Significant, yet not fully understood, racial, ethnic, and nativity-related disparities exist in women's health in the United States1,2; death during and shortly after pregnancy is no exception. Most notably, black women have a three to four times higher risk of dying from pregnancy-related causes than white women.3 For specific mortality causes (eg, ectopic pregnancy), this gap appears to be even greater.4 Research also shows that women of Hispanic origin have better pregnancy outcomes than non-Hispanic white women,5,6 yet limited data exist to explain this “Hispanic paradox.” In addition, a report comparing pregnancy-related mortality of U.S.- and foreign-born women of Hispanic origin documents a considerably higher mortality rate among foreign-born Hispanic women.7
Nearly one third of U.S. women self-identify as members of a racial or ethnic minority group, and it is estimated that this share will increase to more than half by 2045.8 Currently, almost half of U.S. births are to women other than U.S.-born non-Hispanic white.9 As the United States becomes more diverse, understanding racial, ethnic, and nativity-related disparities in maternal health becomes an even higher priority. Examination of such disparities in pregnancy-related mortality can inform the design and delivery of prevention, treatment, and care efforts targeted to specific groups of women, yet, to our knowledge, very few studies explored the associations among race, ethnicity, and nativity and pregnancy-related mortality in the United States. This analysis estimates and compares trends in and causes of pregnancy-related mortality in the United States by race, ethnicity, and nativity from 1993 to 2006.
MATERIALS AND METHODS
We used 1993–2006 data from Centers for Disease Control and Prevention (CDC)'s Pregnancy Mortality Surveillance System. Fifty-two reporting areas (ie, 50 U.S. states, New York, and Washington, DC) are asked to submit to CDC's Division of Reproductive Health deidentified copies of death certificates for all deaths occurring during or within 1 year of pregnancy regardless of the cause of death or the duration of pregnancy as well as matching birth or fetal death certificates if available. Clinically experienced medical epidemiologists review all the information available for each death and record information regarding cause of death, pregnancy outcomes, associated medical conditions, and demographic and obstetric variables available in vital records.
The Pregnancy Mortality Surveillance System requires that a pregnancy-related death satisfies both temporal and causal criteria. Specifically, a pregnancy-related death is defined as the death of a woman during or within 1 year of pregnancy that was caused by a pregnancy complication, a chain of events initiated by pregnancy, or the aggravation of an unrelated condition by the physiologic effects of pregnancy.3 The temporal association between the pregnancy status and death is ascertained in one of the following ways: presence of a selected pregnancy checkbox on the death certificate indicating the woman was pregnant at the time of death or describing an interval between the end of a pregnancy and death; words or codes indicating a pregnancy on the death certificate; presence of a note indicating the duration of complications causing or events leading to death on the death certificate; or availability of a birth or fetal death certificate within 1 year of the woman's death. The causal association between the pregnancy status and death is based on the clinical cause of death, the interval between pregnancy termination and death, and the pathophysiology of pregnancy complications. Deaths attributable to a medical condition exacerbated but not unique to pregnancy are not considered pregnancy-related if the only temporal association between the pregnant status and death is a pregnancy checkbox stating “pregnant within a year”; if the checkbox states “pregnant within 42 days” or “pregnant within 90 days,” the death may be considered pregnancy-related or not depending on the cause of death and the pathophysiologic relationship between the cause and the event of pregnancy.
Women's race, ethnicity, and nativity is based on information recorded on one or more of the data sources (ie, death, birth, or fetal death certificates) received by the CDC; depending on state, this information is recorded using checkboxes, codes, or words that capture women's race, Hispanic origin, and place of birth. For the purpose of this analysis and to make the best use of available data, race and ethnicity were categorized as: non-Hispanic white (to be referred to as white), Hispanic white (to be referred to as Hispanic), black (Hispanic and non-Hispanic), and Asian or Pacific Islander. Of note, as a result of the small number of deaths among Hispanic black women during the study period (n=52), we could not examine U.S.- and foreign-born Hispanic and non-Hispanic black women separately. Data on race were missing for 25 women over the study period. As usually performed with vital statistics data, we imputed these values using the state-specific distribution of race among women who died from pregnancy complications. Data on Hispanic origin were missing for 14 U.S.-born women, and data on nativity (U.S.- compared with foreign-born) were missing for 103 women who died from pregnancy complications between 1993, the first year when nativity information was collected, and 2006. Because nativity has a significant effect on death certificate Hispanic origin classification,10 imputation of missing ethnicity and nativity data is problematic. To be conservative, we excluded the cases with missing nativity data from the analysis and assigned the 14 cases with missing ethnicity information to the majority (ie, white) category.
We calculated annual pregnancy-related mortality ratios (number of pregnancy-related deaths per 100,000 live births) overall and by age groups (younger than 25 years, 25–34, 35 or older) for U.S.- and foreign-born white, Hispanic, black, and Asian or Pacific Islander women, respectively. Categorization of age was guided by the need to have a minimum of 20 deaths in each age, race, ethnicity, and nativity-specific group to provide stable pregnancy-related mortality ratios.11 Denominator data (ie, live births) are from publicly available U.S. natality files from the National Center for Health Statistics.12 Crude risk ratios and 95% confidence intervals (CIs) were calculated to compare U.S.- and foreign-born women within the same race and ethnicity category. Crude and age-standardized risk ratios and 95% CIs were calculated using the U.S.-born white group as a reference. For age standardization, we used the 2000 U.S. census female population as the standard population.13 The Cuzick nonparametric test-for-trend across ordered groups was used to assess the statistical significance of changes in mortality ratios over time.14
For the 2000–2006 period, we assessed cause-specific proportionate mortality (ie, proportion of all deaths resulting from a specific cause) and calculated cause-specific pregnancy-related mortality ratios (ie, number of pregnancy-related deaths attributable to a specific cause per 100,000 live births) for each race, ethnicity, and nativity group with more than 20 cause-specific deaths11 using CDC's 10-group cause of death system in line with previously published reports.3,15,16 The system allocates causes of death into the following categories: hemorrhage, infection, amniotic fluid embolism, thrombotic pulmonary or other embolism, hypertensive disorders of pregnancy, anesthesia complications, cerebrovascular accidents, cardiomyopathy, cardiovascular disease, and other (ie, noncardiovascular) medical conditions.
Kaplan-Meier survival curves were estimated to assess and compare the time between the end of pregnancy and death among race, ethnicity, and nativity groups. A minimum of 30 cases at every sojourn time t is considered needed to arrive at reliable estimates of the Kaplan-Meier survival function.17 For this reason, we could only compare survival curves for U.S.-born white, black, and Hispanic women and foreign-born Hispanic women within 42 days postpartum and between U.S.-born white and black women at 43–180 days postpartum. Log-rank tests were used to statistically compare the estimated survival curves.
All statistical analyses are conducted using STATA 10. For this analysis, the numerator data file (ie, deaths) contains no identifiers and the denominator data file (ie, births) is deidentified and available for public use. Therefore, CDC's institutional review board ruled the study as exempt.
From 1993 to 2006, there were 7,699 pregnancy-related deaths in the United States. After excluding 103 deaths among women for whom the nativity status was unknown, and 109 deaths among race and ethnicity groups not amenable to a nativity comparison (eg, Native American or Alaska Native women), our analytic sample was comprised of 7,487 pregnancy-related deaths among either U.S.- or foreign-born white, Hispanic, black, and Asian and Pacific Islander women (Table 1). Almost 62.0% (7,487–2,862 of 7,487) of deaths were among women other than U.S.-born white and 20.6% (1,539 of 7,487) among foreign-born women specifically. By comparison, only 40.7% of live births during this period were to women other than U.S.-born white, and 19.2% of births were to foreign-born women (data not shown).
The pregnancy-related mortality ratio increased significantly (P<.001) from 11.1 to 15.7 deaths per 100,000 live births from 1993 to 2006, respectively. During the study period, the pregnancy-related mortality ratio was lowest among foreign-born white and U.S.-born Asian or Pacific Islander women (7.5 and 8.7 deaths per 100,000 live births, respectively) and highest among U.S.- and foreign-born black women (35.2 and 32.3 deaths per 100,000 live births, respectively). For all groups of women, pregnancy-related mortality ratios increased with age and were especially high for women older than 35 years (Table 1).
Between 1993–1999 and 2000–2006, pregnancy-related mortality ratios appear to have increased for all race, ethnicity, and nativity groups except for foreign-born Asian or Pacific Islander women (Fig. 1). Between the two time periods, the pregnancy-related mortality ratio increase was statistically significant only among U.S.- and foreign-born white women and U.S.-born black women (30.4%, 29.2%, and 30.4%, respectively; all P<.05).
As shown in Table 2, the risk of dying from pregnancy complications was higher for foreign- relative to U.S.-born Hispanic (relative risk 1.21, 95% CI 1.07–1.36) women; conversely, the risk was lower for foreign- relative to U.S.-born white women (relative risk 0.83, 95% CI 0.69–0.98). However, when compared with U.S.-born white women, only foreign-born white women had a significantly lower risk of dying from pregnancy-related causes after adjusting for age differences. U.S.- and foreign-born Hispanic and Asian or Pacific Islander women were at 1.1–1.5 times higher risk of dying from pregnancy complications than U.S.-born white women; moreover, U.S.- and foreign-born black women were 5.2 and 3.6 times more likely to die from such causes than U.S.-born white women.
Of the 4,244 pregnancy-related deaths during the 2000–2006 period, 62.6% were among women other than U.S.-born white; 21.7% of deaths were among foreign-born women with more than half of these (52.3%) being among foreign-born women of Hispanic origin. There were several important differences in cause-specific proportionate pregnancy-related mortality by race, ethnicity, and nativity during this time period (Table 3). The contribution of cardiovascular disease, cardiomyopathy, and other medical conditions was overall higher among U.S.-born (40.8%) than foreign-born (24.7%) women (P<.05). Of note, cardiomyopathy ranked second among causes of pregnancy-related death in U.S.-born black women and ninth among foreign-born black women. Hemorrhage ranked first or second among all causes of death in all groups of foreign-born women (pregnancy-related mortality ranging between 13.9% and 20.3%), first (pregnancy-related mortality 18.0%) in U.S.-born Hispanic women, but was only sixth in U.S.-born white and black women (pregnancy-related mortality 10.1% and 10.8%, respectively). Infection contributed more deaths among U.S.-born than foreign-born white, Hispanic, and black women. Hypertensive disorders of pregnancy ranked first among foreign-born Hispanic (pregnancy-related mortality 23.1%) and black (pregnancy-related mortality 20.0%) women and third among U.S.-born Hispanic women (pregnancy-related mortality 12.5%).
Both U.S.- and foreign-born black women have notably high cause-specific pregnancy-related mortality ratios (Table 4). When compared with corresponding pregnancy-related mortality ratios among U.S.-born white women, pregnancy-related mortality ratios for hemorrhage, thrombotic pulmonary embolism, hypertensive disorders of pregnancy, cardiomyopathy, and noncardiovascular disease are over four times higher among one or both groups of U.S.- or foreign-born black women. Pregnancy-related mortality ratios for hemorrhage and hypertensive disorders of pregnancy are statistically significantly higher among both U.S.-born (1.85 and 1.28 per 100,000 live births, respectively) and foreign-born (1.71 and 2.84 per 100,000 live births, respectively) Hispanic women than among the U.S.-born white women (1.04 and 0.79 per 100,000 live births, respectively).
The vast majority of women died within 42 days postpartum. This proportion ranged between 77.9% in U.S.-born black women and 90.6% in foreign-born black women. Relative to the 5-day median time to death among U.S.-born white women (Appendix 1, available online at http://links.lww.com/AOG/A305), the median time to death was statistically significantly lower for U.S.-born Hispanic women (4 days; log-rank test P=.007) and foreign-born Hispanic women (3 days; log-rank test P<.001) but not significantly different from that among U.S.-born black women (6 days; log-rank test P=.507). The time of death between 43 and 180 days postpartum (Appendix 2, available online at http://links.lww.com/AOG/A306) was somewhat higher for U.S.-born black than white women, but the difference did not reach statistical significance (log-rank test P=.197).
This study uses national pregnancy-related mortality data to provide an in-depth analysis of the complex mortality patterns in the United States by race, Hispanic ethnicity, and nativity. After adjusting for age differences, only foreign-born white women appear to have a lower risk of pregnancy-related death than their U.S.-born counterparts. Sociodemographic factors such as foreign-born white women's higher education and income levels, their inclination to access health care early in pregnancy based on past practice in their countries of origin2 as well as the “healthy migrant” bias could explain this finding. Our study documents not only the substantial disparity in pregnancy-related mortality between black and white women in the United States, but also a mortality risk difference between U.S.- and foreign-born black women. In line with previous CDC reports,3,15,16 we also find that the risk of dying from pregnancy complications increases with maternal age. Importantly, we show that it is among black women, and especially those born in the United States, that the interaction between race and age plays out most intensely. This interaction could be the effect of pre-existing conditions and pregnancy complications,2 both more prominent among older than younger women, or of the lower access to and use of health care by black women.18 Alternatively, a higher case-fatality rate for severe maternal morbidity among black compared with other groups of women may explain this interaction. Tucker et al19 used national data to calculate prevalence and case-fatality rates for eclampsia, abruptio placentae, placenta previa, and postpartum hemorrhage and found that the higher pregnancy-related mortality from these causes among black women was largely attributable to higher case-fatality rates. Recent research has focused on the site of care as a potential explanation for health disparities in the United States, but results are inconsistent. Ly et al20 used national Medicare data and showed that “black-serving” hospitals performed worse than “non-black-serving” hospitals on six of 11 patient safety indicators studied and that both white and black patients had higher rates of potential safety-threatening events in “black-serving” than in “non-black-serving” hospitals. Gaskin et al21 used data from the Healthcare Cost and Utilization Project and found that risk-adjusted quality-of-care indicators for minorities are not statistically worse than corresponding indicators for whites in the same hospital. Indeed, more research to examine differences in the quality of obstetric care received by the various race, ethnicity, and nativity groups in the United States is needed.
The “Hispanic paradox” has been described in the literature in relation to birth outcomes and all-cause mortality among men and women,5 yet it does not seem to hold with regard to pregnancy-related mortality, and foreign-born Hispanic women are more likely to die from pregnancy complications than their U.S.-born white and Hispanic counterparts. It may be that factors like language barriers,22 concerns by undocumented immigrants over legal action, and lack of familiarity with the U.S. health care system deter Hispanic foreign-born pregnant women from obtaining adequate health care. On the other hand, it should be noted that when compared with the 47.5% higher pregnancy-related mortality ratios among foreign-born relative to U.S.-born Hispanic women reported by CDC based on 1993–1997 Pregnancy Mortality Surveillance System data,7 these 2000–2006 Pregnancy Mortality Surveillance System data show a smaller (19.4%) difference in pregnancy-related mortality ratios between foreign- and U.S.-born Hispanic women. Because we have no reason to believe that Pregnancy Mortality Surveillance System data were prone to differential ascertainment of pregnancy mortality by nativity over time, this difference may be the result of faster acculturation of Hispanics.
As previously described for the U.S. population as a whole,3 this analysis documents important changes in proportionate pregnancy-related mortality for all race, ethnicity, and nativity groups in recent years. We found a decline in the contribution of the “classic” direct causes of pregnancy-related mortality (eg, hemorrhage, sepsis) and the emergence of cardiovascular disease, cardiomyopathy, and other medical conditions as important contributors to pregnancy-related mortality among all U.S.-born non-Hispanic women between 2000 and 2006. Similar changes have been observed in other countries23 and are likely the result of a combination of factors including postponement of pregnancy and higher prevalence of obesity,24 diabetes,25 heart disease,26 and hypertension27 among pregnant women. However, hypertensive disorders of pregnancy, hemorrhage, and embolism remain the main causes of pregnancy-related death among U.S.-born Hispanic and all foreign-born women.
This study is not without limitations. Pregnancy-related mortality appears to have increased significantly from 1993 to 2006. Whether this increase is the result of enhanced identification of maternal deaths or to an actual increase is uncertain. Increasingly more states use checkboxes on death certificates to identify women who died during or within 1 year of the end of a pregnancy and perform linkages between death and birth or fetal death certificates as documented by an increased proportion of death certificates with matching birth or fetal death certificates sent for the Pregnancy Mortality Surveillance System. Nevertheless, pregnancy-related deaths may still be undercounted, especially among foreign-born Hispanic women who may be returning to their country of origin soon after delivery (“salmon bias” hypothesis).28 We could not differentiate between Hispanic and non-Hispanic black women, and, in general, reporting of race and ethnicity on death certificates is prone to information bias. Also of note, although the Pregnancy Mortality Surveillance System is the most complete nationwide source of data on pregnancy-related mortality, it lacks the clinical details only available in the medical records and, in tandem with missing information on potential confounders of the associations of interest (ie, parity, education, mode of delivery, obesity), impedes us from examining adjusted associations.
Overall, racial, ethnic, and nativity-related disparities in pregnancy-related mortality likely stem from a multitude of factors, and understanding these disparities remains a major challenge for clinicians and public health professionals alike. Although not all maternal deaths are preventable, a large proportion of them are.29 Revival of hospital- and state-based maternal deaths reviews and integration of quality-of-care aspects into these reviews may help identify race, ethnicity, and nativity-specific factors playing key roles in women dying from pregnancy complications in the United States in the 21st century.
1. James CV, Salganico A, Thomas M, Ranji U, Lillie-Blanton M, Wyn R. Putting women's health care disparities on the map: examining racial and ethnic disparities at the state level. 2009. Available at: www.kff.org/minorityhealth/upload/7886.pdf
. Retrieved April 20, 2012.
2. Ramadhani TA, Canfield MA, Farag NH, Royle M, Correa A, Waller DK, et al.. Do foreign- and US-born mothers across racial/ethnic groups have a similar risk profile for selected sociodemographic and periconceptional factors? Birth Defects Res A Clin Mol Teratol 2011;91:823–30.
3. Berg CJ, Callaghan WM, Syverson C, Henderson Z. Pregnancy-related mortality in the United States, 1998 to 2005. Obstet Gynecol 2010;116:1302–9.
4. Creanga AA, Shapiro-Mendoza CK, Bish CL, Zane S, Berg CJ, Callaghan WM. Trends in ectopic pregnancy mortality in the United States: 1980–2007. Obstet Gynecol 2011;117:837–43.
5. Bryant AS, Worjoloh A, Caughey AB, Washington AE. Racial/ethnic disparities in obstetric outcomes and care: prevalence and determinants. Am J Obstet Gynecol 2010;202:335–43.
6. Acevedo-Garcia D, Soobader MJ, Berkman LF. Low birthweight among US Hispanic/Latino subgroups: the effect of maternal foreign-born status and education. Soc Sci Med 2007;65:2503–16.
7. Centers for Disease Control and Prevention (CDC). Pregnancy-related deaths among Hispanic, Asian/Pacific Islander, and American Indian/Alaska Native women—United States, 1991–1997. MMWR Morb Mortal Wkly Rep 2001;50:361–4.
8. Racial and ethnic disparities in women's health. ACOG Committee Opinion No. 317. American College of Obstetricians and Gynecologists. Obstet Gynecol 2005;106:889–92.
9. Martin JA, Hamilton BE, Ventura SJ, Osterman MJK, Kirmeyer S, Mathews TJ, et al.. Births: Final data for 2009. Available at: www.cdc.gov/nchs/data/nvsr/nvsr60/nvsr60_01.pdf
. Retrieved April 20, 2012.
10. Arias E, Schauman WS, Eschbach K, Sorlie PD, Backlund E. The validity of race and Hispanic origin reporting on death certificates in the United States. Available at: www.cdc.gov/nchs/data/series/sr_02/sr02_148.pdf
11. Hoyert D. Maternal mortality and related concepts. Available at: www.cdc.gov/nchs/data/series/sr_03/sr03_033.pdf
12. National Center for Health Statistics. Births data files. Available at: www.cdc.gov/nchs/data_access/Vitalstatsonline.htm
. Retrieved April 20, 2012.
13. US Census Bureau. American FactFinder. Available at: www.census.gov/main/www/cen2000.html
. Retrieved April 20, 2012.
14. Cuzick J. A Wilcoxon-type test for trend. Stat Med 1985;4:87–90.
15. Berg CJ, Chang J, Callaghan WM, Whitehead SJ. Pregnancy-related mortality in the United States, 1991–1997. Obstet Gynecol 2003;101:289–96.
16. Callaghan WM, Berg CJ. Pregnancy-related mortality among women aged 35 years and older, United States, 1991–1997. Obstet Gynecol 2003;102:1015–21.
17. Hill T, Lewicki P. Statistics: methods and applications. Available at: www.statsoft.com/textbook/
. Retrieved April 20, 2012.
18. Tucker MJ, Berg CJ, Callaghan WM, Hsia J. The black–white disparity in pregnancy-related mortality from 5 conditions: differences in prevalence and case-fatality rates. Am J Public Health 2007;97:247–51.
19. National Center for Health Statistics. Health, United States, 2009: with special feature on medical technology. Available at: www.cdc.gov/nchs/data/hus/hus09.pdf
. Retrieved April 20, 2012.
20. Ly DP, Lopez L, Isaac T, Jha AK. How do black-serving hospitals perform on patient safety indicators? Implications for national public reporting and pay-for-performance. Med Care 2010;48:1133–7.
21. Gaskin DJ, Spencer CS, Richard P, Anderson GF, Powe NR, Laveist TA. Do hospitals provide lower-quality care to minorities than to whites? Health Aff (Millwood) 2008;27:518–27.
22. Ding H, Hargraves L. Stress-associated poor health among adult immigrants with a language barrier in the United States. J Immigr Minor Health 2009;11:446–52.
23. Lewis G, editor. The Confidential Enquiry into Maternal and Child Health (CEMACH). Saving mothers' lives: reviewing maternal deaths to make motherhood safer—2003–2005. The seventh report on confidential enquiries into maternal deaths in the United Kingdom. Available at: www.mdeireland.com/pub/SML07_Report.pdf
. Retrieved April 20, 2012.
24. Kim SY, Dietz PM, England L, Morrow B, Callaghan WM. Trends in prepregnancy obesity in nine states, 1993–2003. Obesity (Silver Spring) 2007;15:986–93.
25. Albrecht S, Kuklina E, Bansil P, Jamieson D, Whiteman M, Kourtis A, et al.. Diabetes trends among delivery hospitalizations in the United States, 1994–2004. Diabetes Care 2010;33:768–73.
26. Kuklina EV, Callaghan WM. Chronic heart disease and severe obstetric morbidity among hospitalizations for pregnancy in the USA: 1995–2006. BJOG 2011;118:345–52.
27. Kuklina EV, Ayala C, Callaghan WM, Hypertensive disorders and severe obstetric morbidity in the United States: 1998–2006. Obstet Gynecol 2009;113:1299–306.
28. Abraído-Lanza AF, Dohrenwend BP, Ng-Mak DS, Turner JB. The Latino mortality paradox: a test of the 'salmon bias' and healthy migrant hypotheses. Am J Public Health 1999;89:1543–8.
29. Berg CJ, Harper MA, Atkinson SM, Bell EA, Brown HL, Hage ML, et al.. Preventability of pregnancy-related deaths: results of a state-wide review. Obstet Gynecol 2005;106:1228–34.