Clinical and Sociodemographic Factors Associated With Human Donor Milk Supplementation in Term Newborns

In this retrospective study, we investigated supplementation in newborns in the Mother Baby Unit (MBU) at a single quaternary academic medical center in the Pacific Northwest. There are ∼2250 deliveries annually, with a cesarean delivery rate of 30%. The patients admitted to the MBU were primarily white (77%) and non-Hispanic (81%). Most patients had a primary language of English (92%), with 6% of patients primarily Spanish speaking. Phone translation services were available around the clock for assistance with assent in families with limited English proficiency. Sixty-three percent of patients admitted to the MBU in 2019 had commercial insurance. The center’s institutional review board approved this study.

Eligibility criteria for this study included term newborns ≥37 weeks’ gestation admitted to the MBU between March 1, 2017, and April 30, 2019, with a maternal intent to breastfeed and at least one supplemental feeding with either HDM or infant formula. Exclusion criteria included birth weight <2500 g, maternal or newborn contraindications to breastfeeding, supplementation for hypoglycemia (defined as blood glucose level <40 mg/dL within the first 4 hours after birth or <45 mg/dL at ≥4 hours of age), exclusive formula feeding, and nonviable newborns. The electronic medical record (EMR) query used to identify eligible patients was originally designed for a quality-improvement project to decrease supplementation in healthy newborns; as such, hypoglycemic newborns were excluded from the build and subsequently excluded from this subanalysis.

During the study period, our center was working toward Baby-Friendly Hospital Initiative designation. The institutional breastfeeding-initiation rate was ∼80%. HDM had been available in the MBU since October 2013. Supplemental feeding required a physician order; families could choose pasteurized HDM purchased by the hospital from a local Human Milk Banking Association of North America milk bank or infant formula. Use of HDM required signed assent obtained by a nurse or physician. Term newborns were eligible to receive HDM if there was a plan to provide breast milk after discharge and if they had either a medical indication for supplementation (per hospital policy, potential indications include temporary separation from their mother, hypoglycemia, jaundice, dehydration, excessive weight loss, or prematurity) or a recommendation for supplementation from a lactation specialist or physician. HDM was not available to families without intention to breastfeed or those requesting supplementation for personal preference. Both milk options were available to families without charge and were kept accessible to unit staff in the same location within the MBU.

Eligible dyads were identified through a custom EMR query, and select newborn and maternal fields were abstracted into the study database. Abstracted data included newborn feeding data (time of first supplementation and supplemental milk type); newborn medical data, including factors that increase the likelihood of supplementation (birth weight, gestational age at birth, and receipt of phototherapy during birth hospitalization); NICU transfer; 30-day readmission; and maternal data (delivery method, parity, age, race and/or ethnicity, insurance coverage, and primary language). Birth weight was transformed into an ordinal variable and categorized into 4 groups. Gestational age was categorized as early term (37 + 0/7–38 + 6/7 weeks), term (39 + 0/7–40 + 6/7 weeks), or late term (41 + 0/7–42 + 6/7 weeks). Method of delivery was dichotomously split into vaginal (including operative vaginal delivery and vaginal birth after cesarean delivery) or cesarean delivery. Maternal age was separated into 4 categories determined by quartile. Time of first supplementation was coded dichotomously as day shift (7:00 am–6:59 pm) or night shift (7:00 pm–6:59 am).

Descriptive statistics were calculated for each variable. The 2 groups of subjects were compared by using χ² tests for association. A univariate analysis was performed for each independent variable to determine which to include in the final regression model. Only variables with a P value ≤.2 were included in the pooled logistic regression model. Binomial logistic regression was used to evaluate the relationships between clinical and sociodemographic factors and the type of supplementation received. All statistical analyses were conducted by using SPSS Statistics for Macintosh software, version 25 (IBM SPSS Statistics, IBM Corporation), with a significance level of P < .05.

Six hundred two dyads were identified through the EMR query (Fig 1). Of these, 18 received supplementation with both HDM and formula; the maternal and newborn characteristics of this group were similar to those of the overall cohort. Given the small size of this group, they were excluded from further analysis. In the final cohort of 584 dyads, the average maternal age was 31.3 years (SD = 6.1). The majority of women were white (73.3%) and non-Hispanic (75.8%). A slight majority held private insurance at the time of delivery (51.0%). Fifty-three percent of mothers were multiparous, and 54.8% delivered vaginally (Table 1).

Supplementation with infant formula was more frequent than supplementation with HDM and was provided to 337 newborns (57.7%) in this cohort. Most patients received their first supplemental feeding during a night shift (55.5%), and 26.3% received supplementation within 12 hours of delivery.

Multiple demographic characteristics were associated with HDM supplementation compared with formula supplementation. In the univariate analysis, there were associations with receipt of HDM and multiple factors, including timing of the newborn’s first supplementation, gestational age, delivery mode, maternal parity, insurance type, ethnicity, race, receipt of phototherapy, and transfer to the NICU (Table 1). These factors were combined to build the final multivariate logistic regression model. Birth weight category, supplementation within 12 hours of delivery, and 30-day readmission status were not significant in the univariate analysis and were not included in the multivariate model.

On the basis of multivariate analysis, there were multiple newborn characteristics that were associated with receipt of HDM (Table 2). Newborns receiving their first supplemental feeding during the day shift were twice as likely to receive HDM compared with those first supplemented during the night shift (adjusted odds ratio [aOR] 1.9; P = .001). Of newborns receiving phototherapy during their birth hospitalization, HDM was used more frequently than formula (81.3% vs 18.6%; P < .001). After controlling for potential interactions, gestational age and NICU transfers were no longer statistically significant predictors of supplement type.

Multiple maternal characteristics were associated with HDM supplementation. Newborns of primiparous women were more likely to receive HDM supplementation (aOR 1.6; P = .03). Cesarean delivery doubled the odds of a newborn receiving HDM (aOR 2.3; P < .001). Dyads with private insurance were 3.7 times more likely to receive HDM (P < .001). Newborns born to women who self-identified as white and non-Hispanic were also more likely to receive HDM when compared with newborns born to African American (aOR 5.6; P = .007), Asian American (aOR 2.7; P = .007), and Hispanic (aOR 2.9; P = .001) mothers. HDM supplementation was more likely in families who spoke primarily English or Spanish compared with families with a different primary language (aOR 8.5; P = .009). Maternal age did not have a significant impact on supplement type.

There are numerous factors associated with a newborn receiving HDM compared with formula for supplementation during the birth hospitalization. Among demographic variables, newborns of white and non-Hispanic mothers had the greatest likelihood of receiving HDM. Private insurance coverage and a primary language of either English or Spanish also correlated with increased odds of HDM supplementation, although not to the degrees of race and ethnicity. Although these data do not provide evidence of causation, the disparities in supplement type could result from differences in parental attitudes, differences in provider counseling (which may be impacted by ease of counseling in families with limited English proficiency), systems-based differences, or a combination of factors. The literature has revealed that parents selecting a supplement are more likely to choose what is most familiar.¹⁰  Given these findings, it is possible that multiparous mothers and/or mothers from specific sociodemographic backgrounds are more familiar with infant formula and therefore prefer this option. However, drawing a firm conclusion as to the basis of this relationship is beyond the scope of this study. There is also evidence that providers counsel African American and Hispanic mothers and mothers with limited English proficiency differently than white, English-proficient mothers when discussing breastfeeding, which may contribute to parental selection of supplement type.^16,17,18 

Clinically, multiple factors predicted supplementation with HDM. Our data revealed an association between timing of first supplementation and milk type, with newborns first supplemented during the day shift more likely to receive HDM. It is possible that the more frequent use of HDM as initial supplementation during the day shift reflects increased time for counseling and assent and, conversely, a hesitation to engage in the assent process at night. Additionally, this disparity may reflect a misconception in the perceived ease of access to formula during night-shift hours. Both HDM and formula are readily accessible within the unit at all times, but training on the availability and steps for obtaining each milk type may vary between day and night shift and may influence provider-parent counseling. To our knowledge, there is no previous literature in which a relationship between time of first supplementation and milk type is discussed, and further investigation is warranted.

Additionally, both cesarean delivery and phototherapy during admission were associated with a greater likelihood of supplementation with HDM. Although the exact cause of these associations remains unclear, both factors are associated with an increased potential medical need for supplementation. There is evidence that newborns delivered via cesarean delivery have greater weight loss than those delivered vaginally.¹⁹  A newborn requiring phototherapy may benefit from supplementation. In these cases, it is possible that a provider is the one initiating the conversation about supplementation, in contrast to a situation in which a parent is requesting supplementation or assistance for low supply, fatigue, or other concerns. This study was conducted at an academic center moving toward Baby-Friendly Hospital Initiative designation, so it is possible that providers would advocate for HDM when initiating a discussion about the need for supplementation with a family.

Although clinical guidelines from the AAP, World Health Organization, and Academy of Breastfeeding Medicine^9–11  recommend supplementation with HDM when possible, there is currently no literature in which the medical benefit of HDM over formula for supplementation of healthy term newborns is directly compared. The absence of clinical superiority makes it challenging to advocate for one supplement type over the other in this situation. There is good evidence that certain sociodemographic factors, such as race and ethnicity, contribute to differences in the supplement type provided to preterm newborns in the NICU.²⁰  Our findings suggest that similar sociodemographic disparities exist in supplementation of healthy term newborns. Therefore, it is crucial for clinicians to remain mindful of their biases and the complex social dynamics contributing to a dyad’s experience with newborn feeding.

The strengths of this study are twofold. There are no previous reports in which the characteristics of term newborns who receive HDM are described, and therefore this study helps to narrow an important knowledge gap. The large sample size of this study affords precision to these results and offers confidence in applying these results to similar populations. Our study has several limitations. The retrospective nature limited our ability to understand the indications for supplementation as well as parental and provider preferences toward supplementation type. Additionally, the use requirements of infant formula and HDM differ within our hospital setting. Families may use formula for supplementation for personal preference after education, even in the absence of a medical indication. In contrast, hospital policy limits the use of HDM in healthy term newborns to those with an intention to breastfeed and an identified medical indication for supplementation. It is possible that our results reflect different milk use due to the different requirements for the milk options; however, it is likely that the medical indications for supplementation would be equal across sociodemographic groups. Because of the retrospective methodology, we are unable to limit this analysis to only those infants with a medical indication for supplementation.

Additionally, newborns with hypoglycemia were not included in this analysis. This study was a subanalysis of a quality-improvement effort to decrease supplementation in term newborns in the MBU. Because hypoglycemia nearly always requires supplementation, the EMR query was designed to exclude any newborn with hypoglycemia. Lastly, this study was performed within a predominantly white, non-Hispanic population, which may limit generalizability to more diverse patient populations.

The trend of increasing HDM use in well-infant nurseries offers a myriad of future research directions. The clinical impact of HDM supplementation in term newborns has not been well described, and researchers of future studies should investigate the clinical outcomes of healthy term newborns receiving supplementation with HDM compared with infant formula. Given the paucity of literature examining the clinical and social factors contributing to supplementation type in term newborns, we believe these findings offer an important perspective on supplementation practices. Future studies are warranted to prospectively examine the indications and decision-making processes surrounding supplementation.

There are multiple clinical and sociodemographic differences in newborns receiving HDM compared with infant formula for supplementation. These findings highlight the continued need for awareness of the disparities that exist, especially as HDM supplementation becomes more widely available for healthy term newborns.

We thank Dr Dmitry Dukhovny for his execution and review of the final analyses.