American Journal of Epidemiology
ª The Author 2010. Published by Oxford University Press on behalf of the Johns Hopkins Bloomberg School of
Public Health. All rights reserved. For permissions, please e-mail: [email protected].
Vol. 172, No. 2
DOI: 10.1093/aje/kwq109
Advance Access publication:
June 24, 2010
Original Contribution
Exploring Weathering: Effects of Lifelong Economic Environment and Maternal
Age on Low Birth Weight, Small for Gestational Age, and Preterm Birth in AfricanAmerican and White Women
Catherine Love, Richard J. David*, Kristin M. Rankin, and James W. Collins, Jr.
* Correspondence to Dr. Richard J. David, Division of Neonatology, John H. Stroger, Jr., Hospital, 1901 West Harrison Street,
Chicago, IL 60612 (e-mail: [email protected]).
Initially submitted July 6, 2009; accepted for publication December 10, 2010.
White women experience their lowest rate of low birth weight (LBW) in their late 20s; the nadir LBW for AfricanAmerican women is under 20 years with rates rising monotonically thereafter, hypothesized as due to ‘‘weathering’’
or deteriorating health with cumulative disadvantage. Current residential environment affects birth outcomes for all
women, but little is known about the impact of early life environment. The authors linked neighborhood income to
a transgenerational birth file containing infant and maternal birth data, allowing assessment of economic effects
over a woman’s life course. African-American women who were born in poorer neighborhoods and were still poor
as mothers showed significant weathering with regard to LBW and small for gestational age (SGA) but not preterm
birth (PTB). However, African-American women in upper-income areas at both time points had a steady fall in LBW
and SGA rate with age, similar to the pattern seen in white women. No group of white women, even those always
living in poorer neighborhoods, exhibited weathering with regard to LBW, SGA, or PTB. In contrast, the degree of
weathering among African-American women is related to duration of exposure to low-income areas and disappears
for those with a life residence in non-poor neighborhoods.
African Americans; infant, premature; infant, small for gestational age; maternal age; poverty; premature birth;
residence characteristics
Abbreviations: LBW, low birth weight (<2,500 g); PTB, <37 weeks’ gestation; SGA, small for gestational age (<10th percentile for
sex-, race-, and gestational age-specific birth weight).
Editor’s note: An invited commentary on this article appears on page 135, and the authors’ response is published
on page 138.
[their] gestational age (SGA), or because of both processes.
Mechanisms underlying racial differences in LBW and its
components are incompletely understood.
One element that influences the birth outcomes of
African-American and white populations differently is
maternal age. In the white population, the LBW rate has
a nadir between maternal ages 25 and 29 years, in what
has been considered optimal child-bearing years, based on
a developmental paradigm. In 1992, Geronimus (4) demonstrated that, for African-American mothers, the infant mortality rate was actually lowest in the teenage years and
deteriorated steadily with advancing maternal age. Subsequent studies have confirmed worsening birth outcomes with
increasing maternal age in multiple African-American populations (5–10). Geronimus referred to this phenomenon as
In the United States, the rate of low birth weight (LBW)
has retained a striking influence on health outcomes over
several decades (1). LBW accounts for most deaths in the
first year of life, and LBW rates have remained consistently
elevated, as has the black-white disparity in birth outcomes
(1, 2). African-American women were about twice as likely
as non-Hispanic white women (14.0% vs. 7.3%) to have
a LBW infant in 2006 (3). Their risk of infant death was
2.4 times the white rate in 2005 (3). Infants can be LBW
because of preterm birth (PTB), because they are small for
127
Am J Epidemiol 2010;172:127–134
128 Love et al.
‘‘weathering’’ (4) and theorized that it may be a result of the
cumulative impact of socioeconomic disadvantages on the
health of African-American women (4, 5). The weathering
pattern has been reported most often for LBW or infant
mortality (5–9). There are 5 previous reports examining
the weathering phenomenon using preterm birth as an outcome among an adequate sample of African-American
women (10–14) and, to date, only 1 study has addressed
SGA as an outcome displaying this pattern (15).
Investigations have established that the weathering effect
is modified by the mother’s socioeconomic situation and
tends to be most pronounced in women of lower socioeconomic status (5–8, 10). In fact, Geronimus (5) demonstrated
an elimination of weathering among African-American
women living in affluent circumstances, although she did
not report a pattern of risk qualitatively similar to that seen
in white mothers, for whom the best outcomes occur in the
late 20s. To date, only 2 studies have found a weathering
effect in white mothers. Rauh et al. (6) demonstrated an agerelated increase in moderately LBW (1,500–2,500 g) births
to white women receiving Medicaid, and recently Holzman
et al. (10) found an age-related increase in preterm birth
among some high-risk subgroups of white women, although
teens were excluded from this analysis. In short, the impact
of maternal aging on LBW is well established, but socioeconomic modifiers of the relation of maternal age to
LBW and its components are complex and incompletely
understood.
A growing body of evidence indicates that factors early in
a mother’s life also have a significant influence on the outcome of her pregnancies (16, 17), the ‘‘life-course’’ perspective. The creation of a transgenerational birth file linking 2
generations of birth outcomes to economic data from the
census makes possible life-course studies of birth outcomes
(18). Unanswered questions regarding the impact of lifetime
socioeconomic environment on birth outcomes as women
age led us to the following hypotheses: first, that AfricanAmerican women never exposed to poverty would not experience weathering, and second, that white women who are
exposed to poverty throughout their life would display evidence of weathering.
MATERIALS AND METHODS
The Illinois transgenerational birth file has been described in detail elsewhere (18). Briefly, records of livebirths
in Illinois from 1989 to 1991 were linked to the birth certificates of their mothers born in Illinois from 1956 to 1976,
by using the mother’s complete name and date of birth.
Overall, a linkage rate of 78% was achieved in constructing
the Illinois file by using a deterministic approach with
mother’s maiden name and exact date of birth. Unmatched
records were due to slight variations in name spelling or in
dates of birth (18). After this linkage, all identifying information was removed from the transgenerational records.
Records from the Chicago area contained sufficiently detailed geographic coding to allow the determination of the
neighborhood economic context of each birth. Median family income from the census was utilized for this purpose.
Census linking was carried out at the urban community area
or census tract level or at the suburban town, township, or
village level to define residential environment at the level of
the smallest geographic unit available. This was accomplished for all Chicago-area births across the 2 generations
for this analysis (18).
The present study was limited to singleton births to nonHispanic African-American and white women who resided
in Cook County (including the city of Chicago) at the time
of birth. Race was defined by the mother’s race as listed on
the infant’s birth certificate. Maternal age was categorized in
the following way: <20, 20–24, 25–29, and 30–35 years.
The continuous measure of neighborhood income was initially divided into quartiles separately for each generation
based on the distribution of neighborhood incomes for
women of all races in 1956–1976 and 1989–1991. However,
this subgrouping resulted in insufficient numbers in the 2
groups of greatest interest, white women in the lowest quartile for both generations (n ¼ 60) and African-American
women in the highest quartile for both generations (n ¼
197). Therefore, the 2 highest quartiles and the 2 lowest
quartiles were combined, and the resulting halves of the
distribution were referred to as upper and lower income,
respectively. The income cutpoints (i.e., the medians) were
$27,427 for the 1956–1976 generation and $35,427 for the
1989–1991 births, both expressed in 1989 dollars. Using
these definitions for the maternal and infant generations
made it possible to create income groups of sufficient sample size for women experiencing upper-upper, upper-lower,
lower-upper, and lower-lower levels of lifetime neighborhood economic context.
To create variables for preterm birth (gestational age, <37
weeks) and small for gestational age (<10th percentile for
gestational age), gestational age was calculated as the time
in weeks from the last menstrual period, as reported on the
birth certificate, to the date of birth of the infant. When any
element of the last menstrual period was missing, the calculated gestational age was outside the range of 20–44
weeks, or when the birth weight was inconsistent with the
calculated gestational age (19), the clinical estimate of gestation was substituted. SGA was defined on the basis of
a population-based reference that used 1994–1996 US births
to generate the 10th percentile of birth weight for each
gestational age, separately by sex and race (20). This coding
process resulted in less than 0.5% of the sample with missing values for PTB and SGA.
The rates of LBW, SGA, and PTB were calculated by
levels of maternal characteristics, including age. Prenatal
care utilization was determined by using the Kotelchuk index (21). Race-specific multilevel logistic regression
models (22) were constructed with random intercepts to
account for the clustering of individual births (level 1)
within mothers (level 2) and their neighborhoods of residence at the time of the infant’s birth (level 3). Models were
stratified by economic subgroup (lower-lower, lower-upper,
upper-lower, and upper-upper) to test the significance of the
age effect (categorized) on the odds of LBW, SGA, and PTB
for African-American and white women with different experiences of exposure to impoverished environments across
generations. All analyses were performed with SAS, version
Am J Epidemiol 2010;172:127–134
Weathering and Lifelong Economic Environment
20
20
16
12
14.0
11.9
12.3
10
8
6
6.9
4
5.0
4.1
2
18
15.6
4.5
Low Birth Weight, %
Low Birth Weight, %
18
14
129
15.1
16
14
12.5
17.0
13.0
12
10
11.9
10.8
8
6
6.9
4
6.3
2
0
0
<20
20–24
25–29
30–35
Maternal Age, years
Figure 1. Low-birth-weight rates by maternal age in non-Hispanic
African-American (n ¼ 38,995, triangles) and white (n ¼ 31,620,
squares) mothers, Cook County, Illinois, 1989–1991.
9.2, statistical software (SAS Institute, Inc., Cary, North
Carolina).
RESULTS
African-American mothers were significantly more likely
to be younger, less educated, multiparous, unmarried, and
have no or inadequate prenatal care; 67% of AfricanAmerican mothers lived in lower income neighborhoods
in both generations, while 63% of whites lived in upper
income neighborhoods in both generations (Web Table 1).
(This information is described in the first of 2 supplementary
tables; each is referred to as ‘‘Web table’’ in the text and is
posted on the Journal’s website (http://aje.oxfordjournals.
org/).)
All characteristics are significantly associated with each
of the 3 outcomes (LBW, PTB, SGA) within race groups,
but maternal age is the only characteristic that differs in the
direction of the association with LBW and SGA by race,
with increasing rates of both by age group for AfricanAmerican mothers, compared with a decreasing rate with
increasing age for white mothers. Among whites and African Americans, LBW, PTB, and SGA rates all decrease by
categories of lifetime neighborhood economic context, ordered from lower income in both generations, to downwardly mobile, to upwardly mobile, and upper income in
both generations, with only a slight departure from this
pattern for PTB among African Americans (Web Table 2).
We next assessed the effect of maternal age on LBW rates
for African-American and white women, overall and by lifecourse neighborhood context. Figure 1 shows a weathering
pattern for African Americans and a nonweathering pattern
for whites. Figure 2 shows a weathering effect for AfricanAmerican women in the lower-lower group but a pattern
inconsistent with weathering and more similar to the overall
white experience in the upper-upper group.
To better characterize this phenomenon, we analyzed
the components of LBW (i.e., PTB and SGA) separately.
Figure 3 presents the SGA rates for each maternal age group
Am J Epidemiol 2010;172:127–134
<20
20–24
25–29
30–35
Maternal Age, years
Figure 2. Low-birth-weight rates by maternal age in non-Hispanic
African-American mothers experiencing lower-lower (n ¼ 26,043, triangles) versus upper-upper (n ¼ 1,904, squares) neighborhood economic context, Cook County, Illinois, 1989–1991. Lower-lower, lower
half of neighborhood income distribution during maternal and infant
generations; upper-upper, upper half of neighborhood income distribution during maternal and infant generations.
for all African-American and white women. There is evidence for weathering among African-American, but not
white, women. SGA rates for African-American women
were lowest in the teenage years (10.7%) and significantly
increased through the child-bearing years to 13.0% in
women over 30 years of age. For white women, SGA rates
were highest in the teenage years (15.3%) with a progressive
decrease through the 20s to a low of 9.0% at age 30 years or
older. Figure 4 shows that the pattern of the age-SGA relation for upper-upper African Americans is very similar to
the pattern seen for whites in Figure 3, while the weathering
effect is still present among African Americans experiencing lower-lower neighborhood economic context.
In Figure 5, PTB rates for the 4 age groups are shown by
race. Both curves show no evidence of weathering, with
a general downward trend in PTB rate with increasing age
until age 30–35 years, where there is evidence of the start
of a U-shaped curve. A similar pattern is shown for
African-American women in the upper-upper and lowerlower economic subgroups (Figure 6). No statistically significant weathering pattern was demonstrable for LBW,
SGA, or PTB in our white population, either in total or
when disaggregated by life-course economic environment
(Web Figures 1–3). (These supplementary figures are
posted on the Journal’s website (http://aje.oxfordjournals.
org/).)
Tables 1–3 show the results of the 3-level models for
LBW, SGA, and PTB, respectively, accounting for the clustering of births within mothers within neighborhoods. For
all outcomes, the odds ratios and 95% confidence intervals
for each age group, compared with a common referent group
of <20 years of age, are shown for African-American and
white women with different lifetime neighborhood economic contexts. A significant increase in the odds of LBW
(Table 1) and SGA (Table 2) is noted for African Americans
in the 25–29- and 30–35-year age categories but not
for white women exposed to a poorer neighborhood
20
20
18
18
16
14
12
10.7
10
17.7
16
15.3
10.9
10.9
8
12.1
13.0
9.4
9.0
6
Preterm Birth, %
Small for Gestational Age, %
130 Love et al.
16.3
16.3
17.0
6.3
6.5
14
12
10
8
10.8
7.8
6
4
4
2
2
0
0
<20
20–24
25–29
<20
30–35
Maternal Age, years
20–24
25–29
30–35
Maternal Age, years
Figure 5. Preterm birth rates by maternal age in non-Hispanic
African-American (n ¼ 38,818, triangles) and white (n ¼ 31,591,
squares) mothers, Cook County, Illinois, 1989–1991.
environment throughout their lives. A similar trend was also
observed for those living in poorer neighborhoods only as
adults, but the 95% confidence intervals around the odds
ratios for this group included the null. Weathering (defined
here as an odds ratio of increasing magnitude and significantly above 1 for each age category) was not observed in
upwardly mobile, the lower-upper group, or in the upperupper group of African-American women. In fact, the odds
ratios for these groups suggest falling odds of LBW and
SGA with age, of larger magnitude in the upper-upper
group, although the 95% confidence intervals around the
estimates for SGA include the null value. This pattern
among African-American mothers experiencing non-poor
neighborhood context is qualitatively like the pattern observed among the white population. A weathering effect is
not seen for PTB among African-American or white
mothers in any of the economic context groups (Table 2).
In general, the 20–24 or 25–29 group is the lowest risk
group for PTB among African Americans in all strata of
neighborhood context, although the odds ratio for African
Americans aged 20–24 years in the lower-lower group was
the only one to reach statistical significance. Among whites
in the lower-lower and upper-upper groups, all age groups
20 years or older have significantly lower odds of PTB than
teens.
DISCUSSION
The phenomenon of weathering is again demonstrated in
African-American women in the Chicago, Illinois, area population studied. Of interest, however, is the finding that
African-American women exposed to the upper half of
20
20
18
18
16
14
12
10
8
12.9
11.1
14.5
11.3
10.9
8.8
6
8.5
6.9
14
16.8
17.3
17.6
12.7
13.2
15.7
14.5
12
10
8
6
4
4
2
2
0
18.4
16
Preterm Birth, %
Small for Gestational Age, %
Figure 3. Small-for-gestational-age rates by maternal age in
non-Hispanic African-American (n ¼ 38,784, triangles) and white
(n ¼ 31,588, squares) mothers, Cook County, Illinois, 1989–1991.
0
<20
20–24
25–29
30–35
Maternal Age, years
Figure 4. Small-for-gestational-age rates by maternal age in
non-Hispanic African-American mothers experiencing lower-lower
(n ¼ 25,894, triangles) versus upper-upper (n ¼ 1,894, squares)
neighborhood economic context, Cook County, Illinois, 1989–1991.
Lower-lower, lower half of neighborhood income distribution during
maternal and infant generations; upper-upper, upper half of neighborhood income distribution during maternal and infant generations.
<20
20–24
25–29
30–35
Maternal Age, years
Figure 6. Preterm birth rates by maternal age in non-Hispanic
African-American mothers experiencing lower-lower (n ¼ 25,916,
triangles) versus upper-upper (n ¼ 1,897, squares) neighborhood
economic context, Cook County, Illinois, 1989–1991. Lower–lower,
lower half of neighborhood income distribution during maternal and
infant generations; upper-upper, upper half of neighborhood income
distribution during maternal and infant generations.
Am J Epidemiol 2010;172:127–134
Weathering and Lifelong Economic Environment
131
Table 1. Results of Multilevela Logistic Regression Models for the Relation Between Maternal Age and Low Birth Weight Among Non-Hispanic
African-American and White Women, Stratified by Lifetime Neighborhood Economic Context, Cook County, Illinois, 1989–1991
Lifetime Neighborhood Economic Context
Lower-Lower
No.
OR
95% CI
Upper-Lower
No.
OR
95% CI
Lower-Upper
No.
OR
95% CI
Upper-Upper
No.
OR
95% CI
African Americans
Births
26,043
6,338
4,710
1,904
Mothers
22,279
5,521
4,447
1,769
523
445
235
170
Neighborhoods
Maternal age, years
Referent
<20
Referent
Referent
Referent
20–24
1.06
0.96, 1.18
1.12
0.95, 1.33
0.74
0.54, 1.01
0.90
0.65, 1.26
25–29
1.28
1.15, 1.42
1.23
0.90, 1.67
0.84
0.62, 1.13
0.55
0.32, 0.95
30
1.48
1.32, 1.67
1.63
1.13, 2.34
0.98
0.72, 1.34
0.50
0.26, 0.98
Whites
Births
2,093
4,683
4,983
19,861
Mothers
1,945
4,295
4,584
18,075
337
391
324
364
Neighborhoods
Maternal age, years
Referent
<20
Referent
Referentb
Referentb
20–24
0.60
0.31, 1.15
1.08
0.73, 1.59
0.71
0.09, 5.81
0.61
0.43, 0.87
25–29
0.40
0.22, 0.74
1.22
0.83, 1.80
1.32
0.18, 9.80
0.48
0.34, 0.66
30
0.62
0.34, 1.16
1.16
0.77, 1.76
1.32
0.18, 9.87
0.57
0.41, 0.79
Abbreviations: CI, confidence interval; OR, odds ratio.
Births are clustered within mothers, and mothers are clustered within census tracts of residence at the time of the birth.
b
Models for whites in the lower-upper and upper-upper groups were run as 2-level models (births clustered within neighborhoods) because of
convergence issues with the 3-level version of the models. These models were also run with births clustered within mothers, and the results did not
change (data not shown).
a
neighborhood income as both infants and adults showed no
weathering. Rather, they had a pattern of LBW incidence
that was qualitatively similar to that observed in white
women, with a nadir rate of 6.3% in women aged 30–35
years. Weathering in our population was driven by the relation of maternal age to the rate of SGA births, while PTB
rates were not affected in this way. White mothers, even
those with a lifelong residence in lower income neighborhoods, did not display significant weathering. The results
indicated that exposure to lower neighborhood income did
amplify the trend for worsening birth outcomes as AfricanAmerican mothers aged and that the timing of exposure to
lower income may be important, because exposure in adulthood seems to have the strongest effects.
Overall, very few African-American births (4.9%) were
found in the upper-upper economic group, and very few
white births (6.6%) were found in the lower-lower group,
even using the neighborhood median as the cutpoint. Hence,
the population of nonimpoverished African-American
women and relatively poor white women, the groups of
greatest interest for this analysis, were the least available
for study.
The introduction into this analysis of women’s lifetime
neighborhood economic context provided important new
information about the modifying effect of exposure to poor
and nonpoor neighborhoods across the life course on the
Am J Epidemiol 2010;172:127–134
relation between maternal age and birth outcomes, especially for African-American mothers. The transgenerational
birth file used for this analysis allowed for more exploration
into the effects of the socioeconomic environment at critical
periods in a woman’s life, specifically when she is developing in utero and when she gives birth.
The impact of an adverse economic or social environment
over time has been hypothesized to derive from 2 mechanisms, fetal programming and cumulative wear and tear.
Early programming attributes adult morbidity to fetal or
early childhood exposures such as undernutrition, often using infant LBW as a marker (23, 24). These early exposures
have been associated with adult outcomes, including diabetes, hypertension, and diminished reproductive success.
The ‘‘wear and tear’’ mechanism, or weathering effect,
which is manifest by the tendency toward worsening health
status with age, including worse infant outcomes at increasing maternal age, is thought to result from the impact of
cumulative exposure to hardship (4, 5). This mechanism is
supported by the observation of substantially increased incidence of smoking, hypertension, and other pregnancy risk
factors observed in African-American women compared
with white women in the older maternal age groups (5,
25). Many of these intervening risk factors show no difference by race, or their incidence favors African-American
women when comparisons are made at younger ages (5).
132 Love et al.
Table 2. Results of Multilevela Logistic Regression Models for the Relation Between Maternal Age and Small for Gestational Age Among NonHispanic African-American and White Women, Stratified by Lifetime Neighborhood Economic Context, Cook County, Illinois, 1989–1991
Lifetime Neighborhood Economic Context
Lower-Lower
No.
OR
95% CI
Upper-Lower
No.
OR
95% CI
Lower-Upper
No.
OR
95% CI
Upper-Upper
No.
OR
95% CI
African Americans
Births
25,894
6,308
4,688
1,894
Mothers
22,186
5,502
4,429
1,761
523
444
235
170
Neighborhoods
Maternal age, years
Referent
<20
Referent
Referent
Referent
20–24
1.03
0.93, 1.15
1.08
0.90, 1.29
0.84
0.61, 1.17
0.79
0.56, 1.12
25–29
1.21
1.09, 1.35
1.22
0.89, 1.68
0.77
0.57, 1.05
0.76
0.46, 1.26
30
1.40
1.24, 1.58
1.42
0.97, 2.09
0.73
0.53, 1.03
0.61
0.32, 1.15
Whites
Births
2,091
4,670
4,980
19,847
Mothers
1,943
4,283
4,581
18,064
336
391
324
364
Neighborhoods
Maternal age, years
<20
Referent
Referent
Referent
Referent
20–24
0.81
0.47, 1.38
0.75
0.58, 0.96
0.89
0.25, 3.15
0.69
0.54, 0.90
25–29
0.61
0.37, 1.01
0.75
0.58, 0.97
1.02
0.30, 3.44
0.59
0.46, 0.75
30
0.96
0.57, 1.60
0.61
0.46, 0.81
0.92
0.27, 3.11
0.59
0.46, 0.75
Abbreviations: CI, confidence interval; OR, odds ratio.
Births are clustered within mothers, and mothers are clustered within neighborhoods of residence at the time of the birth.
a
Past studies investigating African-American mothers exposed to relative poverty during pregnancy found more
weathering, even after adjustment for maternal health characteristics (5). The authors were able to reproduce this effect
of adult socioeconomic status on weathering among African
Americans, with both the upper-lower and the lower-lower
groups showing an effect of maternal age on SGA rate,
although only the lower-lower group reached statistical significance. The opposite pattern in the lower-upper group
would suggest that fetal programming, while perhaps amplifying the weathering effect, is less important in producing
weathering than the woman’s adult environment.
To date, only a few studies of weathering have sought to
tease apart the component pathways to the rising risk of
LBW with maternal age in African-American women. Only
one study of SGA rates with increasing age was identified,
and it was limited to a very low risk subset of primigravidas
with an SGA rate of only 1%–2% rather than the expected
~10% (15). Weathering reflected in PTB was reported in 3
previous studies when African-American women were stratified into different individual- and neighborhood-level risk
groups (10, 13, 14) but not when they were examined as
a whole (12), as done in the current study.
To our knowledge, the present study is the first to show
a declining rate of any adverse birth outcome with maternal
age for the small subgroup of African-American women who
never resided in low-income neighborhoods, a pattern qual-
itatively like the age risk pattern seen among white women
for all adverse birth outcomes studied. The mechanism underlying this interesting observation remains to be clarified,
but it seems to be driven by environmental effects on intrauterine growth more strongly than length of gestation.
The persistence of an elevated incidence of adverse birth
outcomes at all age groups in African-American compared
with white mothers, even among the group that never lived
in neighborhoods below the median income, in part reflects
incomplete control for racial inequality in neighborhood
economic context as a result of our study design. Even our
economic categories masked racial income inequality,
which persisted within the upper and lower income halves
used. For example, African-American women residing in
neighborhoods below the median of neighborhood income
were nearly 4 times as likely as whites to be in the lowest
quartile. Moreover, it has been demonstrated that some psychosocial stressors such as racial bias, token stress, role
overload, and lower purchasing power can also lead to poor
birth outcomes among African Americans (5, 26). Many of
the most affluent neighborhoods in Cook County are in suburbs where over 90% of the population is white (18).
Stressors directly related to perceived or actual racial discrimination persist despite the neighborhood economic environment (5, 26, 27). The impact of perceived racism has
recently been shown to be stronger in neighborhoods where
African Americans, having escaped the ghetto, comprise
Am J Epidemiol 2010;172:127–134
Weathering and Lifelong Economic Environment
133
Table 3. Results of Multilevela Logistic Regression Models for the Relation Between Maternal Age and Preterm Birth Among Non-Hispanic
African-American and White Women, Stratified by Lifetime Neighborhood Economic Context, Cook County, Illinois, 1989–1991
Lifetime Neighborhood Economic Context
Lower-Lower
No.
OR
95% CI
Upper-Lower
No.
OR
95% CI
Lower-Upper
No.
OR
95% CI
Upper-Upper
No.
OR
95% CI
African Americans
Births
25,916
6,315
4,690
1,897
Mothers
22,201
5,505
4,431
1,764
523
445
235
170
Neighborhoods
Maternal age, years
Referent
<20
Referent
Referent
Referent
20–24
0.90
0.82, 0.98
0.97
0.84, 1.12
0.84
0.63, 1.12
0.91
0.68, 1.22
25–29
0.93
0.85, 1.02
0.85
0.64, 1.13
0.82
0.52, 1.08
0.79
0.51, 1.20
30
0.96
0.86, 1.07
0.99
0.70, 1.40
1.03
0.77, 1.36
0.82
0.50, 1.34
Whites
Births
2,092
4,671
4,980
19,848
Mothers
1,944
4,284
4,581
18,065
336
391
324
364
Neighborhoods
Maternal age, years
Referent
<20
Referent
Referent
Referent
20–24
0.54
0.30, 0.96
1.00
0.74, 1.36
0.46
0.15, 1.41
0.61
0.45, 0.81
25–29
0.50
0.30, 0.85
0.82
0.60, 1.14
0.44
0.15, 1.27
0.51
0.39, 0.67
30
0.48
0.27, 0.84
0.73
0.51, 1.03
0.49
0.17, 1.41
0.56
0.43, 0.74
Abbreviations: CI, confidence interval; OR, odds ratio.
Births are clustered within mothers, and mothers are clustered within census tracts of residence at the time of the birth.
a
a racial minority within a community made up mostly of
whites in their income bracket (28).
We were not able to demonstrate weathering in white
women, even among those living in lower-income neighborhoods throughout their life course. Again, this may in part
reflect persistent inequality of exposure to the lowest income
environments when using the median as a cutpoint. That
feature of the study design was necessitated by the extremely small number of births to white women living in
the lowest quartile income neighborhoods throughout their
lives—only 69 of the 31,591 white births in the cohort population. The racial economic divide becomes much starker
when viewed from the life-course perspective used in the
present study. Few white mothers truly experience the microenvironment of poor inner-city neighborhoods (14, 29,
30). This may explain why weathering has been found in
white mothers only when a more complete picture of socioeconomic context (i.e., living in poor neighborhoods, smoking, publicly insured) was used to describe a subset of white
women with risk profiles similar to those of AfricanAmerican mothers (6, 10).
Unfortunately, both sample and data constraints limited
this study. Because of the cross-sectional nature of the vital
records data, our measurement of ‘‘lifetime’’ neighborhood
economic context is necessarily limited to 2 time points
during the life of the mother. However, having a measure
of economic context at 2 time points is an improvement over
prior research, because it comes closer to measuring wealth
than indicators measured only at the time of delivery (31). In
Am J Epidemiol 2010;172:127–134
addition, there was heterogeneity in the size of the geographic units used to categorize income depending on city
versus suburban dwelling, but due to the nature of social
segregation in the Chicago metropolitan area, misclassification of neighborhood economic context as a result of this is
likely to be minimal (18). The procedure used to exclude
records with implausible gestational ages for birth weight
has been shown to not fully account for errors in the last
menstrual period date (32), which likely resulted in misclassification of the PTB outcome. This may explain the null
findings for the weathering effect on preterm birth among
African Americans. Because of the characteristics of the
linkage, there are no mothers aged over 35 years in the
sample, and the proportion of women over 30 years of age
is lower than that seen in the full 1989–1991 birth cohort. In
addition, the Illinois transgenerational file has a small but
demonstrable skew toward more educated women compared
with the general birth population of Illinois, which could
limit the generalizability of the study’s findings. However,
the rates of LBW and mortality in the file are virtually identical to those of the state as a whole (18), so the size of this
distortion is manageable. The small number of Hispanic,
Asian, and other groups besides whites and African Americans precluded analysis of possible weathering in these
populations.
In conclusion, this study adds to the extensive literature
on racial differences in birth outcomes by showing the impact of lifetime neighborhood economic context on weathering. Current health-care efforts to improve birth outcomes
134 Love et al.
in African-American mothers are focused primarily upon
the period of pregnancy and birth. Our investigation suggests that, indeed, improving women’s adult economic environment can improve birth outcomes at increasing
ages for African Americans. However, to create the possibility of eliminating racial inequalities in birth outcomes,
interventions—including changing economic and social
environments—will have to extend over the life course,
perhaps over multiple generations.
13.
14.
15.
16.
ACKNOWLEDGMENTS
Author affiliations: Division of Neonatology, John H.
Stroger, Jr., Hospital, Chicago, Illinois (Catherine Love,
Richard J. David); Department of Pediatrics, University of
Illinois School of Medicine, Chicago, Illinois (Richard J.
David); Department of Epidemiology, University of Illinois
School of Public Health, Chicago, Illinois (Kristin M.
Rankin); and Department of Pediatrics, Feinberg School
of Medicine, Northwestern University, and Children’s Memorial Hospital, Chicago, Illinois (James W. Collins, Jr.).
This study was funded by a research grant from the March
of Dimes (12-FY04-45 and 12-FY09-159) to J. W. C.
Conflict of interest: none declared.
17.
18.
19.
20.
21.
REFERENCES
1. O’Campo P, Xue X, Wang MC, et al. Neighborhood risk factors for low birthweight in Baltimore: a multilevel analysis.
Am J Public Health. 1997;87(7):1113–1118.
2. David RJ, Collins JW Jr. Bad outcomes in black babies: race or
racism? Ethn Dis. 1991;1(3):236–244.
3. Martin JA, Kung HC, Mathews TJ, et al. Annual summary of
vital statistics: 2006. Pediatrics. 2008;121(4):788–801.
4. Geronimus AT. The weathering hypothesis and the health of
African-American women and infants: evidence and speculations. Ethn Dis. 1992;2(3):207–221.
5. Geronimus AT. Black/white differences in the relationship of
maternal age to birthweight: a population-based test of the
weathering hypothesis. Soc Sci Med. 1996;42(4):589–597.
6. Rauh VA, Andrews HF, Garfinkel RS. The contribution of
maternal age to racial disparities in birthweight: a multilevel
perspective. Am J Public Health. 2001;91(11):1815–1824.
7. Rich-Edwards JW, Buka SL, Brennan RT, et al. Diverging
associations of maternal age with low birthweight for black
and white mothers. Int J Epidemiol. 2003;32(1):83–90.
8. Collins JW Jr, Simon DM, Jackson TA, et al. Advancing maternal age and infant birth weight among urban African
Americans: the effect of neighborhood poverty. Ethn Dis.
2006;16(1):180–186.
9. Buescher PA, Mittal M. Racial disparities in birth outcomes
increase with maternal age: recent data from North Carolina.
N C Med J. 2006;67(1):16–20.
10. Holzman C, Eyster J, Kleyn M, et al. Maternal weathering and
risk of preterm delivery. Am J Public Health. 2009;99(10):
1864–1871.
11. Ekwo EE, Moawad A. Maternal age and preterm births in a black
population. Paediatr Perinat Epidemiol. 2000;14(2):145–151.
12. Ananth CV, Misra DP, Demissie K, et al. Rates of preterm
delivery among black women and white women in the United
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
States over two decades: an age-period-cohort analysis. Am J
Epidemiol. 2001;154(7):657–665.
Schempf AH, Branum AM, Lukacs SL, et al. Maternal age and
parity-associated risks of preterm birth: differences by race/
ethnicity. Paediatr Perinat Epidemiol. 2007;21(1):34–43.
Osypuk TL, Acevedo-Garcia D. Are racial disparities in preterm birth larger in hypersegregated areas? Am J Epidemiol.
2008;167(11):1295–1304.
Sheeder J, Lezottte D, Stevens-Simon C. Maternal age and the
size of white, black, Hispanic, and mixed infants. J Pediatr
Adolesc Gynecol. 2006;19(6):385–389.
Lu MC, Halfon N. Racial and ethnic disparities in birth outcomes: a life-course perspective. Matern Child Health J.
2003;7(1):13–30.
Emanuel I, Filakti H, Alberman E, et al. Intergenerational
studies of human birthweight from the 1958 birth cohort. 1.
Evidence for a multigenerational effect. Br J Obstet Gynaecol.
1992;99(1):67–74.
David R, Rankin K, Lee K, et al. The Illinois transgenerational
birth file: life-course analysis of birth outcomes using vital
records and census data over decades. Matern Child Health J.
2010;14(1):121–132.
Alexander GR, Himes JH, Kaufman RB, et al. A United States
national reference for fetal growth. Obstet Gynecol. 1996;
87(2):163–168.
Alexander GR, Kogan MD, Himes JH. 1994–1996 U.S. singleton birth weight percentiles for gestational age by race,
Hispanic origin, and gender. Matern Child Health J. 1999;
3(4):225–231.
Kotelchuck M. The Adequacy of Prenatal Care Utilization
Index: its US distribution and association with low birthweight. Am J Public Health. 1994;84(9):1486–1489.
Snijders TAB, Bosker RJ. Multilevel Analysis: An Introduction
to Basic and Advanced Multilevel Modeling. Thousand Oaks,
CA: Sage Publications; 1999.
Barker DJ, Winter PD, Osmond C, et al. Weight in infancy and
death from ischaemic heart disease. Lancet. 1989;2(8663):
577–580.
Gluckman PD, Hanson MA, Cooper C, et al. Effect of in utero
and early-life conditions on adult health and disease. N Engl J
Med. 2008;359(1):61–73.
Geronimus AT, Hillemeier MM. Patterns of blood lead levels
in US black and white women of childbearing age. Ethn Dis.
1992;2(3):222–231.
Colen CG, Geronimus AT, Bound J, et al. Maternal upward
socioeconomic mobility and black-white disparities in infant
birthweight. Am J Public Health. 2006;96(11):2032–2039.
Krieger N, Rowley DL, Herman AA, et al. Racism, sexism,
and social class: implications for studies of health, disease, and
well-being. Am J Prev Med. 1993;9(6 suppl):82–122.
Dailey AB, Kasl SV, Holford TR, et al. Neighborhood- and
individual-level socioeconomic variation in reported racial discrimination [abstract]. Am J Epidemiol. 2008;167(suppl):S34.
David R, Collins J Jr.. Disparities in infant mortality: what’s
genetics got to do with it? Am J Public Health. 2007;97(7):
1191–1197.
Massey DS, Denton NA. American Apartheid: Segregation
and the Making of the Underclass. Cambridge, MA: Harvard
University Press; 1993.
Braveman PA, Cubbin C, Egerter S, et al. Socioeconomic
status in health research: one size does not fit all. JAMA.
2005;294(22):2879–2888.
Platt RW, Abrahamowicz M, Kramer MS, et al. Detecting and
eliminating erroneous gestational ages: a normal mixture
model. Stat Med. 2001;20(23):3491–3503.
Am J Epidemiol 2010;172:127–134