 |
|
 |
|
 |
 |
|
 |
 |
 |
 |
 |
 |
 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
ARTICLE |
a School of Social Work, Columbia University, New York, New York
b Department of Pediatrics, Robert Wood Johnson Medical School, New Brunswick, New Jersey
 |
ABSTRACT |
|---|
 TOP ABSTRACT BackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
METHODS. Risk factors and rates of low birth weight were compared across groups for singleton births within each country (white, black, Hispanic, Asian, and American Indian mothers in the United States; white, black, and Asian mothers in England). Crude rates and rates adjusted for socioeconomic status and behaviors were compared. Additional comparisons were limited to native-born mothers.
RESULTS. Racial and ethnic disparities in low birth weight are as large in England as in the United States. Socioeconomic status and behaviors explain little of the variation across racial and ethnic groups in either country.
CONCLUSIONS. Health disadvantages associated with being a minority do not seem to be a uniquely American phenomenon. Universal health care, as provided in the United Kingdom, alone may be insufficient to reduce racial and ethnic disparities in low birth weight.
Key Words: birth weight • racial disparities • ethnic disparities • international comparison
Abbreviations: SES—socioeconomic status • ECLS-B—Early Childhood Longitudinal Survey-Birth Cohort • MCS—New Millennium Cohort Study
Health disparities are an important marker of inequality in a society. In the United States, there are large socioeconomic disparities in health.1 There are also large racial and ethnic health disparities,1 even among infants. In 2001, 13.0% of infants born to black mothers in the United States were low birth weight (<2500 g) compared with 6.7% of those born to white mothers, 6.5% of those born to Hispanic mothers (of any race), 7.3% of those born to American Indian mothers, and 7.4% of those born to Asian and Pacific Islander mothers.2 Although low birth weight is not a direct measure of infant morbidity, it is frequently used as a marker for poor health at birth because it is a leading risk factor for infant mortality and for subsequent morbidity among surviving infants.3 It is also accurately measured, widely available, and useful for subnational and international comparisons.
Despite substantial research on determinants of racial and ethnic disparities in birth outcomes in the United States, much remains to be explained. Cigarette smoking during pregnancy is strongly associated with low birth weight but does not explain the black-white disparity.4 Factors that are related to low socioeconomic status (SES), such as young maternal age and low educational attainment, are associated with low birth weight5 but explain little of the black-white disparity.
The extent to which racial and ethnic disparities in low birth weight reflect socioeconomic inequalities remains an open question, because most representative data on births by race do not include detailed measures of SES and because the association does not seem to hold for certain groups. In the United States, Hispanic mothers as a broad group have favorable birth outcomes despite their low SES, and across virtually all racial and ethnic groups, foreign-born mothers have lower rates of low birth weight than their more advantaged US-born counterparts.6 Proposed explanations for the Hispanic and immigrant advantages involve selective migration on the basis of health7 and protective cultural factors.8–10
Little is known about racial and ethnic disparities in birth outcomes in developed countries other than the United States, because very few countries collect vital statistics data by race and ethnicity. Only recently have national data on births by race become available in the United Kingdom, where a socioeconomic gradient in health similar to that in the United States exists despite substantial differences in the provision of health care and public assistance systems.11,12
The overall rate of low birth weight in the United Kingdom in 2000 was 7.6%,13 the same as the US rate in 2000.5 As in the United States, low birth weight in the United Kingdom is negatively associated with SES.14,15 A recent report indicated that racial and ethnic disparities in low birth weight do exist.16
We used data from 2 new and remarkably comparable nationally representative birth cohort studies to compare disparities in low birth weight in the United States and England. We compared crude disparities and disparities adjusting for SES and prenatal behaviors. We were interested primarily in comparing the levels of within-country disparities. Cross-country comparisons of specific groups may be interesting, but it is critical to keep in mind that groups often referred to with the same label in the United States and England (eg, black or Asian) consist of individuals with very different histories and ancestries.
|
Background |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
In the United States, 75.1% of the population is white. The vast majority of black people in the United States, who represent 12.3% of the nation's population,18 are descendents of African slaves. A minority (but an increasing number) of black people in the United States are immigrants from the Caribbean and sub-Saharan Africa.19 Asian people (primarily Chinese, Filipino, Indian, Vietnamese, Korean, and Japanese) comprise 3.6% of the US population; American Indian, 0.9%; other races, 5.6%; and 
2 races, 2.4%.18 A large share (12.6% and growing) of the US population is Hispanic (of any race), mostly of Mexican descent.20 Finally, 19.9% of the US population is foreign born (5% of white population, 4.9% of black population, 2.3% of American Indian population, 63.1% of Asian population, and 35.8% of Hispanic population).21
In the United Kingdom, 91% of the population is white. The majority of nonwhite people in the United Kingdom are black or Asian. Black people represent 2.3% of the population and are primarily of Caribbean or African ancestry. Asian people from the Indian subcontinent represent 4.6% of the United Kingdom population and are primarily Indian, Pakistani, and Bangladeshi. The remaining 2.2% of the population is of other races or ethnicities (which includes Chinese).22 In 2001, 8.3% of the United Kingdom population was born outside of the United Kingdom (a third of immigrants were born in other European countries).23 Because the bulk of Caribbean, Asian, and African immigration to the United Kingdom has occurred since World War II, most minority individuals in the United Kingdom are first- or second-generation immigrants.
Given the universal access to health care, minimum guaranteed income, and large representation of immigrants (who may be more favorably selected on the basis of health or behaviors than the native born) in minority groups in the United Kingdom, we expected that racial and ethnic disparities in low birth weight in England would not be as pronounced as those in the United States.
|
METHODS |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
15 years old who did not place the child for adoption.24 The US analyses are based on a subsample of 8300 singleton births with nonmissing data on all of the covariates. All unweighted ECLS-B sample sizes are rounded per contractual data user requirements. The MCS is a birth cohort study of 18553 children born in the United Kingdom in 2000–2001. It is representative of all children born in the United Kingdom between 2000 and 2001 who were living in the United Kingdom at 9 months of age and who were eligible to receive the Child Benefit (a universal benefit for all residents of the United Kingdom) at that time. A follow-up survey, which was conducted between September 2003 and April 2005, had an 80% response rate (of mothers who completed initial interviews). We limited our analyses to England because the ethnic minority sample sizes in Northern Ireland, Scotland, and Wales are very small.25 We used a subsample of 10599 singleton births in England with nonmissing data on all of the covariates. The ECLS-B and MCS are highly comparable, because they are from the same year, they are both nationally representative of 9-month infant survivors, and they both include data on many pertinent risk factors.
Measures
The outcome measure is low birth weight (<2500 g), which in the ECLS-B comes from the infants' birth certificates and in the MCS comes from maternal reports of their infant's birth weight. In the United States, maternal reports of birth weight are highly consistent with those from birth certificates.26–28
Racial and ethnic categories that are meaningful or routinely used in each country were used. For the United States, the analysis groups were white, black, Hispanic, Asian, and American Indian. Hispanic mothers were not included in the other racial and ethnic categories. For England, the analysis groups were white, black (mostly of Caribbean and African descent), and Asian from the Indian subcontinent (Indian, Pakistani, and Bangladeshi). The race and ethnicity data came from the main respondent in both surveys, and the birth mother's race or ethnicity was used.
The following analysis variables are comparable in the 2 data sets: maternal age (<20 years, 20–34 years, and 35 years), parity (first birth versus higher-order birth), marital status, nativity (native versus foreign born), any prenatal care received, and any cigarette smoking during pregnancy. Both data sets also include information on alcohol consumption during pregnancy, but that behavior is not included in our analyses because of a high rate of missing data in the ECLS-B. The mother's birthplace comes from the birth certificates for the ECLS-B (island-born Puerto Rican mothers are categorized as native born) and from the follow-up survey for the MCS. Prenatal care and smoking are from self-reports in the MCS and from birth certificates in the ECLS-B. For England, the measure of employment is any work activity during pregnancy, and for the United States, it is any work activity during the 12 months before the birth.
Measures of poverty and education that are meaningful for each country, but not directly comparable, were used. For the United States, if a family's household income fell below the official poverty income threshold (adjusted by household size) for the survey year, that mother was considered poor. For England, the measure of poverty was receipt of means-tested benefits, because there was a high rate of missing data on income in the MCS.
The US education variable consisted of 4 categories: less than high school, high school diploma or General Education Development, some college, and a bachelor's degree or higher. The United Kingdom education variable also consisted of 4 categories: less than O level (0 to 
11 years of education), at least O level but less than A level (12–13 years of education), A level or higher, and other qualifications (eg, education obtained outside of the United Kingdom). As indicated above, these categories were not directly comparable with the US educational classifications.
Analyses
We compared the following across racial and ethnic groups in each country: (1) risk factors for low birth weight; (2) rates of low birth weight (crude, as well as adjusted for infant gender, SES, and behaviors); and (3) rates of low birth weight among native-born mothers. All of the analyses were based on weighted data. The analyses of native-born mothers in the United States were based on the 5900 mothers in the ECLS-B analysis sample who reported in their baseline interview that they were born in the United States. The analyses of native-born mothers in England were based on the 7212 mothers in the MCS analysis sample who completed the follow-up survey and reported at that time that they were born in the United Kingdom. We assessed sensitivity of the findings with an alternative birth weight threshold and with an alternative poverty measure for the United States.
|
RESULTS |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
|
In the United States, most Hispanic and Asian mothers were foreign born, and the vast majority of non-Hispanic white, non-Hispanic black, and American Indian mothers were native born. In England, <5% of white mothers were foreign born, whereas substantial proportions of Asian (61%) and black (43%) mothers were born outside of the United Kingdom.
In both countries, there were large differences in poverty status across racial and ethnic groups. In particular, black mothers were 3 times as likely as white mothers to be classified as poor in both countries based on the country-specific measures that we used. In the United States, Hispanic and American Indian mothers were 3 times as likely as non-Hispanic white mothers to be poor. The poverty rates of Asian mothers were fairly similar to those of white mothers in each country. The racial and ethnic patterns in educational attainment did not follow the patterns in poverty. In the United States, Hispanic mothers had lower rates of high school completion than black mothers, despite a lower poverty rate. In England, black mothers were as likely as white mothers to have an A-level or higher degree despite a substantially higher rate of poverty, and Asian mothers had lower levels of educational attainment than white mothers despite similar poverty levels.
Prenatal care use was widespread in both countries across all of the groups. More than 90% of each racial and ethnic group in each country received some prenatal care. Other prenatal behaviors varied across racial and ethnic groups in both countries, but in a paradoxical way, because white mothers (who had the lowest rates of low birth weight) were more likely to report engaging in behaviors that are associated with low birth weight. In particular, cigarette smoking was high among white mothers in both countries. Asian mothers in both countries reported very low rates of smoking during pregnancy (3% and 5.5%, respectively, in the United States and England).
In sum, in both countries, minority mothers differed from each other and from white mothers in terms of risk factors for low birth weight. In the United States, Asian mothers shared the socioeconomic advantages of non-Hispanic white mothers and also had a low rate of cigarette smoking (ie, they had the most favorable risk factor profile of all of the groups). Black and American Indian mothers were at the greatest risk for low birth weight based on their risk factors, with Hispanic mothers falling between black and white mothers. American Indian mothers were the only group with a higher rate of cigarette smoking than white mothers. In England, Asian mothers had the most favorable risk factor profile of all of the groups, and black mothers, although more likely than white mothers to be poor, had other offsetting factors. They were less likely to smoke and more likely to be immigrants, which, in the United States, is associated with higher birth weight. Overall, there were similar levels of variability in risk factors for low birth weight across racial and ethnic groups in the 2 countries.
Racial and Ethnic Disparities in Low Birth Weight in the United States and England
Racial and ethnic rates of low birth weight in the United States and England are shown in Table 2. For each country, the first row shows the crude rate of low birth weight, the second row shows the percentage of low birth weight adjusted for infant gender and socioeconomic characteristics (standardized to the white mothers in each country), and the third row shows the percentage of low birth weight adjusted for gender, socioeconomic characteristics, smoking, and employment (also standardized to the white mothers).
|
In the United States, adjusting for SES somewhat reduced the disparities between white mothers and those who are black, Hispanic, and American Indian. The SES-adjusted rates of low birth weight were 8.9%, 4.6%, 6.8%, and 5.0% for black, Hispanic, Asian, and American Indian mothers, respectively. Adjusting for SES did little to explain disparities in England; the SES-adjusted rates of low birth weight were 8.9% and 12.2% for black and Asian mothers, respectively. Adjustments for smoking and maternal employment also did not explain disparities in either country. SES and behavior-adjusted rates in the United States were 9.8%, 5.8%, 7.3%, and 4.5% for black, Hispanic, Asian, and American Indian mothers, respectively. The rates in England were 9.6% for black mothers and 12.9% for Asian mothers.
We conducted a corresponding set of comparisons for native-born mothers. The results are shown in Table 3. The racial and ethnic disparities in low birth weight among native-born mothers were similar to those for all mothers in each country. Again, SES explained only some of the disparities in the United States and very little of the disparities in England, and behaviors explained none of the disparities in either country. Full results for all of the mothers and native-born mothers, from logistic regressions, are shown in Appendices 1 (United States) and 2 (England).
|
We also assessed the sensitivity of our findings to a more relaxed definition of poverty in the United States, whether the mother's household income was in the bottom 40% of the income distribution. The results were insensitive to this change (results not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
Unlike most previous research on racial and ethnic disparities in low birth weight in the United States, the new ECLS-B data allowed us go beyond educational attainment in characterizing SES by also including measures of employment and poverty. Thus, aside from the main contribution of this article (the international comparison), we addressed an important research gap on disparities in the United States. We found that the more detailed socioeconomic data did little to explain racial and ethnic disparities in low birth weight in the United States.
The findings that racial and ethnic disparities in low birth weight are similar in magnitude in the 2 countries and that they cannot be explained by differences in SES in either country are revealing. First, they suggest that health disadvantages associated with being a minority are not a uniquely American phenomenon. They also suggest that minority health disadvantages do not solely reflect differential access to lifetime health care, although it is possible that minority status is associated with lower quality or use of care in the United Kingdom. At a minimum, our findings suggest that providing universal health care without ensuring its quality and regular use may not reduce racial and ethnic disparities in low birth weight in the United States.
Given that SES does not explain the disparities in the United States or England, the question of what does explain disparities remains. In the United Kingdom, most minorities have ancestries from countries with very high rates of low birth weight. The rates are 30% in India, 19% in Pakistan, and 30% in Bangladesh. The rates of low birth weight in commonwealth countries in the Caribbean are somewhat higher, on average, than in the United Kingdom, but there is considerable variability among them. For example, the rates in Antigua, the Bahamas, and Jamaica range from 7% to 9%, whereas the rate in Trinidad and Tobago is 23%. Rates in African countries tend to be high (eg, 18% in Senegal, 17% in Chad, 14% in Nigeria and Mozambique, and 11% in Kenya).30 Racial and ethnic disparities in low birth weight may reflect differences in biological predisposition, but it is also possible that minorities in England suffer from residual effects of poor premigration health care, which may have lasting adverse intergenerational effects on health. This potential explanation is not likely to explain the largest disparity in the United States, which is between the black and white population, because most black people have roots in the United States that go back several generations. Another potential explanation, which remains to be tested, is that minorities are disproportionately exposed to adverse environments, including social (eg, discrimination) and physical (eg, poor housing, pollution), that take a toll on their health.
Our study is subject to certain limitations. Because of the small sample sizes of minorities in other parts of the United Kingdom, we focused only on England. There was an insufficient sample to examine disparities in very low birth weight. Low birth weight is only 1 health-related outcome, and, therefore, we cannot generalize our findings to disparities in health more generally.
|
APPENDIX 1 Multiple Logistic Regression Estimates of Associations Between Race and Ethnicity and Low Birth Weight: United States |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
|
|
APPENDIX 2 Multiple Logistic Regression Estimates of Associations Between Race and Ethnicity and Low Birth Weight: England |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
|
|
APPENDIX 3 Percentage <2000 g According to Race and Ethnicity: Singleton Births |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Address correspondence to Julien O. Teitler, PhD, Columbia University, School of Social Work, 1255 Amsterdam Ave, New York, NY 10027. E-mail: jot8{at}columbia.edu
The authors have indicated they have no financial relationships relevant to this article to disclose.
|
REFERENCES |
|---|
|
TOPABSTRACTBackgroundMETHODSRESULTSDISCUSSIONAPPENDIX 1 Multiple Logistic...APPENDIX 2 Multiple Logistic...APPENDIX 3 Percentage <2000...REFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  Y. Kelly, L. Panico, M. Bartley, M. Marmot, J. Nazroo, and A. Sacker Why does birthweight vary among ethnic groups in the UK? Findings from the Millennium Cohort Study J. Public Health Med., July 21, 2008; (2008) fdn057v1. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Bauchner Atoms Arch. Dis. Child., January 1, 2008; 93(1): i - i. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
 |
||
|
||
Home | My Pediatrics | Journal Information | Current Issue | Past Issues | Subscriptions & Services | Contact Us Click here for faster international access © 2007 American Academy of Pediatrics. All rights reserved. |
||
|
||
