Skip to main content
NCBI home page
Maternal & Child Nutrition logoLink to Maternal & Child Nutrition
. 2018 Sep 4;15(1):e12670. doi: 10.1111/mcn.12670

The effect of maternity practices on exclusive breastfeeding rates in U.S. hospitals

Julie A Patterson 1, Nicholas S Keuler 2, Beth H Olson 1,✉
PMCID: PMC7199031  PMID: 30182474

Abstract

The Baby‐friendly Hospital Initiative (BFHI) includes a set of 10 evidenced‐based maternity practices that when used together have been shown to improve breastfeeding outcomes. In 2007, the Centers for Disease Control and Prevention launched the Maternity Practices in Infant Nutrition and Care (mPINC) survey to assess and monitor these and other evidenced‐based maternity practices. The purpose of this study was to explore individual maternity practices measured in the 2013 mPINC survey, along with hospital demographic information, and their relationships with exclusive breastfeeding (EBF) rates, using a sample of United States (U.S.) hospitals. We obtained mPINC survey data from 69 BFHI hospitals and 654 non‐BFHI hospitals in the U.S., and EBF rates from The Joint Commission, a leading hospital accreditation agency. On the basis of linear regression analysis, we found that most maternity practices studied were significantly associated with EBF rates (max adjusted R 2 = 14.9%). We found a parsimonious model with an adjusted R 2 of 47.3%. This study supports the need for a systematic approach in providing breastfeeding support as no one maternity care practice was able to explain the variability in EBF rates as well as a collection of maternity care practices.

Keywords: baby friendly hospital initiative, breastfeeding, exclusive breastfeeding, joint commission, maternity care, maternity practices

Key messages.

  • Most maternity care practices measured in the 2013 Maternity Practices in Infant Nutrition and Care survey and included in our analysis were significantly related to exclusive breastfeeding rates.

  • No one maternity care practice was able to predict exclusive breastfeeding rates as well as a collection of practices.

  • The Baby‐friendly Hospital Initiative, which is a bundle of maternity care practices, is an important initiative to improve exclusive breastfeeding rates.

1. INTRODUCTION

The health benefits of breastfeeding are undisputed (Ip et al., 2007; Victora et al., 2016), yet most women in the United States (U.S.) are not meeting expert recommendations to breastfeed exclusively for about 6 months (American Academy of Pediatrics, 2012; Office of Disease Prevention and Health Promotion, 2018). Health care practices that mothers and newborn infants experience in the hospital have long been recognized as one of the most promising strategies to promote initiation, duration, and exclusivity of breastfeeding (Rollins et al., 2016; World Health Organization, 1989). The Baby‐friendly Hospital Initiative (BFHI) recognizes hospitals' implementation of the Ten Steps to Successful Breastfeeding and compliance with the International Code of Marketing of Breast‐milk Substitutes (World Health Organization, 2017). Globally and in the U.S., BFHI is recognized as improving breastfeeding outcomes (Perez‐Escamilla, Martinez, & Segura‐Perez, 2016). However, 20 years after the inception of BFHI in the U.S., only 25% of annual births occur in BFHI‐certified hospitals (Baby‐Friendly USA, Inc., 2018a). The Maternity Practices in Infant Nutrition and Care (mPINC) survey conducted by the Centers for Disease Control and Prevention (CDC) in the U.S. has been used as a tool to improve hospitals' maternity care practices to optimally support breastfeeding (Grossniklaus et al., 2017), with an overarching goal of assisting hospitals in achieving the BFHI designation (CDC, 2018). The purpose of this study was to evaluate the relationship between individual maternity care practices measured within the mPINC survey that reflect hospitals' implementation of BFHI, along with the attainment of BFHI, and exclusive breastfeeding (EBF) rates. Evaluating individual maternity care practices in the hospital is a meaningful way to prioritize opportunities to improve breastfeeding care to positively influence EBF rates. Although evaluating individual maternity care practices is important, research has shown a dose–response relationship with mothers' exposure to BFHI steps and breastfeeding outcomes (DiGirolamo, Grummer‐Strawn, & Fein, 2008). Therefore, we created models to determine how well we could predict EBF using a combination of maternity care practices, the BFHI designation, and hospitals' demographic characteristics. As far as we are aware, this study is the first to use a standardized in‐hospital definition of EBF applied across a large U.S. hospital sample to evaluate maternity care practices that optimally support EBF.

2. METHODS

2.1. Data sources

Data for this study were obtained from four sources: (a) the CDC (2013a) mPINC survey, (b) The Joint Commission (2016), (c) Baby‐Friendly USA, Inc. (2018a), and (d) the United States Census Bureau (2014). Biennially, the CDC invites the approximately 3,300 birthing hospitals in the U.S. to complete a survey about the maternity care practices in their hospital (Perrine et al., 2015). The mPINC tool gathers information on maternity care practices within seven practice areas (domains). The seven domains include labour and delivery care, feeding of breastfed infants, breastfeeding assistance, mother–baby contact, discharge care, staff training, and structural and organizational aspects of care. Questions with categorical response variables included yes/no responses and Likert scales of four response types from rarely to almost always or few to most. The complete 2013 mPINC survey is available (CDC, 2013b). Hospitals' EBF rates were extracted from The Joint Commission's database for the time period of April 1, 2014, to March 31, 2015. We felt this date range was appropriate as we were interested in looking at how the data collected on hospitals' maternity care practices in 2013 were related to EBF rates in 2014. Effective January 1, 2014, The Joint Commission, a leading hospital accreditation agency in the United States, required hospitals they certify that have ≥1,100 births annually to publically report their EBF rates (The Joint Commission, 2013). This new regulation provides nationwide data on in‐hospital EBF rates using a standardized definition. The Joint Commission defined EBF as “a newborn receiving only breast milk and no other liquids or solids except for drops or syrups consisting of vitamins, minerals or medicines” (The Joint Commission, 2014, 2015b). This publically available data are reported as an aggregated “exclusive breast milk feeding” rate from a sample of eligible infants within the hospital population and expressed as a percentage. This sample excluded infants (a) admitted to the neonatal intensive care unit, (b) diagnosed with galactosaemia, (c) on parenteral nutrition, (d) who have died, (e) with a hospital stay ≥ 120 days, (f) enrolled in clinical trials, (g) transferred to another hospital, or (h) who had some documented reason for not exclusively breastfeeding (The Joint Commission, 2015a). Baby‐friendly USA is the accrediting body and the national authority of BFHI in the United States (Baby‐Friendly USA, Inc., 2018b). As of April 2014, there were 171 BFHI hospitals in the U.S. Using this information, we identified 69 BFHI hospitals in our data set. As the EBF data represent hospital's EBF rate over a 1‐year period, we wanted to ensure that all hospitals had the BFHI designation at the same entry point into the EBF data set. Census tract data from the American Community Survey, conducted by the United States Census Bureau, were applied to the data set (United States Census Bureau, 2014). A census tract is a cluster of areas designed to have a population size of 1,200–8,000 people. This information represents the demographic information for the community surrounding the hospitals and provides a proxy for the patients served by the hospital. Although many factors influence patients' hospital utilization, studies have shown that travel distance is a significant determinant of hospital utilization (Friedman et al., 2015) and the majority of women deliver at their local hospital (Kozhimannil, Casey, Hung, Prasad, & Moscovice, 2016).

2.2. Data set

The 2013 mPINC survey had a total of 2,797 respondents, of which 983 hospitals also appeared in The Joint Commission data set. Figure 1 describes the final hospital sample that included 723 hospitals, with 69 hospitals having achieved the BFHI designation as of April 1, 2014. The 2013 mPINC survey contained 57 questions of which 29 were excluded. Reasons for exclusion included (a) more than 100 hospitals were missing a response to the question (n = 1 question), (b) they were not related to healthy full‐term infants (n = 3 questions), (c) they were related to discharge care (n = 2 questions), (d) they did not pertain to maternity care practices (n = 15 questions), (e) they were duplicate questions (n = 6 questions), or (f) there was insufficient evidence in the literature that the maternity care practice was related to EBF rates (n = 2 questions). Table 1 summarizes the evidence used to identify the maternity care practices that were included in our analysis. For purposes of analysis, we made the following changes to the original survey questions. Survey Question B11 related to facilities' written policy on breastfeeding was scored (0–10) using criteria outlined by the CDC (CDC, 2017b). This new policy variable was treated as a continuous variable, with a higher number indicating more breastfeeding policy elements. The second variable created represented lactation specialist (as defined by each respondent) availability for mPINC Survey Question B10. Hospitals that reported that a lactation specialist was “always” available either 5 days a week or 7 days a week were placed into one of the following categories: “7 days a week” or “5 days a week.” Hospitals that did not report “always” having lactation specialists during those possible time frames were categorized as “not having availability.” The third and fourth variables were created by combining vaginal‐ and caesarean‐specific continuous response survey questions to reflect a weighted average. These variables A13/A07 and A12/A06 account for the caesarean and vaginal birth rates at each hospital. These new predictor variables, A13/A07 and A12/A06, replaced the individual maternity care practices and are reported in Table 2. Only hospitals with complete answers to all 28 questions were included in our analysis. Demographic characteristics were added to represent the population served by each hospital. Demographic variables were calculated and expressed as a proportion of the population of interest. Variables representing race/ethnicity were composed of non‐Hispanic White (White), non‐Hispanic African American (African American), non‐Hispanic Asian (Asian), and Hispanic. The population in poverty was defined as the proportion of the population that is below 200% of the poverty line. The education variables refer only to persons ≥ 25 years old with at least a bachelor's degree (proportion with bachelors) and those who never attended college (proportion without college attendance). Data related to hospitals' BFHI designation, caesarean section rate, and number of deliveries were also added as predictor variables. This study was certified as not human subjects research by the University of Wisconsin–Madison Institutional Review Board.

Figure 1.

Figure 1

Open in a new tab

Flow chart of final hospital sample included in analysis

Table 1.

Research used for selection of Maternity Practices in Infant Nutrition and Care (mPINC) survey questions

Domaina Survey number Survey question Research used for selection of survey questions
Labour and delivery A11, A05 How many mothers are encouraged to hold their infant for 30 min within 1 hr of a vaginal delivery and 2 hr after a caesarean delivery?b Skin‐to‐skin contact improves breastfeeding outcomes (Moore et al., 2016; Stevens, Schmied, Burns, & Dahlen, 2014)
A04, A10 Routine newborn procedures are done while mother holds infant.b Procedures should take place while baby is skin to skin (Crenshaw, 2014; Holmes, McLeod, & Bunik, 2013)
A12, A06 Percent infants put to breast within 1 hr after vaginal delivery and 2 hr after caesarean delivery.c Assist mothers to initiate breastfeeding within 1 hr of birth (American Academy of Pediatrics, 2012; Baby‐Friendly USA, I, 2016; Holmes et al., 2013)
Feeding breastfed infants A07, A13 Breast milk as the first feeding,c Use breast milk supplement only when medically necessary (American Academy of Pediatrics, 2012; Baby‐Friendly USA, Inc., 2016)
A20 Percent of infants supplemented with something other than breast milk.
Breastfeeding assistance A15 Of breastfeeding mothers, how many are taught breastfeeding techniques? Breastfeeding support is associated with improved duration and exclusivity of breastfeeding (McFadden et al., 2017)
A16 How many mothers are taught to recognize infant's first signs of hunger?
A18 How many mother–baby couples are directly observed and assessed by staff for breastfeeding effectiveness?
A17 How often do maternity care staff advise breastfeeding women to limit the length of suckling at each feeding? Limiting length of time on the breast can prevent infant from receiving high‐calorie hind milk (Riordan & Wambach, 2010)
A19 Do staff at facility use tool to assess breastfeeding effectiveness? Scoring tools such as LATCH can identify risk factors for early breastfeeding cessation (Kumar, Mooney, Wieser, & Havstad, 2006; Riordan, 2001)
A24 How many infants are given pacifiers? Pacifier use is associated with decreased EBF rates (Buccini, Perez‐Escamilla, Paulino, Araujo, & Venancio, 2017)
Mother–baby contact A08, A14 Following delivery, infants are routinely taken to nursery.b Preventing mother–baby separation in the hospital is essential to establishing successful breastfeeding (Crenshaw, 2014)
A31 Percentage of infants who remain with their mother: ≥23 hr per day. Rooming‐in is associated with longer duration of breastfeeding (Murray, Ricketts, & Dellaport, 2007)
Staff training B01 Hours nurses spend in breastfeeding training as new employees. Breastfeeding outcomes are related to health care education (Ward & Byrne, 2011)
B05

Hours staff members spent in breastfeeding education in the past year.
B03 How often are staff assessed for level of competency in breastfeeding management and support?
Structural and organizational aspects of care B11 Policy addressing 10 maternity care practice elements. A hospital breastfeeding policy is associated with better breastfeeding outcomes (Li et al., 2014)
A25 Facility receives free infant formula. Gift acceptance is associated with changes in prescribing habits and should be avoided (Lieb & Scheurich, 2014)
A01 Prenatal classes offered by staff or contracted personnel. Attending prenatal education classes is associated with significant increases in breastfeeding rates at 6 months (Rosen, Krueger, Carney, & Graham, 2008)
B10 Availability of lactation specialists. Access to international board‐certified lactation consultants during hospital stay is associated with breastfeeding rates (Castrucci et al., 2006)
A22 Percentage of infants supplemented with formula reason: Doctor's orders Hospital policies requiring providers orders for formula supplementation can help ensure the use of breast milk substitutes is medically necessary (Boban & Zakarija‐Grkovic, 2016; Kellams, Harrel, Omage, Gregory, & Rosen‐Carole, 2017)
C02 Percentage of live births at facility that were caesarean delivered in the past year. Caesarean births are associated with poor breastfeeding outcomes (Dewey et al., 2003; Hobbs et al., 2016)
C01 Number of live births at the facility. Relationship between volume and health outcomes (Halm, Lee, & Chassin, 2002)
Open in a new tab

Note. EBF, exclusive breastfeeding.

a

Domains were created by the Centers for Disease Control (CDC) and provide a structure to report data for individual maternity care practices measured in the mPINC survey.

b

Separate mPINC survey questions for vaginal verses caesarean deliveries.

c

A12 and A06 along with A13 and A07 were combined in Table 2 and reflect weighted averages based on hospitals' caesarean section and vaginal birth rates. Refer to the CDC's website for the complete 2013 mPINC survey (CDC, 2013b).

Table 2.

Partial results from simple linear regression models

Domain Survey question Predictor variable B ± 95% CIa Diff ± 95% CIb Adjc R 2
Feeding breastfed infants A20 Percentage of infants supplemented with something other than breast milk −0.347 ± 0.060d 0.149
A13/A07 Breast milk as first feeding after delivery 0.442 ± 0.087 0.120
Labour and delivery A12/06 Percentage of infants put to breast within 1 hr after vaginal birth or 2 hr after caesarean birth 0.361 ± 0.064d 0.142
A04 Routine newborn procedures after vaginal birth done while mother holds infant (almost always > rarely) 22.165 ± 6.658d 0.126
A11 How many mothers are encouraged to hold their infant for 30 min within 2 hr of caesarean birth? (most > few) 21.013 ± 6.727d 0.108
A05 How many mothers are encouraged to hold their infant for 30 min within 1 hr of vaginal birth? (most > few) 31.430 ± 16.398d 0.090
A10 Routine newborn procedures after caesarean births done while mother holds infant (almost always > rarely) 14.877 ± 6.438d 0.080
Mother–baby contact A08 Following vaginal birth, are babies taken to nursery? (yes < no) −15.373 ± 3.007d 0.120
A14 Following caesarean birth, are babies taken to nursery? (yes < no) −12.403 ± 2.585d 0.109
A31 Percentage of infants remain with mother: 23+ hr per day 0.115 ± 0.032d 0.067
Structural aspects of care A25 Facility receives free infant formula (yes < no) −11.882 ± 3.017d 0.076
C02 Percentage of live births at facility that were caesarean delivered in the past year −0.567 ± 0.179d 0.050
Baby‐friendly hospital designation (BFHI > non‐BFHI) 13.567 ± 4.619d 0.043
B11 Policy addressing 10 maternity care practice elements 1.536 ± 0.635d 0.030
B10 Availability of lactation specialist (7 days a week vs. not available) 8.635 ± 4.310d 0.028
A22 Percentage infants supplemented with formula reason: doctor's orders 0.110 ± 0.106d 0.021
A01 Prenatal classes offered by staff or contracted personnel (yes > no) 11.722 ± 7.048d 0.013
C01 Number live births at the facility (per 1,000 births) −0.006 ± 0.009 0.001
Staff training B01

Hours nurses spend on breastfeeding training as new employees

(18 or more hours > less than 1 hr)

 14.772 ± 10.361 0.038
B05

Hours staff members spent in breastfeeding education in the past year

(3–4 hr > less than 1 hr)

 5.407 ± 6.896 0.001
B03 How often are staff assessed for level of competency in breastfeeding management and support? (less than once a year > at least once a year) 0.535 ± 4.492 −0.003
Breastfeeding assistance A19

Do staff at facility use a tool to assess breastfeeding effectiveness?

(no < yes)

 −9.736 ± 3.464d 0.039
A17 How often do maternity care staff advise breastfeeding women to limit length of feeding? (often < rarely) −8.199 ± 5.997 0.027
A24 How many infants are given pacifiers? (many < few) −7.200 ± 4.778d 0.019
A18 How many mother–baby couples are directly observed by staff for breastfeeding effectiveness? (most > many) 5.790 ± 5.435 0.013
A15 Of breastfeeding mothers, how many are taught breastfeeding techniques? (most > some/many) 6.931 ± 6.043 0.005
A16 How many mothers are taught to recognize infant's first signs of hunger? (most > many) 4.891 ± 5.521 0.007
Demographics Proportion without college attendance −0.587 ± 0.113d 0.125
Proportion African American −0.346 ± 0.072d 0.108
Proportion White 0.222 ± 0.051d 0.090
Proportion with bachelor's degree 0.341 ± 0.100d 0.057
Proportion living below poverty line −0.233 ± 0.069d 0.056
Proportion Hispanic −0.170 ± 0.078d 0.020
Proportion Asian 0.076 ± 0.155 0.000
Open in a new tab

Note. Exclusive breastfeeding rate was the response variable for all models. P values for all pairwise comparisons of the individual levels of categorical variables and for the overall analysis of variance (ANOVA) test are listed in Table S1. Results listed are unadjusted for covariates. Refer to the CDC's website for the complete 2013 mPINC survey (CDC, 2013b). Questions are specific to healthy full‐term infants following an uncomplicated vaginal or caesarean birth.

a

B = regression coefficient for the continuous variables.

b

Diff = largest pairwise difference between category means from the Tukey multiple comparisons correction.

c

Adj R 2 = adjusted R 2 for continuous variables was obtained from the linear regression model and from an ANOVA for the categorical variables.

d

P value is <0.001.

2.3. Statistical analysis

Data analysis was performed with the R statistical package (Version 3.1.1) using RStudio© software (Version 0.99.902) as an interface (R Core Team, 2013). All potential predictors of EBF rate (maternity care practices, demographic characteristics, BFHI) will be referred to as predictor variables. First, separate simple linear regression models were fit with EBF as the response variable and one predictor variable. When the predictor variable was categorical, this took the form of an analysis of variance with post hoc Tukey corrected pairwise comparisons of means.

Then, we used multiple linear regression to determine how much variability in EBF rates could be explained by a collection of variables. These multiple predictor variable models were selected using two different selection criteria. The first model used the Akaike information criterion (AIC) to determine which predictor variables were removed. The variable that lowered the AIC the most at each step was removed. The second model used the least significant P value as the criterion. For both criteria, the model selected was refit at each step with the removed variable eliminated. This process was repeated until the AIC could no longer be reduced by removing a variable (AIC method), or all variables had a P value of <0.05 (P value method). For these model building procedures, potential predictor variables included 26 maternity care practices, the BFHI designation, caesarean section rate, number of births, and hospital demographic data. An exhaustive model search that would have fit every model with every possible combination of predictor variables would have been preferred but was impractical given the large number of potential predictor variables. All models fit were checked graphically for adherence to the assumptions of linear regression. Normality of the residuals was checked using normal Q–Q plots, and constant variance was checked using residuals versus fitted values plots.

3. RESULTS

The final data set included hospitals from all 50 states. The distribution of hospitals that responded to the 2013 mPINC survey was similar to our sample with the exception of California, which was higher by percent in our sample than the mPINC sample. Among the hospitals retained in the data set for analysis, the highest percentage was from California (12%) followed by Texas (7%), Ohio (5%), Florida (5%), and New York (5%). All other states each contributed less than 5%.

Table 2 summarizes the relationship between individual maternity practices measured in the mPINC survey, BFHI designation, and demographic characteristics, and their relationship with EBF rates. Most predictor variables studied were significantly related to EBF rates with a range in adjusted R 2 of 1.3% to 14.9%. The predictor variables in each domain are ranked on the basis of the adjusted R 2. The adjusted R 2 represents each predictor variable's ability to explain the variability in EBF using simple linear regression. The highest overall adjusted R 2 was for mPINC Survey Question A20 that asked facilities what percentage of healthy full‐term infants are supplemented with something other than breast milk (adjusted R 2 = 14.9%). For the continuous variables, the regression coefficient (B) represents the percent change in EBF rates for each 1% change in the predictor variable. For example, for every 1% increase in hospitals reporting breast milk as first feeding after delivery, mPINC Survey Question A13/A07, there was a 0.4% increase in EBF rates. By contrast, for each 1% increase in hospitals reporting infants supplemented with something other than breast milk, mPINC Survey Question A20, there was a 0.3% decrease in EBF rates. All continuous predictor variables included in our analysis were significantly related to EBF rates (P < 0.001) with the exception of the number of live births at each hospital and the proportion of the population who identifies as Asian. For the categorical predictor variables, we reported only the largest category EBF mean differences (Diff) with the lowest P value. For example, hospitals with the BFHI designation had 13.6% higher EBF rates than had hospitals without the BFHI designation. By contrast, hospitals that reported receiving free infant formula had 11.9% lower EBF rates than had hospitals that reported they did not receive free infant formula. Almost all categorical variables were significant with the exception of several maternity care practices in the staff training and the breastfeeding assistance domains. For more in‐depth information on data for each categorical level, please see Table S1. Table 3 summarizes the maternity care practices that remained in the multiple linear regression models selected using each criterion, AIC or P value. The multiple linear regression models selected using the two different selection criteria were largely similar with respect to the predictor variables that remained. The adjusted R 2 for the model selected using AIC was 47.3%. The adjusted R 2 for the model selected using P value was 46.9%. To determine if our ability to predict the variability in EBF rates relied too heavily upon the maternity care practices most directly related to breastfeeding, we again selected a model using AIC, but excluded mPINC Survey Questions A13/07, A12/06, and A20 as potential predictor variables. The resulting model had an adjusted R 2 of 43.5%, a reduction of less than 5% in explanatory ability (model results not shown).

Table 3.

Maternity care practices remaining in the final multiple regression models chosen using two different model selection criteria

Survey number Adjusted R 2 (%)
Model 1a 47.3
A20 Percentage of infants supplemented with something other than breast milk
A13/A07 Breast milk as first feeding after delivery
A12/A06 Percentage of infants put to breast within 1 hr of vaginal birth or 2 hr after caesarean birth
A04 Routine procedures after vaginal birth done while mother holds infant
A31 Percentage of infants who remain with mothers: 23+ hr per day
A08 Following vaginal birth, are babies taken to the nursery?
A14 Following caesarean birth, are babies taken to the nursery?
A22 Percentage of infants supplemented with infant formula: Doctor's orders
A25 Facility receives free infant formula
A01 Prenatal classes offered by staff or contracted personnel
B03 How often are staff assessed for level of competency in breastfeeding management and support?
A19 Do staff at facility use a tool to assess breastfeeding effectiveness?
Model 2b 46.9
All variables from above with the addition of the following:
A10 Routine newborn procedures after caesarean births are done while mother holds infant
B05 Hours staff members spent in breastfeeding education in the past year
Open in a new tab

Note. Both models contained additional predictor variables pertaining to hospital demographics (not shown).

a

Model 1 was selected using Akaike information criterion as the criterion.

b

Model 2 was selected using P value as the criterion.

4. DISCUSSION

To our knowledge, this is the first study to evaluate the relationship between individual maternity care practices and EBF rates, using a standardized in‐hospital feeding definition of EBF, applied across a large sample of hospitals in the United States. This study found that most individual maternity care practices measured in the 2013 mPINC survey and included in our analysis were significantly related to EBF rates. This finding supports that the mPINC survey can be a useful tool for monitoring and assessing individual maternity care practices in order to positively influence EBF rates in birthing facilities. We found that directions of effects for statistically significant maternity care practices were consistent with the literature. For example, we found higher EBF rates when hospitals reported that most versus few mothers were encouraged to hold their infant for 30 min following a caesarean section or vaginal delivery. This is consistent with study findings that show maternity care practices that promote skin‐to‐skin contact have been associated with better breastfeeding outcomes (Moore, Bergman, Anderson, & Medley, 2016). Similarly, we found a positive correlation between breast milk as first feeding following a caesarean section or vaginal birth and EBF rates in the hospital. This is what we would expect on the basis of mothers' physiological response to infants suckling at the breast and the establishment of milk production (Pang & Hartmann, 2007), and research on the relationship between breastfeeding initiation within 1 hr of birth and EBF rates (Perrine, Scanlon, Li, Odom, & Grummer‐Strawn, 2012). Alternatively, maternity care practices measured in the mPINC survey that create barriers to the establishment of a milk production, such as the percentage of infants supplemented with something other than breast milk, were negatively associated with EBF rates. This is consistent with a study that found infants who were supplemented with something other than breast milk in the hospital were less likely to be breastfed at 6 weeks of age (DiGirolamo et al., 2008). The findings related to hospital demographics and EBF rates were consistent with our previous study (Patterson, Keuler, & Olson, 2017). When comparing the adjusted R 2 of the individual maternity practices, we found a higher adjusted R 2 with the maternity care practices measured in the feeding breastfed infants, labour and delivery, and mother–baby contact domains than in the staff training domain. This is what we would expect given the moderate‐ to high‐quality evidence supporting maternity care practices such as skin‐to‐skin contact following birth, initiating breastfeeding within an hour of delivery, and preventing maternal and infant separation, compared with the very low‐quality evidence supporting staff training published in the new breastfeeding guidelines for facilities providing maternity and newborn care (World Health Organization, 2017). Although information on the relationship between individual maternity care practices and EBF rates may be useful in prioritizing changes, it is the collective impact of these practices that we found had the greatest potential to influence EBF rates.

Consistent with the BFHI guidelines, our study reinforces the need for a systematic approach to support breastfeeding in the hospital (World Health Organization, 2017). We found that a collection of maternity care practices was better able to explain the variability in EBF rates than was any one individual maternity care practice. Research has shown a relationship between the number of BFHI practices a mother is exposed to and breastfeeding outcomes (DiGirolamo et al., 2008). Using both model selection criteria, we were able to build a model that explained nearly half of the variability in EBF rates. The majority of this explanatory ability was not primarily due to the maternity care practices directly pertaining to breastfeeding, such as mPINC Survey Question A13/A07 regarding breast milk as first feeding following delivery. This suggests that maternity care practices not directly related to breastfeeding also influence a mother's choice to exclusively breastfeed. Neither one of the multiple linear regression models selected using each of the criteria included BFHI designation. This was not surprising as the BFHI designation was likely represented by some collection of maternity care practices already in the model. The key message from the multiple variable models is that a collection of maternity care practices was able to predict the variability in EBF rates better than any single practice, and no single maternity care practice dominated in the final model. Given the complicated nature of women's decision to breastfeed, we found it remarkable that we were able to explain nearly half of the variability in EBF rates using measurable maternity care practices while accounting for hospital demographic characteristics.

All mothers and infants have the right to high‐quality health care services in order to achieve optimal breastfeeding outcomes (World Health Organization, 2017). An unintended finding in this study was the inconsistencies in hospitals' report of maternity care practices. This is especially pronounced regarding the difference in care experienced by mothers who deliver via caesarean section versus vaginally. For instance, most hospitals in our sample reported routine newborn procedures after a vaginal delivery are completed often or almost always while a mother holds the infant. In contrast, most hospitals in our sample reported routine procedures are completed sometimes or rarely following caesarean delivery. Caesarean births are associated with poor breastfeeding outcomes (Dewey, Nommsen‐Rivers, Heinig, & Cohen, 2003; Hobbs, Mannion, McDonald, Brockway, & Tough, 2016). Our study supports this as we found that for every 1% increase in hospitals' reported caesarean rate, there was a 0.6% decrease in EBF rates, which was significant. Although many factors may be related to hospitals' ability to prevent mother–baby separation following a caesarean delivery, our findings support that the few hospitals in our data set that reported almost always versus rarely completed newborn procedures while mothers hold their infant were associated with EBF rates that were 15% higher (P < 0.001). Identifying opportunities to optimize breastfeeding care for mothers despite delivery modality may be an important strategy to positively influence breastfeeding outcomes.

The implications of this study relative to the newly revised Ten Steps to Successful Breastfeeding (World Health Organization, 2018a, 2018b) are with respect to the way in which the Ten Steps are operationalized, monitored, and assessed. The mPINC survey is a valuable tool to assist hospitals in assessing and monitoring their level of breastfeeding care along with helping hospitals align with the BFHI guidelines. Although there have been large increases in the percentage of hospitals in the U.S. that are BFHI certified, the majority of infants are being delivered in non‐BFHI hospitals (Baby‐Friendly USA, Inc., 2018a). Given the evidence around the BFHI and breastfeeding outcomes (Perez‐Escamilla et al., 2016), public health initiatives and legislation aimed at increasing the number of BFHI hospitals should remain a high priority (Feldman‐Winter et al., 2017; National Conference of State Legislatures, 2017). In addition, expansion of hospital regulatory agencies to evaluate maternity care practices as a part of accreditation may help ensure that all mothers have access to high‐quality maternity care in order to optimize EBF rates.

A strength of this study is that it was a national hospital sample that utilized a standardized definition of EBF in the hospital. We were also able to include a large number of maternity care practices measured in the mPINC survey and explore their relationship with EBF rates in the hospital. Limitations of this study include that U.S. census data for the community surrounding the hospital were used as a proxy for the demographics actually served by the hospital. In addition, hospitals that were marked as non‐BFHI hospitals may have been in various stages of obtaining designation, and therefore, a dichotomous variable may not have fully represented the true impact of the BFHI hospital designation. Lastly, there are potential limitations with data obtained from the mPINC survey. One representative from each hospital is selected to complete the mPINC survey. To ensure that the information obtained by the hospital representative is accurate, the CDC takes additional steps to make sure that the survey is delivered to the person who is most knowledgeable about the maternity care and infant feeding practices (CDC, 2017a).

Future research might explore differences in hospitals implementation of various maternity care practices, especially those strongly associated with better breastfeeding outcomes. A study that evaluates hospitals' ability to implement certain maternity practices but not others may be valuable in identifying solutions for overcoming barriers to providing evidenced‐based care. In addition, qualitative research that explores mothers' experience with the BFHI would be a useful way of evaluating how the programme, from a mother's experience, is effective.

5. CONCLUSION

The mPINC survey is a valuable tool to assist birthing hospitals in identifying opportunities to improve individual maternity care practices in order to positively influence EBF rates. However, consistent with the BFHI, our study findings support that no one maternity care practice is as good as a collection of practices in explaining EBF rates in the hospital. Considering the complicated nature of mothers' decision to exclusively breastfeed, we were able to create a model using a collection of measurable items that was able to predict almost half of the variability in EBF rates. Studies that evaluate barriers and facilitators to implementing evidenced‐based maternity care practices, especially across both delivery modalities, would be valuable in identifying solutions for overcoming barriers to evidenced‐based care.

CONFLICTS OF INTEREST

The authors declare that they have no conflicts of interest.

CONTRIBUTIONS

JAP led the design of the study, statistical analysis, writing, revising, and final approval of manuscript. NSK guided the statistical analysis and contributed to writing and revising as well as final approval of the manuscript. BHO supervised the overall study and contributed to the design, analysis, writing, revising, and final approval of the document.

Supporting information

Supplemental Table S1: Complete results from simple linear regression models. Exclusive breastfeeding rate is the response variable for all models.

Click here for additional data file. (30.6KB, docx)

ACKNOWLEDGMENTS

We thank W. R. Buckingham from the Applied Population Laboratory, University of Wisconsin–Madison, for assistance in obtaining and interpreting census data, and K. Woelky, Department of Nutritional Sciences, University of Wisconsin–Madison, for technical assistance. The authors also acknowledge the CDC's Division of Nutrition, Physical Activity and Obesity for assistance in obtaining the mPINC data.

Patterson JA, Keuler NS, Olson BH. The effect of maternity practices on exclusive breastfeeding rates in U.S. hospitals. Matern Child Nutr. 2019;15:e12670 10.1111/mcn.12670

REFERENCES

  1. American Academy of Pediatrics (2012). Breastfeeding and the use of human milk. Pediatrics, 129(3), e827–e841. 10.1542/peds.2011-3552 [DOI] [PubMed] [Google Scholar]
  2. Baby‐Friendly USA, Inc. (2016). Guidelines and evaluation criteria for facilities seeking baby‐friendly designation. Retrieved from https://www.babyfriendlyusa.org/get-started/the-guidelines-evaluation-criteria. June 20, 2018.
  3. Baby‐Friendly USA, Inc. (2018a). Baby Friendly USA. Find facilities—Designated facilities by state. Retrieved from https://www.babyfriendlyusa.org/find-facilities. June 12, 2018.
  4. Baby‐Friendly USA, Inc. (2018b). Baby‐Friendly Hospital Initiative—About us. Retrieved from https://www.babyfriendlyusa.org/about-us. June 20, 2018.
  5. Boban, M. , & Zakarija‐Grkovic, I. (2016). In‐hospital formula supplementation of healthy newborns: Practices, reasons, and their medical justification. Breastfeeding Medicine, 11, 448–454. 10.1089/bfm.2016.0039 [DOI] [PubMed] [Google Scholar]
  6. Buccini, G. D. S. , Perez‐Escamilla, R. , Paulino, L. M. , Araujo, C. L. , & Venancio, S. I. (2017). Pacifier use and interruption of exclusive breastfeeding: Systematic review and meta‐analysis. Maternal & Child Nutrition, 13(3). 10.1111/mcn.12384 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Castrucci, B. C. , Hoover, K. L. , Lim, S. , & Maus, K. C. (2006). A comparison of breastfeeding rates in an urban birth cohort among women delivering infants at hospitals that employ and do not employ lactation consultants. Journal of Public Health Management and Practice, 12(6), 578-585. [DOI] [PubMed] [Google Scholar]
  8. Centers for Disease Control and Prevention (2013a). Maternity Practices in Infant Nutrition and Care survey. Accessed with permission November 2015.
  9. Centers for Disease Control and Prevention (2013b). Survey: Maternity care practices. Retrieved from https://www.cdc.gov/breastfeeding/data/mpinc/questionnaires.htm. June 20, 2018.
  10. Centers for Disease Control and Prevention (2017a). Methods: Maternity care practices. Retrieved from https://www.cdc.gov/breastfeeding/data/mpinc/methods.html. June 20, 2018.
  11. Centers for Disease Control and Prevention (2017b). Overview of the mPINC scoring algorithm. Retrieved from https://www.cdc.gov/breastfeeding/data/mpinc/pdf/overview-of-the-mpinc-scoring-algorithm.pdf. June 20, 2018.
  12. Centers for Disease Control and Prevention (2018). Maternity Practices in Infant Nutrition and Care (mPINC) survey. Retrieved from https://www.cdc.gov/breastfeeding/data/mpinc/index.htm. June 20, 2018.
  13. Crenshaw, J. T. (2014). Healthy Birth Practice #6: Keep mother and baby together— It's best for mother, baby, and breastfeeding. The Journal of Perinatal Education, 23(4), 211–217. 10.1891/1058-1243.23.4.211 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. 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(3 Pt 1), 607–619. [DOI] [PubMed] [Google Scholar]
  15. DiGirolamo, A. M. , Grummer‐Strawn, L. M. , & Fein, S. B. (2008). Effect of maternity‐care practices on breastfeeding. Pediatrics, 122(Suppl 2), S43–S49. 10.1542/peds.2008-1315e [DOI] [PubMed] [Google Scholar]
  16. Feldman‐Winter, L. , Ustianov, J. , Anastasio, J. , Butts‐Dion, S. , Henrich, P. , Merewood, A. , … Homer, C. (2017). Brest fed beginnings: A nationwide quality improvement initiative to increase breastfeeding. Pediatrics, 140(1), e1–e9. [DOI] [PubMed] [Google Scholar]
  17. Friedman, S. A. , Frayne, S. M. , Berg, E. , Hamilton, A. B. , Washington, D. L. , Saechao, F. , … Phibbs, C. S. (2015). Travel time and attrition from VHA care among women veterans: How far is too far? Medical Care, 53(4 Suppl 1), S15–S22. 10.1097/mlr.0000000000000296 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Grossniklaus, D. A. , Perrine, C. , Macgowan, C. , Scanlon, K. S. , Shealy, K. R. , Murphy, P. , … Grummer‐Strawn, L. M. (2017). Participation in a quality improvement collaborative and change in maternity care practices. The Journal of Perinatal Education, 26(3), 136–143. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Halm, E. A. , Lee, C. , & Chassin, M. R. (2002). Is volume related to outcome in health care? A systematic review and methodologic critique of the literature. Annals of Internal Medicine, 137(6), 511–520. [DOI] [PubMed] [Google Scholar]
  20. Hobbs, A. J. , Mannion, C. A. , McDonald, S. W. , Brockway, M. , & Tough, S. C. (2016). The impact of caesarean section on breastfeeding initiation, duration and difficulties in the first four months postpartum. BMC Pregnancy and Childbirth, 16, 90 10.1186/s12884-016-0876-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Holmes, A. V. , McLeod, A. Y. , & Bunik, M. (2013). ABM Clinical Protocol #5: Peripartum breastfeeding management for the healthy mother and infant at term, revision 2013. Breastfeeding Medicine, 8(6), 469–473. 10.1089/bfm.2013.9979 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Ip, S. , Chung, M. , Raman, G. , Chew, P. , Magula, N. , DeVine, D. , … Lau, J. (2007). Breastfeeding and maternal and infant health outcomes in developed countries. Evid Rep Technol Assess (Full Rep) (153), 1‐186. [PMC free article] [PubMed]
  23. Kellams, A. , Harrel, C. , Omage, S. , Gregory, C. , & Rosen‐Carole, C. (2017). ABM Clinical Protocol #3: Supplementary feedings in the healthy term breastfed neonate, revised 2017. Breastfeeding Medicine, 12, 188–198. 10.1089/bfm.2017.29038.ajk [DOI] [PubMed] [Google Scholar]
  24. Kozhimannil, K. B. , Casey, M. M. , Hung, P. , Prasad, S. , & Moscovice, I. S. (2016). Location of childbirth for rural women: Implications for maternal levels of care. American Journal of Obstetrics and Gynecology, 214(5). 661.e661‐661.e610. doi: 10.1016/j.ajog.2015.11.030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kumar, S. P. , Mooney, R. , Wieser, L. J. , & Havstad, S. (2006). The LATCH scoring system and prediction of breastfeeding duration. Journal of Human Lactation, 22(4), 391–397. 10.1177/0890334406293161 [DOI] [PubMed] [Google Scholar]
  26. Li, C. M. , Li, R. , Ashley, C. G. , Smiley, J. M. , Cohen, J. H. , & Dee, D. L. (2014). Associations of hospital staff training and policies with early breastfeeding practices. Journal of Human Lactation, 30(1), 88–96. 10.1177/0890334413484551 [DOI] [PubMed] [Google Scholar]
  27. Lieb, K. , & Scheurich, A. (2014). Contact between doctors and the pharmaceutical industry, their perceptions, and the effects on prescribing habits. PLoS One, 9(10), e110130 10.1371/journal.pone.0110130 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. McFadden, A. , Gavine, A. , Renfrew, M. J. , Wade, A. , Buchanan, P. , Taylor, J. L. , … MacGillivray, S. (2017). Support for healthy breastfeeding mothers with healthy term babies. Cochrane Database of Systematic Reviews, 2, Cd001141 10.1002/14651858.CD001141.pub5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Moore, E. R. , Bergman, N. , Anderson, G. C. , & Medley, N. (2016). Early skin‐to‐skin contact for mothers and their healthy newborn infants. Cochrane Database of Systematic Reviews, 11, Cd003519 10.1002/14651858.CD003519.pub4 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Murray, E. K. , Ricketts, S. , & Dellaport, J. (2007). Hospital practices that increase breastfeeding duration: Results from a population‐based study. Birth, 34(3), 202–211. 10.1111/j.1523-536X.2007.00172.x [DOI] [PubMed] [Google Scholar]
  31. National Conference of State Legislatures (2017). Breastfeeding state laws. Retrieved from http://www.ncsl.org/research/health/breastfeeding-state-laws.aspx. June 20, 2018.
  32. Office of Disease Prevention and Health Promotion (2018). MICH‐21.5 Increase the proportion of infants who are breastfed exclusively through 6 months. Retrieved from https://www.healthypeople.gov/2020/data-search/Search-the-Data#objid=4863;. June 20, 2018.
  33. Pang, W. W. , & Hartmann, P. E. (2007). Initiation of human lactation: Secretory differentiation and secretory activation. Journal of Mammary Gland Biology and Neoplasia, 12(4), 211–221. 10.1007/s10911-007-9054-4 [DOI] [PubMed] [Google Scholar]
  34. Patterson, J. A. , Keuler, N. S. , & Olson, B. H. (2017). The effect of the baby friendly hospital initiative on exclusive breastfeeding rates in US hospitals. Maternal & Child Nutrition, 14, e12589 10.1111/mcn.12589 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Perez‐Escamilla, R. , Martinez, J. L. , & Segura‐Perez, S. (2016). Impact of the Baby‐friendly Hospital Initiative on breastfeeding and child health outcomes: A systematic review. Maternal & Child Nutrition, 12(3), 402–417. 10.1111/mcn.12294 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Perrine, C. , Galuska, D. , Dohack, J. , Shealy, K. , Murphy, P. , Grummer‐Strawn, L. M. , & Scanlon, K. S. (2015). Vital signs: Improvements in maternity care policies and practices that support breastfeeding—United States, 2007–2013. Retrieved from Atlanta, GA: https://www.cdc.gov/mmwr/preview/mmwrhtml/mm6439a5.htm?s_cid=mm6439a5_w. June 20, 2018. [DOI] [PubMed]
  37. Perrine, C. G. , Scanlon, K. S. , Li, R. , Odom, E. , & Grummer‐Strawn, L. M. (2012). Baby‐friendly hospital practices and meeting exclusive breastfeeding intention. Pediatrics, 130(1), 54–60. 10.1542/peds.2011-3633 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. R Core Team (2013). R: A language and environment for statistical computing (Version 3.1.1) . Vienna, Austria: R Foundation for Statistical Computing; Retrieved from http://www.r-project.org. June 20, 2018 [Google Scholar]
  39. Riordan, J. , Bibb, D. , Miller, M. , & Rawlins, T. (2001). Predicting breastfeeding duration using the LATCH breastfeeding assessment tool. Journal of Human Lactation, 17(1), 20–23. 10.1177/089033440101700105 [DOI] [PubMed] [Google Scholar]
  40. Riordan, J. , & Wambach, K. (2010). Breastfeeding and human lactation ( ed., Vol. Four). Sudbury, MA: Jones and Bartlett Publishers. [Google Scholar]
  41. Rollins, N. C. , Bhandari, N. , Hajeebhoy, N. , Horton, S. , Lutter, C. K. , Martines, J. C. , … Victora, C. G. (2016). Why invest, and what it will take to improve breastfeeding practices? Lancet, 387(10017), 491–504. 10.1016/s0140-6736(15)01044-2 [DOI] [PubMed] [Google Scholar]
  42. Rosen, I. M. , Krueger, M. V. , Carney, L. M. , & Graham, J. A. (2008). Prenatal breastfeeding education and breastfeeding outcomes. MCN: American Journal of Maternal Child Nursing, 33(5), 315–319. 10.1097/01.NMC.0000334900.22215.ec [DOI] [PubMed] [Google Scholar]
  43. Stevens, J. , Schmied, V. , Burns, E. , & Dahlen, H. (2014). Immediate or early skin‐to‐skin contact after a caesarean section: A review of the literature. Maternal & Child Nutrition, 10(4), 456–473. 10.1111/mcn.12128 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. The Joint Commission (2013). Improving performance on perinatal care measures. Benchmark, 11(7), 16–19. Retrieved from http://www.jointcommission.org/assets/1/6/S7_TS_V11_N7.pdf. June 20, 2018. [Google Scholar]
  45. The Joint Commission (2014). Specifications manual for Joint Commission National Quality Measures (v2014A1)—Data element name—Exclusive breast milk feeding. Exclusive Breast Milk Feeding Retrieved from https://manual.jointcommission.org/releases/TJC2014A1/DataElem0273.html. June 20, 2018.
  46. The Joint Commission (2015a). Specifications Manual for Joint Commission National Quality Measures (v2015A1). Retrieved from https://manual.jointcommission.org/releases/TJC2015A1/MIF0170.html. June 20, 2018.
  47. The Joint Commission (2015b). Specifications manual for Joint Commission National Quality Measures (v2015A)—Data element name exclusive breast milk feeding. Exclusive Breast Milk Feeding Retrieved from https://manual.jointcommission.org/releases/TJC2015A1/DataElem0273.html. June 20, 2018.
  48. The Joint Commission (2016). The Joint Commission health care quality data download: Performance measure results. Perinatal care: Exclusive breast milk feeding. Retrieved from: http://www.healthcarequalitydata.org/ (Registration Required)
  49. United States Census Bureau (2014). American community survey: 5‐year data [2010–2014, block groups & larger areas]. Retrieved from: https://data 2.nhgis.org/main
  50. Victora, C. G. , Bahl, R. , Barros, A. J. , Franca, G. V. , Horton, S. , Krasevec, J. , … Rollins, N. C. (2016). Breastfeeding in the 21st century: Epidemiology, mechanisms, and lifelong effect. Lancet, 387(10017), 475–490. 10.1016/s0140-6736(15)01024-7 [DOI] [PubMed] [Google Scholar]
  51. Ward, K. N. , & Byrne, J. P. (2011). A critical review of the impact of continuing breastfeeding education provided to nurses and midwives. Journal of Human Lactation, 27(4), 381–393. 10.1177/0890334411411052 [DOI] [PubMed] [Google Scholar]
  52. World Health Organization (1989). Protecting, promoting and supporting breast‐feeding: The special role of maternity services. A Joint WHO/UNICEF Statement. Retrieved from Switzerland: http://apps.who.int/iris/bitstream/10665/39679/1/9241561300.pdf?ua=1&ua=1 June 20, 2018.
  53. World Health Organization (2017). Guideline: Protecting, promoting and supporting breastfeeding in facilities providing maternity and newborn services. Retrieved from Geneva, Switzerland: http://apps.who.int/iris/bitstream/10665/259386/1/9789241550086-eng.pdf?ua=1. June 20, 2018. [PubMed]
  54. World Health Organization (2018a). Annex 1. Ten Steps to Successful Breastfeeding—Revised 2018 version: comparison to the original Ten Steps and the new 2017 WHO guideline. 44–46.
  55. World Health Organization (2018b). Implementation guidance: Protecting, promoting and supporting breastfeeding in facilities providing maternity and newborn services—The revised Baby‐friendly Hospital Initiative. Geneva.

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Table S1: Complete results from simple linear regression models. Exclusive breastfeeding rate is the response variable for all models.

Click here for additional data file. (30.6KB, docx)

Articles from Maternal & Child Nutrition are provided here courtesy of Wiley

RESOURCES