Encourage, Assess, Transition (EAT): A Quality Improvement Project Implementing a Direct Breastfeeding Protocol for Preterm Hospitalized Infants : Advances in Neonatal Care

Secondary Logo

Journal Logo

Practice Improvements in Neonatal Care

Encourage, Assess, Transition (EAT)

A Quality Improvement Project Implementing a Direct Breastfeeding Protocol for Preterm Hospitalized Infants

Swanson, Nellie Munn DNP, MPH, APRN, CPNP-PC, CLC; Elgersma, Kristin M. DM, MN, MM, RN; McKechnie, Anne Chevalier PhD, RN; McPherson, Patricia L. MSN, RN, NE-BC; Bergeron, Mark J. MD, MPH; Sommerness, Samantha A. DNP, APRN, CNM; Friedrich, Cheri L. DNP, APRN, CPNP-PC, IBCLC; Spatz, Diane L. PhD, RN-BC, FAAN

Editor(s): Harris-Haman, Pamela A. DNP, CRNP, NNP-BC, Section Editor

Author Information
Advances in Neonatal Care 23(2):p 107-119, April 2023. | DOI: 10.1097/ANC.0000000000001037


In 2019, patient care experience data revealed that parents of former preterm infants cared for in our level II to IV neonatal system perceived limited support for their direct breastfeeding (DBF) goals, when compared with the lactation support they received for human milk (HM) feeding. Establishing a DBF relationship between a mother or lactating parent and their infant, defined as the experience of feeding the infant directly from the breast,1 is a pivotal driver of positive healthcare experiences for parents of preterm infants who desire to breastfeed.2,3 Moreover, breastfeeding and DBF are not interchangeable terms when describing a parent's intended feeding method.1 It is well known that DBF is often an exception, rather than standard care for preterm infants hospitalized in the United States. Further, there has been a “call to action” since 2011, to apply evidence-based breastfeeding practices in hospitals to standardize and improve care.4 The Triple Aim, a framework developed by the Institute for Healthcare Improvement in Cambridge, Massachusetts (www.ihi.org), to pursue improved care experience simultaneously with cost and population health outcomes, provides a second imperative to prioritize DBF in preterm infants.


The benefits of exclusive breastfeeding (ie, a diet comprised solely of HM regardless of feeding method) are well established,4,5 and significant for preterm infants (eg, lower morbidity, mortality, and adverse neurodevelopmental outcomes).6 In Minnesota, where the site is located, 95% of infants initiate HM, and exclusivity rates are 62% at 3 months and 39% at 6 months.7 Although term and preterm infants initiate HM at similar rates, preterm birth is an independent risk factor for early HM cessation before 6 months of age.8–10 HM initiation and duration rates, core measures in perinatal care quality, are limited by lack of delineation of actual or intended DBF and HM feeding rates.4,7,11,12 Exclusive and any HM feeding rates at neonatal intensive care unit (NICU) discharge from Vermont Oxford Network for infants born before 30 weeks of gestational age (GA) or very low birth weight (<1500 g) are lower in North America than anywhere in the world, and do not reflect the HM feeding method, or duration beyond discharge.12 Clinical practices that result in no DBF in the hospital are associated with lower HM feeding duration in preterm infants.9,13–17

In the year prior to the project, 79% of parents at the site planned to establish a DBF relationship with their preterm infant (42% DBF only, 37% DBF and bottle feeding) and 5% planned HM by bottle only. Despite the site's supportive lactation practices (eg, breast pumps in each room and International Board Certified Lactation Consultants [IBCLCs]), bottle-feeding preterm infants was the norm. Only 22% of preterm infants were DBF at the first oral meal, and total DBF meals per infant across the stay averaged 13.3. These data were compared to results from 3 retrospective studies. In infants less than 34 weeks GA at birth, 59% to 75% were DBF at the first oral meal; one study reported 10.6% of total meals as DBF18; while a second reported 64% DBF at more than 50% of meals 24 hours before discharge.19 A study of infants from 24 to 35 weeks of GA at birth found 16% DBF at the first oral meal, and 13.8 total DBF meals per infant.13 Conversely, in Denmark 99% of infants born at 24 to 37 weeks of GA initiated DBF, and 68% were exclusively DBF at discharge.20

From the parent perspective, DBF has deep meaning beyond nutritional value, as it facilitates formation of maternal identity, supports maternal mental health, and fosters infant attachment and development.2,3,21,22 Prioritizing HM expression as a means to an end, separate from the dynamic, embodied, DBF relationship desired, can leave mothers feeling marginalized.3,21,22 Studies documenting the experience of mothers in caring for their hospitalized preterm infants revealed that mothers are often discouraged from attempting DBF in the NICU until after bottle-feeding is established, and are advised that preterm infants discharge sooner if bottle-fed2,23,24—suggestions that are not evidence-based.17,20,25–27 Given parents' desire for more DBF support, the “call to action,” and the Triple Aim, we undertook a quality improvement (QI) project to develop and pilot a protocol to standardize DBF practices.

Available Knowledge

Current evidence, including the Spatz 10-step model, Infant Driven Feeding (IDF) concepts, and feeding imprinting,17,28,29 aided in developing the Encourage, Assess, Transition (EAT) protocol.

Spatz 10-Step Model

The Spatz 10-step model for protecting and promoting HM and breastfeeding in vulnerable infants reflects well-established, evidence-based interventions to support DBF in preterm infants.28 The 10 steps are: (1) informed decision to provide HM, (2) establishing an HM supply, (3) proper HM handling and storage, (4) oral care with HM, (5) skin-to-skin (STS) care, (6) practice feeding at breast, (7) transition to DBF, (8) test weighing to measure milk transfer, (9) preparing for discharge, and (10) follow-up. HM expression, preferably within 1 hour of birth, is associated with coming to volume, and HM duration.6,20 STS promotes DBF transition, and is associated with earlier DBF transition in preterm infants.6,30,31 Practice latching at an emptied breast promotes transition to DBF in preterm infants6,20,28,32 and increases DBF at a majority of meals at discharge.33 Test weighing using a standard procedure is a valid and reliable method to measure milk transfer,34,35 and is associated with exclusive DBF at discharge in preterm infants.14 Subjective clinical indices to estimate DBF intake (ie, “minutes protocols”), common in NICUs and at our site, are poorly correlated with HM transfer in preterm infants, resulting in clinically significant discrepancies when compared with test weighing.36,37 Utilizing the 10 steps as a standard, nurse-led care model is an effective strategy to increase DBF transition for vulnerable hospitalized infants.38

Cue-Based, Infant-Driven Feeding

Oral feeding is a complex developmental task for preterm infants, and provides an opportunity for early nurturing relationships between parents and infants. Coordinated suck, swallow, breathe reflexes are needed for safe oral feeding, and signal neurodevelopmental maturation near 34 weeks postmenstrual age (PMA).39 The content-validated IDF scales, previously integrated in the electronic health record (EHR) at the site, assess readiness for oral feeding in preterm infants based on state, tone, and cues.29 Delaying nutritive oral feeding until a majority pattern of IDF scales over 24 hours show readiness, often at a later PMA than previous practice, is a viable approach to move from traditional, provider-driven, scheduled feeding, to an IDF model.40–42 Delayed bottle-feeding and cue-based, IDF models may decrease time to ad lib, or full oral feeding, including when combined with the 10-step model.40–43

Feeding Imprinting

Preterm infants are sensitive to feeding imprinting (ie, developing an oral feeding preference), as they associate early oral feeding experiences with their mother.17 Providing the first oral meal at breast for preterm infants is a critical predictor of continued DBF in the hospital, including DBF a majority of meals, and earlier discharge.13,17–19,44 Moreover, a dose–response relationship has been observed between the number of DBF meals before the first bottle-feeding, transition time to ad lib oral feeds, HM consumption, and earlier discharge.17 Delayed introduction of bottles increases DBF in the hospital, and is protective of exclusive DBF at discharge.14,18,27 Preterm infants who were DBF any amount, or at least once per day in the hospital, achieve longer HM duration, compared with those who were never DBF.13–15,44,45


Despite advances in family-centered care (FCC) intended to improve care through parent–team partnerships,46 NICU feeding routines may negatively influence DBF outcomes.2,3,13,21,22 Our strategy addressed the interplay of clinical practice culture, unit routines, and parental autonomy, using FCC concepts. We posited that standard parent education and a visual pathway checklist would make the evidence-based DBF transition process transparent and predictable for families. Equipped with this knowledge, parents would then feel more empowered to advocate for their DBF goals within our system.

The theory of planned behavior (TPB) by Ajzen47 guided the strategy to standardize decision-making, parent education, and procedures used. The TPB explains how attitudes (ie, knowledge and beliefs about likely outcomes), subjective norms (ie, perceived social pressure), and perceived behavioral control (ie, ability to choose) shape behavioral intent.47 Education to influence knowledge, beliefs, and attitudes about DBF facilitates new practice norms in DBF care.30,38 The protocol was designed as a nurse-driven pathway, endorsed by medical leadership, to standardize care, despite clinician practice variation, patient handoffs, and patient care orders that potentially conflict with a parent's feeding goals.

Specific Aims

The aim of this QI project was to standardize DBF practices to meet parent feeding preferences. Results of this QI pilot are reported using SQUIRE 2.0 guidelines (http://squire-statement.org/).

What This Study Adds
  • For parents who desire a DBF relationship with their preterm infant, standardizing practices with the EAT protocol provides a common language, transparency, predictability, and input.
  • In alignment with the 10-step model, the EAT protocol increased prevalence of DBF in hospitalized preterm infants, without increasing length of stay.
  • This QI study provides further evidence that the method of first oral feeding influences DBF outcomes in hospitalized preterm infants.



As part of the needs assessment, we evaluated the site's current DBF practices, readiness for change, consulted colleagues in neighboring institutions, and identified areas for improvement (see Supplemental Digital Content 1, available at: https://links.lww.com/ANC/A172). There were no standard DBF protocols in place and practices (eg, clinical decision-making, parent education, and procedures used) preceding the first oral meal and during DBF transition, varied among interprofessional clinicians. DBF-sensitive outcome data were not systematically tracked, and clinicians were unaware of baseline DBF statistics or comparison data from other similar populations. Clinicians took for granted that transition to DBF was not a likely outcome in preterm infants, and anecdotal evidence that few preterm infants were DBF at discharge at the site supported these beliefs.

The Spatz 10-step model28 was utilized to varying extent, yet lack of a formalized process led to a gap in translating the DBF-specific steps into practice at the site. Although clinicians with particular interest in lactation and breastfeeding viewed the 10 steps as standard care, it was unclear that the system reliably delivered these interventions. Nonnutritive practice at breast prior to 34 weeks PMA was not routinely encouraged. Initial oral feedings, based on IDF scales, or clinician preference, were not standard, and were sometimes initiated based on readiness at a single interval, rather than a pattern. Oral feedings were offered on schedule, rather than cue-based, and infants born at more than 34 weeks of GA were routinely offered a bottle shortly after birth while separated from the parent. Multiple Medela Baby Weigh scales were available but rarely used, and no standard test weighing procedure, nor standard criteria for use, existed. Many clinicians were skeptical of the validity and reliability of test weighing, and “minutes protocols” were ordered by providers to estimate intake based on time an infant appeared to be latched at breast. Parent educational materials focused on HM expression, and verbal messages about the prospects of DBF a preterm infant varied among clinicians, as there was no unifying standard of DBF care communicated.

The DBF journey for each infant begins with the parent's initial feeding plan, assessed by nurses and providers. Subjective impressions of DBF commitment influenced care. If a parent communicated intensely about exclusive DBF, clinicians were more likely to delay bottles until after a trial of DBF. Yet, parents who experience emergent preterm birth may not communicate authoritatively about their feeding plan. The following de-identified vignette illustrates this phenomenon for a late preterm infant (LPI):

SD, a 34-week infant, was born by emergent cesarean section due to severe pregnancy-induced hypertension. As the birthing parent, JL, lay on the operating table, drapes covering much of her body she stated, “I hoped to breastfeed.” SD's initial glucose was 40 mg/dL, an intravenous line was placed, and mother's or donor HM, oral/enteral was ordered. The nurse called JL in recovery, where she was receiving a magnesium sulfate infusion and experiencing acute pain, to clarify her feeding plan, asking “Would you like us to place a feeding tube, or give SD a bottle?” JL thought a feeding tube would delay discharge, but did not voice this concern, and agreed to a bottle. She later understood that, based on GA and birth weight, SD would be hospitalized for several days or weeks. SD completed several small-volume bottles before requiring a feeding tube. As JL held her STS for the first time, SD fell asleep. JL decided not to try DBF, observing SD suck and swallow when a bottle was offered, despite appearing sleepy. By discharge at 37 weeks PMA, SD had never DBF and received HM and formula due to low HM supply.


Clinical nurses and leaders at every level of the organization were engaged in this QI project, from its inception as an evidence-based practice, Magnet nursing project in 2017. Clinical nurses often raised questions during unit council meetings, about how to better serve parents who desired to DBF their preterm infants. Nursing leaders convened a group of self-identified or peer-nominated “breastfeeding champions” in 2017 to consider ways to improve DBF and lactation care, which resulted in additional education for nurses, and early support for change. In 2019, the first author met 1:1 with approximately 15 interprofessional stakeholders (eg, clinical nurses, providers, lactation, managers, directors, and feeding specialists) and determined there was broad support for change to standardize DBF practices. An expanded leadership team including the neonatal system medical director, unit medical director, patient care manager, system nursing director, and unit council, each endorsed the pilot, which was developed in partial fulfillment of the first author's doctor of nursing practice (DNP) degree.

Ethical Considerations

The hospital's Institutional Review Board reviewed the project, and exempted it from oversight as it was not human subjects research, and no ethical concerns or conflicts of interest were identified.

Project Setting and Team

The project was implemented in a 33-bed level II neonatal unit in a freestanding Magnet-designated, pediatric hospital. The unit is part of a 118-bed neonatal service providing levels II to IV neonatal intensive care adjacent to a level IV perinatal health center. The project team included clinical nurses, an IBCLC, speech–language pathologists, nurse practitioners, neonatologist/medical director, and patient care manager. The entire care team (clinical nurse specialists, feeding team, lactation, unit council, system medical and nursing directors, and a parent advisory group) provided guidance and input on the pilot. The primary team leaders were: (1) the first author, project director; (2) a patient care manager; and (3) a medical director, who worked closely through all phases of Plan–Do–Study–Act (PDSA) cycles.

The EAT Protocol

The EAT protocol (Figure 1) was developed to meet the needs of the site to standardize DBF care by applying IDF concepts and reinforcing the 10-step model, with priority for steps (6) practice feeding at breast, (7) transition to DBF, and (8) test weighing to measure milk transfer. We identified 3 main practice changes to meet the specific aims of the pilot, and organized them in 3 phases. The new practices were embedded in (1) encourage phase: promote practice at breast prior to readiness for nutritive oral feeding; (2) assess phase: begin nutritive oral feeding only after a majority pattern of IDF scores indicate readiness, and parents receive up to 24 hours' notice to provide the feeding; and (3) transition phase: utilize a standard procedure and criteria for test weighing to measure milk transfer at DBF meals. Step-by-step checklists for clinicians, and parents to follow, along with written parent educational materials were developed (see Supplemental Digital Content 2 and 3, available at: https://links.lww.com/ANC/A173 and https://links.lww.com/ANC/A174). The flow diagram, parent checklist, parent educational materials, and star-shaped signs were bundled for clinicians to place in the patient room on admission. Parent educational materials outlined steps to prepare for DBF, signs of feeding readiness, and how milk transfer is measured by test weighing, written to the fifth-grade reading level.

EAT protocol flow diagram. Process steps outlined. EAT indicates Encourage, Assess, Transition.

The pilot was initially intended for infants greater than 32 and less than 37 weeks GA at birth, due to caution in offering practice at breast in very preterm infants. Clinical nurses provided early input that excluding infants less than 32 weeks GA caused confusion and delays in protocol initiation, and that parents of infants less than 32 weeks GA at birth also desired to start the protocol. The primary team leaders determined that removing the lower GA limit could increase use of protocol cares. Guidance for parents to empty the breasts prior to practice at breast for infants less than 32 weeks GA at birth was added. Due to subjectivity of documented feeding plans, any HM goal was the criteria to start EAT, as protocol interventions (ie, early HM expression and STS care) support both lactation and DBF after preterm birth. This facilitated informed parental decision on feeding method, preserving opportunity for DBF at the first oral meal, if desired.

Clinician Education

An educational curriculum was developed for interprofessional staff, and was delivered through written materials, video presentations, and virtual meetings. Due to the coronavirus disease-2019 (COVID-19) pandemic, the pilot was suspended for 4 months, making it possible to share brief educational messages with clinical nurses each week. The messages took 1 to 3 minutes to read, and described the clinical problem of varying DBF support practices, parents' desire for more DBF, and impact of hospital practices (eg, messages given to parents on DBF and method of the first oral feed) on DBF outcomes, and QI strategies. Supporting articles, along with the messages, were shared through regular weekly manager emails, and the unit council's monthly newsletter. Feedback from clinical nurses and the team guided messages from week to week. Two weeks prior to go-live, EAT protocol documents, and a 20-minute video slide presentation summarizing all prior messaging, were sent by email and posted in the unit for all staff to review. The project overview was also presented separately for neonatologists, nurse practitioners, and clinical nurses at 4 virtual staff meetings.

Environmental Cues

Practice cues, designed as visual reminders, were strategically placed in the unit to mitigate the role of clinical practice habit, and promote protocol adherence.48 The reminders included laminated copies of the protocol bundle checklists, which were displayed in high traffic areas (eg, time clocks, computer alcoves, entrances and breakrooms). A tracking list documenting each patient that initiated the protocol was maintained at the front desk by the unit coordinator. When the protocol was initiated, the bundle of protocol documents (see Supplemental Digital Content 2 and 3, available at: https://links.lww.com/ANC/A173 and https://links.lww.com/ANC/A174) was brought to the room, and the star-shaped signs were placed on the patient's door and in-room milk warmer. Each day, star stickers were placed next to the patient's name on provider and charge nurse rounding lists by the unit coordinator. Job aids to check for EAT protocol steps during team rounds were affixed to the provider's mobile computer workstations. A step-by-step checklist for the new test weigh procedure was attached to each of the 4 Medela Baby Weigh scales on the unit to cue practice.

Study of the Interventions

PDSA cycles were utilized to modify implementation strategies throughout the 3-month pilot period. The tracking list was compared to the census of eligible patients each week by the charge nurse, and patient care manager to identify any missed patients. The primary team leaders connected several times each week to review and respond to comments from parents, clinicians, unit council, and the full project team. Progress toward full implementation of the protocol, staff kudos, parent comments, and frequently asked questions were shared with staff via email and shift huddle communications each week. As COVID-19 restrictions eased 2 months after go-live, the project director began weekly on-site rounding to provide real-time 1:1 education, coaching, and problem-solving to promote adherence.

Additional Changes: PDSA

We solicited regular feedback from clinicians during the pilot to guide changes. Frequent comments were EAT was easy to use, reflected care provided by IBCLCs, assisted with continuity of care at shift change, and was well received by families. At times nurses waited for providers to order the protocol, despite the provider endorsed protocol criteria. Challenges were common on night shift, as on-call providers from the level IV NICU who were not as familiar with the pilot, tended to use standard admission orders that did not reflect the planned changes (eg, orders for oral/enteral feedings that were perceived as a directive to bottle an LPI immediately after birth). Night shift nurses felt “stuck” between trying to use the protocol and following orders. As more infants used the protocol, more parents voiced strong preferences that no bottles be given, including when they were not rooming in with their infant. Night nurses voiced concern that communication during parent-team rounds did not address the issue of infants' feeding cues during the night when parents were away. At the end of the pilot, a communication tool was created for use after the first 72 hours of DBF opportunity. The tool guided discussion of the plan for the next 24 hours (ie, to delay, or offer a bottle) during parent team rounds (see Supplemental Digital Content 4, available at: https://links.lww.com/ANC/A175).


Process objectives included: (1) form an interprofessional team; (2) develop a written protocol and visual cues; and (3) develop educational materials for clinicians, and parents. Outcome objectives to evaluate whether the practice changes resulted in improvement were to increase: (1) the rate of DBF at the first oral meal; (2) the total number of DBF meals per infant during hospitalization; and (3) the use of test weights to measure milk transfer, by 50% within 3 months, in infants less than 37 weeks GA at birth whose initial feeding plan included any HM. Outcome measures were defined as follows: method of the first oral meal was the earliest documented oral intake (breast or bottle) and total DBF meals per infant was the sum in the intake summary by discharge. All documented test weigh events in the pilot unit were compared pre- and post-intervention, as these data were not captured per eligible infant. These outcomes were measured by data reports generated from routine EHR documentation, which was unchanged by the protocol. These data allowed comparison of results with other published examples. The first author manually verified the accuracy of the data reports for a sample of 38 patients, and found no systematic errors. The EAT protocol was the only new practice, and any changes in these outcomes were believed to be associated with the pilot. Protocol adherence was tracked by comparing the census of eligible infants with the tracking list of patients who initiated the protocol. Contextual elements assessed included initial feeding plans, PMA at the first DBF, first bottle feeding, and discharge, to identify differences between the pre- and post-populations.


Infants less than 37 weeks GA at birth with any HM goal and present in the unit in the 12 months prior to the pilot (July 2019 to July 2020) served as a baseline comparison group. The post-pilot group included infants present in the unit from go-live in July 2020 to October 2020. Differences in sample characteristics (eg, sex, race, GA at birth, PMA at first DBF and discharge, and initial feeding plan) between the pre- and post-populations were analyzed using the χ3, Fisher exact test, 2-sample t tests, or Wilcoxon rank-sum tests, as appropriate given variable type (categorical or continuous) and normality. Significance was set at P < .05. The project did not aim to test the effectiveness of an intervention, and was not powered to test for significance of primary outcomes. Descriptive statistics (eg, change in number, percentage, mean, and median) were used to report: (a) the proportion of infants who were DBF at their first oral meal; (b) total number of DBF meals during the NICU stay per infant; and (c) frequency of test weighing to measure milk transfer. All analyses were conducted using RStudio (version 2021.09.2/R version 4.0.5, www.rstudio.com).


During the pilot (July to October 2020), 85 infants ranging from 27.7 through 36.7 (median 34.1) weeks of GA at birth had a goal to provide any HM, and 38 (45%) began the protocol. Table 1 summarizes characteristics of the pre- and post-populations. There were no significant differences in GA at birth, PMA at first DBF, or PMA at discharge. PMA at the first bottle-feed occurred significantly later post-intervention (35.4 vs 35 weeks). Initial feeding plans were significantly different post-intervention, as 85% planned to establish a DBF relationship (34% exclusive DBF, 51% DBF and bottle-feeding), and 1% planned HM by bottle only.

TABLE 1. - Summary Characteristics of Pre- and Postintervention Groups (N = 535)
Preintervention (n = 436)
n (%)
Postintervention (n = 99)
n (%)
P Value
Sex P = .020 a
Female 228 (52) 39 (39)
Male 208 (48) 60 (61)
Race P = .507b
Asian 20 (5) 3 (3)
Black/African 94 (22) 18 (18)
Multirace 42 (10) 13 (13)
White 212 (49) 54 (55)
Other/unknown 68 (16) 11 (11)
Language P = .800a
English 404 (93) 91 (92)
Other 32 (7) 8 (8)
Insurance P = .609a
Public 184 (42) 39 (39)
Private/other 252 (58) 60 (61)
Initial feeding plan P = .048 b
Breastfeed only 184 (42) 34 (34)
Breastfeed and supplement 160 (37) 50 (51)
Bottle feed expressed milk 20 (5) 1 (1)
Formula only 37 (9) 10 (10)
No plan specified 35 (8) 4 (4)
Median (IQR)
Median (IQR)
GA at birth, wk 34.1 (3.4)
34.1 (2.5)
P = .920c
Mean (SD) Mean (SD)
PMA at discharge, wk 36.6 (1.5) 36.8 (1.3) P = .1057d
(95% CI: −0.53, 0.05)e
PMA at first DBF, wk 34.9 (1.2) 34.9 (1.2) P = .811d
(95% CI: −0.26, 0.33)e
PMA at first bottle, wk 35.0 (0.98) 35.4 (1.0) P = .002 d
(95% CI: −0.63, −0.17)e
Abbreviations: CI, confidence interval; DBF, direct breastfeed; GA, gestational age; IQR, interquartile range; PMA, postmenstrual age.
aχ2 test.
bFisher exact test.
cWilcoxon rank-sum test.
d2-sample t test.
e95% CI for difference in means.

Rates of DBF at First Oral Feeding Increased

The proportion of preterm infants who experienced their first oral meal at breast increased from 22% to 54% pre- to post-populations—a 145% increase (Figure 2a). By the third month of PDSA, 70% of infants were DBF at their first oral meal. When examining only the 38 infants who initiated EAT, we found that 85% received their first oral meal at breast. Although improvements in DBF at the first oral meal occurred within each GA at birth category, fewer LPIs were DBF at the first oral meal in both pre- and postcohorts (Figure 2b).

Proportion of infants DBF at the first oral meal: (a) run chart by month and (b) mean by GA. DBF indicates direct breastfeed; GA, gestational age.

Total DBF Meals During the NICU Stay Increased

Total DBF meals per preterm infant during the hospital stay increased (Figure 3). Prepilot mean total DBF meals was 13.3 (median 4), and this rose to a mean of 20.3 (median 10) post-pilot—a 53% increase. By the third month of PDSA (October 2020), preterm infants received an average of 26.5 (median 24.5) DBF meals during their stay. Total DBF meals were higher for infants who were DBF at their first oral meal, in both pre- and postcohorts (Figure 4). Pre-pilot, infants who were bottle-fed first had a median (IQR) of 3 (0–10) total DBF meals, compared with 12 (4-21) DBF meals if they were DBF at the first oral meal. Post-pilot, the difference was greater, with infants who were bottle-fed first receiving a total of 1 (0-1) DBF meals compared with 21 (10–40) total DBF meals for those DBF at the first oral meal.

Mean and median total DBF meals per infant, by month. Run chart. DBF indicates direct breastfeed.
Total DBF meals per infant, by method of the first oral meal. Box plots. DBF indicates direct breastfeed.

Use of Test Weights to Measure at-Breast Intake Increased

The use of test weighing to measure milk transfer increased post-pilot. The average number of test weigh events recorded in the pilot unit was 110 per month in the 9 months pre-pilot, and rose to an average of 293 per month post-pilot, a 166% increase. By the end of the pilot in October, 2020 there were 354 test weighs per month, compared with 82 in October 2019—a 332% increase in 1 year.


Considering the paucity of literature describing strategies for implementation of evidence-based DBF practices in NICUs, this project makes a significant contribution. The EAT protocol promoted an infant-driven feeding approach, and ensured opportunity for DBF. Unlike clinician-driven bottle-feeding, DBF relationships are possible only with the infant's full engagement, a grounding premise of an infant-driven approach. To support feeding imprinting and increase DBF meals in the hospital, the first bottle-feedings were delayed. Our observation that PMA at discharge did not differ pre- versus post-populations is notable since neonatal clinicians in the United States often cite concern that supporting a parent's DBF goals will increase length of stay.25 These concerns persist despite evidence that GA, birth weight, and clinician feeding practices, rather than the feeding methods themselves, are predictive of PMA at discharge.20,26,27,33,40–42

Results for each of the outcome measures exceeded our goal of a 50% increase over baseline. Based on improvement trends over time (Figures 2a and 3), it appears that education correlated to an early improvement in outcomes, prior to the July implementation. The link between clinician education and practice change indicates that evidence-based messages are effective to influence knowledge and beliefs that inform DBF care in preterm infants, as reported by others.30 Once in-person rounding was possible in September, improvement increased, highlighting the value of visible leadership, and coaching to enact and sustain QI, as clinicians learn by doing, and integrate changes into routine practice.

Aligning Practices With the Spatz 10-Step Model

The EAT protocol aligned our DBF practices with the Spatz 10-step model by increasing opportunity for practice at breast, and DBF transition with measurement of milk transfer. The proportion of preterm infants who were DBF at the first oral meal was 85% for infants on EAT, and increased dramatically across the pre-/post-population to 70% by the end of the 3-month pilot. Despite these rapid improvements, there were persistent disparities by GA, as fewer LPIs were DBF at the first oral meal, compared with those born at earlier GA. This was a vexing problem that inspired the EAT protocol, and suggests that previous practices may have continued by habit, during the initial implementation of the pilot.

The role of feeding imprinting in DBF outcomes is evident across the literature.14,15,17,19,27,33,44,45 Our results further illustrate these phenomena, and underscore the significant impact of DBF at the first oral meal on DBF across the hospital stay. Improvement in total DBF meals per infant occurred in tandem with increased DBF at the first oral meal, and peaked at a mean of 26.5 (median 24.5) at the end of the pilot, nearly doubling the mean from the baseline period. Bottle-feeding at the first oral meal was associated with fewer total DBF meals, pre- and post-population, despite 86% HM initiation, including 34% desiring an exclusive DBF relationship, and 51% planning to DBF and offer bottles. If bottles are routinely offered before DBF, preterm infants and parents may not experience improved DBF outcomes.

Test weighing is the only empirically validated tool for measuring milk transfer during DBF.34,35 Our results indicate that test weighing, previously rare and inconsistent at our site, became standard. This is meaningful, as test weighing ensures precise nutritional supplementation37 during the vulnerable DBF transition period, when underfeeding can result in inadequate weight gain and a drowsy state, common antecedents to DBF cessation in preterm infants. Increased test weighing occurred along with increased DBF meals, consistent with research linking test weighing with increased exclusive DBF.14,20

Although 85% of preterm infants were eligible to receive protocol care, early adherence based on tracking receipt of the protocol bundle (ie checklists, educational materials for parents) was low. Given the observed changes in DBF indicators across the preterm infant population, it is likely that EAT resulted in a change in the clinician's DBF care, despite low documented adherence rates. Despite efforts to simplify criteria, adherence required clinician choice, which may have been influenced by non-evidence-based beliefs, as fidelity to a parent's DBF goals was at times commensurate with the intensity of parent communication about DBF. The challenges clinicians faced in implementing the protocol and integrating changes into routine practice are not surprising given the barriers we faced in implementation, including those related to COVID-19. We could not measure uptake of individual protocol interventions, due to a lack of such EHR tools to capture DBF-sensitive indicators. We missed opportunities to cue practice through passive dissemination of evidence, an effective strategy to integrate clinician knowledge, actual practice, and patient outcomes in QI.49 Integrating protocol interventions into EHR standard order sets and outcome standards, which drive expected care and documentation, may increase adherence by cueing practice.

Parent Experience

The EAT protocol provided parents with evidence-based education and anticipatory guidance that empowered them to advocate for their feeding goals. Standardizing practice across the interprofessional team reinforced a common language in DBF care, providing predictability and family engagement, which is consistent with other nurse-led QI projects.38 Previous parent feedback that more support for DBF in the hospital was desired provided context for our baseline prevalence of DBF. Considering these DBF indicators as a proxy for satisfaction with DBF care, EAT improved the experience of families who desired to establish a DBF relationship with their preterm infant. This is significant given qualitative research linking support for DBF with positive impressions of care in the NICU.2,3,21 In summary, the EAT protocol created new practice norms, and made our system more sensitive to parents' DBF goals.

Further Call to Action

This QI project filled an important gap in translating evidence to practice, and provides a further call to action for neonatal clinicians to improve DBF support practices for preterm infants. Our results demonstrate that QI to standardize DBF care for preterm infants with the EAT protocol is achievable, and made a major difference in DBF rates in preterm infants at the site. Although the EAT protocol included interventions that were believed to be in place, the DBF-specific practices needed to be integrated into a unified standard of care. Once this occurred, it increased the sensitivity of the system to the feeding plans of the majority of parents of preterm infants served. Without opportunity to DBF in the hospital, preterm infants and parents are unlikely to receive the necessary support to establish DBF post-discharge. Furthermore, the COVID-19 pandemic has dramatically decreased parents' access to in-person support groups and breastfeeding care in the community setting.50 For parents with a goal to DBF their infant, it is critical that clinicians prioritize DBF, and provide both educational and tangible assistance for DBF in the hospital. Given the importance of DBF outcome measures as indicators of neonatal care quality, reforms in informatics infrastructure to capture and report DBF-sensitive indicators are also needed.


Limitations were our small patient sample, short follow-up, and incomplete understanding of the extent of implementation, as we could not measure adherence to each intervention. Completion rates for the staff education were unclear, due to the pilot occurring off-cycle of mandatory education. Early on, clinicians reported a lack of understanding of how and when to use the protocol, despite education offered in several formats. Unique limitations imposed by the global COVID-19 pandemic prevented in-person contact with clinicians, and impacted measurement, and reporting of outcomes in real time, which limited fidelity to PDSA methods. Generalizability is limited to the project site, although it is plausible that implementation at other sites may result in similar improvements.

- Summary of Recommendations for Practice and Research
What we know:
  • The Spatz 10-step model for protecting and promoting HM and breastfeeding in vulnerable infants represents essential practices in DBF for preterm infants.

  • Opportunity to establish a DBF relationship is desired by parents of preterm infants, and is a pivotal driver of positive NICU experiences.

  • DBF outcomes, including HM duration beyond discharge, are sensitive to hospital practices (ie, clinical decision-making, parent education, and procedures used).

  • GA and birth weight are strong predictors of length of stay in preterm hospitalized infants.

What needs to be studied:
  • Definitions of the extent of DBF (ie, proportion of all oral meals as DBF, total DBF meals or days before a first bottle, and DBF at discharge).

  • Breastfeeding initiation and HM duration, delineated by intended and actual DBF rates and HM feeding rates.

  • The parent experience with the EAT protocol.

  • Methods to improve effectiveness of QI interventions on DBF practices.

  • The impact of specific clinical practices, including messaging to parents on DBF outcomes.

  • Timing of nonnutritive breastfeeding practice in very preterm and extremely preterm infants.

  • Oral feeding alternatives for LPI to protect the DBF relationship (eg, cup feeding and supplemental nursing systems).

What we can do today:
  • Examine organizational practices to identify iatrogenic barriers to DBF in vulnerable, hospitalized infants.

  • Outline the support available for DBF in ways that are transparent and predictable for parents.

  • Neutralize messaging to parents about the prospects of DBF in preterm infants, focusing on the ways the healthcare team can support DBF, if that is the parent's goal.

  • Encourage DBF, if desired by the parent, and delay introduction of bottles during DBF transition.

  • Engage parents in feeding, and facilitate opportunity to provide the first oral feed by the method of their choice.


The EAT protocol integrated evidence for DBF preterm infants and aligned care with the Spatz 10-step model for protecting and promoting HM and breastfeeding in vulnerable infants.28 Improved DBF outcomes were observed in preterm infants on a rapid timeline. EAT increased the use of standard DBF practices (eg, clinical decision-making, parent education, and procedures used) across the interprofessional team. Moreover, EAT promoted a common language, transparency, predictability, and input, for parents who desire a DBF relationship with their preterm infant. Modifications of EAT and EHR tools are needed to promote passive dissemination of evidence, increase adherence, and sustain change. Future research should validate the reported benefits of EAT, including increased DBF at the first oral meal, total DBF meals, and use of test weighing, in hospitalized preterm infants, and mechanisms to sustain change.


The authors thank Andrew Berndt, Alice Chernich, Kathleen Coates, Lee Diedrick, Tom George, Robyn Gizzi, Ashley Johnson, Laureen Love, Kristin McCullough, Heidi Poser, Jill Reinarts-Simonsen, Natalie Reker, Jennifer Rivera, Jeanine Schweiss, and the entire interprofessional team who provide 24/7 care to preterm infants and families, and without whom, this QI project would not have been possible.


1. Elgersma KM, Sommerness SA. What does it mean to be breastfed? A concept analysis in the context of healthcare research, clinical practice, and the parent perspective. J Perinat Neonatal Nurs. 2021;35(4):305–312. doi:10.1097/JPN.0000000000000572.
2. Ikonen R, Paavilainen E, Kaunonen M. Preterm infants' mothers' experiences with milk expression and breastfeeding: an integrative review. Adv Neonatal Care. 2015;15(6):394–406. doi:10.1097/ANC.0000000000000232.
3. Palmquist AEL, Holdren SM, Fair CD. “It was all taken away”: lactation, embodiment, and resistance among mothers caring for their very-low-birth-weight infants in the neonatal intensive care unit. Soc Sci Med. 2020;244:112648. doi:10.1016/j.socscimed.2019.112648.
4. US Department of Health and Human Services. The Surgeon General's Call to Action to Support Breastfeeding. Rockville, MD: Office of the Surgeon General; Centers for Disease Control and Prevention; Office on Women's Health; 2011. https://www.ncbi.nlm.nih.gov/books/NBK52682/?report.
5. Victora CG, Bahl R, Barros AJD, et al. Breastfeeding in the 21st century: epidemiology, mechanisms, and lifelong effect. Lancet. 2016;387(10017):475–490. doi:10.1016/S0140-6736(15)01024-7.
6. Parker MG, Stellwagen LM, Noble L, et al. Promoting human milk and breastfeeding for the very low birth weight infant. Pediatrics. 2021;148(5):e2021054272. doi:10.1542/peds.2021-054272.
7. Centers for Disease Control and Prevention. Breastfeeding Report Card, United States, 2020. https://www.cdc.gov/breastfeeding/data/reportcard.htm.
8. Colaizy TT, Saftlas AF, Morriss FH. Maternal intention to breast-feed and breast-feeding outcomes in term and preterm infants: Pregnancy Risk Assessment Monitoring System (PRAMS), 2000-2003. Public Health Nutr. 2012;15(4):702–710. doi:10.1017/S1368980011002229.
9. Hannan KE, Juhl AL, Hwang SS. Impact of NICU admission on Colorado-born late preterm infants: breastfeeding initiation, continuation and in-hospital breastfeeding practices. J Perinatol. 2018;38(5):557–566. doi:10.1038/s41372-018-0042-x.
10. Jónsdóttir RB, Jónsdóttir H, Skúladóttir A, Thorkelsson T, Flacking R. Breastfeeding progression in late preterm infants from birth to one month. Matern Child Nutr. 2020;16(1):e12893. doi:10.1111/mcn.12893.
11. Chiang KV, Sharma AJ, Nelson JM, Olson CK, Perrine CG. Receipt of breast milk by gestational age—United States, 2017. MMWR Morb Mortal Wkly Rep. 2019;68(22):489–493. doi:10.15585/mmwr.mm6822a1.
12. Vermont Oxford Network. Infants Receiving Human Milk at Discharge Varies Internationally. https://public.vtoxford.org/nicu-by-the-numbers/infants-receiving-human-milk-at-discharge-varies-internationally/
13. Pineda R. Direct breast-feeding in the neonatal intensive care unit: is it important? J Perinatol. 2011;31(8):540–545. doi:10.1038/jp.2010.205.
14. Maastrup R, Hansen BM, Kronborg H, et al. Factors associated with exclusive breastfeeding of preterm infants. Results from a prospective national cohort study. PLoS One. 2014;9(2):e89077. doi:10.1371/journal.pone.0089077.
15. Pinchevski-Kadir S, Shust-Barequet S, Zajicek M, et al. Direct feeding at the breast is associated with breast milk feeding duration among preterm infants. Nutrients. 2017;9(11):1202. doi:10.3390/nu9111202.
16. Bonnet C, Blondel B, Piedvache A, et al. Low breastfeeding continuation to 6 months for very preterm infants: a European multiregional cohort study. Matern Child Nutr. 2019;15(1):e12657. doi:10.1111/mcn.12657.
17. Suberi M, Morag I, Strauss T, Geva R. Feeding imprinting: the extreme test case of premature infants born with very low birth weight. Child Dev. 2018;89(5):1553–1566. doi:10.1111/cdev.12923.
18. Briere CE, McGrath JM, Cong X, Brownell E, Cusson R. Direct-breastfeeding premature infants in the neonatal intensive care unit. J Hum Lact. 2015;31(3):386–392. doi:10.1177/0890334415581798.
19. Casey L, Fucile S, Dow KE. Determinants of successful direct breastfeeding at hospital discharge in high-risk premature infants. Breastfeed Med. 2018;13(5):346–351. doi:10.1089/bfm.2017.0209.
20. Maastrup R, Hansen BM, Kronborg H, et al. Breastfeeding progression in preterm infants is influenced by factors in infants, mothers and clinical practice: the results of a national cohort study with high breastfeeding initiation rates. PLoS One. 2014;9(9):e108208. doi:10.1371/journal.pone.0108208.
21. Holdren S, Fair C, Lehtonen L. A qualitative cross-cultural analysis of NICU care culture and infant feeding in Finland and the U.S. BMC Pregnancy Childbirth. 2019;19(1):345. doi:10.1186/s12884-019-2505-2.
22. Brown A. What do women lose if they are prevented from meeting their breastfeeding goals? Clin Lactation. 2018;9(4):200–207. doi:10.1891/2158-0782.9.4.200.
23. Niela-Vilén H, Axelin A, Melender HL, Salanterä S. Aiming to be a breastfeeding mother in a neonatal intensive care unit and at home: a thematic analysis of peer-support group discussion in social media: breastfeeding mother in an NICU and at home. Matern Child Nutr. 2015;11(4):712–726. doi:10.1111/mcn.12108.
24. Parker MG, Hwang SS, Forbes ES, Colvin BN, Brown KR, Colson ER. Use of the theory of planned behavior framework to understand breastfeeding decision-making among mothers of preterm infants. Breastfeed Med. 2020;15(10):608–615. doi:10.1089/bfm.2020.0127.
25. Briere CE. Breastfed or bottle-fed: who goes home sooner? Adv Neonatal Care. 2015;15(1):65–69. doi:10.1097/ANC.0000000000000159.
26. Wener E, Dow KE, Fucile S. Evaluation of methods of breast or bottle feeding on length of hospitalization of preterm infants. Breastfeed Med. 2021;16(11):899–903. doi:10.1089/bfm.2021.0066.
27. Allen E, Rumbold AR, Keir A, Collins CT, Gillis J, Suganuma H. Avoidance of bottles during the establishment of breastfeeds in preterm infants. Cochrane Database Syst Rev. 2021;10(10):CD005252. doi:10.1002/14651858.CD005252.pub5.
28. Spatz DL. Beyond BFHI: the Spatz 10-step and breastfeeding resource nurse model to improve human milk and breastfeeding outcomes. J Perinat Neonatal Nurs. 2018;32(2):164–174. doi:10.1097/JPN.0000000000000339.
29. Waitzman KA, Ludwig SM, Nelson CLA. Contributing to content validity of the Infant-Driven Feeding Scales© through Delphi surveys. Newborn Infant Nurs Rev. 2014;14(3):88–91. doi:10.1053/j.nainr.2014.06.010.
30. Maastrup R, Rom AL, Walloee S, Sandfeld HB, Kronborg H. Improved exclusive breastfeeding rates in preterm infants after a neonatal nurse training program focusing on six breastfeeding-supportive clinical practices. PLoS One. 2021;16(2):e0245273. doi:10.1371/journal.pone.0245273.
31. Oras P, Thernström Blomqvist Y, Hedberg Nyqvist K, et al. Skin-to-skin contact is associated with earlier breastfeeding attainment in preterm infants. Acta Paediatr. 2016;105(7):783–789. doi:10.1111/apa.13431.
32. Nyqvist KH. Early attainment of breastfeeding competence in very preterm infants. Acta Paediatr. 2008;97(6):776–781. doi:10.1111/j.1651-2227.2008.00810.x.
33. Fucile S, Wener E, Dow K. Enhancing breastfeeding establishment in preterm infants: a randomized clinical trial of two non-nutritive sucking approaches. Early Hum Dev. 2021;156:105347. doi:10.1016/j.earlhumdev.2021.105347.
34. Haase B, Barreira J, Murphy PK, Mueller M, Rhodes J. The development of an accurate test weighing technique for preterm and high-risk hospitalized infants. Breastfeed Med. 2009;4(3):151–156. doi:10.1089/bfm.2008.0125.
35. Rankin MW, Jimenez EY, Caraco M, Collinson M, Lostetter L, DuPont TL. Validation of test weighing protocol to estimate enteral feeding volumes in preterm infants. J Pediatr. 2016;178:108–112. doi:10.1016/j.jpeds.2016.08.011.
36. Greenslade S, Miller J, Tonkin E, Marshall P, Collins C. Estimating the Dietary Intake of Breastfeeding Preterm Infants. Int J Environ Res Public Health. 2015;12(5):5408–5419. doi:10.3390/ijerph120505408.
37. Perrella SL, Nancarrow K, Rea A, Murray K, Geddes DT, Simmer KN. Estimates of preterm infants' breastfeeding transfer volumes are not reliably accurate. Adv Neonatal Care. 2020;20(5):E93–E99. doi:10.1097/ANC.0000000000000721.
38. Edwards TM, Spatz DL. An innovative model for achieving breast-feeding success in infants with complex surgical anomalies. J Perinat Neonatal Nurs. 2010;24(3):246–253; quiz 254-255. doi:10.1097/JPN.0b013e3181e8d517.
39. Shaker CS. Infant-guided, co-regulated feeding in the neonatal intensive care unit. Part I: theoretical underpinnings for neuroprotection and safety. Semin Speech Lang. 2017;38(2):96–105. doi:10.1055/s-0037-1599107.
40. Gelfer P, McCarthy A, Spruill CT. Infant driven feeding for preterm infants: learning through experience. Newborn Infant Nurs Rev. 2015;15(2):64–67. doi:10.1053/j.nainr.2015.04.004.
41. Thomas T, Goodman R, Jacob A, Grabher D. Implementation of cue-based feeding to improve preterm infant feeding outcomes and promote parents' involvement. J Obstet Gynecol Neonatal Nurs. 2021;50(3):328–339. doi:10.1016/j.jogn.2021.02.002.
42. Wellington A, Perlman JM. Infant-driven feeding in premature infants: a quality improvement project. Arch Dis Child Fetal Neonatal Ed. 2015;100(6):F495–500. doi:10.1136/archdischild-2015-308296.
43. Chrupcala KA, Edwards TM, Spatz DL. A continuous quality improvement project to implement infant-driven feeding as a standard of practice in the newborn/infant intensive care unit. J Obstet Gynecol Neonatal Nurs. 2015;44(5):654–664. doi:10.1111/1552-6909.12727.
44. Casavant SG, McGrath JM, Burke G, Briere CE. Caregiving factors affecting breastfeeding duration within a neonatal intensive care unit. Adv Neonatal Care. 2015;15(6):421–428. doi:10.1097/ANC.0000000000000234.
45. Briere CE, McGrath JM, Cong X, Brownell E, Cusson R. Direct-breastfeeding in the neonatal intensive care unit and breastfeeding duration for premature infants. Appl Nurs Res. 2016;32:47–51. doi:10.1016/j.apnr.2016.04.004.
46. Franck LS, O'Brien K. The evolution of family-centered care: from supporting parent-delivered interventions to a model of family integrated care. Birth Defects Res. 2019;111(15):1044–1059. doi:10.1002/bdr2.1521.
47. Ajzen I. The theory of planned behavior. Organ Behav Hum Decis Process. 1991;50:179–211.
48. Nilsen P, Roback K, Broström A, Ellström PE. Creatures of habit: accounting for the role of habit in implementation research on clinical behaviour change. Implement Sci. 2012;7:53. doi:10.1186/1748-5908-7-53.
49. Donaldson NE, Rutledge DN, Ashley J. Outcomes of adoption: measuring evidence uptake by individuals and organizations. Worldviews Evid Based Nurs. 2004;1 suppl 1:S41–S51. doi:10.1111/j.1524-475X.2004.04048.x.
50. Spatz DL. The COVID-19 pandemic and breastfeeding: concerns & positive opportunities. MCN Am J Matern Child Nurs. 2021;46(4):238. doi:10.1097/NMC.0000000000000732.

direct breastfeeding; human milk; maternal autonomy; neonatal intensive care; patient-centered care; patient preference; premature infant; quality improvement; Spatz 10-step; test weigh

Supplemental Digital Content

© 2022 by The National Association of Neonatal Nurses.