SUPPLEMENT ARTICLE

Effect of Maternity-Care Practices on Breastfeeding

Ann M. DiGirolamo, PhD, MPHa, Laurence M. Grummer-Strawn, PhDb, Sara B. Fein, PhDc

aHubert Department of Global Health, Emory University, Atlanta, Georgia; bNational Center for Chronic Disease Prevention and Health Promotion, Centers for Disease

Control and Prevention, Atlanta, Georgia; cCenter for Food Safety and Applied Nutrition, Food and Drug Administration, US Department of Health and Human Services,
College Park, Maryland

The authors have indicated they have no financial relationships relevant to this article to disclose.

ABSTRACT
OBJECTIVE. Our goal was to assess the impact of “Baby-Friendly” hospital practices and
other maternity-care practices experienced by mothers on breastfeeding duration.
www.pediatrics.org/cgi/doi/10.1542/
METHODS. This analysis of the Infant Feeding Practices Study II focused on mothers who peds.2008-1315e
initiated breastfeeding and intended prenatally to breastfeed for ⬎2 months, with doi:10.1542/peds.2008-1315e
complete data on all variables (n ⫽ 1907). Predictor variables included indicators of The findings and conclusions in this article
6 “Baby-Friendly” practices (breastfeeding initiation within 1 hour of birth, giving are those of the authors and do not
only breast milk, rooming in, breastfeeding on demand, no pacifiers, fostering necessarily represent the official position of
the Centers for Disease Control and
breastfeeding support groups) along with several other maternity-care practices. The Prevention or the Food and Drug
main outcome measure was breastfeeding termination before 6 weeks. Administration.

RESULTS. Only 8.1% of the mothers experienced all 6 “Baby-Friendly” practices. The Key Words
breast feeding, maternity, hospital
practices most consistently associated with breastfeeding beyond 6 weeks were
Abbreviation
initiation within 1 hour of birth, giving only breast milk, and not using pacifiers. IFPS—Infant Feeding Practices Study
Bringing the infant to the room for feeding at night if not rooming in and not giving Accepted for publication Jun 4, 2008
pain medications to the mother during delivery were also protective against early Address correspondence to Ann M.
breastfeeding termination. Compared with the mothers who experienced all 6 DiGirolamo, PhD, MPH, Emory University,
“Baby-Friendly” practices, mothers who experienced none were ⬃13 times more Hubert Department of Global Health, 1518
Clifton Rd, NE, Atlanta, GA 30307. E-mail:
likely to stop breastfeeding early. Additional practices decreased the risk for early [email protected]
termination. PEDIATRICS (ISSN Numbers: Print, 0031-4005;
Online, 1098-4275); published in the public
CONCLUSIONS. Increased “Baby-Friendly” hospital practices, along with several other domain by the American Academy of
maternity-care practices, improve the chances of breastfeeding beyond 6 weeks. The Pediatrics
need to work with hospitals to implement these practices continues to exist, as
illustrated by the small proportion of mothers who reported experiencing all 6 of the “Baby-Friendly” hospital
practices measured in this study. Pediatrics 2008;122:S43–S49

B REASTFEEDING PROVIDES MANY benefits to both infants and mothers, including optimal nutrients for infant
growth and development, enhancing infants’ immunologic defenses, and facilitating mother-infant attachment
and mothers’ recovery from childbirth.1,2 However, despite the known benefits of breastfeeding, a substantial
proportion of mothers do not breastfeed their infants or breastfeed for ⬍6 months postpartum. In 2004, 73.8% of US
mothers breastfed during the early postpartum period, and 41.5% continued to breastfeed at 6 months postpartum.3
Although these findings represent a steady increase over the years in the percentage of women breastfeeding, the
data still fall short of the national Healthy People 2010 goals of 75% women breastfeeding during the early
postpartum period and 50% breastfeeding 6 months postpartum.4
Certain maternity-care practices in hospital settings have been shown to promote breastfeeding. In particular, the
Baby-Friendly Hospital Initiative, a global movement launched in 1991 by the World Health Organization and the
United Nations Children’s Fund (UNICEF), has been associated with positive breastfeeding outcomes both nationally
and internationally.5–9 This initiative includes 10 steps to successful breastfeeding, including specific recommenda-
tions for maternity-care practices (Table 1). A 2001 national study that used data from the original Infant Feeding
Practices Study (IFPS I) evaluated the influence of 5 of the 10 “Baby-Friendly” practices on breastfeeding. The study
demonstrated significant associations between 2 of the 5 practices measured (initiating breastfeeding within 1 hour
of birth and giving no food or drink other than breast milk) and breastfeeding and illustrated the cumulative effects
of these 5 steps on positive breastfeeding outcomes.10 The study also revealed that only a small percentage (7%) of
women reported experiencing all 5 of the “Baby-Friendly” practices measured.
The purpose of the current study is to examine the current prevalence and the individual and cumulative
influences of a greater number of “Baby-Friendly” hospital practices on breastfeeding duration among mothers who
intended to breastfeed for at least 2 months postpartum. Specifically, using data from the IFPS II, the study provides
an opportunity to assess changes in the prevalence of reported “Baby-Friendly” hospital practices since the admin-

PEDIATRICS Volume 122, Supplement 2, October 2008 S43

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
 TABLE 1 The Baby-Friendly Hospital Initiative’s 10 Steps to and 1-month postnatal questionnaires, along with the
Successful Breastfeeding4,5 and Prevalence of Reported data on actual breastfeeding duration, and were lim-
Practices Among Women Who Initiated Breastfeeding in ited to the sample of women with complete data on all
variables (n ⫽ 1907).
the IFPS II (N ⴝ 1907)
“Baby-Friendly” Practice Prevalence,
% Definition of Variables
1. Have a written breastfeeding policy that is routinely NM The outcome variable was breastfeeding for ⬍6 weeks
communicated to all health care staff postpartum. Breastfeeding duration was defined as the
2. Train all health care staff in skills necessary to NM total length of time in weeks that infants were fed breast
implement this policy milk. This variable was based on age of the infant in
3. Inform all pregnant women about the benefits and NM weeks when the mother completely stopped breastfeed-
management of breastfeeding ing or pumping milk. The outcome of breastfeeding for
4. Help mothers initiate breastfeeding within 1 h of birtha 59.6 ⬍6 weeks was chosen because it is a time period less
5. Show mothers how to breastfeed and how to maintain NM
than, but close to, the mother’s stated intention to
lactation, even if they should be separated from their
infants
breastfeed for at least 2 months. The main independent
6. Give newborn infants no food or drink other than 51.5 variables were indicators of 6 of the 10 “Baby-Friendly”
breast milk unless medically indicateda,b hospital practices, those most easily measured through
7. Practice “rooming in” by allowing mothers and infants 56.8 mothers’ reports; the steps measured in the study in-
to remain together 24 h/da cluded breastfeeding initiation within 1 hour of birth,
8. Encourage breastfeeding on demanda,c 66.3 only breast milk given, rooming in, breastfeeding on
9. Give no artificial teats, pacifiers, dummies, or soothers 44.2 demand, no pacifiers given, and providing information
to breastfeeding infantsa,d on breastfeeding support. We attempted to measure a
10. Foster the establishment of breastfeeding support 72.0 seventh “Baby-Friendly” practice (show mothers how to
groups and refer mothers to them on discharge from
breastfeed and maintain lactation); however, after ex-
the hospital or clinica,e
amination of the data, the indicator seemed to be more
NM indicates not measured.
a Practices that were measured in the current study. The sample is limited to women who
reflective of providing help to those women experienc-
initiated breastfeeding and intended to breastfeed for ⬎2 months, with complete data on all ing initial problems with breastfeeding rather than a
variables. general practice of education and demonstration to all
b Question asks whether infant was fed water, formula, or sugar water at any time while in the
women. Therefore, this variable was dropped from ad-
hospital; practice coded as “yes” if the answer was “no” to all. Medical indication was not
ditional analysis. We also examined certain other mater-
assessed.
c Practice coded as “yes” if infant was fed whenever he or she cried or seemed hungry and staff nity-care practices that might impact breastfeeding, in-
encouraged this. cluding not providing a formula sample or coupon in a
d Question asks whether infant was given a pacifier while in the hospital or birth center; practice
hospital gift pack, bringing infants to the room at night
coded as “yes” if answer was “no.” for feeding if not rooming in, type of delivery (ie, vaginal
e Practice coded as “yes” if mother reported being given information about breastfeeding sup-

port groups or services before she went home from the hospital or birth center.
versus cesarean), not giving pain medications during
delivery, and the presence of labor support such as a
doula. Mothers were asked about their hospital experi-
ence on the first questionnaire sent after the delivery at
istration of the original IFPS 11 years ago and to confirm ⬃1 month postpartum.
the cumulative effects of a more complete measurement Control variables included demographic characteris-
of these practices on breastfeeding. In addition, the new tics, prenatal maternal smoking, number of friends and
survey allowed us to assess the impact of additional relatives who breastfed, mother’s prenatal intentions to
maternity-care practices not currently designated as be- work after birth, and prenatal attitudes toward breast-
ing “Baby-Friendly” on breastfeeding duration. feeding (ie, strength of agreement or disagreement with
the following statements: “Infant formula is as good as
METHODS breast milk” and “If a child was breastfed, he or she will
be less likely to become obese”).
Study Population
The IFPS II is a longitudinal survey of pregnant women
and new mothers conducted by the US Food and Drug Statistical Analyses
Administration. The sample is from a nationally distrib- SAS 9.1 (SAS Institute, Inc, Cary, NC) was used for all
uted consumer opinion panel of 500 000 households. All analyses. Procedures used included frequencies, ␹2 tests,
questionnaires were administered by mail except a short and logistic regression. For multivariate models with
birth screener telephone interview. Questionnaires were “Baby-Friendly” practices, we used the following proce-
sent prenatally and 10 times postnatally. Qualifying cri- dures. In model 1, analyses controlled for all “Baby-
teria were used to achieve the sample goals of healthy Friendly” practices experienced. Model 1 allowed us to
term or near-term singleton infants (see ref 11 for addi- examine the individual impact of each of the “Baby-
tional study details). The current study focused on Friendly” practices while controlling for the other practices.
mothers from the IFPS II who initiated breastfeeding In model 2, analyses included variables in model 1 plus
and intended prenatally to breastfeed for ⬎2 months. various demographic variables shown to be associated with
Analyses were performed on data from the prenatal breastfeeding duration. Model 3 included variables in

S44 DIGIROLAMO et al
Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
model 2 plus additional behavioral and attitudinal variables TABLE 2 Sample Characteristics for Maternity-Care Practices
also shown to be associated with breastfeeding duration to According to Timing of Breastfeeding Termination: IFPS II
assess the full impact of “Baby-Friendly” practices while (N ⴝ 1907)
controlling for relevant demographic, behavioral, and atti-
Variable Breastfed for ⬍6 wk Breastfed for ⱖ6 wk
tudinal variables. Analyses that examined additional ma-
(n ⫽ 274) (n ⫽ 1633)
ternity-care practices used crude associations followed by
Gender of child, %
associations adjusted for demographic, behavioral, and at-
Male 50.0 49.7
titudinal variables.
Female 50.0 50.3
We also counted how many of the 6 “Baby-Friendly” Age of mother, mean (SD), ya 27.0 (5.8) 29.5 (5.1)
practices were experienced. We modeled the effect of Race/ethnicity of the mother, %
each number of “Baby-Friendly” practices experienced White 85.8 84.9
compared with the referent group of experiencing all 6 Black 4.7 3.8
practices. We used a similar model-building strategy to Hispanic 6.9 6.3
examine effects while controlling for relevant demo- Asian/Pacific Islander/other 2.6 5.0
graphic, behavioral, and attitudinal variables. Education of mother, %a
ⱕ4 y of high school 29.2 14.5
1–3 y of college 48.5 39.3
RESULTS
ⱖ4 y of college 22.3 46.2
At least half of the mothers reported experiencing each
Household income, %a
of the “Baby-Friendly” hospital practices measured in Less than $20 000 22.6 10.0
the survey (Table 1). Very few women reported experi- $20 000–$39 999 31.0 28.8
encing all 6 (8.1%) or none (1.6%) of the hospital $40 000–$74 999 29.6 40.3
practices, with approximately half of the women expe- $75 000 or more 16.8 20.9
riencing either 3 or 4 of the “Baby-Friendly” practices. Marital status, %a
The overall prevalence of women who terminated Married 65.7 85.0
breastfeeding by 6 weeks postpartum was 14.4%. As Widowed 0.4 0.2
shown in Table 2, early breastfeeding termination was Divorced or separated 4.7 3.0
Never married 29.2 11.8
associated with younger age, lower education and in-
Previous births, %a
come, being unmarried, primiparity, smoking prena-
0 46.7 26.5
tally, having fewer friends and relatives who breastfed, ⱖ1 53.3 73.5
intending to work after birth, and having less-favorable No. of cigarettes smoked per day
attitudes toward breastfeeding. prenatally, %a
Of the 6 “Baby-Friendly” practices, 4 (breastfeeding 0 87.6 94.8
initiation within 1 hour, only breast milk given, breast- 1–5 4.7 1.8
feeding on demand, and no pacifiers given) showed a 6–10 4.7 1.7
significant protective effect against early termination of ⱖ11 2.9 1.7
breastfeeding (ie, ⬍6 weeks) in crude analyses (Table 3). No. of friends and relatives who
breastfed, %a
Results remained significant for 3 of the practices
None or don’t know 18.6 9.1
(breastfeeding initiation within 1 hour, only breast milk,
1–2 27.4 17.0
and no pacifiers) after controlling for the other “Baby- 3–5 31.0 28.1
Friendly” practices (Table 4, model 1). Controlling for ⱖ5 23.0 45.9
various demographic factors associated with breastfeed- Mother’s prenatal intention to
ing duration (Table 4, model 2) and certain relevant work after birth, %a
behavioral and attitudinal variables (Table 4, model 3) Yes 63.9 56.9
only slightly diminished the effect sizes for these prac- No 36.1 43.1
tices. Prenatal attitudes toward
Additional hospital practices showing a protective re- breastfeeding, %
Infant formula as good as
lationship with breastfeeding duration included bringing
breast milka
the infant to the mother’s room at night for feeding
Agree or unsure 44.5 25.8
among mothers not rooming in and not giving any pain Disagree 55.5 74.2
medications to the mother during childbirth (Table 5). If breastfed, child less likely to
Other practices, such as not providing a formula sample become obesea
or coupon in a hospital gift pack, support during labor, or Agree 31.0 48.0
type of delivery, were not significantly associated with Disagree or unsure 69.0 52.0
breastfeeding outcome (Table 5). These results remained The sample was limited to women who initiated breastfeeding and intended to breastfeed for
stable even after controlling for various demographic, ⬎2 months, with complete data on all variables.
a Significant differences between 2 groups of breastfeeding duration: P ⬍ .05.
behavioral, and attitudinal variables.
Our analyses of the association between the number
of “Baby-Friendly” hospital practices experienced (0 – 6)
and early breastfeeding termination suggested that who intended to breastfeed for at least 2 months, one
women experiencing fewer practices were more likely to third of them had stopped breastfeeding before 6 weeks
terminate breastfeeding before 6 weeks. Among women if they had experienced none of the “Baby-Friendly”

PEDIATRICS Volume 122, Supplement 2, October 2008 S45

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
 TABLE 3 Prevalence of Stopping Breastfeeding Before 6 Weeks According to Reported “Baby-Friendly”
Hospital Practice (N ⴝ 1907)
“Baby-Friendly” Practice n Breastfed for OR (95% CI)a
⬍6 wk, %
Step 4: breastfeeding initiation within 1 h
Yes 1137 10.9 0.51 (0.39–0.65)b
No 770 19.5 1.00
Step 6: only breast milk given
Yes 982 8.5 0.35 (0.27–0.47)b
No 925 20.7 1.00
Step 7: rooming in
Yes 1082 13.4 0.83 (0.65–1.08)
No 825 15.6 1.00
Step 8: breastfeeding on demand
Yes 1265 12.5 0.65 (0.50–0.84)b
No 642 18.1 1.00
Step 9: no pacifiers given
Yes 843 10.2 0.53 (0.40–0.70)b
No 1064 17.7 1.00
Step 10: provide information on breastfeeding support
Yes 1373 13.6 0.79 (0.60–1.05)
No 534 16.5 1.00
a OR indicates odds ratio (odds of stopping breastfeeding before 6 weeks according to hospital practice); CI, confidence interval; no (did not

experience practice; 1.00), referent group.

b The odds ratio was significant at P ⬍ .05.

practices, compared with 3% of those who experienced hospital practices in influencing breastfeeding outcome.
all 6 “Baby-Friendly” practices (Fig 1). Compared with Among women who intended to breastfeed for at least 2
mothers experiencing all 6 “Baby-Friendly” practices, months, 14.4% breastfed for ⬍6 weeks. “Baby-Friendly”
mothers experiencing none were ⬃13 times more likely hospital practices that were consistently associated with
to stop breastfeeding early (Table 6). Adjusted analyses longer breastfeeding duration included breastfeeding
confirmed a clear dose-response relation even after con- initiation within 1 hour of birth, giving only breast milk,
trolling for a variety of demographic, behavioral, and and not giving any pacifiers, even after controlling for
attitudinal variables (Table 6). Compared with the ref- several relevant demographic, behavioral, and attitudi-
erence group of mothers who experienced all 6 of the nal variables. These results are similar to those found in
“Baby-Friendly” practices, mothers who experienced other studies12 and in the IFPS I, in which breastfeeding
fewer practices were more at risk for early breastfeeding initiation within 1 hour of birth and giving only breast
termination. milk were both significantly associated with longer
breastfeeding duration.10 The earlier survey only mea-
DISCUSSION sured 5 of the 10 “Baby-Friendly” steps; in contrast, the
The current study provides additional support for the current survey measured 6 of the 10 steps (those most
importance of certain “Baby-Friendly” maternity-care likely to be able to be reported on by the mother).

TABLE 4 Multivariate Models Predicting Breastfeeding for <6 Weeks According to Type of
“Baby-Friendly” Hospital Practice Experienced (N ⴝ 1907)
“Baby-Friendly” Practice Model 1, aOR (95% CI)a Model 2, aOR (95% CI)b Model 3, aOR (95% CI)c
Breastfeeding initiation within 1 h 0.63 (0.48–0.83) d 0.69 (0.52–0.91) d 0.71 (0.53–0.95)d
Only breast milk given 0.43 (0.32–0.58)d 0.44 (0.32–0.59)d 0.47 (0.34–0.64)d
Rooming in 1.08 (0.82–1.43) 0.90 (0.67–1.21) 0.93 (0.69–1.26)
Breastfeeding on demand 0.76 (0.58–1.00) 0.85 (0.64–1.14) 0.82 (0.61–1.10)
No pacifiers given 0.63 (0.48–0.84)d 0.68 (0.51–0.91)d 0.73 (0.54–0.99)d
Provide information on breastfeeding 0.87 (0.65–1.15) 0.90 (0.66–1.21) 0.97 (0.71–1.32)
support
a Adjusted odds ratio (aOR) and 95% Wald confidence interval (CI) based on logistic regression; analysis controlling for other “Baby-Friendly”

hospital practices.
b Adjusted odds ratio (aOR) and 95% Wald confidence interval (CI) based on logistic regression; analysis includes those variables in model 1 plus

child gender, maternal age, maternal race, maternal education, household income, marital status, number of previous children, and number of
cigarettes smoked.
c Adjusted odds ratio (aOR) and 95% Wald confidence interval (CI) based on logistic regression; analysis includes those variables in model 2 plus

number of friends and relatives who breastfed, mother’s prenatal intentions to work after birth, and prenatal attitudes toward breastfeeding (ie,
formula as good as breast milk; if child breastfed, less likely to be obese).
d The odds ratio was significant at P ⬍ .05.

S46 DIGIROLAMO et al
Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
 TABLE 5 Prevalence of Stopping Breastfeeding Before 6 Weeks According to Additional Hospital
Practices Reported (N ⴝ 1907)
Hospital Practice n Breastfed for OR (95% CI)a aOR (95% CI)b
⬍6 wk, %
No formula sample or coupon given in hospital
gift pack
Yes 309 12.3 0.81 (0.56–1.17) 0.89 (0.60–1.31)
No 1598 14.8 1.00 1.00
Infant brought to room at night for feeding if
not rooming inc
Yes 687 13.8 0.48 (0.31–0.75)d 0.48 (0.29–0.78)d
No 136 25.0 1.00 1.00
Type of delivery
Cesarean 518 14.3 0.99 (0.74–1.32) 0.97 (0.71–1.33)
Vaginal 1389 14.4 1.00 1.00
No pain medications given
Yes 317 7.6 0.44 (0.28–0.68)d 0.56 (0.35–0.89)d
No 1590 15.7 1.00 1.00
Labor support (doula)
Yes 57 10.5 0.69 (0.30–1.63) 0.97 (0.39–2.41)
No 1850 14.5 1.00
a OR indicates odds ratio (odds of stopping breastfeeding before 6 weeks by hospital practice); CI, confidence interval; no (did not experience

practice) or vaginal delivery, referent group.

b Adjusted odds ratio and 95% Wald confidence interval based on logistic regression; analysis controlling for child gender, maternal age, maternal

race, maternal education, household income, marital status, number of previous children, number of cigarettes smoked, number of friends and
relatives who breastfed, mother’s prenatal intentions to work after birth, and prenatal attitudes toward breastfeeding (ie, formula as good as
breast milk; if child breastfed, less likely to be obese).
c Total n ⫽ 823 because only asked of those mothers not rooming in with their children.

d The odds ratio was significant at P ⬍ .05.

Comparison of the current data with that collected in the measured in the IFPS I. This practice, however, did not
IFPS I administered 11 years ago10 shows that women seem to influence breastfeeding duration.
who intend to breastfeed are currently receiving more The current study also provides information on addi-
“Baby-Friendly” practices within hospitals than they tional practices, not included in the original 10 steps,
were in the past. A higher prevalence was found in the which were important in increasing breastfeeding dura-
most recent survey for each of the 5 practices measured tion. Mothers who did not receive any pain medication
in both studies, with the largest increases occurring for during labor or delivery were more likely to continue
rooming in (from 44.9% to 56.8%) and not giving pac- breastfeeding beyond 6 weeks, an association that re-
ifiers (from 31.4% to 44.2%). A large percentage of mained stable even after controlling for relevant demo-
mothers in the current study reported receiving infor- graphic, behavioral, and attitudinal variables. However,
mation on breastfeeding support (72%), a practice not type of delivery (ie, vaginal versus cesarean) did not
seem to influence breastfeeding duration. Others have
suggested that pain medication during delivery, specifi-
35 cally epidural analgesia, may be a barrier to breastfeed-
% of mothers who stopped breastfeeding

ing success.13,14 There is some evidence that pain medi-

30 30.0 cation may cross the placenta, making the infant drowsy
26.9
and decreasing the sucking reflexes.15 Infants of mothers
25
who were not medicated, therefore, were also not med-
21.5
20
before 6 wk

icated and perhaps were better able to suck effectively

15.5 on the breast. In addition, mothers may have been better
15
13.7 able to focus on and have positive experiences with
10 breastfeeding if they were not medicated during deliv-
6.2
3.2
ery. Those who perform future studies may want to
5 further explore some of the reasons behind the relation
between pain medication and decreased breastfeeding
0
duration.
0 1 2 3 4 5 6
No. of “Baby-Friendly” practices reported
If the mother and child were not rooming in, mothers
were more likely to continue feeding after 6 weeks if the
FIGURE 1
Among women who initiated breastfeeding and intended to breastfeed for ⬎2 months,
child was brought to the mother’s room for feeding at
percentage who stopped breastfeeding before 6 weeks according to the number of night. The “Baby-Friendly” practice of rooming in was
Baby-Friendly Hospital Initiative practices they experienced. not significantly associated with breastfeeding duration

PEDIATRICS Volume 122, Supplement 2, October 2008 S47

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
 TABLE 6 Multivariate Models for Breastfeeding for <6 Weeks According to Number of “Baby-Friendly”
Hospital Practices (N ⴝ 1907)
No. of Practices Prevalence: No. of Model 1, aOR (CI)a Model 2, aOR (CI)b Model 3, aOR
Experienced Practices Experienced, (CI)c
% (n)
0 1.6 (30) 12.86 (3.93–42.04)d 9.13 (2.65–31.48)d 7.60 (2.13–27.14)d
1 6.8 (130) 11.05 (4.18–29.20)d 8.95 (3.32–24.13)d 7.51 (2.75–20.51)d
2 15.9 (303) 8.19 (3.23–20.81)d 6.93 (2.69–17.85)d 6.41 (2.47–16.67)d
3 25.1 (478) 5.49 (2.18–13.85)d 4.68 (1.83–11.96)d 4.12 (1.60–10.62)d
4 24.8 (473) 4.78 (1.89–12.10)d 3.80 (1.48–9.75)d 3.81 (1.47–9.85)d
5 17.7 (338) 1.99 (0.74–5.37) 1.81 (0.66–4.96) 1.91 (0.69–5.30)
6 8.1 (155) 1.00 1.00 1.00
a Adjusted odds ratio (aOR) and 95% Wald confidence interval (CI) based on logistic regression; referent group, experienced all 6 “Baby-Friendly”

hospital practices.
b Adjusted odds ratio (aOR) and 95% Wald confidence interval (CI) based on logistic regression; analysis controlling for child gender, maternal

age, maternal race, maternal education, household income, marital status, number of previous children, and number of cigarettes smoked.
c Adjusted odds ratio (aOR) and 95% Wald confidence interval (CI) based on logistic regression; analysis includes those variables in model 2 plus

number of friends and relatives who breastfed, mother’s prenatal intentions to work after birth, and prenatal attitudes toward breastfeeding (ie,
formula as good as breast milk; if child breastfed, less likely to be obese).
d The odds ratio was significant at P ⬍ .05.

(Table 3); however, most of those mothers not rooming third had stopped breastfeeding before 6 weeks if they
in had their child brought to them at night for feeding had experienced none of the “Baby-Friendly” practices,
(Table 5), which may have affected these results. It compared with only 3% of those who experienced all 6
seems that the critical practice here may be ensuring that “Baby-Friendly” practices. Replicating a dose-response
the infant breastfeeds at night rather than rooming in, relationship a second time further confirms the impor-
per se. Given that rooming in (step 7) may not be pos- tance of encouraging hospitals to adopt as many of these
sible in all situations, and that what seems important is practices as possible. This notion was emphasized by
having the infant with the mother at night for feedings, Radford and Southall in their Pediatrics commentary,19
this may be an important addition to step 7 for hospitals in which they acknowledged that adopting “Baby-
to keep in mind. Friendly” practices within the hospital takes time and
We did not find an association between not providing cited additional evidence for a dose-response effect ob-
a formula sample or coupon in a hospital gift pack and served during different stages of adoption of the 10 steps.
increased breastfeeding duration. These results are con- Given that in the current study a very small percentage
sistent with a Cochrane review in 2000,16 which sug- (8.1%) of women reported experiencing all 6 “Baby-
gested that there was no evidence to support a signifi- Friendly” practices measured, the need still exists to
cant effect of commercial hospital discharge packs (with increase efforts aimed at implementing these strategies
formula or promotional materials) on early termination within the hospital environment. As of August 2007,
of nonexclusive breastfeeding. However, evidence from there were only 59 “Baby-Friendly” designated hospitals
this review did suggest that hospital discharge packs and birth centers in the United States,20 further suggest-
seem to reduce the number of women exclusively ing the need to facilitate adoption of these practices
breastfeeding at 6 and 13 weeks postpartum when com- within the United States.
pared with no intervention or a noncommercial pack in This study has several limitations. First, results are
which samples of infant formula had been removed. based on mothers’ reports of their experiences within
Researchers who perform future analyses may want to the hospital, which addressed only 1 aspect of “Baby-
examine the effects of this and other hospital practices Friendly” practices: the mother’s perception. Mothers’
on duration of exclusive breastfeeding to determine if perceptions of their experiences in the hospital may
different results are found. prove to be somewhat different from the hospitals’ per-
Several studies have reported positive breastfeeding ceptions of the practices they support. Second, we did
outcomes among women who give birth in hospitals not have data on the first 3 steps related to policies
that endorse the 10 steps.17,18 Although support was not established within the hospital or the degree to which
statistically significant for all of the individual practices the hospital staff adhered to these practices and recom-
measured in this study, evidence was found for the mendations. A survey aimed at obtaining hospital and
cumulative effects of experiencing these practices on birth center reports of maternity-care practices is cur-
breastfeeding duration. As found in the IFPS I,10 women rently being implemented, which may provide addi-
who experienced fewer “Baby-Friendly” practices were tional information on the prevalence of these practices
more likely to cease breastfeeding by 6 weeks with a within US hospitals.21 Researchers who perform studies
dose-related response relationship, which suggests that in the future may want to obtain information from both
the cumulative effect of the practices, rather than each mothers and hospital staff to obtain a full sense of what
individual practice, is important. Even among mothers may be happening within the hospital setting. Third, the
who intended to breastfeed for at least 2 months, one way in which the practices were measured may not fully

S48 DIGIROLAMO et al
Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
capture what is intended in the 10 Steps to Successful and Policy Recommendations [joint publication of Pan American
Breastfeeding, and wording of the actual items and mea- Health Organization, World Bank, and Tropical Metabolism
surement of the steps may affect the nature of the re- Research Unit (University of the West Indies)]. Washington,
sults. However, the fact that our results are fairly con- DC: PAHO; 1998:43– 68
sistent with those found in the literature and with the 3. Centers for Disease Control and Prevention. National Immuniza-
tion Survey 2004. Available at: www.cdc.gov/breastfeeding/data/
IFPS I lends additional support to the validity of the
NIS㛭data/data㛭2004.htm. Accessed February 19, 2008
findings. Fourth, the associations between hospital prac-
4. US Department of Health and Human Services. Healthy People
tices and early breastfeeding termination may have been 2010: Understanding and Improving Health. 2nd ed. Washington,
affected by other factors about the women that were not DC: US Government Printing Office; 2000
controlled for in this study. This also may have affected 5. World Health Organization/United Nations Children’s Fund.
the magnitude of the effects shown. Finally, the sample Protecting, Promoting and Supporting Breastfeeding: The Special Role
may not be nationally representative, because the results of Maternity Services. Geneva, Switzerland: World Health
are based on information from a consumer mail survey, Organization; 1989
with the sample likely underrepresenting mothers from 6. World Health Organization. Evidence for the Ten Steps to Successful
a lower socioeconomic status. Breastfeeding. Geneva, Switzerland: World Health Organization;
1998
CONCLUSIONS 7. Lazarov M, Feldman A. WHO-UNICEF Ten Steps to Successful
Breastfeeding. Washington, DC: US Committee for the United
Our study provides a strong case for the importance of
Nations Children’s Fund: 1995
promoting certain maternity-care practices to move
8. Powers NG, Naylor AJ, Wester RA. Hospital policies: crucial to
closer to the goals for breastfeeding duration. In addi- breastfeeding success. Semin Perinatol. 1994;18(6):517–524
tion, the results illustrate the need to work with hospi- 9. Saadeh R, Akre J. Ten steps to successful breastfeeding: a
tals to increase their implementation of “Baby-Friendly” summary of the rationale and scientific evidence. Birth. 1996;
practices, with the goal of including as many of these 23(3):154 –160
practices as possible. The results reinforce recommenda- 10. DiGirolamo A, Grummer-Strawn L, Fein S. Maternity care
tions made by us in a previous article10 to help hospitals practices: implications for breastfeeding. Birth. 2001;28(2):
implement these practices, even if gradually, and to 94 –100
reinforce their use rather than approaching these goals 11. Fein SB, Labiner-Wolfe J, Shealy KR, Li R, Chen J, Grummer-
as an all-or-none endeavor. These recommendations are Strawn LM. Infant Feeding Practices Study II: study methods.
further supported by the commentary by Radford and Pediatrics. 2008;122(suppl 2):S28 –S35
Southall19 on lessons learned for successful application of 12. World Health Organization. Evidence for the Ten Steps to Successful
the Baby-Friendly Hospital Initiative. More attention Breastfeeding. Geneva, Switzerland: World Health Organization;
1998
should also be given to other practices shown to influ-
13. Crowell M, Hill P, Humenich S. Relationship between obstetric
ence breastfeeding duration that may not currently be
analgesia and time of effective breast feeding. J Nurse Midwifery.
represented in the original 10 steps (eg, bring the infant 1994;39(3):150 –156
to the room at night for feeding if not rooming in; avoid 14. Righard L, Alade M. Sucking technique and its success in
giving pain medications during delivery if possible). Fu- breastfeeding. Birth. 1992;19(4):185–189
ture studies should assess whether a similar pattern of 15. Radzyminski S. The effect of ultra low dose epidural analgesia
results is found with mothers more at risk for early on newborn breastfeeding behaviors. J Obstet Gynecol Neonatal
breastfeeding termination (eg, less-educated mothers, Nurs. 2003;32(3):322–331
mothers of lower socioeconomic status) and should at- 16. Donnelly A, Snowden HM, Renfrew MJ, Woolridge MW.
tempt to obtain in-depth information on all 10 of the Commercial hospital discharge packs for breastfeeding women.
“Baby-Friendly” steps. Cochrane Database Syst Rev. 2000;(2):CD002075
17. Broadfoot M, Britten J, Tappin D, MacKenzie J. The Baby
Friendly Hospital Initiative and breastfeeding rates in Scotland.
ACKNOWLEDGMENTS
Arch Dis Child Fetal Neonatal Ed. 2005;90(2):F114 –F116
This study was funded by the Food and Drug Adminis- 18. Merten S, Dravta J, Ackerman-Liebrich U. Do Baby-Friendly
tration, Centers for Disease Control and Prevention, Of- hospitals influence breastfeeding duration on a national level?
fice of Women’s Health, National Institutes of Health, Pediatrics. 2005;116(5). Available at: www.pediatrics.org/cgi/
and Maternal and Child Health Bureau in the US De- content/full/116/5/e702
partment of Health and Human Services. 19. Radford A, Southall DP. Successful application of the Baby-
Friendly Hospital Initiative contains lessons that must be ap-
REFERENCES plied to the control of formula feeding in hospitals in industri-
1. American Academy of Pediatrics, Work Group on Breastfeed- alized countries. Pediatrics. 2001;108(3):766 –768
ing. Breastfeeding and the use of human milk. Pediatrics. 1997; 20. Baby-Friendly USA. Available at: www.babyfriendlyusa.org/
100(6):1035–1039 eng. Accessed February 19, 2008
2. Andraca I, Peirano P, Uauy R. Nutrition and care in the preterm 21. Centers for Disease Control and Prevention. CDC National
and neonatal periods and later development: human milk is Survey of Maternity Care Practices in Infant Nutrition and Care
best for optimal mental development. In: Grantham-McGregor (mPINC). Available at: www.mpincsurvey.com:80. Accessed
SM. Nutrition, Health, and Child Development: Research Advances March 3, 2008

PEDIATRICS Volume 122, Supplement 2, October 2008 S49

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020
 Effect of Maternity-Care Practices on Breastfeeding
Ann M. DiGirolamo, Laurence M. Grummer-Strawn and Sara B. Fein
Pediatrics 2008;122;S43
DOI: 10.1542/peds.2008-1315e

Updated Information & including high resolution figures, can be found at:
Services http://pediatrics.aappublications.org/content/122/Supplement_2/S43
References This article cites 10 articles, 4 of which you can access for free at:
http://pediatrics.aappublications.org/content/122/Supplement_2/S43#
BIBL
Subspecialty Collections This article, along with others on similar topics, appears in the
following collection(s):
Nutrition
http://www.aappublications.org/cgi/collection/nutrition_sub
Breastfeeding
http://www.aappublications.org/cgi/collection/breastfeeding_sub
Permissions & Licensing Information about reproducing this article in parts (figures, tables) or
in its entirety can be found online at:
http://www.aappublications.org/site/misc/Permissions.xhtml
Reprints Information about ordering reprints can be found online:
http://www.aappublications.org/site/misc/reprints.xhtml

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020

 Effect of Maternity-Care Practices on Breastfeeding
Ann M. DiGirolamo, Laurence M. Grummer-Strawn and Sara B. Fein
Pediatrics 2008;122;S43
DOI: 10.1542/peds.2008-1315e

The online version of this article, along with updated information and services, is
located on the World Wide Web at:
http://pediatrics.aappublications.org/content/122/Supplement_2/S43

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, 345 Park Avenue, Itasca, Illinois, 60143. Copyright © 2008
by the American Academy of Pediatrics. All rights reserved. Print ISSN: 1073-0397.

Downloaded from www.aappublications.org/news at Indonesia:AAP Sponsored on November 7, 2020