Best, best
bs_bs_banner
OR I G I N A L A R T I C L E
Predictors of breastfeeding non‐initiation in the NICU
Brooke Gertz | Emily DeFranco
OB/GYN Department, University of Cincinnati,
Cincinnati OH, USA
Correspondence
Brooke Gertz, Medical Sciences Building,
Room 4407 231 Albert Sabin Way Cincinnati,
OH 45267‐0526, USA. Email: brown3b8@mail.uc.edu
Funding information
Cincinnati Children's Hospital Medical Center
Perinatal Institute; March of Dimes Founda-
tion, Grant/Award Number: 22‐FY14‐470
Abstract
This study compared predictors of breastfeeding non‐initiation between infants who
were and were not admitted to the NICU so that interventions can target high‐risk
mothers whose infants desperately need breastmilk. This was a population‐based ret-
rospective cohort study of singleton Ohio live births using birth certificates, 2006–
2015. In babies who were and were not admitted to the NICU, a multivariable logistic
regression model assessed the association between breastfeeding non‐initiation and
predictors relating to the mother, neonate, and labour and delivery events while
adjusting for covariables. Of 1,463,506 births, 76,855 infants were admitted to the
NICU (5.8% of study population), and breastfeeding was not initiated in 39.4% of
them, compared with 31.5% of infants in the newborn nursery, p < 0.001. Apart from
abnormal newborn conditions, smoking during pregnancy was the most significant
risk factor for not breastfeeding in the NICU (RR 1.91 [95% CI 1.82–2.02]) and new-
born nursery (RR 2.10 [95% CI 2.08–2.13]), followed by socioeconomic factors and
multiparity. Limited prenatal visits (≤5) were a significantly higher risk factor in the
NICU (RR 1.41 [95% CI 1.34–1.49]) than in the newborn nursery (RR 1.24 [95% CI
1.22–1.26]). Intentional home birth and use of infertility treatment were associated
with breastfeeding initiation. The rate of breastfeeding initiation is lower in infants
admitted to the NICU than those who are not, especially among mothers with limited
prenatal care. Interventions should target mothers who smoke because they are least
likely to breastfeed, and their babies, who are prone to serious health conditions,
could especially benefit from breastmilk.
KEYWORDS
breastfeeding, breastfeeding initiation, breast milk, NICU, prenatal care, smoking
1 | INTRODUCTION
Because of increased awareness of the benefits premature infants
receive from breastfeeding that can ultimately lower hospital costs,
recent studies have sought to identify populations less likely to
breastfeed to target educational interventions and support efforts
(Johnson, Patel, Bigger, Engstrom, &Meier, 2014). In 2011, the Surgeon
General announced a call to action to support breastfeeding (Office of
the Surgeon General, 2011). During the 10‐year study period,
breastfeeding rates significantly increased in both the wellborn and
NICU branches of the study cohort. However, breastfeeding rates in
theNICU only climbed to 56%.We can do better, especially considering
the unique value of breastfeeding for mother and child. Some of the
immediate health benefits of breastfeeding for the infant include
decreased rates of respiratory infections, otitis media, and gastrointes-
tinal infections due to passive immunity transferred from the mother's
milk (Bartick et al., 2017; Leon‐Cava, Lutter, Ross, & Martin, 2002;
Victora et al., 2016). Infants who are breastfed also have a lower risk
of sudden infant death syndrome (Bartick et al., 2017). Breastfeeding
is also associated with maternal health benefits. When mothers initiate
breastfeeding immediately after delivery, it facilitates uterine contrac-
tion and expulsion of the placenta, which reduce postpartum bleeding
Received: 8 July 2018 Revised: 23 January 2019 Accepted: 28 January 2019
DOI: 10.1111/mcn.12797
Matern Child Nutr. 2019;15:e12797.
https://doi.org/10.1111/mcn.12797
© 2019 John Wiley & Sons Ltdwileyonlinelibrary.com/journal/mcn 1 of 12
(Leon‐Cava et al., 2002). Long‐term maternal benefits include lower
rates of ovarian cancer, premenopausal breast cancer, hypertension,
diabetes, and myocardial infarction (Bartick et al., 2017; Leon‐Cava
et al., 2002; Victora et al., 2016). Additionally, breastfeeding helps to
naturally space births, which reduces the pregnancy risks associated
with short birth intervals (Bartick et al., 2017; Leon‐Cava et al., 2002).
Breastfeeding is especially important for babies admitted to the
NICU, who have susceptible immune systems and specific nutritional
needs. Among this especially vulnerable population, breastfeeding
has been associated with lower risks of retinal detachment and necro-
tizing enterocolitis as well as improved neurodevelopmental outcomes
(Bharwani et al., 2016; Isaacs et al., 2010; Maffei & Schanler, 2017;
Okamoto et al., 2007; Sisk, Lovelady, Dillard, Gruber, & O'Shea,
2007; Vohr et al., 2006).
Despite these findings, breastfeeding rates are significantly lower
in developed countries such as the United States than in low‐income
countries (Victora et al., 2016), and in many U.S. hospitals, women
have been even less likely to breastfeed if their infant is admitted to
the NICU (Callen & Pinelli, 2005; Kachoria & Oza‐Frank, 2015). This
study aims to identify factors that influence the rate of breastfeeding
non‐initiation in babies admitted to the NICU across all hospitals in
Ohio to estimate U.S. trends. Including data from all hospitals in Ohio
ensures a large and socioeconomically diverse sample that can provide
insight into NICU breastfeeding initiation trends across the United
States despite birth certificate variability that prevents the execution
of a more comprehensive, state‐wide study. Breastfeeding data from
the CDC National Immunization Survey estimate breastfeeding rates
by state and confirm that the breastfeeding rates in Ohio approximate
the national average during the study interval, although these data
likely overestimates national breastfeeding rates by relying on tele-
phone sampling, information recall, and self‐reported data (National
Center for Chronic Disease Prevention and Health Promotion, 2018).
The study objective is to identify a population of mothers who may
particularly benefit from breastfeeding education and support inter-
ventions in a hospital environment with accessible breastfeeding
resources already in place. Furthermore, identifying risk factors for
not breastfeeding can help tailor intervention programs to best reach
as many mothers as possible. For infants in the NICU, any intervention
that has the potential to improve their health outcomes is extremely
important. Because breastfeeding has already been proven to posi-
tively impact infants in numerous ways, it is critical that more focus
is placed on advocating for breastfeeding in the NICU. This study will
identify if there are any maternal risk factors for breastfeeding non‐
initiation that are different between the NICU and newborn nursery
and will thus need additional attention in the NICU setting.
2 | METHODS
This was a population‐based retrospective cohort study of all live
births in Ohio from 2006 to 2015 based on deidentified vital statistics
birth certificate records from the Ohio Department of Health. The
protocol for this study was approved by the Ohio Department of
Health human subjects institutional review board. This study was
exempt from review by the institutional review board at the University
of Cincinnati, Cincinnati, Ohio.
2.1 | Sample
Births were excluded from the study prior to analysis if the infant was
not living or if data were missing about whether the infant was living
at the time the birth certificate was generated. Infants born before
23 weeks gestational age, and infants born with anencephaly were
excluded from study because these conditions are considered not
compatible with life and typically result in mortality within hours of
birth, thus prohibiting assessment of the outcome, breastfeeding initi-
ation. Multiple gestation births were also excluded from the study to
prevent double counting of an individual mother and to eliminate a
potential confounding variable. Births to Ohio resident mothers out-
side of Ohio were not included in the study due to interstate birth cer-
tificate variability. Analyses were limited to births with complete
breastfeeding and NICU admission data.
Key messages
• Although ill infants in the NICU could especially benefit
from breastmilk, breastfeeding initiation rates remain
lower in the NICU than in the newborn nursery. The
NICU is especially well suited for educational
interventions because mothers are in frequent contact
with the healthcare staff and oftentimes have access to
lactation consultants and other breastfeeding resources.
To better utilize this opportunity to promote
breastfeeding, hospitals should be aware of the
population of mothers who are unlikely to breastfeed so
that educational and support interventions can be
tailored to help overcome the barriers preventing that
population from breastfeeding.
• Breastfeeding intervention efforts should target mothers
who smoke because not only are they least likely to
breastfeed but also their babies are prone to serious
health conditions. Aside from anomalies and abnormal
newborn conditions, smoking during pregnancy is the
most significant risk factor for not breastfeeding in the
NICU and the newborn nursery.
• Hospitals should identify mothers with low educational
attainment and limited prenatal care so they can be
educated perhaps for the first time about the benefits
of breastfeeding. Limited prenatal care is a significantly
higher risk factor for not breastfeeding in the NICU
than in the newborn nursery. Gastroschisis and
extremely low birth weight are protective for
breastfeeding, likely because physicians advocate for
these patients in particular, which show that education
can protect infants who need breastmilk.
2 of 12 GERTZ AND DEFRANCO bs_bs_banner
2.2 | NICU admission and breastfeeding
The first part of the study analysed the cohort of infants admitted to
the NICU. On the birth certificate, “NICU admission” is an option
under the category “abnormal conditions of the newborn” that is
checked if it applies to the infant. The primary outcome was no
breastfeeding initiation, which is addressed on the birth certificate
by the question, “Is the infant being breastfed at discharge? (yes/
no)” (National Center for Health Statistics, 2003). This question does
not gather information about whether the breastfeeding was exclusive
or supplemented or whether it involved direct breastfeeding, pumped
milk, or donor milk. There are no data about breastfeeding duration
after hospital discharge. Therefore, this study analyses initiation of
any of the above forms of breastfeeding within the first 48 hr postpar-
tum, when the birth certificates are typically generated.
2.3 | Measures
The predictors of breastfeeding non‐initiation fell into three broad cat-
egories relating to the mother, the neonate, and labour and delivery
events. The predictors of breastfeeding non‐initiation fell into three
broad categories relating to the mother, the neonate, and labour and
delivery events, and they are enumerated in the Supporting Information
section. The maternal variables that were analysed included age,
race/ethnicity, county of residence, completed education level, Medic-
aid recipience, marital status, limited prenatal care, smoking during
pregnancy, prepregnancy weight, parity, pregnancy interval, previous
preterm birth, previous C‐section, prepregnancy or gestational diabe-
tes, chronic or gestational hypertension, eclampsia, infection during
pregnancy, and use of infertility treatment. Counties of maternal resi-
dence were categorized as urban, suburban, or rural according to the
2013 rural–urban continuum code made by the United States Depart-
ment of Agriculture (Economic Research Service, 2016). Metro
counties ranked 1–3 by the continuum code were considered urban;
nonmetro counties ranked 4–6 were considered suburban, and non-
metro counties ranked 7–9 were considered rural. Infections of
interest included gonorrhoea, syphilis, chlamydia, hepatitis B, or hepati-
tis C. Variables related to labour and delivery events included spinal or
epidural anaesthesia administration, onset of labour, induction or aug-
mentation of labour, tocolysis, cervical cerclage, antibiotic administra-
tion, maternal complication, method of delivery, and place where birth
occurred. Maternal complications included maternal transfusion, third
or fourth degree perineal laceration, ruptured uterus, unplanned
hysterectomy, admission to intensive care unit, or unplanned operating
room procedure following delivery. The neonatal variables analysed
were birthweight, gestational age at delivery, gender, low Apgar score,
assisted ventilation, surfactant replacement therapy, antibiotic adminis-
tration, seizure or serious neurological dysfunction, significant birth
injury, and various congenital anomalies. Significant birth injuries
included skeletal fracture(s), peripheral nerve injury, or soft tissue/solid
organ haemorrhage, which required intervention, as recorded in the
certificate of live birth. All births during the study period utilized the
2003 version of the U.S. birth certificate (National Center for Health
Statistics, 2003). Standardized guidelines for completing the birth
certificate are based on the National Vital Statistics Guide to
Completing the Facility Worksheets for the Certificate of Live Birth
and Report of Fetal Death (National Center for Health Statistics, 2016).
2.4 | Statistical analysis
Statistical analyses were performed using STATA v15.0 software
(StataCorp, College Station, TX). Demographic characteristics were com-
pared between neonates who were breastfed or not using the χ2 or t
test. Associations between the above predictors and breastfeeding
non‐initiation were computed using a multivariable logistic regression
model. Factors that were highly predictive of not breastfeeding based
on statistical significance in univariate comparisons were included in a
multivariable logistic regression model to adjust for potential confound-
ing risks. These covariables included maternal age, race, education, Med-
icaid recipience, marital status, prenatal care, smoking habits, BMI, parity,
presence of infection during pregnancy, and use of infertility treatment,
as well as infant birth weight and place where birth occurred. Factors
were removed from the multivariable logistic regression model via back-
ward selection if they no longer had significance when added to the
model. Birth interval was not included in the adjusted model because of
collinearity with parity, and gestational age was not included in the
adjusted model because of collinearity with infant birth weight (birth
weight was a stronger and more accurate measure than gestational
age). Abnormal conditions of the newborn and congenital anomalies
were not included in the adjusted model because of variability in the
ability of the infant to receive nutrition by mouth. However, crude
relative risks of breastfeeding non‐initiation were calculated for these
conditions. Results of the statistical analyses are reported as crude and
adjusted relative risks ratios with 95% confidence intervals.
We then compared predictors of breastfeeding non‐initiation in an
exposure group of babies admitted to the NICU with a reference
group of babies who remained in the newborn nursery in a similar
fashion. Predictors were considered significantly different between
the exposure and reference groups if there was no overlap between
the confidence intervals for each of the relative risks.
3 | RESULTS
The total number of Ohio live births from 2006 to 2015 was
1,463,506. After exclusions, 1,317,535 births were analysed in this
study. There were minimal missing data (<5%) for all variables except
use of infertility treatment (30.3%) and spinal or epidural anaesthesia
administration (10.1%). As shown in Figure 1, 76,855 (5.8%) infants
were admitted to the NICU, and 1,240,680 (94.2%) infants were not.
Breastfeeding was not initiated in 390,972 (31.5%) infants in the ref-
erence group (no NICU admission), compared with 30,283 (39.4%)
infants in the exposure group (NICU admission), p < 0.001.
Most mothers in the exposure and reference groups were non‐
smoking multiparas between 20 and 34 years of age with at least
some college education. The majority (76.1%) of mothers were non‐
GERTZ AND DEFRANCO 3 of 12 bs_bs_banner
Hispanic White, and the largest minority (15.9%) were non‐Hispanic
Black. Within this socioeconomically diverse population, almost four
in every 10 women received Medicaid, and over four in every 10
women were unmarried. Additional characteristics of the study popu-
lation are shown in Tables 1, 2, and 3.
Within the adjusted model, the greatest risk factors for breast-
feeding non‐initiation in the NICU were smoking during pregnancy
(aRR 1.91 [95% CI 1.82–2.02]), multiparity (aRR 1.72 [95% CI 1.64–
1.79]), being unmarried (aRR 1.68 [95%CI 1.60–1.77]), Medicaid
recipience (aRR 1.51 [95% CI 1.44–1.58]), no high school diploma
(aRR 1.48 [95% CI 1.40–1.57]), and ≤5 prenatal visits (aRR 1.41
[95% CI 1.34–1.49]). Other maternal risk factors for breastfeeding
non‐initiation in the NICU were infection present and/or treated dur-
ing pregnancy (aRR 1.32 [95% CI 1.21–1.44]), young maternal age
<20 years (aRR 1.30 [95% CI 1.21–1.40]), and maternal obesity with
prepregnancy BMI ≥40.0 (aRR 1.26 [95% CI 1.16–1.37]). The most
significant risk factors for breastfeeding non‐initiation are represented
by the adjusted model in Table 4.
Limited prenatal visits (≤5) were a significantly higher risk factor
for breastfeeding non‐initiation in the NICU (aRR 1.41 [95% CI
FIGURE 1 Flow diagram of study cohort
4 of 12 GERTZ AND DEFRANCO bs_bs_banner
TABLE 1 Maternal characteristics: exposure and reference groups
Maternal characteristics Newborn nursery NICU admission p value
Demographic
Age (year): <0.001
<20 112,604 (9.1%) 7,908 (10.3%)
20–34 980,281 (79.0%) 58,542 (76.2%)
≥35 147,753 (11.9%) 10,399 (13.5%)
Race: <0.001
Non‐Hispanic White 950,130 (76.6%) 53,098 (69.1%)
Non‐Hispanic Black 191,247 (15.4%) 17,786 (23.1%)
Hispanic 57,021 (4.6%) 3,460 (4.5%)
Other 33,827 (2.7%) 1,855 (2.4%)
Socioeconomic
County of residence:a <0.001
Urban 960,485 (77.4%) 62,621 (81.5%)
Suburban 225,796 (18.2%) 10,393 (13.5%)
Rural 25,285 (2.0%) 862 (1.1%)
Out‐of‐state 20,699 (1.7%) 2,188 (2.9%)
Education level: <0.001
No high school diploma 193,090 (15.6%) 14,178 (18.5%)
High school diploma 315,845 (25.5%) 21,362 (27.8%)
Some college 274,698 (22.1%) 17,944 (23.4%)
College degree 449,945 (36.3%) 22,669 (29.5%)
Medicaid 463,573 (37.4%) 35,009 (45.6%) <0.001
Unmarried 521,875 (42.1%) 40,305 (52.4%) <0.001
Prenatal care
Limited prenatal visits (≤5) 91,844 (7.4%) 13,295 (17.3%) <0.001
Smoked during pregnancy 217,165 (17.5%) 17,533 (22.8%) <0.001
Maternal health indicators
Prepregnancy BMI: <0.001
<18.5 (underweight) 50,051 (4.0%) 3,643 (4.7%)
18.5–24.9 (normal) 566,636 (45.7%) 31,757 (41.3%)
25.0–29.9 (overweight) 288,524 (23.3%) 17,182 (22.4%)
30.0–39.9 (obese) 229,123 (18.5%) 15,660 (20.4%)
≥40.0 (morbidly obese) 56,244 (4.5%) 4,933 (6.4%)
Parity: <0.001
Primiparous 484,865 (39.1%) 34,075 (44.3%)
Multiparous 742,463 (59.8%) 41,922 (54.6%)
Birth interval (months): <0.001
<18 81,809 (6.6%) 5,867 (7.6%)
18–23 101,754 (8.2%) 4,809 (6.3%)
24–59 349,527 (28.2%) 17,041 (22.2%)
59–71 43,970 (3.5%) 2,616 (3.4%)
>72 106,779 (8.6%) 7,687 (10.0%)
Previous preterm birth 44,648 (3.6%) 8,039 (10.5%) <0.001
Previous C‐section 162,307 (13.1%) 13,252 (17.2%) <0.001
(Continues)
GERTZ AND DEFRANCO 5 of 12 bs_bs_banner
TABLE 1 (Continued)
Maternal characteristics Newborn nursery NICU admission p value
Diabetes:
Prepregnancy 8,489 (0.7%) 2,302 (3.0%) <0.001
Gestational 69,817 (5.6%) 6,777 (8.8%) <0.001
Hypertension:
Chronic 22,767 (1.8%) 4,057 (5.3%) <0.001
Gestational 57,934 (4.7%) 8,774 (11.4%) <0.001
Eclampsia 5,302 (0.4%) 2,372 (3.1%) <0.001
Infection present and/or
treated during pregnancy:b 40,583 (3.3%) 4,682 (6.1%) <0.001
Current pregnancy result
of infertility treatment
9,614 (0.8%) 838 (1.1%) <0.001
aCounty of residence info provided by United States Department of Agriculture. bInfections included gonorrhoea, syphilis, chlamydia, hepatitis B, or hepatitis C.
There was minimal missing data, which are not represented here but were part of the study calculations.
TABLE 2 Labour and delivery event characteristics: exposure and reference groups
Labour and delivery events Newborn nursery NICU admission p value
Spinal or epidural anaesthesia 832,495 (67.10%) 53,325 (69.4%) <0.001
Onset of labour: <0.001
PROM 30,912 (2.5%) 12,079 (15.7%)
Precipitous (without PROM) 39,969 (3.2%) 2,638 (3.4%)
3–20 hr (without PROM) 1,155,020 (93.1%) 60,808 (79.1%)
Prolonged (without PROM) 14,743 (1.2%) 1,328 (1.7%)
Induction of labour 407,776 (32.9%) 17,312 (22.5%) <0.001
Augmentation of labour 315,482 (25.4%) 15,416 (20.1%) <0.001
Tocolysis 21,246 (1.7%) 5,707 (7.4%)
Cervical cerclage 3,061 (0.3%) 831 (1.1%)
Antibiotics administered 336,779 (27.1%) 37,144 (48.3%) <0.001
Maternal complication:a 32,363 (2.6%) 3,948 (5.1%)
Method of delivery: <0.001
Vaginal 896,015 (72.2%) 39,085 (50.9%)
Caesarean 344,013 (27.7%) 37,724 (49.1%)
Place where birth occurred: <0.001
Hospital 1,222,710 (98.6%) 76,476 (99.5%)
Freestanding birth centre 5,204 (0.4%) 20 (0.03%)
Home (intended) 10,310 (0.8%) 68 (0.09%)
Home (not intended or intention unknown) 1,427 (0.1%) 192 (0.3%)
Clinic/doctor's office 301 (0.02%) 30 (0.04%)
Other 722 (0.06%) 69 (0.09%)
Note. PROM: premature rupture of membranes. aMaternal complications during labour and delivery included maternal transfusion, third or fourth degree perineal laceration, ruptured uterus, unplanned
hysterectomy, admission to intensive care unit, or unplanned operating room procedure following delivery.
6 of 12 GERTZ AND DEFRANCO bs_bs_banner
1.34–1.49]) than in the newborn nursery (aRR 1.24 [95% CI 1.22–
1.26]). Women who smoked during pregnancy were also at the
greatest risk of breastfeeding non‐initiation in the newborn nursery
(aRR 2.10 [95% CI 2.08–2.13]). Crude relative risks for the reference
group are reported in the supporting information section as well as
the crude results for the exposure group for comparison.
4 | DISCUSSION
In this contemporary cohort of live births, we found that significantly
fewer mothers initiated breastfeeding for their newborns if they were
admitted to NICU. There is a paucity of information about national rates
of breastfeeding initiation for ill newborns admitted to the NICU, but
TABLE 3 Neonatal characteristics: exposure and reference groups
Neonatal characteristics Newborn nursery NICU admission p value
Birthweight (grams): <0.001
<1,000 (ELBW) 406 (0.03%) 3,910 (5.1%)
1,000–1,499 (VLBW) 1,150 (0.09%) 5,949 (7.7%)
1,500–2,499 (LBW) 45,423 (3.7%) 24,456 (31.8%)
2,500–3,999 (NBW) 1,085,257 (87.5%) 38,296 (49.8%)
≥4,000 (HBW) 107,987 (8.7%) 3,978 (5.2%)
Gestational age at delivery (weeks): <0.001
23–27 (extremely preterm) 2,062 (0.2%) 4,000 (5.2%)
28–31 (very preterm) 5,801 (0.5%) 7,536 (9.8%)
32–36 (moderate to late preterm) 82,227 (6.6%) 28,587 (37.2%)
≥37 weeks (term) 1,149,566 (92.7%) 36,542 (47.6%)
Gender: <0.001
Male 630,822 (50.8%) 43,038 (56.0%)
Female 609,856 (49.2%) 33,816 (44.0%)
Apgar score less than 7 (at 5 min)a 15,963 (1.3%) 10,778 (14.0%) <0.001
Abnormal conditions of the newborn
Assisted ventilation required 35,500 (2.9%) 21,616 (28.13%) <0.001
Surfactant replacement therapy 617 (0.1%) 3,860 (5.02%) <0.001
Antibiotics for suspected neonatal sepsis 5,998 (0.5%) 15,322 (19.9%) <0.001
Seizure or serious neurologic dysfunction 75 (0.01%) 259 (0.3%) <0.001
Significant birth injuryb 659 (0.05%) 172 (0.2%) <0.001
Congenital anomalies
Meningomyelocele/spina bifida 85 (0.01%) 227 (0.3%) <0.001
Cyanotic congenital heart disease 104 (0.01%) 654 (0.9%) <0.001
Congenital diaphragmatic hernia 32 (0.00%) 153 (0.2%) <0.001
Omphalocele 31 (0.00%) 126 (0.2%) <0.001
Gastroschisis 47 (0.00%) 480 (0.6%) <0.001
Limb reduction defectc 155 (0.01%) 66 (0.1%) <0.001
Cleft lip with or without cleft palate 566 (0.05%) 210 (0.3%) <0.001
Cleft palate alone 412 (0.03%) 219 (0.3%) <0.001
Down syndrome (confirmed or pending) 428 (0.03%) 382 (0.5%) <0.001
Hypospadias 848 (0.07%) 146 (0.2%) <0.001
Note. ELBW: extremely low birth weight; VLBW: very low birth weight; LBW: low birth weight; NBW: normal birth weight; HBW: high birth weight. aApgar score: score given to newborn immediately after delivery to determine its overall well‐being. Scale from 0 to 10; <7 indicates that the newborn needs medical attention. bSignificant birth injuries included skeletal fracture(s), peripheral nerve injury, or soft tissue/solid organ haemorrhage, which required intervention. cLimb reduction defect: excludes congenital amputation and dwarfing syndrome.
GERTZ AND DEFRANCO 7 of 12 bs_bs_banner
one study used data from 27 states to determine breastfeeding trends
in the NICU from 2000 to 2003. The study showed that although NICU
admission may have a positive influence on breastfeeding continuation,
mothers of NICU infants were less likely to initiate breastfeeding than
mothers of nonadmitted infants (Colaizy & Morriss, 2008). Although
the gap between breastfeeding non‐initiation in NICU‐admitted and
nonadmitted infants may be narrowing, our results show that
breastfeeding is still not initiated in almost four in 10 infants admitted
to the NICU (Kachoria & Oza‐Frank, 2015). For many mothers,
breastfeeding may not be a priority amidst the stress that comes along
with having a child in theNICU, and a stressful labour and delivery expe-
rience is a risk factor for delayed onset of lactogenesis (difficulty with
initial milk production; Bernaix, Schmidt, Jamerson, Seiter, & Smith,
2006; Dewey, Nommsen‐Rivers, Heinig, & Cohen, 2003).
Some studies have described racial, cultural, and socioeconomic
factors as predictors of not breastfeeding but have not specifically
examined predictors of breastfeeding non‐initiation among the sickest
infants who may benefit from breastmilk the most, infants admitted to
the NICU (Celi, Rich‐Edwards, Richardson, Kleinman, & Gillman, 2005;
Kelly, Watt, & Nazroo, 2006; McDowell, Wang, & Kennedy‐
Stephenson, 2008). Consistent with results for the general population
analysed in these studies, socioeconomically disadvantaged mothers
were significantly less likely to initiate breastfeeding in the NICU as
well. However, while non‐Hispanic Black race was a risk factor for
breastfeeding non‐initiation in the United States in these studies as
well as in the reference group of this study, it was not a risk factor
in the adjusted model for infants admitted to the NICU. In both the
reference and exposure groups of this study, smoking during preg-
nancy was the highest risk factor for breastfeeding non‐initiation,
even after adjustment for potential confounding variables. Interven-
tion efforts should particularly target mothers who smoke because
not only are they least likely to breastfeed but also their babies are
prone to serious health conditions and could thus especially benefit
from breastmilk. Mothers may be deterred from breastfeeding while
smoking because of the idea that it is harmful for the baby, which
may have some merit, according to studies that have shown that nic-
otine is several times more concentrated in the breastmilk of smokers
than plasma (Napierala, Mazela, Merritt, & Florek, 2016). However,
smoking during pregnancy also has negative effects on the developing
fetus, such as fetal growth restriction, but most mothers who begin
pregnancy smoking continue to smoke during pregnancy anyways
(Blatt, Moore, Chen, Van Hook, & DeFranco, 2015). One study found
that quitting smoking during pregnancy was associated with subse-
quent breastfeeding and that low‐income women who intended to
breastfeed were more likely to quit smoking during pregnancy
(Carswell et al., 2018). Thus, identifying mothers who smoke while
they are early in the pregnancy to support cessation and to educate
about breastfeeding benefits may have a positive impact on both life-
style choices. Part of the education process for women who smoke
may involve the medicinal use of donor human milk to meet the nutri-
tional needs of the infant when the mother is struggling to produce
enough of her own milk, which is a challenging feat for mothers of
TABLE 4 Adjusted relative risks
Newborn nursery NICU
Predictor of not breastfeeding aRR 95% CI aRR 95% CI
Maternal characteristics
Young age (<20 years) 1.54 1.51–1.57 1.30 1.21–1.40
Black race 1.07 1.06–1.09 0.94 0.89–1.00
No high school diploma 1.57 1.54–1.59 1.48 1.40–1.57
Medicaid recipient 1.69 1.67–1.71 1.51 1.44–1.58
Unmarried 1.98 1.95–2.01 1.68 1.60–1.77
Limited prenatal visits (≤5) 1.24 1.22–1.26 1.41 1.34–1.49
Smoked during pregnancy 2.10 2.08–2.13 1.91 1.82–2.02
Underweight BMI (<18.5) 1.14 1.12–1.17 1.09 0.99–1.20
Morbidly obese BMI (≥40.0) 1.40 1.37–1.44 1.26 1.16–1.37
Multiparous 1.81 1.79–1.84 1.72 1.64–1.79
Infection present and/or treated during pregnancy:a 1.17 1.14–1.21 1.32 1.21–1.44
Current pregnancy result of infertility treatment 0.77 0.72–0.81 0.88 0.74–1.06
Labour and delivery events
Intentional home birth 0.03 0.02–0.04 0.42 0.17–1.02
Neonatal characteristics
Extremely low birth weight (<1,000 g) 1.51 1.12–2.03 0.77 0.69–0.85
aInfections included gonorrhoea, syphilis, chlamydia, hepatitis B, or hepatitis C.
There was minimal missing data, which are not represented here but were part of the study calculations.
Values that are significant are bolded.
8 of 12 GERTZ AND DEFRANCO bs_bs_banner
preterm infants (Callen & Pinelli, 2005). Advocation of supplemental
donor milk in this population may lead to a need to expand the use
of donor human milk services, which has been shown to increase the
use of mother's own milk at discharge (Haslam et al., 2018; Parker,
Burnham, Mao, Philipp, & Merewood, 2016).
Many of the risk factors for breastfeeding non‐initiation are also risk
factors for preterm birth, such as young maternal age, low socioeco-
nomic and educational status, smoking, low BMI, intrauterine infection,
and use of reproductive technology (Goldenberg, Culhane, Iams, &
Romero, 2008; Vogel et al., 2018). Despite the association between
these factors and preterm birth, they are independently associated with
breastfeeding non‐initiation after adjustment in a multivariable logistic
regression model using gestational age and infant birth weight to repre-
sent prematurity.
Limited prenatal care (≤5 prenatal visits) was a greater risk factor
for breastfeeding non‐initiation in the NICU than in the newborn nurs-
ery, perhaps because mothers with less formal education may not
know about the specific benefits of breastmilk for NICU babies. It is
critical for the healthcare team to identify mothers with limited prena-
tal care, so they can be educated for potentially the first time about
the benefits of breastfeeding, especially for sick infants. Babies with
gastroschisis and extremely low or very low birth weight had an
increased likelihood of being breastfed in the NICU, likely because
physicians know they could benefit tremendously from breastmilk
and advocate for them. Opportunities exist to educate mothers of all
babies in the NICU to breastfeed.
Intentional home birth was also associated with an increased likeli-
hood of being breastfed. There may be factors shared among women
who choose home birth that also influence their choice to breastfeed.
The results of the regression model demonstrating an association
between the two do not infer that there is a cause and effect between
either the decision to give birth at home or to breastfeed, but rather, the
decision to do both may be inherently correlated. This correlation does
not necessarily negate the influence of post‐delivery counselling in the
NICU on breastfeeding nor the importance of breastfeeding support
services to NICU mothers to initiate and sustain breastfeeding. It is
beyond the scope of this study to measure differences in outcomes of
breastfeeding in NICU mothers based on the environment of birth
and unknown resources associated with it.
Evidence suggests that some interventions specifically targeting
mothers in the NICU have been successful (Fugate, Ivonne, Terri,
Miladinovic, & Spatz, 2015; McInnes & Chambers, 2008; Meier,
Engstrom, Mingolelli, Miracle, & Kiesling, 2004; Merewood et al.,
2006; Pinelli, Atkinson, & Saigal, 2001). Education and support
approaches range from meetings with lactation consultants and other
healthcare professionals to peer counselling from mothers in the com-
munity and programs for fathers (Furman, Killpack, Matthews, Davis,
& O'Riordan, 2016; McInnes & Chambers, 2008; Meier et al., 2004;
Merewood et al., 2006; Pinelli et al., 2001). Some of these interventions
involved socioeconomically advantaged mothers with strong support
networks who already have an increased likelihood of breastfeeding,
whereas others reached less advantaged inner‐city populations
(Balogun et al., 2016; Furman et al., 2016; Howe‐Heyman &
Lutenbacher, 2016; Humphreys, Thompson, &Miner, 1998;Merewood
et al., 2006; Pinelli et al., 2001).Motherswith higher socioeconomic sta-
tus and mothers who already had intent to breastfeed benefitted less
from intervention programs than inner‐city mothers who did not ini-
tially intend to breastfeed (Furman et al., 2016; Humphreys et al.,
1998; Merewood et al., 2006; Pinelli et al., 2001). Furthermore, the
most effective interventionmethod varied depending onmaternal char-
acteristics. High‐risk inner‐city mothers benefitted more from peer
involvement than healthcare staff support alone (Furman et al., 2016;
Humphreys et al., 1998). Despite efforts to promote breastfeeding ini-
tiation, rates in the NICU remain below rates in the newborn nursery,
suggesting that more targeted approaches are needed. Furthermore,
many of these interventions focused on breastfeeding duration as a pri-
mary outcome, but research shows that early breastfeeding initiation is
a key predictor of optimal breastfeeding behavior and decreases infant
mortality (Debes, Kohli, Walker, Edmond, & Mullany, 2013; Dewey
et al., 2003; Edmond et al., 2006). By developing a systematic way to
identify mothers who are least likely to intend to breastfeed and by
alerting healthcare professionals to high‐risk populations for
breastfeeding non‐initiation, interventions can promote breastfeeding
in these mothers even before the baby is born or within the baby's first
hours of life.
A strength of our study is that it analysed a wide array of
breastfeeding non‐initiation predictors to identify risk factors that
have not been assessed in other studies, such as smoking during preg-
nancy, congenital anomalies, and county of maternal residence. One
previously published study from Ohio assessed the association
between breastfeeding initiation in the NICU and gestational age but
did not analyse any other predictors (Kachoria & Oza‐Frank, 2015).
Additional strengths of this study include a large sample size and a
population‐based cohort, which we believe make the results and impli-
cations of the study generalizable across the United States, especially
because breastfeeding rates in Ohio approximate national
breastfeeding rates. Studies that have analysed breastfeeding initia-
tion in the NICU either studied a nondiverse population from one hos-
pital and/or took place in Europe, where breastfeeding culture and
rates are much different than those in the United States (Herich
et al., 2017; Lessen & Crivelli‐Kovach, 2007; Nielan‐Vilén, Melender,
Axelin, Loyttyniemi, & Salantera, 2016; Wallwiener et al., 2016).
Because this study included birth data from all hospitals in the state
of Ohio, it analysed a socioeconomically diverse population represen-
tative of much of the United States. This study would be difficult to
execute across multiple U.S. states due to birth certificate variability
and variation in birth certificate data collection policies. One previ-
ously published study from Ohio assessed the association between
breastfeeding initiation in the NICU and gestational age but did not
analyse any other predictors (Kachoria & Oza‐Frank, 2015). Another
strength of our study is that it analysed a wide array of breastfeeding
non‐initiation predictors to identify risk factors that have not been
assessed in other studies, such as smoking during pregnancy, congen-
ital anomalies, and county of maternal residence.
This study was limited by the type and quality of information
recorded on Ohio birth certificate records. The primary outcome,
GERTZ AND DEFRANCO 9 of 12 bs_bs_banner
breastfeeding initiation, was assessed by a “yes” or “no” response to
the question, “Is the infant being breastfed at discharge?” This ques-
tion does not specify whether the breastfeeding was exclusive or sup-
plemented or whether it involved direct breastfeeding, pumped milk,
or donor milk. Many infants who are born prematurely have physio-
logical barriers that inhibit their ability to breastfeed directly, which
requires coordination of sucking, swallowing, and breathing using
muscles and skills that mature with gestational age (Callen & Pinelli,
2005). These infants must receive nutrition by other methods, such
as by hand or through gavage feeding. Because donor milk has bene-
fits over preterm formula, it is the preferred alternative for preterm
infants when mother's own milk is not available (Sisk et al., 2017).
The delivery of human milk through methods other than the breast
has been referred to as breastmilk feeding, which this study also con-
siders to be breastfeeding.
The birth certificate question, “Is the infant being breastfed at dis-
charge?” does not provide information about breastfeeding duration
either. However, this study not only focused on in‐hospital
breastfeeding initiation because of available data but also because of
the aforementioned benefits specific to early breastfeeding initiation.
In fact, the World Health Organization now not only recommends
exclusive infant breastfeeding for the first 6 months of life but also
recommends breastfeeding initiation within the first hour of life. Many
of the risk factors for breastfeeding non‐initiation are the same risk
factors for early termination of breastfeeding, so interventions that
promote breastfeeding in the hospital should provide support for
mothers after discharge (Almqvist‐Tangen, Bergman, Dahlgren,
Roswall, & Alm, 2012; Thulier & Mercer, 2009; Valentine et al., 2019).
Exploration of the interplay between breastfeeding and additional
variables such as drug and alcohol use was limited by the maternal
information available on the birth certificate, but variables such as
drug and alcohol use may not be accurately reported by mothers any-
ways. Although it is not possible to ensure that all birth certificate data
are reported accurately, NICU admission and breastfeeding initiation
are objective measures likely easily captured by vital statistics staff
at the time the birth certificate is prepared. Self‐reported variables,
such as prepregnancy BMI, may contain inaccuracies, but they were
not the primary focus of this study, and the frequency of data
reporting errors likely does not differ notably between exposure and
reference groups. Despite the chance that smoking during pregnancy
may have been underreported, it still had a significant impact on
breastfeeding rates. There are quality improvement efforts underway
by the Ohio Perinatal Quality Collaborative to improve the data accu-
racy of the birth certificate, and it is a topic of ongoing research (Ohio
Perinatal Quality Collaborative, 2018; Lannon et al., 2017).
5 | CONCLUSIONS
Despite recent research and media efforts to highlight the benefits of
breastfeeding, initiation rates in the NICU remain low. After thor-
oughly accounting for confounding variables, we found that socioeco-
nomically disadvantaged mothers who smoked during pregnancy are
the least likely to initiate breastfeeding at hospital discharge. Multi-
parity, young age, and limited prenatal care are also risk factors for
breastfeeding non‐initiation. Hospital intervention efforts should cater
support and education programs toward the needs of these mothers
so that their sick infants have the best possible chance of a healthy
infancy and childhood by being breastfed.
ACKNOWLEDGMENTS
This manuscript includes data that were presented as abstracts in an oral
presentation at the ASPEN 2018 Nutrition Science and Practice Confer-
ence, January 22–25, Las Vegas, Nevada (Abstract #2832657) and in a
poster at the 2018 ACOG Annual Clinical and Scientific Meeting, April
27–30, Austin, Tx (Presentation #27 K). This study includes data
provided by the Ohio Department of Health, which should not be
considered an endorsement of this study or its conclusions. Cincinnati
Children's Hospital Medical Center Perinatal Institute; March of Dimes
Foundation, Grant/Award Number: 22‐FY14‐470.
CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.
CONTRIBUTIONS
ED conceived of the presented idea. BG developed the theory, gath-
ered the data, and performed the computations. ED verified the ana-
lytical methods and helped organize the findings of this work. BG
composed the manuscript, and ED edited it.
ORCID
Brooke Gertz https://orcid.org/0000-0003-3941-7914
REFERENCES
Almqvist‐Tangen, G., Bergman, S., Dahlgren, J., Roswall, J., & Alm, B. (2012). Factors associated with discontinuation of breastfeeding
before 1 month of age. Acta Paediatrica, 101, 55–60. https://doi.org/ 10.1111/j.1651‐2227.2011.02405.x
Balogun, O. O., O'Sullivan, E. J., McFadden, A., Ota, E., Gavine, A., Garner,
C. D., … MacGillivray, S. (2016). Interventions for promoting the initia- tion of breastfeeding. Cochrane Database of Systematic Reviews, (11).
https://doi.org/10.1002/14651858.CD001688.pub3
Bartick, M. C., Schwarz, E. B., Green, B. D., Jegier, B. J., Reinhold, A. G.,
Colaizy, T. T., … Stuebe, A. M. (2017). Suboptimal breastfeeding in the United States: Maternal and pediatric health outcomes and costs.
Maternal & Child Nutrition, 13, e12366. https://doi.org/10.1111/
mcn.12366
Bernaix, L. W., Schmidt, C. A., Jamerson, P. A., Seiter, L., & Smith, J. (2006).
The NICU experience of lactation and its relationship to family man-
agement style. The American Journal of Maternal Child Nursing, 31,
95–100. https://doi.org/10.1097/00005721‐200603000‐00008
Bharwani, S. K., Green, B. F., Pezzullo, J. C., Bharwani, S. S., Bharwani, S. S.,
& Dhanireddy, R. (2016). Systematic review and meta‐analysis of human milk intake and retinopathy of prematurity: A significant
update. Journal of Perinatology, 36, 913–920. https://doi.org/ 10.1038/jp.2016.98
Blatt, K., Moore, E., Chen, A., Van Hook, J., & DeFranco, E. A. (2015). Asso-
ciation of reported trimester‐specific smoking cessation with fetal
10 of 12 GERTZ AND DEFRANCO bs_bs_banner
growth restriction. Obstetrics and Gynecology, 125, 1452–1459. https://doi.org/10.1097/AOG.0000000000000679
Callen, J., & Pinelli, J. (2005). A review of the literature examining the ben-
efits and challenges, incidence and duration, and barriers to
breastfeeding in preterm infants. Advances in Neonatal Care, 5,
72–88. https://doi.org/10.1016/j.adnc.2004.12.003
Carswell, A. L., Ward, K. D., Vander Weg, M. W., Scarinci, I. C., Girsch, L.,
Read, M., … Chen, W. (2018). Prospective associations of breastfeeding and smoking cessation among low‐income pregnant women. Maternal & Child Nutrition, 14, e12622. https://doi.org/10.1111/mcn.12622
Celi, A. C., Rich‐Edwards, J. W., Richardson, M. K., Kleinman, K. P., & Gillman, M. W. (2005). Immigration, race/ethnicity, and social and eco-
nomic factors as predictors of breastfeeding initiation. Archives of
Pediatrics and Adolescent Medicine, 159, 255–260. https://doi.org/ 10.1001/archpedi.159.3.255
Colaizy, T. T., & Morriss, F. H. (2008). Positive effect of NICU admission on
breastfeeding of preterm US infants in 2000 to 2003. Journal of Perina-
tology, 28, 505–510. https://doi.org/10.1038/jp.2008.32
Debes, A. K., Kohli, A., Walker, N., Edmond, K., & Mullany, L. C. (2013).
Time to initiation of breastfeeding and neonatal mortality and morbid-
ity: A systematic review. BMC Public Health, 13, 519.
Dewey, K. G., Nommsen‐Rivers, L. A., Heinig, M. J., & Cohen, R. J. (2003). Risk factors for suboptimal infant breastfeeding behavior, delayed
onset of lactation, and excess neonatal weight loss. Pediatrics, 112,
607–619. https://doi.org/10.1542/peds.112.3.607
Economic Research Service of US Department of Agriculture (2016). Rural‐ Urban Continuum Codes (2013 version). [Data file]. Retrieved from
https://www.ers.usda.gov/data‐products/rural‐urban‐continuum‐ codes.aspx
Edmond, K. M., Zandoh, C., Quigley, M. A., Amenga‐Etego, S., Owusu‐ Agyei, S., & Kirkwood, B. R. (2006). Delayed breastfeeding initiation
increases risk of neonatal mortality. Pediatrics, 117, e380–e386. https://doi.org/10.1542/peds.2005‐1496
Fugate, K., Ivonne, H., Terri, A., Miladinovic, B., & Spatz, D. (2015). Improv-
ing human milk and breastfeeding practices in the NICU. Journal of
Obstetric, Gynecologic & Neonatal Nursing, 44, 426–438. https://doi. org/10.1111/1552‐6909.12563
Furman, L., Killpack, S., Matthews, L., Davis, V., & O'Riordan, M. A. (2016).
Engaging inner‐city fathers in breastfeeding support. Breastfeeding Medicine, 11, 15–20. https://doi.org/10.1089/bfm.2015.0092
Goldenberg, R. L., Culhane, J. F., Iams, J. D., & Romero, R. (2008). Epidemi-
ology and causes of preterm birth. The Lancet, 371, 75–84. https://doi. org/10.1016/S0140‐6736(08)60074‐4
Haslam, M. D., Gardner, F., Greecher, C., Stephens, H., Marvin, M., Hozella,
A., & Doheny, K. (2018). The impact of a donor milk program on breast
milk consumption, infection rates, and necrotizing enterocolitis in hos-
pitalized neonates. Pediatrics, 142, 686.
Herich, L. C., Cuttini, M., Croci, I., Franco, F., Di Lallo, D., Baronciani, D., … Zeitlin, J. (2017). Maternal education is associated with disparities in
breastfeeding at time of discharge but not at initiation of enteral feed-
ing in the neonatal intensive care unit. Journal of Pediatrics, 182, 59–65. https://doi.org/10.1016/j.jpeds.2016.10.046
Howe‐Heyman, A., & Lutenbacher, M. (2016). The baby‐friendly hospital initiative as an intervention to improve breastfeeding rates: A review
of the literature. Journal of Midwifery & Women's Health, 61, 77–102. https://doi.org/10.1111/jmwh.12376
Humphreys, A. S., Thompson, N. J., & Miner, K. R. (1998). Intention to
breastfeed in low‐income pregnant women: The role of social support and previous experience. Birth, 25, 169–174. https://doi.org/ 10.1046/j.1523‐536X.1998.00169.x
Isaacs, E. B., Fischl, B. R., Quinn, B. T., Chong, W. K., Gadian, D. G., & Lucas,
A. (2010). Impact of breast milk on intelligence quotient, brain size, and
white matter development. Pediatric Research, 67, 357–362. https:// doi.org/10.1203/PDR.0b013e3181d026da
Johnson, T. J., Patel, A. L., Bigger, H. R., Engstrom, J. L., & Meier, P. P.
(2014). Economic benefits and costs of human milk feedings: A strat-
egy to reduce the risk of prematurity‐related morbidities in very‐low‐ birth‐weight infants. Advances in Nutrition, 5, 207–212. https://doi. org/10.3945/an.113.004788
Kachoria, R., & Oza‐Frank, R. (2015). Trends in breastfeeding initiation in the NICU by gestational age in Ohio, 2006‐2012. Birth, 42, 56–61. https://doi.org/10.1111/birt.12146
Kelly, Y. J., Watt, R. G., & Nazroo, J. Y. (2006). Racial/ethnic differences in
breastfeeding initiation and continuation in the United Kingdom and
comparison with findings in the United States. Pediatrics, 118,
e1428–e1435. https://doi.org/10.1542/peds.2006‐0714
Lannon, C., Kaplan, H. C., Friar, K., Fuller, S., Ford, S., White, B., … Iams, J. (2017). Using a state birth registry as a quality improvement tool. The
American Journal of Perinatology, 34, 958–965.
Leon‐Cava, N., Lutter, C., Ross, J., & Martin, L. (2002). Quantifying the ben- efits of breastfeeding: A summary of the evidence. Washington, D.C.:
PAHO.
Lessen, R., & Crivelli‐Kovach, A. (2007). Prediction of initiation and dura- tion of breast‐feeding for neonates admitted to the neonatal intensive care unit. The Journal of Perinatal & Neonatal Nursing, 21,
256–266. https://doi.org/10.1097/01.JPN.0000285817.51645.73
Maffei, D., & Schanler, R. J. (2017). Human milk is the feeding strategy to
prevent necrotizing enterocolitis! Seminars in Perinatology, 41, 36–40. https://doi.org/10.1053/j.semperi.2016.09.016
McDowell, M. M., Wang, C., & Kennedy‐Stephenson, J. (2008). Breastfeeding in the United States: Findings from the national health
and nutrition examination surveys, 1999–2006. In NCHS data briefs, no 5. Hyattsville, MD: National Center for Health Statistics.
McInnes, R. J., & Chambers, J. (2008). Infants admitted to neonatal units— Interventions to improve breastfeeding outcomes: A systemic review
1990‐2007. Maternal & Child Nutrition, 4, 235–263. https://doi.org/ 10.1111/j.1740‐8709.2008.00150.x
Meier, P. P., Engstrom, J. L., Mingolelli, S. S., Miracle, D. J., & Kiesling, S.
(2004). The Rush mothers' milk club: Breastfeeding interventions for
mothers with very‐low‐birth‐weight infants. Journal of Obstetric, Gyne- cologic & Neonatal Nursing, 33, 164–174. https://doi.org/10.1177/ 0884217504263280
Merewood, A., Chamberlain, L. B., Cook, J. T., Philipp, B. L., Malone, K., &
Bauchner, H. (2006). The effect of peer counselors on breastfeeding
rates in the neonatal intensive care unit. Archives of Pediatrics & Adoles-
cent Medicine, 160, 681–685. https://doi.org/10.1001/archpedi. 160.7.681
Napierala, M., Mazela, J., Merritt, T. A., & Florek, E. (2016). Tobacco
smoking and breastfeeding: effect on lactation process, breast milk
composition and infant development. A critical review. Environmental
Research, 151, 321–338. https://doi.org/10.1016/j.envres.2016. 08.002
National Center for Chronic Disease Prevention and Health Promotion
(2018). Results: Breastfeeding rates. Atlanta, GA: US Department of
Health and Human Services, Centers for Disease Control and
Prevention.
National Center for Health Statistics (2003). Revisions of the U.S. standard
certificates of live birth. Hyattsville, MD: US Department of Health
and Human Services, Centers for Disease Control and Prevention.
National Center for Health Statistics (2016). Guide to completing the facility
worksheets for the certificate of live birth and report of fetal death (2003
GERTZ AND DEFRANCO 11 of 12 bs_bs_banner
revision). Hyattsville, MD: US Department of Health and Human Ser-
vices, Centers for Disease Control and Prevention.
Nielan‐Vilén, H., Melender, H. L., Axelin, A., Loyttyniemi, E., & Salantera, S. (2016). Predictors of breastfeeding initiation and frequency for preterm
infants in the NICU. Journal of Obstetric, Gynecologic, and Neonatal
Nursing, 45, 346–358. https://doi.org/10.1016/j.jogn.2016.01.006
Office of the Surgeon General (US) (2011). The Surgeon General's call to
action to support breastfeeding. Rockville, MD: Office on Women's
Health, Centers for Disease Control and Prevention.
Ohio Perinatal Quality Collaborative (OPQC). (2018). Birth registry
accuracy. Retrieved from https://opqc.net/projects/BirthRegistry
AccuracyResources.
Okamoto, T., Shirai, M., Kikubo, M., Takahashi, S., Kajino, M., Takase, M., … Oki, J. (2007). Human milk reduces the risk of retinal detachment in
extremely low‐birthweight infants. Pediatrics International, 49, 894–897. https://doi.org/10.1111/j.1442‐200X.2007.02483.x
Parker, M. G. K., Burnham, L., Mao, W., Philipp, B. L., & Merewood, A.
(2016). Implementation of a donor milk program is associated with
greater consumption of mother's own milk among VLBW infants in a
US, level 3 NICU. Journal of Human Lactation, 32, 221–228. https:// doi.org/10.1177/0890334415598305
Pinelli, J., Atkinson, S. A., & Saigal, S. (2001). Randomized trial of
breastfeeding support in very low‐birth‐weight infants. Archives of Pediatrics & Adolescent Medicine, 155, 548–553. https://doi.org/ 10.1001/archpedi.155.5.548
Sisk, P. M., Lambeth, T. M., Rojas, M. A., Lightbourne, T., Barahona, M.,
Anthony, E., & Auringer, S. T. (2017). Necrotizing enterocolitis and
growth in preterm infants fed predominantly maternal milk, pasteurized
donor milk, or preterm formula: A retrospective study. American Journal
of Perinatology, 34, 676–683. https://doi.org/10.1055/s‐0036‐ 1597326
Sisk, P. M., Lovelady, C. A., Dillard, R. G., Gruber, K. J., & O'Shea, T. M.
(2007). Early human milk feeding is associated with a lower risk of nec-
rotizing enterocolitis in very low birth weight infants. Journal of
Perinatology, 27, 428–433. https://doi.org/10.1038/sj.jp.7211758
Thulier, D., & Mercer, J. (2009). Variables associated with breastfeeding
duration. Journal of Obstetric, Gynecologic & Neonatal Nursing, 38,
259–268. https://doi.org/10.1111/j.1552‐6909.2009.01021.x
Valentine, G. C., Burgess, A., Ford, S., Gandhi, M., Hagan, J., Hurst, N., … Hair, A. (2019). Mother's own milk use during hospitalization predicts
later expressed breast milk feeds at hospital discharge. American Jour-
nal of Obstetrics and Gynecology, 220, S603.
Victora, C. G., Bahl, R., Barros, A. J. D., Franca, G. V. A., Horton, S.,
Krasevec, J., … Rollins, N. C. (2016). Breastfeeding in the 21st century: Epidemiology, mechanisms, and lifelong effect. The Lancet, 387,
475–490. https://doi.org/10.1016/S0140‐6736(15)01024‐7
Vogel, J. P., Chawanpaiboon, S., Moller, A., Watananirun, K., Bonet, M., &
Lumbiganon, P. (2018). The global epidemiology of preterm birth. Best
Practice & Research Clinical Obstetrics & Gynaecology ( ed., Vol. 52)
(pp. 3–12). https://doi.org/10.1016/j.bpobgyn.2018.04.003
Vohr, B. R., Poindexter, B. B., Dusick, A. M., McKinley, L. T., Wright, L. L.,
Langer, J. C., & Poole, W. K. (2006). Beneficial effects of breast milk
in the neonatal intensive care unit on the developmental outcome of
extremely low birth weight infants at 18 months of age. Pediatrics,
118, 115–123.
Wallwiener, S., Muller, M., Doster, A., Plewniok, K., Wallwiener, C. W.,
Fluhr, H., … Reck, C. (2016). Predictors of impaired breastfeeding initi- ation and maintenance in a diverse sample: What is important?
Maternal‐Fetal Medicine, 294, 455–466.
SUPPORTING INFORMATION
Additional supporting information may be found online in the
Supporting Information section at the end of the article.
How to cite this article: Gertz B, DeFranco E. Predictors of
breastfeeding non‐initiation in the NICU. Matern Child Nutr.
2019;15:e12797. https://doi.org/10.1111/mcn.12797
12 of 12 GERTZ AND DEFRANCO bs_bs_banner
Copyright of Maternal & Child Nutrition is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.