What This Study Adds

  

Are Starting and Continuing Breastfeeding Related to Educational Background?

The Generation R Study

Lenie van Rossem, Anke Oenema, Eric A. P. Steegers, Henriëtte A. Moll, Vincent W.

V. Jaddoe, Albert Hofman, Johan P. Mackenbach and Hein Raat

  Pediatrics 2009;123;e1017-e1027

DOI: 10.1542/peds.2008-2663

  

The online version of this article, along with updated information and services, is

located on the World Wide Web at:

  PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly publication, it has been published continuously since 1948. PEDIATRICS is owned, published, and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elk Grove Village, Illinois, 60007. Copyright © 2009 by the American Academy of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275. ARTICLE

Are Starting and Continuing Breastfeeding Related to

Educational Background? The Generation R Study a,b b c d Lenie van Rossem, MSc , Anke Oenema, PhD, MPH , Eric A. P. Steegers, MD, PhD , Henrie¨tte A. Moll, MD, PhD , a,d,e e b b Vincent W. V. Jaddoe, MD, PhD , Albert Hofman, MD, PhD , Johan P. Mackenbach, MD, PhD , Hein Raat, MD, PhD a b c

  Generation R Study Group, Department of Public Health, Division of Obstetrics and Prenatal Medicine, Department of Obstetrics and Gynecology, and Departments of

  d e

  Pediatrics and Epidemiology, Erasmus MC-University Medical Centre Rotterdam, Rotterdam, Netherlands The authors have indicated they have no financial relationships relevant to this article to disclose.

  What’s Known on This Subject What This Study Adds

In Western countries, socially disadvantaged mothers start breastfeeding less often and Socially disadvantaged mothers are less likely to start breastfeeding and to continue it

also have a shorter breastfeeding duration. It is unknown which factors explain the for 2 months. Lifestyle-related and birth characteristics are important modifying factors

association between social disadvantage and breastfeeding. in the association between educational level and breastfeeding.

  ABSTRACT OBJECTIVE.

  To assess the effect of a woman’s educational level on starting and continu- ing breastfeeding and to assess the role of sociodemographic, lifestyle-related, psy- www.pediatrics.org/cgi/doi/10.1542/ chosocial, and birth characteristics in this association. peds.2008-2663 METHODS.

  We used the data of 2914 participants in a population-based prospective doi:10.1542/peds.2008-2663 cohort study. Information on educational level, breastfeeding, sociodemographic

  Key Words (maternal age, single parenthood, parity, job status), lifestyle-related (BMI, smoking, breastfeeding, socioeconomic status alcohol use), psychosocial (whether the pregnancy was planned, stress), and birth

  Abbreviations OR— odds ratio

  (gestational age, birth weight, cesarean delivery, place and type of delivery) charac- CI— confidence interval teristics were obtained between pregnancy and 12 months postpartum. Odds ratios

  FAD—Family Assessment Device and 95% confidence intervals of starting and continuing breastfeeding for educa-

  MABS—Mother and Baby Scales tional level were obtained by logistic regression, adjusted for each group of covariates

  Accepted for publication Feb 17, 2009 and for all covariates simultaneously. Address correspondence to Lenie van Rossem, MSc, Erasmus MC-University Medical Centre

  RESULTS.

  Of 1031 highest-educated mothers, 985 (95.5%) started breastfeeding; the Rotterdam, Generation R Study Group, PO Box 2040, Room Ae-007, 3000 CA, Rotterdam, percentage was 73.1% (255 of 349) in the lowest-educated mothers. At 6 months, Netherlands.

  39.3% (405 of 1031) of highest-educated mothers and 15.2% (53 of 349) of lowest- E-mail: l.vanrossem@erasmusmc.nl educated mothers were still breastfeeding. Educationally related differences were

  PEDIATRICS (ISSN Numbers: Print, 0031-4005; present in starting breastfeeding and the continuation of breastfeeding until 2 Online, 1098-4275). Copyright © 2009 by the

  American Academy of Pediatrics months but not in breastfeeding continuation between 2 and 6 months. Lifestyle- related and birth characteristics attenuated the association between educational level and breastfeeding, but the association was hardly affected by sociodemographic and psychosocial characteristics.

  CONCLUSIONS.

  Decisions to breastfeed were underlain by differences in educational background. The underlying path-

ways require further research. For the time being, interventions on promoting breastfeeding should start early in

pregnancy and should increase their focus on low-educated women. Pediatrics 2009;123:e1017–e1027

  REASTFEEDING HAS CONVINCINGLY been established as the best type of feeding for a child, bringing short-term

  1

  2

  3 health benefits such as a lower risk for otitis media, gastroenteritis, and respiratory infections. In the

  B

  4

  5

long-term, breastfeeding may also reduce the risk for obesity and diabetes mellitus. To benefit from these health

  6 effects, the World Health Organization recommends 6 months of exclusive breastfeeding.

  Various studies have found that socially disadvantaged mothers start breastfeeding less often and also breastfeed 7–11

for a shorter period. Although none of these studies explained such social inequalities in breastfeeding, the

implication is clear: at the start of their lives, children from socially disadvantaged families may already run a higher

risk of several diseases.

  12–18 The literature on the determinants of breastfeeding has consistently identified maternal smoking and lower

  

17

maternal age as predictors of lower breastfeeding rates. To establish the role of these and other characteristics in

explaining the social inequalities relative to breastfeeding, we used the Generation R Study, a birth cohort study, to

collect data on breastfeeding and on relevant demographic, psychosocial, lifestyle-related, and birth characteristics.

  Our study had 2 objectives. The first was to establish whether there were any differences between high-educated

and low-educated mothers with regard to (1) starting breastfeeding, (2) continuing it for 2 months after birth, and

(3) continuing it for between 2 and 6 months after birth. The second objective was to establish how the known determinants of breastfeeding affected the association between mother’s educational level and starting and continuing breastfeeding.

  For the second objective, we conceptualized that ed- ucational level is the most distal factor determining breastfeeding behavior, and sociodemographic, lifestyle- related, psychosocial, and birth characteristics are more proximal factors for breastfeeding initiation or continu- ation.

  Maternal education was used as an indicator of socioeconomic status, because it reflects not only ma- terial resources but also noneconomic characteristics such as general and health-related knowledge, which in turn influence health behavior, and problem-solving skills.

  Lifestyle-Related Characteristics Maternal height was assessed at the research center at enrollment; prepregnancy weight was self-reported.

  31,32 Planned pregnancy (yes, no) was assessed by using a prenatal questionnaire.

  30 Postnatal stress was measured by using the 13-item subscale “lack of confidence in care- taking” from the Mother and Infant Scales (MABS), which was administered 2 months after birth (cutoff point at mean ⫹ 1 SD).

  Psychosocial Characteristics Prenatal stress was measured by using the Family As- sessment Device (FAD), a 12-item questionnaire on gen- eral family function (pathologic score: ⬎2.17; nonpatho- logic score: ⱕ2.17).

  Sociodemographic Characteristics Parity (primiparity, multiparity) and maternal age (⬍30, 30 –35, and ⱖ35 years) were measured at intake. Single parenting (no partner, partner) and job status after birth (not employed, employed) was established by using a questionnaire.

  12–18,29 was used to select potential mediators for the association between maternal educational level and breastfeeding (categories are in parentheses).

  Covariates Literature on the determinants of breastfeeding

  23 the continuation of any breast- feeding for 2 months after birth, and the continuation of breastfeeding between 2 months and 6 months after birth. Data on starting breastfeeding were collected from delivery reports, and data on breastfeeding initiation and continuation were derived from postal questionnaires at 2, 6, and 12 months after birth. Information on exclusive breastfeeding was available for 2 months. Questionnaire items are included in Appendix 2.

  28 Breastfeeding Three indicators of breastfeeding were constructed: starting breastfeeding,

  Educational Level Level of maternal education was established at enroll- ment and categorized as follows: 1 (low: no education, or primary school: ⱕ3 years of general secondary school); 2 (midlow: ⬎3 years of general secondary school); 3 (midhigh: higher vocational training, under- graduate programs, or bachelor’s degree); and 4 (high: higher academic education) (Appendix 1).

  Because of the different circumstances regarding breastfeeding and delivery, twin pregnancies (115) were excluded. To avoid clustering, our analyses excluded data on the second (n ⫽ 365) or third (n ⫽ 8) pregnancy of the women who were participating in the Generation R Study with ⬎1 child (10.2%). Our analyses also ex- cluded participants with missing data on any of the following: educational level (n ⫽ 14), starting breast- feeding (n ⫽ 130), breastfeeding at 2 months after birth (n ⫽ 186), or breastfeeding at 6 months after birth (n ⫽ 55). Eventually, 2914 subjects were available for analyses. Thus, after exclusion of the data of mothers who gave birth to twins, and of mothers regarding their second or third pregnancy in the study, complete data on determinant and outcome was available for 88.3% of Dutch mothers who gave postnatal consent.

  given by a total of 3787 Dutch women with live-born children.

  This study was restricted to Dutch participants. A woman was classified as being of Dutch ethnicity when she reported that both her parents had been born in the Netherlands.

  Study Population The Generation R Study is conducted in Rotterdam, Netherlands. The study area is defined by postal codes and covers more than half of the city’s inhabitants (⬃600 000). Collaboration was established with all 8 midwifery practices, 3 hospitals, and 16 child health centers located in this area. Invitations to participate in the study were made to all pregnant mothers who had an expected delivery date between April 2002 and Jan- uary 2006 and who lived in the study area.

  in early pregnancy (at a gestational age of ⬍18 weeks), which was the case in 69% of subjects, but was possible until the birth of the child. The study was conducted in accordance with the guidelines proposed in the World Medical Association Declaration of Helsinki and was ap- proved by the Medical Ethical Committee at Erasmus MC, University Medical Centre Rotterdam. Written con- sent was obtained from all participants.

  METHODS Study Design This study was embedded in the Generation R Study, a population-based, prospective cohort study from fetal life until young adulthood. The Generation R Study was designed to identify early determinants of growth, de- velopment, and health.

  stricted to women with Dutch ethnicity to eliminate the effect of ethnicity from the effect of educational level on breastfeeding.

  21 found that so- cioeconomic status is not associated with breastfeeding continuation in nonwhite women.

  21–25 For example, Griffiths et al

  19,20 The association between socioeconomic status and infant feeding practice may vary by ethnic group.

21 This study was re-

26 Enrollment ideally took place

27 Consent for postnatal follow-up was

  BMI was calculated (normal weight [BMI ⬍ 25 kg/m

  Methodologic Considerations This study used maternal educational level as single indi- cator of maternal socioeconomic status.

  2 ], obese [BMI ⱖ 30 kg/m

  2 ]).

33 Information on maternal smoking (nonsmok-

  ing during pregnancy, smoking during pregnancy) and alcohol use (nonuser during pregnancy, user during pregnancy) was retrieved from the prenatal questionnaire.

  Birth Characteristics Birth characteristics were obtained from delivery records and included birth weight (⬍2500 g, ⱖ2500 g), cesarean delivery (yes, no), and gestational age (⬍37 weeks, ⱖ37 weeks). In the Netherlands, deliveries by women who are not at increased risk of obstetric and medical com- plications are conducted under the responsibility of a community midwife, either at home or in the hospital. Delivery by women with an increased risk (eg, twin pregnancy, gestational hypertension) takes place in the hospital under the responsibility of an obstetrician. Place and type of delivery was obtained from delivery reports (home birth; hospital delivery, low risk; hospital deliv- ery, high risk).

  Statistical Analyses We established the frequency distributions of breast- feeding and covariates for breastfeeding according to educational level. ␹

  2 tests were used to test the differ- ences in mothers’ educational level for covariates for breastfeeding.

  Multivariable logistic regression was used to test the association between educational level, breastfeeding indicators, and mediators. Model 1 was the association between educational level and breastfeeding indicators. We then added successively to model 1 all sociodemo- graphic characteristics (model 2), all lifestyle-related characteristics (model 3), all psychosocial characteris- tics (model 4), and all birth characteristics (model 5). In model 6, all covariates were simultaneously added to model 1. All covariates were treated as categorical variables.

  If the percentage of missing values of covariates in the study population did not exceed 5%, subjects with miss- ing values on that covariate were assigned to the most prevalent category for that variable (single parenting, place and type of delivery, parity). If ⬎5% was missing on a particular covariate, a separate “missing” category was included in the analyses (job status, maternal BMI, maternal smoking, alcohol use, method of delivery, planned pregnancy, maternal stress).

  2 ], overweight [BMI ⫽ 25–30 kg/m

  Overall, 89.1% (2596 of 2914) of women started breastfeeding, which was lower for mothers in education category 1 (73.1% [255 of 349]) than for mothers in education category 4 (95.5% [985 of 1031]) (P ⬍ .001) (Fig 1, Table 2). Of the 2596 mothers who started breast- feeding, 1923 (74.1%) were still breastfeeding when their infants were aged 2 months (Fig 1); 1042 (40.1%) of infants were exclusively breastfed. During this period, more mothers in education category 4 than mothers in education category 1 continued breastfeeding (odds ratio [OR]: 6.36 [95% confidence interval (CI): 4.71– 8.60]) (Table 3), and more mothers in education category 4 continued exclusive breastfeeding (OR: 2.78 [95% CI: 2.04 –3.80]). Of the 1923 mothers who were still breast- feeding when their infants were aged 2 months, 904 (47%) continued breastfeeding until 6 months (Fig 1). The continuation of breastfeeding between 2 and 6 months did not differ between mothers in education category 4 and education category 1 (OR: 1.17 [95% CI: 0.79 –1.72]) (data not shown). At 6 months, 39.3% (405 of 1031) of the highest-educated mothers and 15.2% (53 of 349) of the lowest-educated mothers were still breast- feeding. All covariates, except prenatal stress and pre- term birth, were associated with at least 1 of 3 breast- feeding indicators (Appendix 3).

  The association between educational level and start- ing breastfeeding was attenuated most by the addition of the lifestyle-related characteristics followed by the birth characteristics. Sociodemographic and psychosocial characteristics had no appreciable effect on the ORs for starting breastfeeding (Table 2). A similar effect was seen in the association between educational level and con- tinuing breastfeeding for 2 months (Table 3).

  DISCUSSION This cohort study shows that although there was a graded inverse gradient between educational level and starting breastfeeding and continuing it until 2 months after birth, there was no such gradient for continuing breastfeeding between 2 and 6 months after birth. These associations are partly explained by lifestyle-related and birth characteristics. Nonetheless, associations between breastfeeding practices and a woman’s educational back- ground remain largely unexplained.

34 The maximum percentage of missing values was 25.0%.

  35 Socioeconomic status is a multifactor construct, whose most common in- dicators are educational level, income level, and occupa- tional class.

  20,35 As a socioeconomic indicator, level of edu- cation can also be applied to teenage and unemployed mothers, unlike occupational class. However, educational level does not entirely capture the material and financial aspects of socioeconomic status.

  19,20 Although fathers also influence breastfeeding decisions,

  8 fathers’ educational level was not taken into account. We chose to study the association between mothers’ educational level and breast-

  All statistical analyses were performed by using SPSS 15.0 for Windows (SPSS Inc, Chicago, IL).

  RESULTS The mothers’ mean age was 31.7 years (SD: 4.22). The infants’ mean birth weight was 3490 g (SD: 551) and lay below 2500 g for 4.1% of the infants. Mean gestational age was 40.0 weeks (range: 27.1– 43.4 weeks); 4.5% of the infants were born before 37 weeks of gestation (Ta- ble 1). All covariates in Table 1, except cesarean delivery and gestational age, were significantly associated with educational level. feeding, because she is the main caregiver for the child and the only person who can breastfeed. However, fathers’ educational level was associated with all breastfeeding in- dicators, but less strongly than mothers’ educational level. In addition, fathers’ educational level did not differ much from mothers’ educational level within 1 family (data not shown).

  There was some selection toward a study population that was relatively highly educated and somewhat healthier.

  12.9 Job status after delivery Employed

  13.4

  17.0

  18.9 ⬍.01

  Missing

  12.9

  16.0

  14.1

  10.5

  62.1

  16.2

  40.7

  56.5

  65.9

  70.4 ⬍.001

  Missing

  25.0

  31.8

  26.6

  23.3

  12.0

  Yes

  ⬍37

  3.9

  84.1 ⬍.001

  Missing

  5.8

  5.4

  5.4

  6.7

  5.6 Prenatal stress (FAD)

  c

  Yes

  6.9

  d

  4.3

  4.2

  2.3 ⬍.001

  Missing

  9.8

  12.9

  11.6

  7.1

  9.4 Postnatal stress (MABS)

  22.9 Birth characteristics Gestational age, wk

  4.5

  75.1

  24.2 ⬍.001

  6.6

  8.2

  7.8

  7.4 Place and type of delivery Home

  20.5

  10.6

  15.5

  24.5

  Hospital, low-risk

  13.2 11.4 .62 Missing

  16.2

  22.9

  17.3

  13.9

  14.9 Hospital, high-risk

  63.3

  66.5

  67.2

  61.6

  7.6

  14.1

  6.9

  7.2

  4.7

  4.1 4.0 .13 ⱖ

  37

  95.5

  93.1

  95.3

  95.9

  96.0 Birth weight, g ⬍2500

  4.1

  4.4

  14.0

  3.9

  2.9 ⬍.01

  ⱖ 2500

  95.9

  92.8

  95.6

  96.1

  97.1 Cesarean delivery Yes

  12.9

  79.7

  66.5

  26 For selective participation to explain our re- sults, nonresponders would have been more often of low socioeconomic status and more often breastfeeding. This is unlikely, although we cannot ascertain this.

  96.4 Parity Primiparae

  7.0

  4.4

  3.6 ⬍.001

  No

  93.9

  84.8

  93.0

  95.6

  67.0

  6.1

  60.5

  68.8

  69.5

  66.0 ⬍.05

  Multiparae

  33.0

  39.5

  31.2

  30.5

  15.2

  22.6 Single parenting Yes

  2

  16.1 ⬍.001

  This study was restricted to Dutch women. In accor- dance with the Dutch Standard Classification,

  27 we as- signed a Dutch ethnicity to a participant if both her

  TABLE 1 Subject Characteristics for Total Study Population and According to Educational Level (N ⴝ 2914) Total, % Mothers’ Educational Level, % a P b 1 (N ⴝ 349) 2 (N ⴝ 747) 3 (N ⴝ 787) 4 (N ⴝ 1031)

  Maternal characteristics Age, y

  ⬍30

  30.3

  53.3

  41.0

  28.5

  30–35

  21.5

  49.8

  30.7

  42.4

  50.1

  61.3 ⱖ

  35

  20.0

  16.0

  16.6

  34.0 Prepregnant BMI ⬍25 kg/m

  67.4

  78.5

  21.4

  4.7

  2.2 ⬍.001

  Missing

  13.2

  10.6

  14.6

  13.2

  13.0 Alcohol use in pregnancy Yes

  8.6

  28.1

  14.1

  23.0

  29.9 ⬍.001

  Missing

  13.0

  10.3

  14.5

  12.7

  13.0 Planned pregnancy Yes

  10.4

  8.1

  53.9

  11.9 ⱖ 30 kg/m

  61.8

  70.1

  73.8 ⬍.001

  25–30 kg/m

  2

  14.1

  18.6

  16.5

  12.8

  2

  12.4 Continued smoking in pregnancy Yes

  5.4

  11.7

  9.2

  3.7

  1.8 Missing

  13.1

  15.8

  12.4

  13.3

  60.8 _a Scale: 1 ⫽ lowest, 4 ⫽ highest. _b P for ␹ ² tests. _c Yes indicates a score of ⬎2.17 (clinical cutoff score). _d Subscale ⫽ lack of confidence in caretaking, yes ⫽ score 1 SD above mean. parents had been born in the Netherlands. However, when identifying immigrant descent in Dutch residents, this classification goes no further than the second gen- eration. The number of third-generation immigrants is nonetheless likely to have been very small and not to have affected our conclusions. Non-Dutch women in the Generation R Study had a substantially different distri- bution on educational level, and some categories were not represented, which makes it difficult to distinguish between the effect of ethnicity and educational level. However, ethnicity may also be an important marker of social disadvantage, and its association with breastfeed- ing should be further investigated.

  When women recall breastfeeding only shortly after it has finished, as in our study, self-reported information on breastfeeding has shown itself to be valid and reli- able.

36 However, because we lacked information on post-

37 Misclassification or underreporting of the

  44 The Role of Smoking Alterations in infants’ sleep patterns have been found after they had ingested nicotine from breastfeeding. It is

  FIGURE 1 Percentage of mothers who breastfed during their child’s first 6 months of life, according to mother’s educational level (N ⫽ 2914).

  80 100 Low educational level Mid-low educational level Mid-high educational level High educational level

  60

  40

  20

  6 Percentage

  5

  4

  3

  2

  1

  Child's age, mo

  44,45 Also, obese women often have large breasts, which can create diffi- culties in attaching the infant to the breast. Finally, obese women more often belong to environments with different health beliefs. These last 2 mechanisms could also be interrelated; the delay in lactogenesis can be due to a lower intention to breastfeed.

  natal smoking behavior, we used smoking during preg- nancy as a proxy for postnatal smoking. Consequently, nonsmoking mothers during pregnancy who restarted smoking after delivery were misclassified as “nonsmok- ing.” In addition, smoking during pregnancy is likely to be underreported, which was slightly the case in our population.

  7–11 However, the educationally re- lated differences were greater than those reported in earlier studies.

  determinants of breastfeeding, and in particular smoking status, may have led to residual confounding resulting in a lack of an explanation for the association we observed between maternal education and breastfeeding.

  Health gain from breastfeeding is optimal when an infant is exclusively breastfed for 6 months.

  exclusive breastfeeding were available for the first 2 months after birth. Educationally related differences were seen in both breastfeeding and exclusive breast- feeding. However, health benefits are also reported for breastfeeding initiation and several durations of (partly) breastfeeding compared with never breastfed.

  39 Comparison With Other Studies Although the breastfeeding rates in our study were fairly high relative to those in earlier reports,

  16,40 they were consistent with the upward trend toward breastfeeding that has been taking place in Western countries since the 1990s.

  41 Our study is also consistent with earlier reports stat- ing that rates of breastfeeding are higher in women with a higher socioeconomic status than in those with a lower socioeconomic status.

  7–11 For example, when comparing moth- ers who had at least a high school diploma with those who had no such diploma, Dubois and Girard

  44 First, obese women more often have medical complications. However, this is probably not the under- lying mechanism in our study, because we adjusted for several perinatal characteristics. Second, psychological factors may play a role; obese mothers are more often depressed and depressed mothers are less likely to breastfeed. From our data, prenatal and postnatal stress did not influence the association between educational level and breastfeeding. A third mechanism may be an- atomically/physiologically related. A delay in lactogene- sis has been found in obese women.

  7 found an OR of 3.54 (95% CI: 2.56 – 4.89) for starting breastfeed- ing. When we repeated our analyses, categorizing ma- ternal education into 3 levels similar to those of Dubois and Girard,

  7 the OR for the highest versus the lowest educated mothers was 5.99 (95% CI: 3.40 –10.58). A recent study that found no socioeconomic differences in breastfeeding claimed that when overall breastfeeding prevalence exceeds 80%, there will no longer be any socioeconomic differences.

  42 Our results would tend to refute this; despite a very high overall breastfeeding initiation rate, we found substantial educationally re- lated differences in breastfeeding rates.

  Explaining the Educational Level/Breastfeeding Association We considered 4 groups, including 13 covariates in total, of potential mediators. Lifestyle-related and birth char- acteristics can be considered as mediators, because they considerably attenuated the association between educa- tional level and breastfeeding.

  43 Next, we discuss the role of potential mediating covariates in the association be- tween educational level and breastfeeding. However, because the causal mechanisms regarding the effects on breastfeeding are unknown, the covariates in our study might also be considered as additional markers of social disadvantage.

  The Role of Obesity It is unclear why obese mothers breastfeed less than nonobese mothers. Four possible mechanisms have been described.

38 Data on

  hypothesized that decreased (quality of) sleep results in dissatisfied infants, which is then thought by the moth- ers to be due to lack of breastfeeding.

  1.00

  Psychosocial factors Planned pregnancy

  Planned

  1.00

  1.00 Unplanned 1.03 (0.76–1.41) 1.03 (0.73–1.44)

  Missing 1.08 (0.60–1.94) 1.51 (0.73–3.12)

  Prenatal stress (FAD)

  c

  No

  1.00

  1.00 Yes 0.87 (0.61–1.23) 1.67 (0.87–3.20)

  Missing 0.51 (0.37–0.70) 1.66 (1.05–2.64)

  Birth characteristics Gestational age, wk

  ⱖ

  37

  1.00 ⬍37

  1.00

  1.07 (0.55–2.09) 0.97 (0.50–1.90) Birth weight, g

  ⱖ 2500

  1.00

  1.00 ⬍2500

  0.68 (0.35–1.30) 0.74 (0.38–1.44) Cesarean delivery

  No

  1.00

  1.00 Yes 0.79 (0.55–1.13) 0.78 (0.54–1.13)

  Missing 0.73 (0.47–1.15) 0.71 (0.44–1.13)

  Place and type of delivery Home

  1.00

  1.00 Hospital, low risk 0.49 (0.32–0.75) 0.45 (0.29–0.69)

  Hospital, high risk 0.75 (0.51–1.11) 0.67 (0.45–1.00) _a

  

Values shown are ORs (95% CIs). Model 1: basic model ⫽ association between educational level and breastfeeding; model 2: basic model ⫹ sociodemographic factors simultaneously;

model 3: basic model ⫹ lifestyle-related factors simultaneously; model 4: basic model ⫹ psychosocial factors simultaneously; model 5: basic model ⫹ birth characteristics simultaneously;

model 6: basic model ⫹ models 2, 3, 4, and 5. _b Scale: 1 ⫽ lowest, 4 ⫽ highest. _c Yes indicates a score of ⬎2.17 (clinical cutoff score).

  1.00 Smoking in pregnancy 0.47 (0.33–0.66) 0.45 (0.31–0.64) Missing 0.62 (0.44–0.87) 0.52 (0.36–0.75)

  Smoking Nonsmoking

  mothers are more likely to perceive their milk supply as insufficient and are less likely to seek help with breast- feeding difficulties than nonsmokers.

  1.00 30–35 1.00 (0.75–1.32)

  46–49 However, this

  TABLE 2 Logistic Regression for Starting Breastfeeding (N ⴝ 2914) Model 1 a Model 2 a Model 3 a Model 4 a Model 5 a Model 6 a

  Mothers’ educational level

  b

  1 (reference)

  1.00

  1.00

  1.00

  1.00

  1.00

  1.00 2 1.77 (1.30–2.39) 1.73 (1.27–2.37) 1.54 (1.12–2.11) 1.75 (1.29–2.38) 1.70 (1.25–2.31) 1.47 (1.06–2.04) 3 5.55 (3.82–8.07) 5.39 (3.65–7.96) 4.38 (2.96–6.46) 5.40 (3.70–7.89) 5.16 (3.54–7.52) 4.03 (2.68–6.06) 4 7.89 (5.41–11.53) 7.85 (5.24–11.75) 5.93 (3.98–8.85) 7.86 (5.34–11.55) 7.31 (4.99–10.71) 5.51 (3.59–8.45)

  Sociodemographic factors Maternal age, y

  ⬍30

  1.00

  1.04 (0.78–1.38) ⱖ 35 1.46 (1.00–2.15)

  0.46 (0.31–0.70) 0.50 (0.33–0.77) Missing 0.91 (0.63–1.32) 0.81 (0.53–1.26)

  1.00

  1.00 Overweight 0.86 (0.61–1.21) 0.91 (0.64–1.29) Obese

  1.00

  Normal weight

  Lifestyle-related factors Prepregnant BMI

  Missing 0.57 (0.38–0.84) 0.55 (0.37–0.82)

  1.00 Employed 0.73 (0.50–1.06) 0.74 (0.50–1.09)

  Job status Not employed

  1.56 (1.06–2.32) Single parenting

  1.00 Multiparae 0.63 (0.49–0.82) 0.58 (0.44–0.77)

  1.00

  Parity Primiparae

  1.00 Yes 0.91 (0.58–1.43) 0.89 (0.55–1.45)

  1.00

  No

46 Indeed, smoking

  TABLE 3 Logistic Regression for Continuing Breastfeeding Until Baby Is Aged 2 Months (N ⴝ 2596) Model 1 a Model 2 a Model 3 a Model 4 a Model 5 a Model 6 a

  37

  1.00 Alcohol use 1.12 (0.88–1.42) 1.01 (0.79–1.28) Missing 2.14 (0.44–10.30) 1.94 (0.38–9.78)

  Psychosocial factors Planned pregnancy

  Planned

  1.00

  1.00 Unplanned 1.03 (0.80–1.32) 0.97 (0.74–1.28)

  Missing 1.15 (0.73–1.80) 1.07 (0.61–1.89)

  Postnatal stress (MABS)

  c

  No

  1.00

  1.00 Yes 0.85 (0.66–1.09) 0.82 (0.63–1.06)

  Missing 1.24 (0.92–1.68) 1.26 (0.92–1.72)

  Birth characteristics Gestational age, wk

  ⱖ

  1.00

  Alcohol use No alcohol use

  1.00 ⬍37

  1.07 (0.63–1.83) 1.13 (0.66–1.93) Birth weight, g

  ⱖ 2500

  1.00

  1.00 ⬍2500

  1.06 (0.60–1.88) 1.00 (0.56–1.79) Cesarean delivery

  No

  1.00

  1.00 Yes 0.72 (0.55–0.94) 0.71 (0.54–0.94)

  Missing 1.57 (1.02–2.42) 1.44 (0.92–2.23)

  Place and type of delivery Home

  1.00

  1.00 Hospital, low-risk 0.72 (0.52–1.01) 0.71 (0.50–0.99)

  Hospital, high-risk 0.63 (0.48–0.82) 0.62 (0.47–0.82) _a

  1.00

  1.00 Smoking in pregnancy 0.52 (0.37–0.74) 0.47 (0.33–0.67) Missing 0.57 (0.12–2.71) 0.55 (0.11–2.69)

  Mothers’ educational level

  No

  b

  1 (Reference)

  1.00

  1.00

  1.00

  1.00

  1.00

  1.00 2 2.11 (1.57–2.84) 2.16 (1.60–2.92) 1.84 (1.35–2.49) 2.12 (1.58–2.85) 2.07 (1.54–2.80) 1.84 (1.34–2.52) 3 3.82 (2.83–5.15) 3.85 (2.81–5.26) 3.01 (2.20–4.12) 3.87 (2.86–5.24) 3.68 (2.71–4.98) 2.94 (2.11–4.10) 4 6.36 (4.71–8.60) 6.33 (4.59–8.73) 4.81 (3.49–6.62) 6.44 (4.75–8.74) 6.14 (4.52–8.33) 4.70 (3.33–6.64)

  Sociodemographic factors Maternal age, y

  ⬍30

  1.00

  1.00 30–35 1.19 (0.96–1.47)

  1.26 (1.01–1.57) ⱖ 35 1.72 (1.29–2.28)

  1.85 (1.38–2.49) Single parenting

  1.00

  1.00

  1.00 Yes 0.93 (0.64–1.36) 0.94 (0.63–1.41)

  Parity Primiparae

  1.00

  1.00 Multiparae 1.05 (0.85–1.30) 0.96 (0.77–1.21)

  Job status Not employed

  1.00

  1.00 Employed 0.79 (0.59–1.05) 0.78 (0.58–1.04)

  Missing 0.90 (0.66–1.24) 0.95 (0.68–1.31)

  Lifestyle-related factors Prepregnant BMI

  Normal weight

  1.00

  1.00 Overweight 0.64 (0.49–0.82) 0.66 (0.51–0.86) Obese

  0.37 (0.25–0.55) 0.39 (0.26–0.58) Missing 0.86 (0.64–1.14) 0.87 (0.62–1.23)

  Smoking Nonsmoking

  

Values shown are ORs (95% CIs). Model 1: basic model ⫽ association between educational level and breastfeeding; model 2: basic model ⫹ sociodemographic factors simultaneously;

model 3: basic model ⫹ lifestyle-related factors simultaneously; model 4: basic model ⫹ psychosocial factors simultaneously; model 5: basic model ⫹ birth characteristics simultaneously; model 6: basic model ⫹ models 2, 3, 4, and 5. _b Scale: 1 ⫽ lowest, 4 ⫽ highest. _c Subscale ⫽ lack of confidence in caretaking, yes ⫽ score 1 SD above mean. does not fully explain the association we found between smoking and starting breastfeeding. Literature describes that smokers also have a lower intention to breastfeed.

47 This suggests that smoking in our study may be a proxy

  for other more motivational factors rather than for chemical factors.

  18. Giglia RC, Binns CW, Alfonso HS, Scott JA, Oddy WH. The effect of alcohol intake on breastfeeding duration in Australian women. Acta Paediatr. 2008;97(5):624 – 629

  14. Hendricks K, Briefel R, Novak T, Ziegler P. Maternal and child characteristics associated with infant and toddler feeding prac- tices. J Am Diet Assoc. 2006;106(1 suppl 1):S135–S148 15. Kohlhuber M, Rebhan B, Schwegler U, Koletzko B, Fromme H. Breastfeeding rates and duration in Germany: a Bavarian co- hort study. Br J Nutr. 2008;99(5):1127–1132

  16. Lanting CI, Van Wouwe JP, Reijneveld SA. Infant milk feeding practices in the Netherlands and associated factors. Acta Paedi-

  atr. 2005;94(7):935–942

  17. Scott JA, Binns CW. Factors associated with the initiation and duration of breastfeeding: a review of the literature. Breastfeed

  Rev. 1999;7(1):5–16

  munity Health. 2006;60(1):7–12

  19. 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 20. Galobardes B, Shaw M, Lawlor DA, Lynch JW, Davey Smith G. Indicators of socioeconomic position (part 1). J Epidemiol Com-

  12. Callen J, Pinelli J. Incidence and duration of breastfeeding for term infants in Canada, United States, Europe, and Australia: a literature review. Birth. 2004;31(4):285–292

  21. Griffiths LJ, Tate AR, Dezateux C; Millennium Cohort Study Child Health Group. Do early infant feeding practices vary by maternal ethnic group? Public Health Nutr. 2007;10(9):957–964

  22. Kelly YJ, Watt RG, Nazroo JY. Racial/ethnic differences in breastfeeding initiation and continuation in the United King- dom and comparison with findings in the United States. Pedi-

  atrics. 2006;118(5). Available at: www.pediatrics.org/cgi/

  content/full/118/5/e1428

  23. Singh GK, Kogan MD, Dee DL. Nativity/immigrant status, race/ethnicity, and socioeconomic determinants of breastfeed- ing initiation and duration in the United States, 2003. Pediatrics. 2007;119(suppl 1):S38 –S46

  13. Grjibovski AM, Yngve A, Bygren LO, Sjostrom M. Socio- demographic determinants of initiation and duration of breast- feeding in northwest Russia. Acta Paediatr. 2005;94(5):588 –594

  11. Heck KE, Braveman P, Cubbin C, Chavez GF, Kiely JL. Socio- economic status and breastfeeding initiation among California mothers. Public Health Rep. 2006;121(1):51–59

  The Role of Place and Type of Delivery Relative to those who had delivered at home, fewer women who delivered in the hospital started breastfeed- ing. More research is needed to elucidate this finding. But, in any case, hospitals should give optimal guidance to mothers on breastfeeding.

  50–52 The effects of smoking behavior and maternal obesity on the association between educational level and breast- feeding may be partly explained by motivational fac- tors.

  10. Celi AC, Rich-Edwards JW, Richardson MK, Kleinman KP, Gillman MW. Immigration, race/ethnicity, and social and eco- nomic factors as predictors of breastfeeding initiation. Arch

  9. Kelly YJ, Watt RG. Breast-feeding initiation and exclusive du- ration at 6 months by social class—results from the Millennium Cohort Study. Public Health Nutr. 2005;8(4):417– 421

  8. Glenn LL, Quillin SI. Opposing effects of maternal and paternal socioeconomic status on neonatal feeding method, place of sleep, and maternal sleep time. J Perinat Neonatal Nurs. 2007; 21(2):165–172

  7. Dubois L, Girard M. Social inequalities in infant feeding during the first year of life. The Longitudinal Study of Child Develop- ment in Quebec (LSCDQ 1998 –2002). Public Health Nutr. 2003; 6(8):773–783

  5. Owen CG, Martin RM, Whincup PH, Smith GD, Cook DG. Does breastfeeding influence risk of type 2 diabetes in later life? A quantitative analysis of published evidence. Am J Clin Nutr. 2006;84(5):1043–1054 6. World Health Organization, United Nations Children’s Fund. Global strategy for infant and young child feeding. Available at: www.paho.org/english/ad/fch/ca/GSIYCF㛭infantfeeding㛭eng.pdf. Accessed March 24, 2009

  2005;115(5):1367–1377

  The Role of Motivational Determinants of Breastfeeding We had no information on the motivational determi- nants of breastfeeding. From the literature, it is known that breastfeeding behavior can be predicted by atti- tudes, perceived control, and social support, including social norms, peer influence, and intergenerational fac- tors (elements of the theory of planned behavior).

  Pediatr Adolesc Med. 2005;159(3):255–260

53 CONCLUSIONS

  We acknowledge the contribution of participating mothers, general practitioners, hospitals, and midwives in Rotterdam.

  ACKNOWLEDGMENTS This study was supported by grant 7110.0002 from the Netherlands Organization for Health Research and De- velopment (ZonMW).

  However, more research is needed on obesity and smoking behavior and their association with motiva- tional factors that underlie social inequalities in breast- feeding.

  Compared with their less-educated counterparts, more mothers with a higher level of education started breast- feeding and more continued for the first 2 months after birth. Because breastfeeding benefits health both in childhood and adulthood, increasing breastfeeding by less-educated mothers would help reduce socially based health inequalities, potentially achieving a broad public health impact.

  REFERENCES

  2. Howie PW, Forsyth JS, Ogston SA, Clark A, Florey CD. Protec- tive effect of breast feeding against infection. BMJ. 1990; 300(6716):11–16 3. Wright AL, Holberg CJ, Martinez FD, Morgan WJ, Taussig LM. Breast feeding and lower respiratory tract illness in the first year of life. Group Health Medical Associates. BMJ. 1989; 299(6705):946 –949 4. Owen CG, Martin RM, Whincup PH, Smith GD, Cook DG. Effect of infant feeding on the risk of obesity across the life course: a quantitative review of published evidence. Pediatrics.

  1. Duncan B, Ey J, Holberg CJ, Wright AL, Martinez FD, Taussig LM. Exclusive breast-feeding for at least 4 months protects against otitis media. Pediatrics. 1993;91(5):867– 872

  24. Gibson-Davis CM, Brooks-Gunn J. Couples’ immigration sta- tus and ethnicity as determinants of breastfeeding. Am J Public

  nancy Childbirth. 2007;7:9

  39. Ip S, Chung M, Raman G, et al. Breastfeeding and maternal and infant health outcomes in developed countries. Evid Rep

  Technol Assess (Full Rep). 2007;(153):1–186

  40. Dyson L, McCormick F, Renfrew MJ. Interventions for pro- moting the initiation of breastfeeding. Cochrane Database Syst

  Rev. 2005;(2):CD001688

  41. Ryan AS, Wenjun Z, Acosta A. Breastfeeding continues to increase into the new millennium. Pediatrics. 2002;110(6): 1103–1109

  42. Scott JA, Binns CW, Graham KI, Oddy WH. Temporal changes in the determinants of breastfeeding initiation. Birth. 2006; 33(1):37– 45