whereas the minority population of interest was considered present if individuals classified themselves as that minority group or any combination of races that also included the minority population of interest (52). This distinction is particularly relevant for whites because residential patterns, socioeconomics, and health outcomes vary considerably between non-Hispanic whites and Hispanic whites. However, using the same criteria for blacks does not substantially change conclusions regarding residential pattern (55). Categorizing infant deaths by using the same criteria as the US Census Bureau does assures that the exposure measurement reflects the actual comparison group.

Measures
Our outcome of interest was infant death. The infant mortality rate represents the number of infant deaths up to 1 year of life per 1,000 livebirths. The postneonatal mortality rate represents the number of deaths occurring between 28 days and 1 year per 1,000 neonatal survivors. External postneonatal mortality deaths are a subset and were classified according to International Statistical Classification of Diseases and Related Health Problems, Tenth Revision, code. We used code R95 for sudden infant death syndrome, codes V01–X59 for accidents, and codes X85–Y09 for assault. A separate analysis of neonatal mortality was not conducted because neonatal mortality is largely due to maternal factors. Although neonatal mortality accounts for two-thirds of the infant mortality rate, we analyzed overall infant mortality rate for comparison with previous ecologic studies.

Segregation is a multidimensional construct representing 5 geospatial patterns—dissimilarity, isolation, clustering, centralization, and concentration (56). Housing patterns, or segregation index scores representing these 5 domains, are calculated by the US Census Bureau for only those places with 10 or more census tracts, a population of 10,000 or more, and 100 or more of the minority population of interest. Acevedo-Garcia et al. (43) recommends multilevel analysis of the effect of segregation, and Polednak (48) recommends examining geographies smaller than the Metropolitan Statistical Area level (48). Therefore, city-level data were chosen as the smallest unit of analysis available. The measure of segregation for this study was the isolation index, which indicates that negative health outcomes for blacks may be due to the isolation from mainstream society and has the clearest pathway to health outcomes and social spaces to which infants are exposed (43, 57, 58). The isolation index is a weighted average for a given metropolitan area, ranging from 0 to 1, with 1 indicating higher isolation (52). It captures the extent to which blacks are surrounded by other blacks or nonblacks. Here, the index is highly correlated (r = 0.90) with the more commonly used dissimilarity index. The isolation index was dichotomized at 0.60 to emphasize the extreme isolation experienced by blacks in those cities and is supported by previous sociologic work (10, 45, 56, 59). Sensitivity analysis was conducted, altering the cutpoint between 0.50 and 0.70, with little substantive difference noted.

Covariates that we expected to differ between levels of segregation and that occurred prior to the infant's exposure to segregation were used for the predictive model in propensity score matching methods. The intent was to approximate experimental design, balancing individual-level covariates that may confound the relation between segregation and infant mortality. The hypothetical experiment would be to randomly assign women to reside in segregated cites or not after giving birth and then observe the difference in infant deaths. To compare the 2 hypothetical populations, we examined covariate balance of the characteristics of mother and infant. As such, the covariates used in this observational study were mother's age (years), education group (0–8, 9–11, 12, 13–15, and ≥16 years of schooling), and marital status (unmarried or not); number of previous births and number of infants at this birth; prenatal care utilization (adequate, intermediate, inadequate); father's age (years); birth weight (≤2,500 g or >2,500 g); gestational age (<37 weeks or ≥37 weeks); and year of delivery. Because postneonatal mortality and external causes of death are attributed to social and environmental conditions, matching on these covariates enables "control" of maternal factors along the infant's life course. Hot deck imputation was used for "unknown or not stated" response categories in the Linked Birth/Infant Death records for birth weight, prenatal care utilization, and maternal education because heterogeneity within those response categories was expected (26, 60). The missing values fall within the acceptable imputation range of less than 10% missing (60). Hot deck imputation is the most common method in demographic surveys, is used by the US Census Bureau for correcting item nonresponse, and is a relatively simple procedure to implement (61).

Analytic approach
We used propensity score matching methods to improve exchangeability by balancing covariates across exposure groups (segregated and nonsegregated), thus reducing observable bias while maintaining support of the data (62, 63). The difference in disease frequency under the 2 exposure distributions is an estimate of the independent causal effect of segregation on black infant mortality (64).

For each infant, we used logistic regression to estimate the predicted probability of an infant's exposure to segregation given the observed covariates and interaction terms. We matched infants with the same predicted probability of exposure (i.e., propensity score) to segregation—only some were actually exposed and some were not—by using the PSMATCH2 module (65) of the Stata statistical software program (66). Since matching is the most exact method for balancing covariates (63), exposed infants were matched 1:1 with replacement to unexposed infants with the same predicted probability of exposure to segregation within a range of ±0.01. Standardized differences were calculated to assess covariate balance across exposure groups, with a goal of a standardized difference of less than 10% indicating sufficient balance (67). The estimated number of deaths that would have occurred among infants born in segregated places if instead they had been born in nonsegregated places is the average effect of the treatment on the treated (65). Bootstrap methods were used to calculate appropriate standard errors. Sensitivity analysis was conducted by using Rosenbaum's test bounds (68) as a way to assess hidden bias.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
There were 677,777 black infants born between 2000 and 2002 in the 64 US places with a population greater than 250,000. A total of 9,003 black infant deaths occurred (overall infant mortality rate = 13.3 per 1,000 livebirths).

Table 1 displays the crude black infant mortality rates by maternal characteristics for the study population as a 3-year average. Twenty-nine places were classified as segregated (isolation index ≥0.60), and 35 places were classified as nonsegregated (isolation index <0.60). The mean percentage of the population living below the poverty level was 21.6% in segregated cities compared with 14.5% in nonsegregated cities. Of the total population, we found that 42.8% of residents in segregated cities are black compared with 12.2% in nonsegregated cities.


View this table:
[in this window]
[in a new window]

  		Table 1. Infant Mortality Rates per 1,000 Livebirths for US Black Infants by Maternal and Infant Characteristics, Linked Birth/Infant Death Records, 2000–2002

 
Propensity score matching yielded 100% matching of infants exposed to segregation to 1 control in nonsegregated cities. Table 2 shows covariate imbalance of select covariates as defined for the predictive model, level of imbalance prior to and after matching, and percentage bias reduction for all black infants. Because the sample size for this study was quite large and small differences may present as statistically different, we instead focused on the standardized difference. Only 1 covariate level was greater than 10% prior to matching (mother's education). All covariates were balanced (<10%) after matching. Sensitivity analysis using Rosenbaum's test bounds (68) yielded tight confidence bounds around the log odds of differential assignment due to unobserved factors and the very small Hodges-Lehmann point estimate, revealing that unmeasured confounding was inconsequential.


View this table:
[in this window]
[in a new window]

  		Table 2. Select Covariate Imbalance Across Exposure Groups Prior to and After Matching for All US Black Study Infants, 2000–2002

 
There were 1.12 excess infant deaths per 1,000 livebirths among black infants due to living in a segregated city compared with a nonsegregated city, although the difference was not statistically significant. In fact, no statistical difference in the number of infant deaths per 1,000 livebirths was found for black infants residing in segregated cities compared with those born in nonsegregated places regarding overall infant death (rate difference = 1.12, Z = 1.35, P = 0.178), postneonatal deaths (rate difference = 0.99, Z = 1.63, P = 0.103), or external causes of death (rate difference = 0.53, Z = 1.39, P = 0.164). Refer to Table 3 for the rate difference in the number of infant deaths per exposure category per 1,000 livebirths.


View this table:
[in this window]
[in a new window]

  		Table 3. Independent Effect of Racial Residential Segregation on Black Infant Deaths per 100,000 Livebirths, US Cities With a Population >250,000, 2000–2002

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
We estimated the independent effect of living in a segregated place on black infant mortality, focusing on the postneonatal period. We hypothesized that segregation has a harmful effect on black infant survival due to macrosocial, social and environmental, and individual-level risk exposures. However, we found limited statistical evidence that segregation independently caused excess deaths of black infants when comparing exchangeable populations of black infants born in segregated and nonsegregated cities.

Our descriptive findings confirm that blacks have a high infant mortality rate regardless of maternal characteristics. Compared with nonsegregated cities, segregated cities, as defined in this study, innately have higher populations of black residents but also have a higher number of residents living in poverty. A vast literature describes the effects of poverty on infant survival and the joint relation between segregation and poverty (4, 42, 47, 57, 69–72). Economic deprivation often accompanies segregation, accentuating the common mechanisms of each. Middle-class blacks are not immune to the joint effect because the majority of middle-class black households are located within 4 blocks of impoverished neighborhoods and thus are still exposed to degrees of deprivation (4, 42, 47, 57, 69–72).

Our findings contradict previously published research on racial residential segregation and infant mortality (9–13, 16, 19, 20, 49, 50). Yankauer's landmark study in 1950 (13) demonstrated the steady and significant increase in nonwhite infant mortality as the proportion of nonwhite infant births per area increased. In a follow-up study, he found that, although infant mortality rate declined overall, the disadvantage for the nonwhite population remained unchanged and perhaps worsened (12).

After a gap of nearly 40 years, LaVeist (19, 20, 49) conducted a series of studies focused on 176 cities with populations greater than 50,000 residents and 10% or more black residents using the dissimilarity index. His studies consistently found that segregation was positively associated with black infant mortality rate and postneonatal mortality rate after adjusting for socioeconomic factors and geographic region. Polednak (10, 11, 48) conducted 3 studies by using standard Metropolitan Statistical Areas and the dissimilarity index. Segregation was found to have an independent association with infant mortality and was the most important predictor of infant mortality. Bird (8) assessed whether state-level structural variables modify the relation between states’ infant mortality for blacks and whites and found that racial composition, education, poverty, and segregation contributed to black infant mortality at the state level. Finally, Guest et al. (16) looked at both infant mortality and working-age mortality in Chicago, Illinois, by using the isolation index and found that segregation predominantly increases black mortality through differential socioeconomic patterns but also appears to have an independent "deleterious" effect.

These previous studies contributed to the scientific literature, but all were ecologic. Ecologic study designs may suffer the ecologic fallacy (73). Our use of individual-level covariates advanced the work in this area and may account for the different results. Our use of propensity score matching methods offers a potentially more rigorous design. We also conducted a comparable analysis by using traditional logistic regression and found no significant difference in infant mortality rate between blacks living in segregated and nonsegregated cities. The value of propensity score matching methods is that it makes explicit the exchangeability between comparison groups. Additionally, to our knowledge only 1 study (16) previously used the isolation index. The choice of segregation dimension may matter in health research (43).

We do not claim that segregation has no effect on the black infant mortality rate but rather that it is difficult, if not impossible, to disentangle the individual effects from contextual effects. Our results suggest that the effects of segregation as defined here are not larger than the characteristics of the individuals when comparing exchangeable populations or that they are not independent from the individuals who live in the segregated cities. There is no group without the specific individuals who constitute it (74), and distinct categories of contextual versus compositional effects themselves are contrived distinctions (75). What differentiates a segregated city from a nonsegregated city is the cumulative effect of the people and the policies, both formal and informal, that are in place.

There are limitations to our approach. The choice of covariates predicting segregation may be on the causal path, and separating the characteristics of the mother and infant ignores a life course perspective. We attempted to address both of these issues by focusing primarily on postneonatal mortality, which is more distal to conditions of birth. Measurement of segregation, as with many social measures, is a function of the cutpoints chosen, the measurement tool used, and the assumptions of the meaningfulness of segregation at and near the cutpoints. Another limitation is restriction of the analysis to the city level. It could be that the relation between segregation and birth outcomes varies depending on the size of the geographic area and number of residents. Patterns of segregation are seen at the Metropolitan Statistical Area level, resulting in possible selection bias because urban dwellers may differ from suburban dwellers. At the same time, the city-level segregation may not accurately reflect the experiences of blacks within the city. Analysis at the census tract level may provide a different estimate of individual-level exposures. We were unable to account for heterogeneity of exposure, both within city and over time. We were able to account for the city in which the infant was born but not if that infant then moved, possibly altering the infant's social and environmental exposures.

Finally, propensity score methods also have limitations. Even though they enabled us to more closely mimic experimental design, minimize emphasis on P values for model specification, and make inference more transparent than traditional regression techniques do, the 2 most important limitations are unmeasured confounding and poor overlap. Propensity score methods do not account for unmeasured confounding, as would perfect randomization. There is no direct way to test unmeasured confounding, but our sensitivity analysis showed it to be largely inconsequential here. Subjects without a match, no overlap or propensity scores, were dropped from the analysis, leading to selection bias. Despite these limitations, our data were of high quality, measurement of outcomes was reliable, and the analytic approach was rigorous. This study offers an important advancement to both methods and the understanding of contextual effects on health.

In conclusion, future work is needed in many areas, including understanding the effect of segregation at different geographic levels (i.e., Metropolitan Statistical Area), further exploring the mechanisms, understanding the experience and exposures of black women and infants in nonsegregated areas, and incorporating historical migratory patterns in the making of segregation and its effect on health. The infant mortality rate continues to require public health and policy attention. Although this study suggests no independent effect of racial residential segregation on black infant mortality rate, we emphasize the difficulties separating composition from context and cautious interpretation.


   ACKNOWLEDGMENTS
 
Author affiliations: Division of Epidemiology and Community Health, School of Public Health, University of Minnesota, Minneapolis, Minnesota (Mary O. Hearst, J. Michael Oakes); and Health Services Research and Policy, School of Public Health, University of Minnesota, Minneapolis, Minnesota (Pamela J. Johnson).

This study was supported by a doctoral dissertation fellowship from the University of Minnesota Graduate School.

The authors thank Prof. Myron Orfield, Dr. James Pankow, and Dr. Eileen Harwood for their guidance as dissertation committee members.

Conflict of interest: none declared.


   NOTES
 
Editor's note: An invited commentary on this article is published on page 1255.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Wegman ME. Infant mortality in the 20th century, dramatic but uneven progress. J Nutr (2001) 131(2):401S–408S.[Free Full Text]
  2. Moss N, Carver K. The effect of WIC and Medicaid on infant mortality in the United States. Am J Public Health (1998) 88(9):1354–1361.[Abstract/Free Full Text]
  3. Brooks CH. Social, economic, and biologic correlates of infant mortality in city neighborhoods. J Health Soc Behav (1980) 21(1):2–11.[CrossRef][Web of Science][Medline]
  4. Collins CA, Williams DR. Segregation and mortality: the deadly effects of racism? Sociol Forum (1999) 14:495–523.[CrossRef]
  5. Buka SL, Brennan RT, Rich-Edwards JW, et al. Neighborhood support and the birth weight of urban infants. Am J Epidemiol (2003) 157(1):1–8.[Abstract/Free Full Text]
  6. Collins JW Jr, David RJ. The differential effect of traditional risk factors on infant birthweight among blacks and whites in Chicago. Am J Public Health (1990) 80(6):679–681.[Abstract/Free Full Text]
  7. Fang J, Madhavan S, Alderman MH. Low birth weight: race and maternal nativity—impact of community income [electronic article]. Pediatrics (1999) 103(1):E5.[CrossRef][Medline]
  8. Bird ST. Separate black and white infant mortality models: differences in the importance of structural variables. Soc Sci Med (1995) 41(11):1507–1512.[CrossRef][Web of Science][Medline]
  9. Bird ST, Bauman KE. The relationship between structural and health services variables and state-level infant mortality in the United States. Am J Public Health (1995) 85(1):26–29.[Abstract/Free Full Text]
  10. Polednak AP. Trends in US urban black infant mortality, by degree of residential segregation. Am J Public Health (1996) 86(5):723–726.[Abstract/Free Full Text]
  11. Polednak AP. Black-white differences in infant mortality in 38 standard metropolitan statistical areas. Am J Public Health (1991) 81(11):1480–1482.[Abstract/Free Full Text]
  12. Yankauer A, Allaway NC. The relation of indices of fetal and infant loss to residential segregation: a follow-up report. Am Soc Rev. (1958) 23:573–578.[CrossRef]
  13. Yankauer A Jr. The relationship of fetal and infant mortality to residential segregation: an inquiry into social epidemiology. Am Soc Rev. (1950) 15:644–648.[CrossRef]
  14. Willie CV, Rothney WB. Racial, ethnic, and income factors in the epidemiology of neonatal mortality. Am Soc Rev. (1962) 27:522–526.[CrossRef]
  15. Jiobu RM. Urban determinants of racial differentiation in infant mortality. Demography (1972) 9(4):603–615.[CrossRef][Web of Science][Medline]
  16. Guest AM, Almgren G, Hussey JM. The ecology of race and socioeconomic distress: infant and working-age mortality in Chicago. Demography (1998) 35(1):23–34.[Web of Science][Medline]
  17. Hogue CJ, Vasquez C. Toward a strategic approach for reducing disparities in infant mortality. Am J Public Health (2002) 92(4):552–556.[Abstract/Free Full Text]
  18. Hummer RA. Racial differentials in infant mortality in the U.S.: an examination of social and health determinants. Soc Forces (1993) 72:529–554.[CrossRef][Web of Science]
  19. LaVeist TA. Simulating the effects of poverty on the race disparity in postneonatal mortality. J Public Health Policy (1990) 11(4):463–473.[CrossRef][Medline]
  20. LaVeist TA. Segregation, poverty, and empowerment: health consequences for African Americans. Milbank Q (1993) 71(1):41–64.[CrossRef][Web of Science][Medline]
  21. Mathews TJ, Menacker F, MacDorman MF. Infant Mortality Statistics From the 2000 Period Linked Birth/Infant Death Data Set. National Vital Statistics Reports (2002) vol 50(no. 12). Hyattsville, MD: National Center for Health Statistics.
  22. Mathews TJ, Menacker F, MacDorman MF. Infant Mortality Statistics From the 2001 Period Linked Birth/Infant Death Data Set. National Vital Statistics Reports (2003) vol 52(no. 2). Hyattsville, MD: National Center for Health Statistics.
  23. Mathews TJ, Menacker F, MacDorman MF. Infant Mortality Statistics From the 2002 Period Linked Birth/Infant Death Data Set. National Vital Statistics Reports (2004) vol 53(no. 10). Hyattsville, MD: National Center for Health Statistics.
  24. Singh G, Yu S. Infant mortality in the United States: trends, differentials, and projections, 1950 through 2010. Am J Public Health (1995) 85(7):957–964.[Abstract/Free Full Text]
  25. Schoendorf KC, Hogue CJ, Kleinman JC, et al. Mortality among infants of black as compared with white college-educated parents. N Engl J Med (1992) 326(23):1522–1526.[Abstract]
  26. Rothman KJ, Greenland S. Modern Epidemiology (1998) 2nd ed. Philadelphia, PA: Lippincott-Raven, Publishers.
  27. Hauck FR, Moore CM, Herman SM, et al. The contribution of prone sleeping position to the racial disparity in sudden infant death syndrome: the Chicago Infant Mortality Study. Pediatrics (2002) 110(4):772–780.[Abstract/Free Full Text]
  28. Beckwith JB. Defining the sudden infant death syndrome. Arch Pediatr Adolesc Med (2003) 157(3):286–290.[Abstract/Free Full Text]
  29. Shields LB, Hunsaker DM, Muldoon S, et al. Risk factors associated with sudden unexplained infant death: a prospective study of infant care practices in Kentucky [electronic article]. Pediatrics (2005) 116(1):e13–e20.[Abstract/Free Full Text]
  30. Gibson E, Fleming N, Fleming D, et al. Sudden infant death syndrome rates subsequent to the American Academy of Pediatrics supine sleep position. Med Care (1998) 36(6):938–942.[CrossRef][Web of Science][Medline]
  31. Centerwall BS. Race, socioeconomic status, and domestic homicide, Atlanta, 1971–72. Am J Public Health (1984) 74(8):813–815.[Abstract/Free Full Text]
  32. Brenner RA, Overpeck MD, Trumble AC, et al. Deaths attributable to injuries in infants, United States, 1983–1991. Pediatrics (1999) 103(5 pt 1):968–974.[Abstract/Free Full Text]
  33. Jain A, Khoshnood B, Lee KS, et al. Injury related infant death: the impact of race and birth weight. Inj Prev (2001) 7(2):135–140.[Abstract/Free Full Text]
  34. Anderson R, Smith B. Deaths: Leading Causes for 2002. National Vital Statistics Reports (2005) vol. 53(no. 17). Hyattsville, MD: National Center for Health Statistics.
  35. Anderson R. Deaths: Leading Causes for 2000. National Vital Statistics Report (2002) vol. 50(no 16). Hyattsville, MD: National Center for Health Statistics.
  36. Anderson R, Smith B. Deaths: Leading Causes for 2001. National Vital Statistics Reports (2003) vol. 52(no 9). Hyattsville, MD: National Center for Health Statistics.
  37. Tomashek KM, Hsia J, Iyasu S. Trends in postneonatal mortality attributable to injury, United States, 1988–1998. Pediatrics (2003) 111(5 pt 2):1219–1225.[Abstract/Free Full Text]
  38. Unger B, Kemp JS, Wilkins D, et al. Racial disparity and modifiable risk factors among infants dying suddenly and unexpectedly [electronic article]. Pediatrics (2003) 111(2):E127–E131.[CrossRef][Medline]
  39. Drago DA, Dannenberg AL. Infant mechanical suffocation deaths in the United States, 1980–1997 [electronic article]. Pediatrics (1999) 103(5):e59.[Abstract/Free Full Text]
  40. Shenassa ED, Stubbendick A, Brown MJ. Social disparities in housing and related pediatric injury: a multilevel study. Am J Public Health (2004) 94(4):633–639.[Abstract/Free Full Text]
  41. Collins JW Jr, Schulte N. Infant health: race, risk, and residence. In: Neighborhoods and Health—Kawachi I, Berkman L, eds. (2003) New York, NY: Oxford University Press.
  42. Acevedo-Garcia D, Lochner K. Residential segregation and health. In: Neighborhoods and Health—Kawachi I, Berkman L, eds. (2003) New York, NY: Oxford University Press.
  43. Acevedo-Garcia D, Lochner KA, Osypuk TL, et al. Future directions in residential segregation and health research: a multilevel approach. Am J Public Health (2003) 93(2):215–221.[Abstract/Free Full Text]
  44. Williams DR, Collins C. Racial residential segregation: a fundamental cause of racial disparities in health. Public Health Rep (2001) 116(5):404–416.[Web of Science][Medline]
  45. Reardon S. A conceptual framework for measuring segregation and its association with population outcomes. In: Methods in Social Epidemiology—Oakes J, Kaufmann J, eds. (2006) San Francisco, CA: Jossey-Bass.
  46. Orfield M, Luce T. Minority suburbanization and racial change: stable integration, neighborhood transition, and the need for regional approaches [abstract]. Presented at the Race and Regionalism Conference, Minneapolis, Minnesota, May 6–7, 2005.
  47. Massey DS, Condran GA, Denton NA. The effect of residential segregation on black social and economic well-being. Soc Forces (1987) 66:29–56.[CrossRef][Web of Science]
  48. Polednak AP. Segregation, Poverty, and Mortality in Urban African Americans (1997) New York, NY: Oxford University Press.
  49. LaVeist TA. Linking residential segregation to the infant-mortality race disparity in U.S. cities. Sociol Soc Res. (1989) 73:90–94.[Web of Science]
  50. Polednak AP. Poverty, residential segregation, and black/white mortality ratios in urban areas. J Health Care Poor Underserved (1993) 4(4):363–373.[Medline]
  51. National Center for Health Statistics. NCHS's Vital Statistics Perinatal Mortality Data—1995–2002. In: National Bureau of Economic Research, ed. Atlanta, GA; 2005. Updated October 13, 2005. Accessed December 2005. (http://www.nber.org/data/vital-statistics-perinatal-mortality-data.html).
  52. US Census. U.S. Census Bureau Housing Patterns. Washington, DC; 2005. Updated June 27, 2005. Accessed January, 2006. (http://www.census.gov/hhes/www/housing/housing_patterns/housing_patterns.html).
  53. Centers for Disease Control and Prevention. Linked Birth/Infant Death Data Set: 2001 Period Data Public-use Documentation. In: National Vital Statistics System (NVSS), ed. Mortality Data from the National Statistics System. Atlanta, GA, CDC; 2004. Accessed January 2006. (http://www.cdc.gov/nchs/about/major/dvs/linked/01linked.htm).
  54. US Geographical Survey. Geographic Names Information System: FIPS55 Data. Reston, VA; 2005. Updated December 28, 2005. Accessed January 2006. (http://geonames.usgs.gov/fips55.html).
  55. Iceland J, Weinberg D, Steinmetz E. Racial and Ethnic Residential Segregation in the United States: 1980–2000 (2002) Atlanta, GA: US Census Bureau. 1–31.
  56. Massey DS, Denton NA. The dimensions of residential segregation. Soc Forces (1988) 67:281–315.[CrossRef][Web of Science]
  57. Grady SC. Racial disparities in low birthweight and the contribution of residential segregation: a multilevel analysis. Soc Sci Med (2006) 63(12):3013–3029.[CrossRef][Web of Science][Medline]
  58. Acevedo-Garcia D. Residential segregation and the epidemiology of infectious diseases. Soc Sci Med (2000) 51(8):1143–1161.[CrossRef][Web of Science][Medline]
  59. Massey DS, Denton NA. Hypersegregation in U.S. metropolitan areas: black and Hispanic segregation along five dimensions. Demography (1989) 26(3):373–391.[Web of Science][Medline]
  60. Kalton G, Kish G. Two Efficient Random Imputation Procedures. Survey Research Methods Section (1981) Alexandria, VA: American Statistical Association.
  61. Davern M, Blewett L, Bershadsky B, et al. Missing the mark? Imputation bias in the current population survey's state income and health insurance coverage estimates. J Off Stat (2004) 20:519–549.
  62. Oakes JM, Johnson PJ. Propensity score matching for social epidemiology. In: Methods in Social Epidemiology—Oakes JM, Kaufman JA, eds. (2006) San Francisco, CA: Jossey-Bass.
  63. Rubin DB. Estimating causal effects from large data sets using propensity scores. Ann Intern Med (1997) 127(8 pt 2):757–763.[Abstract/Free Full Text]
  64. Maldonado G, Greenland S. Estimating causal effects. Int J Epidemiol (2002) 31(2):422–429.[Free Full Text]
  65. Lueven E, Sianesi B. PSMATCH2: Stata Module to Perform Full Mahalanobis and Propensity Score Matching, Common Support Graphing, and Covariate Imbalance Testing (2005) Chestnut Hill, MA: Department of Economics, Boston College.
  66. Stata Corporation. Stata statistical software, release 9.0. (2005) College Station, TX: Stata Corporation.
  67. D'Agostino R. Tutorial in biostatistics: propensity scores methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med (1998) 17:2265–2281.[CrossRef][Web of Science][Medline]
  68. Rosenbaum PR. Heterogeneity and causality: unit heterogeneity and design sensitivity in observational studies. Am Stat (2005) 59:147–152.[CrossRef]
  69. Wilson W. The Truly Disadvantaged: The Inner City, the Underclass and Public Policy (1987) Chicago, IL: University of Chicago Press.
  70. Orfield G. Segregation and Its Consequences. Race & Regionalism (2005) Minneapolis, MN: Institute on Race and Poverty.
  71. Orfield M. American Metropolitics: The New Suburban Reality (2002) Washington, DC: The Brookings Institute.
  72. Massey DS. American apartheid: segregation and the making of the underclass. Am J Soc. (1990) 96:329–357.[CrossRef]
  73. Morgenstern H. Uses of ecologic analysis in epidemiologic research. Am J Public Health (1982) 72(12):1336–1344.[Abstract/Free Full Text]
  74. Oakes JM, Kaufman JA. Introduction: advancing methods in social epidemiology. In: Methods in Social Epidemiology—Oakes JM, Kaufman J, eds. (2006) San Francisco, CA: Jossey-Bass.
  75. Kaufman J. The (Mis)Conception of Neighborhood Effects. Proceedings of the American Statistical Association, Health Policy Statistics Section [CD-ROM] (2006) Alexandria, VA.

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?

Related articles in Am. J. Epidemiol.:

Invited Commentary: Residential Segregation and Health—The Complexity of Modeling Separate Social Contexts
Dolores Acevedo-Garcia and Theresa L. Osypuk
Am. J. Epidemiol. 2008 168: 1255-1258. [Abstract] [FREE Full Text]  



This article has been cited by other articles:


Home page
 Am J EpidemiolHome page
L. C. Messer, J. M. Oakes, and S. Mason
Effects of Socioeconomic and Racial Residential Segregation on Preterm Birth: A Cautionary Tale of Structural Confounding
Am. J. Epidemiol., February 5, 2010; (2010) kwp435v1.
[Abstract] [Full Text] [PDF]

Home page
 Epidemiol RevHome page
M. R. Kramer and C. R. Hogue
Is Segregation Bad for Your Health?
Epidemiol. Rev., November 1, 2009; 31(1): 178 - 194.
[Abstract] [Full Text] [PDF]

Home page
 Int J EpidemiolHome page
J M. Oakes
Commentary: Individual, ecological and multilevel fallacies
Int. J. Epidemiol., April 1, 2009; 38(2): 361 - 368.
[Full Text] [PDF]

Home page
 Am J EpidemiolHome page
D. Acevedo-Garcia and T. L. Osypuk
Invited Commentary: Residential Segregation and Health--The Complexity of Modeling Separate Social Contexts
Am. J. Epidemiol., December 1, 2008; 168(11): 1255 - 1258.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
168/11/1247    most recent
kwn291v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Related articles in Am. J. Epidemiol.
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (1)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Hearst, M. O.
Right arrow Articles by Johnson, P. J.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Hearst, M. O.
Right arrow Articles by Johnson, P. J.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?