Download Socioeconomic position and the risk of preterm birth—a study within

Published by Oxford University Press on behalf of the International Epidemiological Association
ß The Author 2008; all rights reserved. Advance Access publication 24 June 2008
International Journal of Epidemiology 2008;37:1109–1120
doi:10.1093/ije/dyn112
Socioeconomic position and the risk of
preterm birth—a study within the
Danish National Birth Cohort
Camilla Schmidt Morgen,1* Christina Bjørk,1 Per Kragh Andersen,2 Laust Hvas Mortensen1 and
Anne-Marie Nybo Andersen1,3
Accepted
24 April 2008
Background Low socioeconomic position is generally associated with increased
risk of preterm birth, but it remains unclear whether the inequality
depends on the socioeconomic measure used, if the associations
differ according to the degree of prematurity, and how individual
level risk factors mediate the association.
Methods
The hazard ratios (HR) of preterm birth associated with five different measures of socioeconomic position and three degrees of
preterm birth were analysed in a dataset of 75 890 singleton
pregnancies (1996–2002) from the Danish National Birth Cohort.
This, and the mediating role of selected individual level risk factors
(smoking, alcohol consumption, binge drinking, pre-pregnancy
body mass index, gestational weight gain) were estimated, using
Cox regression analyses.
Results
Mothers with <10 years of education had an elevated risk of
preterm birth compared with mothers with 412 years of education
and the association interacted with parity, while income and
occupation affected the risk to a lesser degree. The adjusted HR for
less educated nulliparous and parous women were 1.22 (95% CI
1.04–1.42) and 1.56 (95% CI 1.31–1.87), respectively, compared
with women with 412 years of education. For parous women with
<10 years of education inclusion of smoking in the model
decreased the HR of preterm birth to 1.43 (95% CI 1.19–1.72).
Conclusions Maternal educational level was the strongest predictor of preterm
birth among five socioeconomic measures and the gradient did not
differ significantly according to the degree of preterm birth. For
parous women smoking explained some of the educational gradient
but in general the selected risk factors only reduced the relative
educational gradient in preterm birth marginally.
Keywords
1
2
3
Premature labour, gestational
pregnancy, obstetrics
National Institute of Public Health, Copenhagen, Denmark.
Department of Biostatistics, University of Copenhagen,
Denmark.
Epidemiology, Institute of Public Health, University of
Southern Denmark, Denmark.
age,
socioeconomic
position,
* Corresponding author. National Institute of Public Health,
University of Southern Denmark, Oester Farimagsgade 5,
DK-1399 Copenhagen K, Denmark. E-mail: [email protected]
1109
1110
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Introduction
Preterm birth is associated with high risk of infant
mortality and morbidity and has in most cases an
unknown aetiology.1,2 The risk of preterm birth has
been shown to be elevated among the socioeconomic
disadvantaged women when measured by maternal
educational level.3–10 Few studies have investigated
more than one socioeconomic indicator3,7,11 as well as
the socioeconomic position of the father.8 Heterogeneity of the association between different socioeconomic
characteristics and preterm birth might shed light on
the mechanisms behind a possible social gradient.
Extremely preterm birth is associated with high risks
of death, chronic lung disease,12 cerebral palsy,13
retinopathy of prematurity14 and other kinds of disabilities,15 while the moderately preterm born child
faces a risk of disability that is almost similar to the
child born term.16,17 Because of these profoundly
different consequences according to the degree of
prematurity, interventions aiming at reducing the risk
of death and severe disability among the children born
preterm should focus on reducing the number of
clinically important cases.18 However, many studies
have failed to report more than one degree of prematurity (<37 full weeks of gestation) when considering
socioeconomic inequality in preterm birth. Previously
it has been shown that people of disadvantaged
socioeconomic position are more likely to live a sedentary lifestyle, to be overweight and to be smokers.19,20
Therefore, it has been hypothesized that socioeconomic
disadvantage affects health through a higher prevalence of individual level risk factors among those
with lower levels of education21 but it has to our
knowledge not previously been quantified in the case
of preterm birth.
In the present study, we analysed data from 75 980
women with a singleton pregnancy in Denmark
between 1996 and 2002. The aim of this article was
to clarify and compare how five different indicators of
socioeconomic position were associated with preterm
birth in order to create a better understanding for
the mechanisms linking socioeconomic inequality
with preterm birth. Further, the aim was to clarify
whether the associations differed according to the
degree of preterm birth and to explore to what extend
five selected risk factors could explain some of the
socioeconomic inequality in preterm birth. We hypothesized education, occupation and household income
to represent different aspects of socioeconomic position22,23 and, therefore, to have different associations
with preterm birth. We hypothesized smoking,16,24–27
alcohol consumption,28–31 pre-pregnancy body mass
index (BMI)32 and gestational weight33 gain to be
possible intermediate variables linking socioeconomic disadvantage with preterm birth. These indicators were chosen because of their possible
association with socioeconomic position34–38 and
preterm birth18,24,27–33 and because of their statistical
association with preterm birth in our data material.
Our findings may improve the understanding of
the factors and processes that mediate the socioeconomic disparities in preterm birth and may create
a basis for future interventions or studies with the
purpose to reduce the socioeconomic inequality in
preterm birth.
Methods
This study was carried out within the Danish National
Birth Cohort (DNBC), a nationwide ongoing study
of pregnant women and their offspring. Between 1996
and 2002 women were invited to the cohort at their
first antenatal visit at the general practitioner. It is
estimated that about 30% of all Danish pregnant
women in the years between 1996 and 2002 were
recruited to the cohort. The women were included
if they intended to carry the pregnancy to term, had a
permanent address in Denmark and spoke Danish
well enough to participate in telephone interviews. The
women provided information on exposures during
pregnancy by means of computer-assisted telephone
interviews and the first interview was scheduled to
take place in pregnancy week 12 (range 7–37). More
details about the cohort are presented elsewhere.39–41
A total of 100 418 pregnant women were enrolled
in the cohort. For this study, we initially included
the 90 165 pregnancies for which we had a first
pregnancy interview. Subsequently, we excluded 1985
women with multiple pregnancies, 918 women with a
pregnancy terminated before 22 completed gestational
weeks, 5428 women because of missing data on the
covariates used in the analysis, 5833 pregnancies
because the women participated in the cohort with
more than one pregnancy (the first pregnancy was
included), 21 women because the interview took place
later than 37 weeks of gestation. Thus, 75 980 were
eligible for analyses in this study. A total of 97.4% of
the women were of Danish origin, 2% were from
Scandinavian or other OECD countries and 0.5% from
the rest of the world (0.1% unknown).
Measurement of outcome
The outcome measure of interest was gestational age
at delivery. This information was based on information
from the National Discharge Register. These estimates
were predominantly based on ultrasound examination
before 24 weeks of gestation, since all Danish women
are invited for a scan at this stage. Preterm delivery
was defined as birth after 22 completed gestational
weeks (after 153 days) and before 37 completed
gestational weeks (before 259 days) and was subdivided into three degrees of prematurity: extremely
preterm (22–27 completed weeks), very preterm
(28–31 completed weeks) and moderately preterm
birth (32–36 completed weeks).
SOCIOECONOMIC POSITION AND THE RISK OF PRETERM BIRTH
Measurement of exposure
Individual information of socioeconomic measures for
each year was obtained from the Integrated Database
for Labour Market Research in the year before the
birth. The national educational codes were categorized
according to the international ISCED classification
system and were converted into three educational
groups reflecting the highest number of years of
completed academic educational attainment [9 years
or less (pre-primary, primary and lower secondary),
10–12 years (upper secondary, post-secondary) and
13 years or more (tertiary)]. Occupation was categorized as: self-employed, employed (blue collar, lower
and upper white collar workers), unemployed, student, disability-retired and unknown. Income was
measured as the disposable household-size income
the year before birth and calculated as the sum of the
parents income adjusted for the number of adults and
children in the household according to an OECD
method.42 Income was grouped into six categories
by percentiles: (i) <5 percentile, (ii) 5–25 percentile,
(iii) 425–50 percentile, (iv) 450–75 percentile,
(v) 475–95 percentile and (vi) 495 percentile based
on the income of all Danish households with a child
born between 1996 and 2002.
Measurement of potential confounders
and effect modifiers
The covariates included were: maternal and paternal
age (<25, 25–29, 30–34, 35–39, 540), parity (0, 1þ),
fertility treatment (yes, no), bleeding during pregnancy (yes, no), maternal cohabitation (living alone,
cohabitant), maternal height (4160 cm, 4160 to
<170 cm, 5170 cm) and chronic disease (defined as
having any of the following; high blood pressure,
metabolic disorder, muscle and skeletal diseases,
mental disorder, diseases in the urinary system,
diseases in the abdominal region, urinary bladder
infection more than five times or having had a
cervical cone biopsy).
Potentially mediating individual level
risk factors
The selected risk factors were cigarette smoking (not
smoking, 0–10 g of tobacco/day, 410 g of tobacco/day),
alcohol drinking (non-drinking, <1 unit of alcohol/
week, 1–2 units/week, 3–5 units/week and 45 units/
week), binge drinking (yes, no), BMI (<18.5,
18.5–24.99, 25–29.99, 30–34.99 and 35þ according
to WHO guidelines43) and gestational weight
gain (<0.2 kg/week, 0.2 to <0.5 kg/week and
50.5 kg/week).
Statistical analyses
The hazard ratios (HR) of preterm delivery according
to five different measures of socioeconomic position
were estimated using Cox regression. Gestational age
in days was used as the underlying time. We used
1111
a model with delayed entry, so women entered the
cohort at the day she completed 22 weeks of gestation
(154 days) or on the day of her first pregnancy
interview, whatever came last. The follow-up ended at
birth, emigration, maternal death or by the time she
completed 37 weeks of gestation (258 days), whichever came first. Deliveries that occurred after 258 days
were censored at that time. We conducted five sets of
analyses estimating the relations between maternal
educational level, paternal educational level, maternal
occupation, paternal occupation and household
income, respectively, and preterm birth. Since parity
modified the effect of maternal educational level and
paternal occupation these analyses were conducted
for nulliparous and parous separately. We estimated
the influence of socioeconomic position on extremely,
very and moderate preterm birth, respectively, by
including an interaction term between the five
indicators of socioeconomic position and time, defined
as 22–27 completed weeks of gestation, 28–31 completed weeks of gestation and 32–36 completed weeks
of gestation. Variables that changed the estimate
(done by stepwise backwards elimination) by 45%
were included in the analysis.44 We tested for interactions, which a priori were considered plausible, i.e.
interactions between the five socioeconomic measures
and cohabitation, fertility treatment, parity, paternal
and maternal age, respectively.
The crude analyses were repeated in a subpopulation
of women with no record of any chronic disease and
in an unselected population of mothers and fathers
of all Danish children born between 1996 and 2002.
The HR of preterm birth were estimated for the group
with unknown exposure level for each of the five
socioeconomic indicators. The potential intermediate
variables were tested for association with education
on a 5% significance level by the w2 test and the
variables were tested for relation to preterm birth in
Cox regression analyses. The mediating role of risk
factors was assessed by comparing the adjusted
estimates of the association between maternal educational level and preterm birth before and after including the variables in the Cox regression model. All
statistical analyses were performed using the SAS
software package version 9.1.
Results
The mean gestational age was 279 days (range
154–308 days) and the overall proportion of preterm
birth was 5.0% (3.797/7.980). Among the preterm
births, the proportion of extremely preterm was 7.7%,
the proportion of very preterm was 10.4% and the
proportion of moderately preterm births was 81.9%.
Based on the percentage wise distribution of characteristics of the mother and father in relation to
preterm birth, it seemed that women with all types of
preterm birth was more often lower educated, having
the child with a man with a low educational level,
1112
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
living in a household with an income below the 25–50
percentile, in fertility treatment prior to the current
pregnancy, smoking during pregnancy, cohabitant
and bleeding during pregnancy prior to the interview.
Furthermore, women with a preterm birth seemed to
have a greater proportion of nulliparous and women
with a chronic disease. Women with a preterm birth
were not different from women with term births
regarding maternal age, paternal age, maternal
occupation, paternal occupation, maternal height,
mean alcohol consumption, episodes with binge
drinking during pregnancy, mean pre-pregnancy
BMI and average weight gain in the first trimester
(Tables 1 and 2).
Table 1 Socioeconomic characteristics in relation to gestational age, n ¼ 75 980
Characteristics
Education, mother
1–10 years
410–12 years
412 years
Missing
Gestational age in full weeks at delivery
22–27 (n ¼ 293)
28–31 (n ¼ 396)
32–36 (n ¼ 3108)
37–43 (n ¼ 72 183)
57 (19.5)
68 (17.2)
491 (15.8)
9540 (13.2)
145 (49.5)
205 (51.8)
1576 (51.7)
35 491 (49.2)
89 (30.4)
117 (29.6)
1019 (32.8)
26 762 (37.1)
2 (0.7)
6 (1.5)
22 (0.7)
390 (0.5)
Education, father
1–10 years
410–12 years
412 years
25 (8.5)
66 (16.7)
542 (17.4)
11424 (15.8)
66 (22.5)
179 (45.2)
1585 (51.0)
36 801 (51.0)
30 (10.2)
84 (21.2)
804 (25.9)
22 007 (30.5)
168 (57.3)
61 (15.4)
134 (4.3)
1037 (1.4)
4 (1.4)
6 (1.5)
43 (1.4)
914 (1.3)
3 (2.0)
13 (3.4)
54 (1.8)
1463 (2.0)
127 (84.1)
314 (81.1)
2550 (82.7)
60 299 (83.6)
3 (2.0)
8 (2.1)
106 (3.4)
2413 (3.3)
Student
10 (6.6)
23 (5.9)
174 (5.6)
4086 (5.7)
Retired
2 (1.3)
3 (0.8)
20 (0.7)
192 (0.3)
6 (2.1)
25 (6.3)
177 (5.7)
3627 (5.0)
142 (48.5)
10 (2.5)
27 (0.9)
103 (0.1)
Father not known
Missing
Maternal occupation
Self-employed
Employed
Unemployed
Other (3835)
Missing (283)
Paternal occupation
Self-employed
10 (3.4)
27 (7.0)
240 (7.8)
5734 (8.0)
106 (36.2)
284 (73.4)
2538 (82.3)
60 709 (84.2)
Unemployed
1 (0.3)
7 (1.8)
36 (1.2)
1072 (1.5)
Student
4 (1.4)
8 (2.1)
83 (2.7)
1860 (2.6)
Employed
Retired
Other occupation
Father not known
Missing
0 (0)
0 (0)
12 (0.4)
190 (0.3)
4 (1.4)
2 (1.8)
55 (1.8)
1317 (1.8)
168 (57.3)
61 (15.4)
134 (4.3)
1037 (1.4)
0 (0)
2 (0.5)
10 (0.3)
264 (0.4)
Household income
<5 percentile
14 (4.8)
15 (3.8)
91 (2.9)
1936 (2.7)
5–25 percentile
49 (16.7)
73 (18.4)
424 (13.6)
8948 (12.4)
25–50 percentile
84 (28.7)
83 (21.0)
694 (22.3)
17 456 (24.2)
50–75 percentile
76 (25.9)
97 (24.5)
881 (28.4)
20 785 (28.8)
75–95 percentile
56 (19.1)
109 (27.5)
819 (26.4)
18 313 (25.4)
14 (4.8)
17 (4.3)
191 (6.2)
4648 (6.4)
0 (0)
2 (0.5)
8 (0.3)
97 (0.1)
495 percentile
Missing
SOCIOECONOMIC POSITION AND THE RISK OF PRETERM BIRTH
1113
Table 2 Maternal and paternal characteristics in relation to gestational age, n ¼ 75 980
Characteristics
Age, mother (years)
Age, father (years)
Gestational age in full weeks at delivery
22–27 (n ¼ 293)
28–31 (n ¼ 396) 32–36 (n ¼ 3108)
29.7 (17–43)
29.1 (16–44)
29.0 (16–46)
33.5 (22–54)
31.8 (19–48)
31.7 (18–58)
37–43 (n ¼ 72 183)
29.1 (15–46)
32.2 (16–71)
Parity
0
180 (61.4)
244 (61.6)
1938 (62.4)
35 769 (49.5)
1þ
113 (38.6)
152 (38.4)
1170 (37.6)
36 414 (50.5)
40 (13.7)
39 (9.9)
273 (8.8)
4166 (5.8)
7 (2.4)
10 (2.5)
73 (2.4)
1496 (5.0)
113 (38.6)
130 (32.8)
835 (26.9)
13 918 (19.3)
93 (31.7)
118 (29.8)
909 (29.3)
16 977 (23.5)
168 (146–192)
168 (150–186)
168 (145–190)
169 (142–198)
196 (66.9)
256 (64.7)
2195 (70.6)
53 334 (73.9)
40–10 g tobacco/day
47 (16.0)
54 (13.6)
478 (15.4)
10 434 (14.5)
410 g tobacco/day
50 (17.1)
86 (21.7)
435 (14.0)
8415 (11.7)
127 (43.3)
160 (40.4)
1274 (41.0)
32 175 (44.6)
0.6 (0–7)
0.5 (0–9)
0.6 (0–15)
0.6 (0–36)
75 (25.6)
97 (24.5)
764 (24.6)
18 511 (25.6)
24.0 (16.5–40.4)
24.3 (16.2–47.8)
23.7 (15.3–45.4)
23.6 (13.9–64.4)
0.2 (0.4 to 1.0)
0.2 (0.9 to 2.2)
0.2 (1.5 to 2.6)
0.2 (1.9 to 4.8)
Fertility treatment
Yes
Cohabitation
Living alone
Bleeding during pregnancy (prior to interview)
Yes
Chronic diseasea
Yes
Height, mother (cm)
Smoking status
Non-smoker
Alcohol consumption
Drinking alcohol during pregnancy (yes)
Mean consumption per week (units)
Binge drinking
Yes
BMI
Mean pre-pregnancy
Weight gain
Average kg/week, 1. trimester
a
Chronic disease was defined as having any of the following: high blood pressure, metabolic disorder, muscle and skeletal diseases,
mental disorder, diseases in the urinary system, diseases in the abdominal region, urinary bladder infection more than five times or
having had a conisatio.
Associations between socioeconomic
position and preterm birth
An educational level below 13 years was associated
with an elevated HR of preterm birth. Additionally, a
likelihood ratio test for trend revealed a significant
trend in the results (P < 0.001). The gradients in HR
of preterm birth differed with parity, such that and
the educational gradient was steepest among women
who had given birth before. There was a slightly
elevated risk of preterm birth among couples where
the father had a low educational level and a likelihood ratio test for trend showed that paternal education had an independent effect on preterm birth.
Disability-retired mothers had an elevated HR of
preterm birth compared with employed mothers.
Income displayed a slight gradient in risk of preterm birth (linear variable in the model, P ¼ 0.02),
with a decreased risk of preterm birth for women who
were living in low-income households (Table 3).
Repeating the analyses in the subpopulation with no
record of chronic disease (n ¼ 57 883) and in the subpopulation of women without record of fertility treatment (n ¼ 71 452) did not change the estimates (data
not shown), except for the effect of disability retired
women. The association between receiving disability
pension and preterm birth was not present [the
estimate was reduced from 2.24 (95% CI 1.50–3.35) to
0.37 (95% CI 0.05–2.66)] in this subpopulation.
1114
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Table 3 Adjusted HR of the association between socioeconomic position, measured by education, occupation,
household income and preterm birth
Exposure
Maternal educationb
Table 3 Continued
Delivery before 37 completed
weeks of gestation, n ¼ 75 980
Unadjusted
Adjusted
HR (95% CI)
HRa (95% CI)
Nulliparous
1–10 years
1.23 (1.07–1.40) 1.22 (1.04–1.42)
10–12 years
1.16 (1.05–1.27) 1.16 (1.05–1.28)
412 years/ref.
1
Exposure
Household incomee
Parous
1–10 years
1.73 (1.48–2.02) 1.56 (1.31–1.87)
10–12 years
1.21 (1.07–1.37) 1.23 (1.08–1.41)
412 years/ref.
Paternal education
1
c
1
Nulliparous and parous
1–10 years
1.21 (1.10–1.33) 1.13 (1.02–1.26)
10–12 years
1.09 (1.01–1.17) 1.07 (0.98–1.16)
412 years/ref.
1
1
Maternal occupation Nulliparous and parous
Self-employed
0.98 (0.77–1.24) 0.96 (0.76–1.23)
Employed/ref.
1
0.99 (0.83–1.20) 0.97 (0.81–1.17)
Student
1.01 (0.87–1.16) 0.98 (0.84–1.14)
Retired
2.61 (1.76–3.87) 2.24 (1.50–3.35)
Other occupation
1.16 (1.01–1.33) 1.03 (0.89–1.15)
Paternal occupationd Nulliparous
1.02 (0.86–1.20) 0.98 (0.82–1.16)
Employed/ref.
1
0.61 (0.39–0.93) 0.57 (0.37–0.87)
Student
0.81 (0.63–1.03) 0.83 (0.64–1.07)
Retired
0.96 (0.40–2.32) 0.76 (0.32–1.85)
Other occupation
1.26 (1.04–1.51) 1.29 (1.05–1.58)
Parous
Employed/ref.
1.09 (0.91–1.31) 1.00 (0.82–1.21)
1
0.89 (0.71–1.12)
1.12 (1.00–1.24)
0.97 (0.86–1.10)
25–50 percentile
0.95 (0.86–1.04)
0.91 (0.83–1.00)
1
1
75–95 percentile
1.06 (0.97–1.16)
1.11 (1.01–1.21)
495 percentile
0.96 (0.83–1.11)
1.05 (0.90–1.22)
a
Mutually adjusted for maternal educational level, paternal
educational level, maternal occupation, paternal occupation,
household income, maternal age and maternal cohabitation.
b
Since parity modified the effect of maternal education on
preterm birth, these analyses were conducted for nulliparous
and parous separately.
c
A likelihood ratio test for trend: 2 Log L ¼ 38.3, degrees of
freedom ¼ 1, P < 0.005.
d
Parity modified the effect of paternal occupation on preterm
birth, and these analyses were conducted for nulliparous and
parous separately.
e
Likelihood ratio test for trend: 2 Log L ¼ 7.6, degrees of
freedom ¼ 4, P ¼ 0,1, P ¼ 0.002 for the linear variable.
extremely preterm births (Table 4). The results revealed
no consistent patterns for differences in the gradients
for occupation and household income and the test for
proportional hazard showed no significant difference in
the associations between the five measures of socioeconomic position, and extremely, very and moderately
preterm birth, respectively (results not shown).
1
Unemployed
Self-employed
1.05 (0.85–1.29)
5–25 percentile
1
Unemployed
Self-employed
<5 percentile
50–75 percentile/ref.
1
Delivery before 37 completed
weeks of gestation, n ¼ 75 980
Unadjusted
Adjusted
HR (95% CI)
HRa (95% CI)
Nulliparous and parous
1
Unemployed
1.31 (0.87–2.00) 1.11 (0.73–1.68)
Student
1.50 (1.03–2.19) 1.44 (0.97–2.12)
Retired
1.88 (0.89–3.94) 1.27 (0.60–2.68)
Other occupation
1.80 (1.40–2.31) 1.37 (1.04–1.80)
(continued)
Associations between socioeconomic
position and extremely, very and
moderately preterm birth
Results from the analyses with preterm birth grouped
into extremely, very and moderately preterm birth,
respectively, showed that the educational gradient was
steeper among the group of women with very and
The mediating role of health behaviours
Since maternal educational level was the clearest
predictor of preterm birth among five socioeconomic
measures, analyses regarding mediation were only
made on the association between maternal education
and the risk of preterm birth (Table 5). The overall
reduction in the HR of preterm birth after including
smoking, alcohol consumption, binge drinking prepregnancy BMI and gestational weight gain in the
model was higher among parous than among nulliparous women. Including all five mediators lowered
the risk estimate by 19% [(HRwithout the mediator in the
model
HRwith the mediator in the model)/HRwithout the
mediator in the model
1]45 for nulliparous with the
lowest educational level and 30% for parous women
with the lowest educational level. The elevated risk of
preterm birth among nulliparous women was reduced
only marginally when the intermediate variables were
included in the regression model separately, but
smoking reduced the HR of preterm birth with 23%
for parous and low educated (<10 years) women and
with 13% for parous women with 10–12 years of
education.
SOCIOECONOMIC POSITION AND THE RISK OF PRETERM BIRTH
1115
Table 4 Adjusted HR for the association between socioeconomic position and extremely, very and moderate preterm
birtha, n ¼ 75 980
Completed weeks of gestation at birth
Maternal education, nulliparous womenb
HR (95% CI)
22–27 weeks
HR (95% CI)
28–31 weeks
HR (95% CI)
32–36 weeks
<10 years
1.60 (0.74–3.46)
2.33 (1.44–3.78)
1.49 (1.23–1.80)
10–12 years
1.13 (0.60–2.15)
1.56 (1.03–2.36)
1.21 (1.05–1.39)
1
1
1
412 years
Maternal education, parous women
<10 years
2.29 (1.28–4.10)
1.32 (0.85–2.03)
1.16 (0.99–1.37)
10–12 years
1.36 (0.88–2.17)
1.27 (0.94–1.71)
1.13 (1.02–1.26)
1
1
<10 years
0.84 (0.52–1.34)
1.08 (0.79–1.146)
1.12 (1.00–1.26)
10–12 years
0.70 (0.49–1.01)
0.95 (0.75–1.20)
1.08 (0.99–1.18)
1
1
1
0.98 (0.31–3.08)
1.67 (0.94–2.98)
2.02 (1.29–3.16)
1
1
1
Unemployed
0.61 (0.19–1.91)
0.70 (0.34–1.40)
1.05 (0.86–1.28)
Student
0.96 (0.49–1.90)
0.91 (0.58–1.44)
0.89 (0.76–1.05)
Retired
3.92 (0.97–15.82)
2.54 (0.81–7.94)
1.15 (0.98–1.36)
0.79 (0.35–1.79)
1.28 (0.84–1.97)
0.78 (0.28–2.18)
1.22 (0.71–2.11)
412 years
Paternal education
412 years
Maternal occupation
Self-employed
Employed
Other occupational status
Paternal occupation, parous womanc
Self-employed
Employed
1.47 (0.70–3.12)
1
1
1
1.21 (0.17–8.80)
2.04 (0.75–5.56)
1.00 (0.63–1.60)
Student
–
1.61 (0.51–5.09)
1.49 (0.99–2.24)
Retired
–
Unemployed
Other occupational status
1.20 (0.89–1.63)
1.55 (0.48–5.04)
2.82 (1.56–5.08)
–
0.98 (0.39–2.42)
0.82 (0.46–1.48)
0.88 (0.36–2.12)
1
1
1
(0.53–4.00)
0.83 (0.27–2.60)
0.56 (0.35–0.89)
1.45
0.51 (0.19–1.37)
0.83 (0.64–1.09)
7.03 (4.33–11.42)
1.83 (1.16–3.03)
<5 percentile
1.45 (0.62–3.42)
1.37 (0.75–2.51)
0.94 (0.73–1.20)
5–25 percentile
0.77 (0.42–1.41)
1.57 (1.13–2.19)
1.10 (0.97–1.26)
25–50 percentile
1.22 (0.79–1.87)
1.09 (0.80–1.48)
1.02 (0.92–1.23)
50–75 percentile
1
1
1
75–95 percentile
0.79 (0.50–1.24)
1.15 (0.87–1.53)
0.94 (0.85–1.04)
495 percentile
0.91 (0.45–1.80)
0.69 (0.41–1.18)
0.84 (0.71–0.98)
Paternal occupation, nulliparous woman
Self-employed
Employed
Unemployed
Student
Retired
Unknown
0.98 (0.77–1.25)
Household income
a
Mutually adjusted for maternal educational level, paternal educational level, maternal occupation, paternal labour market
attachment, household income, maternal age and cohabitation.
b
Results shown for nulliparous and parous women separately because parity modified the effect of maternal educational level on
preterm birth, nulliparous: n ¼ 38 131, parous: n ¼ 37 849.
c
Results shown for nulliparous and parous separately because parity modified the effect of paternal occupation on preterm birth.
1
Mutually adjusted for paternal educational level, maternal occupation, paternal occupation, household income, maternal age and maternal cohabitation.
a
1.19 (1.04–1.36)
1.23 (1.08–1.41)
1.23 (1.08–1.41)
1
1
1.24 (1.08–1.41)
1.23 (1.08–1.41)
1
1
1.20 (1.05–1.37)
1.23 (1.08–1.41)
10–12 years
1
1.54 (1.29–1.85)
1.55 (1.29–1.86)
1.56 (1.31–1.87)
1.54 (1.29–1.85)
1.43 (1.19–1.72)
1.56 (1.31–1.87)
<10 years
Parous, n ¼ 37 849
Discussion
412 years
1.39 (1.16–1.68)
1
1.14 (1.03–1.26)
1.15 (1.04–1.27)
1
1
1.15 (1.04–1.27)
1.15 (1.04–1.27)
1
1
1.15 (1.04–1.27)
1.15 (1.05–1.27)
1
1
1.15 (1.05–1.27)
10–12 years
412 years
1.17 (1.00–1.37)
1.21 (1.03–1.41)
1.20 (1.03–1.40)
1.21 (1.04–1.41)
1.21 (1.04–1.41)
1.21 (1.04–1.42)
1.22 (1.04–1.42)
<10 years
Adjusted HR,
smoking
included
Adjusted HR, alcohol
consumption included
Adjusted HR, binge
drinking included
Adjusted HR,
pre-pregnancy
BMI included
Adjusted HR,
gestational
weight gain
included
All five risk
factors included
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
Adjusted HR for
preterm birth
according to
maternal
educational levela
Nulliparous, n ¼ 38 131
Table 5 HR estimates for the adjusted associations between maternal educational level and preterm birth before and after including five intermediate variables
1116
This study, using birth outcomes from 75 980 pregnancies, showed that maternal educational level was
the indicator of socioeconomic position that most
clearly displayed a social gradient in preterm birth
and that the gradient was steepest among parous
women. The trend test results displayed a slightly
increased risk of preterm birth if the father of the
child had a low educational level and a slightly
decreased risk if the household income was low.
There was no statistically significant difference in the
associations according to the degree of preterm birth.
Smoking explained some of the relative educational
gradient in preterm birth among parous women and a
minor part of the association for nulliparous women.
Alcohol consumption, binge drinking, pre-pregnancy
BMI and gestational weight gain explained only a
minor part of the association between maternal educational level and preterm birth and only for women
with an educational attainment below 10 years.
Previous studies based on populations from
Europe,46 Norway,9 Sweden,4,5 Finland8 and Canada6
have found strong evidence for socioeconomic differences in preterm birth, when measured by maternal
educational level. Other studies reported a greater
elevated HR of preterm birth among women with the
lowest educational level than found in the present
study, which may be explained by a higher level of
education in the reference group and different degrees
of socioeconomic inequalities in different countries.
Two of the previous studies investigating the difference in associations between maternal educational
level and preterm birth according to the degree of
preterm birth (<32 weeks and below 37 weeks)
likewise showed a steeper gradient in very preterm
birth than in moderately preterm birth.3,6 In one study,
the authors found an inverse association but the
outcome in this study was small for gestational age
(SGA) children born preterm, which can not be compared directly with preterm births because the aetiology of SGA might differ from that of preterm birth.10
In a study based on a Dutch population, the authors
did not find any association between maternal educational level and preterm birth but the estimates
showed a tendency to a gradient and the insignificant results could be due to a small sample size
(2027 women and 108 preterm births).
The results of no association between occupation and
preterm birth is in agreement with the earlier findings from Denmark47 and from Finland,8,48 but in a
study by Ancel et al.46 with data from all over Europe,
unemployed women had an elevated risk of preterm
birth compared with employed women. The difference
could be due to a modifying effect from the welfare models in the Nordic countries with high social
security for people outside the labour market. The
elevated risk of preterm birth among disability-retired
women was not surprising since the illness (for
instance chronic diseases) that entitled the woman to
SOCIOECONOMIC POSITION AND THE RISK OF PRETERM BIRTH
the National Supplementary Disability Pension could
be related to the risk of preterm birth. The additional
analyses on a population with no record of chronic
disease showed a much lowered HR of preterm birth
for the disability retired women [the estimate was
reduced from 2.24 (95% CI 1.50–3.35) to 0.37 (95% CI
0.05–2.66)] indicating that it was the disease rather
than the attachment to the labour market that
contributed to the elevated HR of preterm birth for
the disability-retired women.
The results on the tendency to a reduced HR of
preterm birth for women living in low-income households are surprising. The CIs did all include 1 and the
trend test was not statistically significant, but as the
model with a linear variable for income described
the data best and in this model, there was some
statistical support for positive trend in the association
between income and preterm birth i.e. lower income
was associated with lower risk of preterm birth, once
adjustment was made for the other socioeconomic
indicators. From the results, it is not possible to
conclude that a low income has a protective effect,
however, lower household income does not seem to be
a risk factor for preterm birth in this study. The
association between income at individual level and
preterm birth is to our knowledge not previously
investigated. A study by Gudmundsson et al.49 reported
no elevated risk for women from deprived neighbourhoods and Luo et al.7 reported a slightly elevated risk for
woman from areas with income in the lowest quintile.
This is contrary to the findings in this study but these
studies were not directly comparable with this study
because of the area-based exposure measure.50
The literature on the role of individual level risk
factors as mediating factors explaining socioeconomic
inequality in preterm birth is sparse. Kramer et al.35
have addressed the contribution of individual level risk
factors to socioeconomic inequality in preterm birth. In
this study, the authors suggested bacterial vaginosis
and cigarette smoking were the quantitatively most
important mediators on the association between
socioeconomic position and preterm birth. We had no
good clinical information regarding bacterial vaginosis
and urinary tract infection, which is a limitation in this
study, as these can display a social gradient and
explain a proportion of the socioeconomic inequality
found in preterm birth. Smoking was the most important mediator in this study; we need, however, access
to a wider array of potentially mediating variables in
order to conclude that it is the most important. With
regards to birth weight, smoking has been shown to
mediate the association between social deprivation
and pre- and full-term low birth weight.51 In a study
by Stephansson et al.,52 the relative increased risk
of stillbirth among women with low socioeconomic
status could not be explained by differences in lifestyle
factors such as smoking and BMI, which support the
findings of the present study although the results are
not directly comparable.
1117
This study was based on a large population with
complete follow-up. The proportion with missing
information was relatively low. A possible limitation
of this study is the selection to the DNBC, which
could introduce selection bias. In order to examine
the magnitude of selection bias, we repeated the
unadjusted analysis in the source population consisting
of 506 598 (all) Danish births between 1996 and 2002.
The results showed a greater elevated risk of preterm
birth among the socioeconomic disadvantaged in
the general Danish population but the relative HR
(Relative HR ¼ HR in DNBC/HR in the Danish population) differed only between 4% and 12% from the
unadjusted results presented in this study (results
not shown). We believe the selection to the cohort to
have affected the internal validity only slightly,
which is supported by results from a Danish study
from 2006,53 but it may reduce the generalizability of
the study.
The group of women and men with no information
on education or occupation and births with unknown
father had a high relative elevated risk of preterm
birth compared with the reference groups (412 years
of education, employed, income in the 50–75 percentile, data not shown) but including these groups in
the analyses did only change the estimates for the
association between paternal educational level and
preterm birth (upwards with 9%). The interpretation
of the elevated risks is unclear, since unknown status
of the father may both represent unwanted pregnancies and pregnancies where the women is no longer
in relationship with the infants father, in addition to
lack of registration if the infant died shortly after
the birth.
The bias due to possible self-under or over reporting
of socioeconomic position is minimized in this study,
since exposure was registry-based information. However, information on covariates and health behaviours
such as fertility treatment, bleeding during pregnancy,
chronic disease, smoking, drinking etc. was obtained
by self-reports. Literature on the accuracy on selfreported health behaviours suggest that, although
most people report honestly, the respondents tend
to underreport characteristics that are considered to
be undesirable or negative54 and this could be more
pronounced among women with a low educational
level.55 If underreporting of smoking is more common
among the women with a low educational level, the
proportion of socioeconomic inequality in preterm
birth explained by smoking would be underestimated.
The method used to investigate the mediating role
of individual level risk factors has its limitations,
since this proportion is dependent on the choice of
reference group for the exposure variables. Thus,
the calculated percent-wise change in the estimates
should be interpreted with caution. We did repeat
the analyses using other reference groups and
this revealed the same tendencies as shown in this
article.
1118
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
The educational gradient in preterm birth might
be explained by different use of the health care
system, even though the access to care is free and in
principle equal in Denmark, if well-educated women
optimise their use of the health services through
better communication with health professionals.
However, the evidence of the effect of prenatal care
on preterm birth is diverse.56–59 Furthermore, education can be regarded as marker of a life long exposure
because education reflects a number of possible key
exposures—among other things—the childhood intellectual and material resources, and education may in
this way be a more precise indicator of socioeconomic
conditions across the life course.
The slightly positive income gradient in preterm
birth could indicate that material wealth does not
have a substantial influence on preterm birth or that
a possible inverse income gradient is modified by the
welfare system.
In conclusion, maternal educational level was the
strongest predictor of preterm birth among five socioeconomic measures. Income displayed a slightly positive gradient indicating that women from low-income
household had a decreased risk of preterm birth.
The educational gradient was steepest among the
more ‘severe’ types of preterm birth, but the differences were not statistically significant. Smoking was
the only factor that contributed to explaining the
educational gradient in preterm birth and this was
only persistent among parous women.
Acknowledgements
The Danish National Research Foundation has
established the Danish Epidemiology Science
Centre that initiated and created the Danish
National Birth Cohort. The cohort is furthermore a
result of a major grant from this Foundation.
Additional support for the Danish National Birth
Cohort is obtained from the Pharmacy Foundation,
the Egmont Foundation, the March of Dimes Birth
Defects Foundation, the Augustinus Foundation
and the Health Foundation. The authors wish to
thank Sarah Fredsted Villadsen for valuable comments to this article.
Conflict of interest: None declared.
KEY MESSAGES
Preterm birth has been shown to be elevated among the socioeconomically disadvantaged.
Among five socioeconomic measures, maternal education displays the clearest gradient in preterm
birth.
The maternal educational gradient in preterm birth does not differ significantly among extremely, very
and moderate preterm birth.
Smoking explains a minor part of the educational gradient in preterm birth.
Alcohol consumption, binge drinking, pre-pregnancy BMI and gestational weight gain only reduces the
educational gradient in preterm birth marginally.
References
1
2
3
4
5
Basso O, Olsen J, Christensen K. Study of environmental,
social, and paternal factors in preterm delivery using sibs
and half sibs. A population-based study in Denmark.
J Epidemiol Community Health 1999;53:20–23.
Haram K, Mortensen JH, Wollen AL. Preterm
delivery: an overview. Acta Obstet Gynecol Scand
2003;82:687–704.
Ancel PY, Saurel-Cubizolles MJ, Di Renzo GC, Papiernik E,
Breart G. Very and moderate preterm births: are
the risk factors different? Br J Obstet Gynaecol
1999;106:1162–70.
Clausson B, Cnattingius S, Axelsson O. Preterm and term
births of small for gestational age infants: a populationbased study of risk factors among nulliparous women.
Br J Obstet Gynaecol 1998;105:1011–17.
Koupilova I, Vagero D, Leon DA et al. Social variation in
size at birth and preterm delivery in the Czech Republic
and Sweden, 1989–91. Paediatr Perinat Epidemiol
1998;12:7–24.
6
7
8
9
10
11
Kramer MS, McLean FH, Eason EL, Usher RH. Maternal
nutrition and spontaneous preterm birth. Am J Epidemiol
1992;136:574–83.
Luo ZC, Wilkins R, Kramer MS. Effect of neighbourhood
income and maternal education on birth outcomes: a
population-based study. CMAJ 2006;174:1415–20.
Olsen P, Laara E, Rantakallio P, Jarvelin MR, Sarpola A,
Hartikainen AL. Epidemiology of preterm delivery in two
birth cohorts with an interval of 20 years. Am J Epidemiol
1995;142:1184–93.
Thompson JM, Irgens LM, Rasmussen S, Daltveit AK.
Secular trends in socio-economic status and the implications for preterm birth. Paediatr Perinat Epidemiol
2006;20:182–87.
Verkerk PH, Zaadstra BM, Reerink JD, Herngreen WP,
Verloove-Vanhorick SP. Social class, ethnicity and other
risk factors for small for gestational age and preterm
delivery in The Netherlands. Eur J Obstet Gynecol Reprod
Biol 1994;53:129–34.
Wildschut HI, Nas T, Golding J. Are sociodemographic
factors predictive of preterm birth? A reappraisal of the
SOCIOECONOMIC POSITION AND THE RISK OF PRETERM BIRTH
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
1958 British Perinatal Mortality Survey. Br J Obstet
Gynaecol 1997;104:57–63.
Hentze TI, Hansen BM, Jonsbo F, Greisen G. Chronic
lung disease in a cohort of children born before the 28th
gestational week. Incidence and etiological factors. Ugeskr
Laeger 2006;168:2243–47.
Himmelmann K, Hagberg G, Beckung E, Hagberg B,
Uvebrant P. The changing panorama of cerebral palsy in
Sweden. IX. Prevalence and origin in the birth-year
period 1995–1998. Acta Paediatr 2005;94:287–94.
Markestad T, Kaaresen PI, Ronnestad A et al. Early death,
morbidity, and need of treatment among extremely
premature infants. Pediatrics 2005;115:1289–98.
Marlow N, Wolke D, Bracewell MA, Samara M.
Neurologic and developmental disability at six years of
age after extremely preterm birth. N Engl J Med 2005;
352:9–19.
Morken NH, Kallen K, Hagberg H, Jacobsson B. Preterm
birth in Sweden 1973–2001: rate, subgroups, and effect of
changing patterns in multiple births, maternal age, and
smoking. Acta Obstet Gynecol Scand 2005;84:558–65.
Tucker J, McGuire W. Epidemiology of preterm birth.
Br Med J 2004;329:675–78.
Smith GC, Shah I, White IR, Pell JP, Crossley JA,
Dobbie R. Maternal and biochemical predictors of
spontaneous preterm birth among nulliparous women: a
systematic analysis in relation to the degree of prematurity. Int J Epidemiol 2006;35:1169–77.
Osler M. Social class and health behaviour in Danish
adults: a longitudinal study. Public Health 1993;107:
251–60.
Wagenknecht LE, Perkins LL, Cutter GR et al. Cigarette
smoking behavior is strongly related to educational
status: the CARDIA study. Prev Med 1990;19:158–69.
Lantz PM, House JS, Lepkowski JM, Williams DR,
Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults. JAMA
1998;279:1703–8.
Braveman PA, Cubbin C, Egerter S et al. Socioeconomic
status in health research - one size does not fit all. JAMA
2005;294:2879–88.
Lynch J, Kaplan G. Socioeconomic position. In:
Berkman L, Kawachi I (eds). Social Epidemiology.
New York: Oxford University Press, 2000. pp. 13–35.
Cnattingius S. The epidemiology of smoking during
pregnancy: smoking prevalence, maternal characteristics, and pregnancy outcomes. Nicotine Tob Res 2004;6
(Suppl 2):S125–40.
Hadley CB, Main DM, Gabbe SG. Risk factors for preterm
premature rupture of the fetal membranes. Am J Perinatol
1990;7:374–79.
Hoffman DR, Romero R, Johnston JM. Detection
of platelet-activating factor in amniotic fluid of
complicated
pregnancies.
Am
J
Obstet
Gynecol
1990;162:525–28.
Kyrklund-Blomberg NB, Granath F, Cnattingius S.
Maternal smoking and causes of very preterm birth.
Acta Obstet Gynecol Scand 2005;84:572–77.
Albertsen K, Andersen AM, Olsen J, Gronbaek M. Alcohol
consumption during pregnancy and the risk of preterm
delivery. Am J Epidemiol 2004;159:155–61.
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
1119
Berkowitz GS, Holford TR, Berkowitz RL. Effects of
cigarette smoking, alcohol, coffee and tea consumption
on preterm delivery. Early Hum Dev 1982;7:239–50.
Kesmodel U, Olsen SF, Secher NJ. Does alcohol increase
the risk of preterm delivery? Epidemiology 2000;11:512–18.
Lundsberg LS, Bracken MB, Saftlas AF. Low-to-moderate
gestational alcohol use and intrauterine growth retardation, low birthweight, and preterm delivery. Ann Epidemiol
1997;7:498–508.
Naeye RL. Maternal body weight and pregnancy outcome.
Am J Clin Nutr 1990;52:273–79.
Nohr EA, Bech BH, Vaeth M, Rasmussen KM,
Henriksen TB, Olsen J. Obesity, gestational weight gain
and preterm birth: a study within the Danish National
Birth Cohort. Paediatr Perinat Epidemiol 2007;21:5–14.
Casswell S, Pledger M, Hooper R. Socioeconomic status
and drinking patterns in young adults. Addiction 2003;98:
601–10.
Kramer MS, Seguin L, Lydon J, Goulet L. Socio-economic
disparities in pregnancy outcome: why do the poor fare so
poorly? [In process citation]. Paediatr Perinat Epidemiol
2000;14:194–210.
Kristensen PL, Wedderkopp N, Moller NC, Andersen LB,
Bai CN, Froberg K. Tracking and prevalence of cardiovascular disease risk factors across socio-economic
classes: a longitudinal substudy of the European Youth
Heart Study. BMC Public Health 2006;6:6–20.
Stamatakis E, Primatesta P, Chinn S, Rona R,
Falascheti E. Overweight and obesity trends from 1974
to 2003 in English children: what is the role of socioeconomic factors? Arch Dis Child 2005;90:999–1004.
Stunkard AJ, Sorensen TI. Obesity and socioeconomic status–a complex relation. N Engl J Med
1993;329:1036–37.
Nybo Andersen AM, Olsen J. Do interviewers’ health
beliefs and habits modify responses to sensitive questions? A study using data Collected from pregnant
women by means of computer-assisted telephone interviews. Am J Epidemiol 2002;155:95–100.
Olsen J, Melbye M, Olsen SF et al. The Danish National
Birth Cohort–its background, structure and aim. Scand J
Public Health 2001;29:300–7.
Olsen J. The Danish National Birth Cohort–a data source
for studying preterm birth. Acta Obstet Gynecol Scand
2005;84:539–40.
OECD modified scale. Available at: http://www.oecd.org/
dataoecd/61/52/35411111.pdf. (Accessed on 27 February
2007).
World Health Organisation. Physical status: the use and
interpretation of anthropometry. Geneva: WHO, 1995.
Greenland S, Rothman KJ. Introduction to stratified
analysis. In: Rothman KJ, Greenland S (eds). Modern
Epidemiology. Philidelphia, USA: Lippincott Williams &
Wilkins, 1998. pp. 253–79.
Ditlevsen S, Keiding N, Christensen U, Damsgaard MT,
Lynch J. Mediation proportion. Epidemiology 2005;16:592.
Ancel PY, Saurel-Cubizolles MJ, Di Renzo GC,
Papiernik E, Breart G. Social differences of very preterm
birth in Europe: interaction with obstetric history.
Europop Group. Am J Epidemiol 1999;149:908–15.
Henriksen TB, Savitz DA, Hedegaard M, Secher NJ.
Employment during pregnancy in relation to risk factors
1120
48
49
50
51
52
53
INTERNATIONAL JOURNAL OF EPIDEMIOLOGY
and pregnancy outcome. Br J Obstet Gynaecol 1994;101:
858–65.
Hartikainen-Sorri AL, Sorri M. Occupational and sociomedical factors in preterm birth. Obstet Gynecol 1989;
74:13–16.
Gudmundsson
S,
Bjorgvinsdottir
L,
Molin
J,
Gunnarsson G, Marsal K. Socioeconomic status and
perinatal outcome according to residence area in the
city of Malmo. Acta Obstet Gynecol Scand 1997;76:318–23.
Andersen AM, Mortensen LH. Socioeconomic inequality
in birth outcomes: what do the indicators tell us,
and where do we find the data? CMAJ 2006;174:1429–30.
Reime B, Ratner PA, Tomaselli-Reime SN, Kelly A,
Schuecking BA, Wenzlaff P. The role of mediating factors
in the association between social deprivation and low
birth weight in Germany. Soc Sci Med 2006;62:1731–44.
Stephansson O, Dickman PW, Johansson AL, Cnattingius S.
The influence of socioeconomic status on stillbirth risk in
Sweden. Int J Epidemiol 2001;30:1296–301.
Nohr EA, Frydenberg M, Henriksen TB, Olsen J. Does low
participation in cohort studies induce bias? Epidemiology
2006;17:413–18.
54
55
56
57
58
59
Baranowski T. Methodologic issues in self-report of
health behavior. J Sch Health 1985;55:179–82.
Parna K, Rahu M, Youngman LD, Rahu K, NygardKibur M, Koupil I. Self-reported and serum cotininevalidated smoking in pregnant women in Estonia. Matern
Child Health J 2005;9:385–92.
Galobardes B, Shaw M, Lawlor DA, Lynch JW, Davey SG.
Indicators of socioeconomic position (part 1). J Epidemiol
Community Health 2006;60:7–12.
Gomez-Olmedo M, Delgado-Rodriguez M, BuenoCavanillas A, Molina-Font JA, Galvez-Vargas R.
Prenatal care and prevention of preterm birth. A casecontrol study in southern Spain. Eur J Epidemiol
1996;12:37–44.
Scholl TO, Miller LK, Salmon RW, Cofsky MC, Shearer J.
Prenatal care adequacy and the outcome of adolescent
pregnancy: effects on weight gain, preterm delivery, and
birth weight. Obstet Gynecol 1987;69:312–16.
White DE, Fraser-Lee NJ, Tough S, Newburn-Cook CV.
The content of prenatal care and its relationship to
preterm birth in Alberta, Canada. Health Care Women Int
2006;27:777–92.