Keywords
infant - ethnicity - gestational age - infant mortality - educational status
Introduction
Infant mortality rate (IMR) is an important indicator of the overall health of a population.
A study examining data from 180 countries showed a strong linear relationship between
IMR and disability-adjusted life expectancy.[1] Since 1995, the IMR has had a downward trend in the United States and fell by 15%
from 2005 to 2017.[2] Despite the fall in total IMR over the years, a difference in IMR between Black
and White infants persists.[3] Studies of the U.S. IMR rates show that the relative gap between IMR for Black infants
versus White infants has grown. Between 1916 and 2017, the IMR among White infants
was reduced by 3.1% each year, but the IMR among Black infants declined by 2.6% annually.[4] This discrepancy in IMR reduction had a significant impact on the Black–White IMR
ratio. In 1916, the IMR for Black infants was 87% higher than the rate for White infants,
while in 2017, the Black IMR was 122% higher than the White IMR. In the United States,
the proportion of Black births in a state is strongly associated with the state's
overall IMR. This highlights the significant impact that racial disparities can have
at the state level.[5]
It is known that race and ethnicity are social constructs, not biological variables,
which serve as a proxy for differential exposure to social and structural conditions
that influence health, including unequal access to health-promoting resources.[6]
[7] Along with race, maternal education level has been shown to have an independent
association with pregnancy outcomes. The Fundamental Cause Theory proposes that preventative
women's health screening was consistently related to educational level, and women
with higher education levels were screened more often when compared with their lower
education counterparts.[8] Maternal education is one of the largest socioeconomic factors associated with disparities
in preterm birth rates between Black and White infants in the United States, accounting
for approximately 11.3% of the difference.[9] However, while higher maternal educational attainment is associated with fewer preterm
births among White women, higher educational attainment may not be associated with
fewer preterm births among Black women, suggesting differential effects of education
by race.[10] However, these studies have not assessed the association of maternal education and
race with IMR.
The purpose of this study was to analyze the relationship between maternal race and
maternal education to determine if there was a significant association between gestational
age (GA)-specific IMRs as well as IMRs at a population-level by GA.[5]
[11] We hypothesized that the disparities in IMR among Black and Hispanic infants compared
with White infants occur at each specific GA from 27 to 41 weeks' gestation, despite
the level of maternal education attained.[3] Alternatively, disparities in IMR by maternal education and race may be related
to differences in the GA distribution, and this may be more accurately reflected using
the fetus-at-risk approach. We hypothesized that the disparities in IMR would be present
across all GAs using the fetus-at-risk approach.[5]
[11]
Materials and Methods
Cohort Selection
The participants in this study are infants registered in the Centers for Disease Prevention
and Control (CDC) WONDER expanded linked birth and infant death records database.
Deidentified data were obtained from the CDC WONDER database in agreement with the
data use restrictions, which limit reporting of subnational data or numbers of infants
with less than 10 events. We extracted data from the database using computer code
(R version 4.0) to define the inclusion and exclusion criteria and ensure reproducibility.
The analysis included live-born infants from 2017 to 2019. We included the following
weeks of gestation: extremely preterm (22–27 weeks), very preterm (28–31 weeks), late
preterm (32–36 weeks), and term (37–41 weeks).
Exposure
We included infants whose maternal race was recorded as non-Hispanic Black, non-Hispanic
White, or Hispanic in the database, with a known maternal education status, and a
maternal age greater than or equal to 20 to allow sufficient years of age for higher
levels of educational attainment. Given that non-Hispanic Whites constitute 58.4%
of the United States population, Hispanics 19.5%, and Blacks 13.7%, this study focuses
on these groups, which together represent approximately 92% of the national population.[12] We excluded infants with major congenital anomalies as well as infants whose cause
of death included congenital malformations, deformations, and chromosomal anomalies.
The study was approved as not human subjects research by our Institutional Review
Board. Strengthening the Reporting of Observational Studies in Epidemiology guidelines
were used.[13]
Outcomes
This cohort study compared IMR by maternal race and educational level at each GA.
The primary outcome measure was the GA-specific IMR comparing infants with maternal
race of non-Hispanic Black with non-Hispanic White. Secondary outcome measures included
the GA-specific IMR comparing infants with maternal ethnicity Hispanic, including
both Black and White Hispanic infants, with non-Hispanic White. We also compared the
population-adjusted IMR at each GA by maternal race/ethnicity and education to account
for differences in GA distribution. We calculated the population-adjusted IMR based
on the “fetus-at-risk” calculation by subtracting infants already delivered at lower
GAs to determine the denominator for subsequent GAs.[5]
[11] Maternal education levels were categorized based on the mother's highest level of
education, and they are defined by the four following groups: no high school diploma
or general educational development (GED) testing; high school graduate or GED completed;
some college credit but not a degree; and college graduate (including associate degree,
bachelor's degree, master's degree, or doctorate).
Statistical Analysis
We used the sample of all infants who met the inclusion criteria for this population-based
study. For each GA and education level, data were compared by maternal race of non-Hispanic
Black, non-Hispanic White, or Hispanic to determine IMR, defined as the number of
deaths within the first year after birth per 1,000 live-born infants. We calculated
relative risk and 95% confidence intervals using White infants as the reference category.
Forest plots were created for each comparison, displaying the point estimate and 95%
confidence intervals. Findings with a p < 0.05 were considered statistically significant. Differences in IMR are presented
without adjustment for known confounders to illustrate the unmitigated burden of disparities,
as controlling for these factors would potentially mask the structural determinants
underlying these inequities.
Ethics Approval
This study used deidentified patient data from a publicly available database, and
patient consent was not required. The study was approved as not human subjects research
by our Institutional Review Board.
Results
There were 9,356,130 infants that were born during the study period; 1,504,230 (16.1%)
Black infants; 2,406,327 (25.7%) Hispanic infants; and 5,445,573 (58.2%) White infants.
There were higher rates of college graduation among mothers of White infants compared
with mothers of Black or Hispanic infants ([Table 1]). There were higher rates of mothers whose highest level of education was either
a high school diploma or GED among mothers of Black and Hispanic infants compared
with White infants. Mothers of Hispanic infants had a higher rate of not completing
high school compared with mothers of Black and White infants.
Table 1
Highest maternal education achieved by race/ethnicity
|
Level of education
|
White; n = 5,445,573
|
Black; n = 1,504,230
|
Hispanic; n = 2,406,327
|
|
College graduate, n (%)
|
2,948,949 (54.1)
|
410,763 (27.3)
|
537,037 (22.3)
|
|
Some college, no degree, n (%)
|
1,079,485 (19.8)
|
404,008 (26.9)
|
502,518 (20.9)
|
|
High school graduate or GED completed, n (%)
|
1,124,727 (20.7)
|
528,154 (35.1)
|
781,531 (32.5)
|
|
12th grade or less with no diploma, n (%)
|
292,412 (5.4)
|
161,305 (10.7)
|
585,241 (24.3)
|
Abbreviation: GED, general educational development.
Among college graduates, Black term infants (37–41 weeks' gestation) had a higher
gestational-age-specific IMR, whereas extremely preterm infants at 23 and 24 weeks'
gestation had a lower IMR compared with White infants ([Fig. 1]; [Table 2]). Similar patterns were seen when comparing Black Infants and White infants with
mothers who had completed some college without receiving a degree, as well as among
those with a high school diploma or GED ([Fig. 1]). For Black mothers without a high school diploma or GED, infants born at 23 and
24 weeks' also had lower IMRs compared with infants born to White mothers, while those
born at 34, 38, and 40 weeks' had higher IMRs ([Fig. 1]; [Table 2]).
Fig. 1 Relative risk and 95% confidence intervals for the gestational age-specific infant
mortality rate (IMR) from 41 to 22 weeks by race/ethnicity. Black and Hispanic infants
are compared with White Infants subdivided by maternal educational level.
Table 2
Gestational age-specific infant mortality rates (IMR) subdivided by race and highest
educational level
|
GA
|
Black
|
White
|
Hispanic
|
|
12th grade or less with no diploma
|
High school graduate or GED completed
|
Some college credit, but no degree
|
Associate, bachelor's, master's, or doctorate degree
|
12th grade or less with no diploma
|
High school graduate or GED completed
|
Some college credit, but no degree
|
Associate, bachelor's, master's, or doctorate degree
|
12th grade or less with no diploma
|
High school graduate or GED completed
|
Some college credit, but not a degree
|
Associate, bachelor's, master's, or doctorate degree
|
|
22
|
838.9
|
809.1
|
817.6
|
762.4
|
825.4
|
830.3
|
820.4
|
838.7
|
799.1
|
846.8
|
820.3
|
774.2
|
|
23
|
388.2
|
472.5
|
418.6
|
431.2
|
582.9
|
565.3
|
554.2
|
561.8
|
483.8
|
492.6
|
465.2
|
453.1
|
|
24
|
236.2
|
260.1
|
260.4
|
285.2
|
363.3
|
341.1
|
347.5
|
345.1
|
291.6
|
279.1
|
230.3
|
279.0
|
|
25
|
211.3
|
189.0
|
168.4
|
163.7
|
198.5
|
217.9
|
213.9
|
194.4
|
182.6
|
159.3
|
153.7
|
143.2
|
|
26
|
110.1
|
108.9
|
108.0
|
83.1
|
144.0
|
140.0
|
116.9
|
126.0
|
108.0
|
106.3
|
116.6
|
115.5
|
|
27
|
94.2
|
86.0
|
67.7
|
77.6
|
65.1
|
106.1
|
68.9
|
79.7
|
72.0
|
58.1
|
60.3
|
69.4
|
|
28
|
54.8
|
50.3
|
39.2
|
42.5
|
46.3
|
56.5
|
42.3
|
39.3
|
43.6
|
52.5
|
27.2
|
27.6
|
|
29
|
43.1
|
44.6
|
32.7
|
22.0
|
38.5
|
43.0
|
32.6
|
29.1
|
23.5
|
28.5
|
30.2
|
36.3
|
|
30
|
36.7
|
24.1
|
23.9
|
14.2
|
27.0
|
29.2
|
20.4
|
20.1
|
29.6
|
24.2
|
14.0
|
14.5
|
|
31
|
30.7
|
19.0
|
21.4
|
11.3
|
22.1
|
20.2
|
16.3
|
12.6
|
16.5
|
15.4
|
11.2
|
11.4
|
|
32
|
20.8
|
17.8
|
13.3
|
11.1
|
17.1
|
14.0
|
11.3
|
7.7
|
11.0
|
10.4
|
12.3
|
5.9
|
|
33
|
23.1
|
13.5
|
11.4
|
5.8
|
14.4
|
9.9
|
8.7
|
6.0
|
8.0
|
7.1
|
6.2
|
4.3
|
|
34
|
13.2
|
10.2
|
6.9
|
5.2
|
10.8
|
8.0
|
5.6
|
3.3
|
5.7
|
5.6
|
4.7
|
3.0
|
|
35
|
9.3
|
7.2
|
5.6
|
4.2
|
9.0
|
6.6
|
5.1
|
2.1
|
3.3
|
4.3
|
2.5
|
1.3
|
|
36
|
6.2
|
6.1
|
4.6
|
2.8
|
6.4
|
4.8
|
3.4
|
1.7
|
2.8
|
2.6
|
2.3
|
1.2
|
|
37
|
6.6
|
4.4
|
3.3
|
2.0
|
5.2
|
3.4
|
2.6
|
1.0
|
1.7
|
1.7
|
1.6
|
1.1
|
|
38
|
4.7
|
3.5
|
3.1
|
1.4
|
4.0
|
2.4
|
1.8
|
0.7
|
1.2
|
1.3
|
1.0
|
0.6
|
|
39
|
3.9
|
3.0
|
2.3
|
1.0
|
3.2
|
2.0
|
1.3
|
0.6
|
1.0
|
0.9
|
0.8
|
0.6
|
|
40
|
2.8
|
2.8
|
2.0
|
0.9
|
2.9
|
1.7
|
1.1
|
0.5
|
1.0
|
0.9
|
0.9
|
0.5
|
|
41
|
1.8
|
2.2
|
2.2
|
0.8
|
2.0
|
1.1
|
1.1
|
0.6
|
0.8
|
1.1
|
0.9
|
0.4
|
Abbreviations: GA, gestational age; GED, general educational development.
For infants born to Hispanic mothers, there was a gestational-age-specific survival
advantage seen across all educational levels among infants at the lowest GAs of 23
and 24 weeks, as well as some other extremely preterm GAs at different levels of education
([Fig. 1]; [Table 2]). At the lowest level of education, Hispanic infants born from 34 weeks' gestation
or later had lower IMR at each GA when compared with White infants. Patterns were
similar among those whose mother's highest level of education was either a High school
diploma or completing some college, but did not include all GAs from 34 weeks or higher.
Among Hispanic infants whose parents had any college degree, there was no difference
in gestational-age-specific IMR from 26 weeks' gestation or higher.
When adjusted for population-at-risk, Black infants had a higher risk of infant mortality
at all GAs among those whose mothers completed their high-school education and those
with some college when compared with White infants whose mothers had the same level
of education ([Fig. 2]). At the highest level of educational attainment, there were significant differences
at most but not all GAs. Among those at the lowest level of educational attainment,
there were differences in survival at lower GAs but not among term infants and some
late preterm groups. Differences were largest among those at the lowest gestations,
reflecting the higher rate of extremely and very preterm birth among Black infants
in the United States.
Fig. 2 Relative risk and 95% confidence intervals for the gestational age-specific infant
mortality rate (IMR) from 41 to 22 weeks by race/ethnicity adjusted for the population-at-risk.
Black and Hispanic infants are compared with White Infants subdivided by maternal
educational level.
Hispanic infants at the lowest GAs had a higher risk of IMR based on the population-at-risk
analyses, suggesting a higher risk of preterm birth at the lowest gestations affecting
the Hispanic populations with high-school or higher education levels compared with
White infants. However, at higher GAs, the results for the population-at-risk analyses
comparing Hispanic with White infants did not differ substantially from the results
for the GA-specific IMR analyses across educational groups.
Discussion
This population-based study found that the association of educational attainment with
IMR varies by race/ethnicity. The disparities in IMR among term Black and White infants
are less pronounced at the lowest education levels compared with higher education
levels. Paradoxically, the lower IMR favoring Hispanic infants at lower levels of
education is no longer present at the highest education levels. Furthermore, the lower
GA-specific IMR among Black and Hispanic infants at the lowest gestations was reversed
when analyzed based on the fetus-at-risk approach.[5]
[11] These data suggest that the relatively small survival benefit among individual Black
and Hispanic infants born at the lowest gestations is more than offset by the higher
number of births, and therefore, deaths at a population level.
Whereas studies examining disparities at a population level report worsening disparities
in IMR, studies examining outcomes among extremely preterm infants found either decreasing
disparities or no disparities in care practices, major morbidities, and mortality.[14]
[15] However, these studies adjusted for multiple baseline clinical risk factors, including
GA, birth weight, sex, and multiple births, and maternal demographic differences,
including maternal education, marital status, and insurance, between groups. Furthermore,
it has been noted that comparing outcomes among preterm birth cohorts can lead to
bias depending on whether data are analyzed by group or by fetus-at-risk to account
for the entire population.[11] This, in turn, may explain the “so-called” preterm infant paradox,[16] where Black extremely preterm infants have been reported to have a survival advantage
at an individual level compared with White infants but have higher rates of infant
mortality at each GA when analyzed at a population level.[3] While Black infants born extremely preterm experience higher rates of mortality,
there is evidence that clinical severity alone does not fully account for these disparities.
A recent cohort study found that maternal race and ethnicity were associated with
a lower likelihood of redirection-of-care discussions and actions, independent of
education and insurance status.[17] These differences may reflect variation in cultural values, religious beliefs, understanding
of prognosis, or clinician bias. Such findings suggest that observed disparities in
extremely preterm infant outcomes may stem from unequal access to or engagement in
end-of-life decision-making processes.
Our study agrees with data from the 2007 to 2008 NCHS Cohort Linked Live Birth—Infant
Death Files, which found that the risk of overall infant mortality was lower for Hispanic
infants when compared with non-Hispanic White and non-Hispanic Black infants, respectively.[18] Several studies have also shown that infants born to Hispanic mothers appear to
have a paradoxical advantage when it comes to infant health and mortality when compared
with their non-Hispanic counterparts.[19]
[20] In our study, Hispanic infants had a lower risk of death, which was most pronounced
at lower levels of educational attainment, while there were limited differences at
the highest levels of educational attainment. This is in contrast to the higher risk
of death among Black infants at higher levels of educational attainment, with more
limited differences among those at the lowest levels of educational attainment.
There are known differences in IMR within racial/ethnic groups by level of educational
attainment, suggesting that higher levels of education may be protective. In California,
from 2007 to 2015, mothers with a high school education or less accounted for 48.5%
of the total births but 59% of infant deaths.[21] College education or higher was associated with lower IMR. Furthermore, infants
of Black mothers with the lowest level of education had the highest risk of mortality
compared with White infants or infants of Black mothers with college-education or
higher (both p < 0.001). However, Black infants of college-educated or higher mothers still had
higher infant mortality than White infants of college-educated or higher mothers (p < 0.001). These differences are in line with our study, which showed that college
education is not sufficient to fully attenuate the racial disparities in IMR among
Black infants in the United States.
Our study analyzed national data, but results from individual states in the United
States may differ. A 15-year U.S. study that followed the IMR rates of each state
found that only 13 of the 50 states made a statistically significant reduction in
the Black–White IMR ratio. The rest of the states showed that while both Black and
White IMR rates were decreasing, the disparity in the Black–White IMR ratio was sustained
from 2000 to 2012.[22] In a study from Wisconsin, the IMR among Black infants born to mothers with high-school
education or less was higher than those of White mothers or Black mothers with higher
levels of education, but this study did not examine results by GA which is the largest
determinant of IMR.[23] Older U.S. data from 1998 to 2002 reported an educational gradient in IMR with greater
protection from higher education for White infants relative to Black infants.[24] In contrast, infants born to Hispanic mothers may have similar IMR when compared
with infants of White Mothers after adjustment for maternal education.[25] Taken together with our investigation, current data suggest that while education
is protective against IMR, the effect size varies by race/ethnicity.
Low birth weight, birth, and prematurity account for much of the higher IMR in the
United States.[26] There is also an association between maternal education and low birth weight births.
A meta-analysis found that high levels of maternal education are associated with a
33% lower risk of low birth weight.[27] In a study from the 1980s using a national database to examine infant outcomes,
infants of Black mothers with college education were more than twice as likely to
have low birth weight births when compared with infants with White college-educated
mothers.[28] Furthermore, there was no difference in preterm birth rates among the most socially
disadvantaged and lowest-educated Black and White mothers. Similar results have been
reported in a study using higher levels of education in a model of socioeconomic status
to examine the association between race and preterm birth.[29] It has also been reported that the protective association of maternal education
on preterm birth may be decreasing over time.[30] In this study, the differences in the population-level fetus-at-risk analyses suggested
that an important difference in preterm birth rates drives overall disparities in
IMR.
The persistent disparities in GA-specific IMRs among late preterm and term Black infants,
even after considering educational attainment, have been the subject of extensive
research. The “weathering hypothesis” proposes that chronic stress experienced throughout
a Black individual's life may negatively impact reproductive health, contributing
to higher rates of prematurity and infant mortality within this population.[31] Stress can impact biological aging, and Black women had telomere lengths consistent
with being approximately 7.5 years older than White women.[32] Furthermore, maternal mortality is associated with higher rates of infant mortality.
A study examining the relationship between maternal and infant health found that countries
that made progress in reducing maternal mortality also experienced improvements in
infant and child mortality rates during the same period.[33] Black women are more than three times more likely to die during the postnatal period
compared with their White counterparts, highlighting significant racial disparities
in maternal health outcomes.[34] The risk of maternal death rises with increasing BMI, and Black women have disproportionately
high obesity rates even at higher levels of income, whereas higher income was protective
against obesity in White and Hispanic women.[35]
[36] The racial disparity in pregnancy-related morbidity and mortality may contribute
to the persistently higher rates of infant mortality among Black infants. These data
suggest that the disparity in Black infant mortality reflects a complex, multifactorial
mix of social determinants of health that are not mitigated by maternal educational
attainment.
The Hispanic Paradox, characterized by better health outcomes and lower IMRs among
Hispanic immigrants in the United States despite socioeconomic disadvantages, has
been subject to various explanatory hypotheses. One of these is immigration selectivity,
positing that Hispanic immigrants may be positively selective on good health, with
healthier individuals more likely to migrate than those with poorer health. This selectivity
results in a higher concentration of healthy individuals within the immigrant population,
potentially contributing to improved health outcomes.[37]
[38] Furthermore, researchers have studied cultural patterns, which suggest that certain
cultural aspects within Hispanic communities promote healthy behaviors and strong
family ties, contributing to relatively favorable health and mortality patterns.[37] Our study found that as Hispanic mothers achieved higher levels of education, the
rate of infant mortality was not lower compared with White infants. Our fetus-at-risk
analysis suggests that Hispanic infants born to mothers with higher education are
at higher risk of preterm birth. These findings may be explained by the negative impact
of acculturation, the process by which individuals or groups adopt and incorporate
elements of a new culture into their own culture. It occurs when individuals from
one cultural background come into continuous contact with another culture and gradually
adopt aspects of that culture while retaining some aspects of their original cultural
identity. As immigrants and their descendants become more acculturated, they may be
more prone to adopt less healthy behaviors.[39] Studies exploring acculturation and maternal health behaviors found that mothers
with a higher degree of U.S. cultural adaptation are more likely to adopt unhealthy
behaviors such as smoking during pregnancy, potentially contributing to a decline
in health outcomes.[40]
[41] Analysis of the Latino National survey showed that educational attainment can predict
English Mastery, a core acculturation marker, even after correcting for generational
standing and income, suggesting that higher education may coincide with greater acculturation.[42]
[43]
Limitations
In this population-based study, we did not adjust for differences in baseline characteristics,
including pregnancy complications between groups, due to database restrictions on
access to individual patient data. The large sample size and the matching for GA in
our comparison groups may have reduced the risk of confounding, but remaining biases
may have impacted the estimated effect sizes reported in this study. We did not have
access to data on the day of birth within each week category, and it is possible this
differed by race/ethnicity. We did not have data on fetal loss, and previous studies
have reported higher fetal mortality rates impacting Black women, such that our study
may underestimate racial/ethnic differences using the fetus-at-risk approach.[44] In addition, redirection of care practices may differ among infants by race/ethnicity,
but these data were not available.[17] We did not adjust for multiple testing, and given the large number of comparisons,
it is possible that some results may have been significant or not significant by chance.
This study compared differences across maternal education levels but did not examine
differences across paternal education levels. We did not consider the association
between education and mortality rates among infants born to women of races/ethnicities
other than Black, White, and Hispanic mothers, which represent approximately 8% of
births in the United States.
Conclusion
Disparities in GA-specific IMRs among late preterm and term Black infants persist
despite educational attainment. Hispanic infants have a lower gestational-age-specific
infant mortality at lower levels of educational attainment, consistent with the Hispanic
Paradox.
