Best, best

msgbemmy
Predictorsofbreastfeedingnon-initiationintheNICU.pdf

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.