Maternity Care Practices and Their Role in U.S. Breastfeeding Disparities

Abstract

Introduction:

Breastfeeding provides substantial health benefits for both children and mothers; yet the U.S. rates remain suboptimal, with disparities shaped by structural, social, and policy factors. This study examines how the U.S. hospital maternity care practices influence breastfeeding initiation, with particular attention to their impact across different population groups.

Methods:

We analyzed cross-sectional county-level variations in hospital maternity care quality and breastfeeding initiation from 2017 to 2022 using data from the National Vital Statistics System and the Maternity Practices in Infant Nutrition and Care (mPINC) surveys. We employed a linear probability model to assess these relationships.

Results:

Higher county mPINC scores are significantly associated with increased breastfeeding initiation, with each additional point linked to a 0.10 percentage point (pp) increase (p < 0.001). This association varies by race/ethnicity. Each additional mPINC point corresponds to a 0.25 pp increase for non-Hispanic Black mothers (p < 0.001) and a 0.14 pp increase for non-Hispanic American Indian/Alaska Native mothers (p < 0.001), approximately three and two times higher, respectively, than the increase for non-Hispanic White mothers. The effect of better maternity practices also differs by county type, with a 0.08 pp increase in metro areas (p < 0.001) and a 0.17 pp increase in nonmetro areas (p < 0.001).

Conclusions:

Higher quality hospital maternity care practices are associated with increased breastfeeding initiation, particularly among population groups with historically lower breastfeeding rates. Enhancing maternity care policies and practices may help reduce long-standing breastfeeding disparities.

Introduction

Breastfeeding, a nearly universally available practice, significantly enhances both maternal and child health.¹ While more than 85% of American mothers initiate breastfeeding, disparities persist, with lower initiation rates among economically disadvantaged populations and notable variations across racial and ethnic groups. Non-Hispanic White (NHW) and non-Hispanic Asian (NHA) mothers have higher initiation rates, while lower rates are observed among non-Hispanic Black (NHB) and non-Hispanic American Indian or Alaskan Native (NHAI/AN) mothers.^2,3 Structural, social, and policy factors, such as inadequate community or workplace lactation support,⁴ lack of or limited paid maternity leave, and extensive marketing of infant formula⁵ contribute to these disparities.

Effective approaches to reducing breastfeeding gaps are essential for equitable access to breastfeeding benefits. Supportive maternity care practices in the U.S. hospitals can improve breastfeeding outcomes,⁶ but access to such care is often unequal, with less support available in rural areas or areas with a higher proportion of NHB residents.^7,8 These factors contribute to notable geographic disparities in breastfeeding initiation rates, such as those in the Delta and Appalachian regions, exacerbating other structural, social, and policy inequities.^2,9–11 Prior research on hospital-based maternity care practices focused on their overall impact, with limited consideration of variations in breastfeeding across population groups or rural/urban areas.^6,12–14 Exceptions include a study of hospitals in Mississippi, Louisiana, Tennessee, and Texas, which found a decrease in racial disparities in breastfeeding initiation in hospitals that adhered to the Ten Steps to Successful Breastfeeding.¹⁵ Another study found that hospitals in neighborhoods with a high percentage of Black residents and poverty had lower in-hospital exclusive breastfeeding (EBF) rates. Baby-friendly designation did not reduce EBF disparities between neighborhoods with low versus high percentage of Black residents, while it was associated with reduced EBF disparities between neighborhoods with low versus high relative poverty status.¹⁶

This study examines the relationship between the U.S. hospital maternity care practices at the county level and breastfeeding initiation rates, focusing on variations across population groups. Specifically, it aims to: (1) evaluate associations between hospital-based maternity care practices and overall breastfeeding initiation and (2) test for differences in these associations among population groups.

Methods

Data

Data for this analysis was obtained from multiple national sources, including the National Vital Statistics System (NVSS), the Centers for Disease Control and Prevention’s (CDC) biennial Maternity Practices in Infant Nutrition and Care Survey (mPINC), the U.S. Department of Agriculture’s 2023 Rural-Urban Continuum Codes (RUCC), and the U.S. Census Bureau’s Small Area Income and Poverty Estimates (SAIPE). The restricted-use NVSS provides a detailed record of all live births in the United States, including county identifiers for each infant’s birth location.^17,18 The mPINC survey assesses infant feeding along with maternity care policies and practices through a biennial census of all the U.S. hospitals with maternity care; restricted-use data includes five-digit hospital zip code for hospital locations. Participation in the survey is voluntary, but the high response rate enables data collection for 70–75% of eligible U.S. hospitals. Detailed data collection methods of the mPINC survey are described elsewhere.^19,20 The RUCC dataset categorizes metropolitan counties by population size of their metropolitan area and nonmetropolitan counties by the degree of urbanization and adjacency to metropolitan areas.²¹ Finally, the SAIPE produces single-year estimates of household income and poverty at the county level.²²

Measures

The main variable of interest was hospital policies and practices supportive of breastfeeding, as measured by the mPINC scores. The mPINC survey covers six subdomains of maternity care services, including immediate postpartum care, rooming-in, feeding practices, feeding education and support, discharge support, and institutional management, which were scored from 0 to 100 to generate six subdomain scores. These were averaged to calculate a total mPINC score for every participating hospital, ranging from 0 to 100, with higher scores indicating better maternity care practices and policies.²³ Given the biennial nature of the mPINC survey, the 2018 mPINC score was assigned in the birth certificate NVSS data to the 2017–2018 period, 2020 to the 2019–2020 period, and 2022 to the 2021–2022 period (Supplementary Appendix, Section 1). We used five-digit hospital zip code identifiers to assign hospitals to their counties and calculated the county-level weighted average mPINC score based on the survey-reported number of births at each hospital (Supplementary Appendix, Section 2). Due to nonresponse, the data were not consistently available for all hospitals across the three biennial waves. As a result, hospitals contributing to the county weighted average mPINC score vary across periods. The county composition also varies over time (Supplementary Appendix, Section 3).

Our analysis of the 2017–2022 NVSS data focused on infants born in hospitals who were alive when their birth certificate was completed, not transferred to another facility within 24 hours of birth, with mothers not admitted to an intensive care unit, and with complete covariate data. California, Michigan, and Utah were excluded due to lack of breastfeeding data or data quality concerns.²⁴ We linked infant-level data from the NVSS to county-level average mPINC scores using the year and county of birth, as the NVSS does not include hospital-specific information (n = 15,719,673 infants in 1,386 counties; Supplementary Appendix, Supplementary Fig. A1).

The outcome was breastfeeding initiation, determined from a birth certificate’s item about the infant’s breastfeeding status at discharge. Data were extracted by the clerk completing the birth certificate from medical records.^25,26 Race was derived from mothers’ self-reported race and ethnicity (i.e., Hispanic or non-Hispanic ethnicity, American Indian or Alaska Native, Black or African American, Asian, Native Hawaiian or Pacific Islander, White or Multiracial). A set of independent variables was added to reduce confounding, including mothers’ self-reported age, marital status, mother’s place of birth (United States or elsewhere), smoking 3 months before and during pregnancy, prepregnancy weight status based on the body mass index (BMI) (i.e., underweight BMI: <18.5, normal BMI: 18.5–24.9, overweight BMI: 25.0–29.9, obesity I BMI: 35.0–34.9, obesity II BMI: 35.0–39.9, and extreme obesity III BMI: ≥40.0), participation in the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), and educational level (less than high school, high school, some college, college or more). The payment source for delivery (private insurance, Medicaid, other) and route of delivery (vaginal spontaneous, vaginal assisted by forceps or vacuum, and cesarean delivery) were extracted from medical records. Infant characteristics such as birth order, gestational age in weeks, and birth weight were gathered from medical records.^25,26 Counties were classified as metro or nonmetro based on the RUCC dataset. Finally, we controlled for county-level year covariates from the SAIPE dataset, including median household income (in dollars) to account for overall economic prosperity, which could reflect resources available to hospitals that influence maternity care, and poverty rate estimates (for those under age 18) to account for potential within-county inequalities. These factors are critical predictors of breastfeeding rates.²⁷

Statistical analysis

We used a linear probability model (LPM) to evaluate the likelihood of an infant being breastfed at discharge, based on the average mPINC score for the infant’s birth county. The LPM estimated average marginal effects, offering computational simplicity while yielding results comparable with other generalized linear model with large samples.²⁸ The model included adjustments for individual and county-level covariates, as described previously. State fixed effects accounted for state-level maternity policies, and year fixed effects addressed common temporal factors such as the COVID-19 pandemic, infant formula shortages, and inflation. Robust standard errors were applied (Supplementary Appendix, Section 4).

Heterogeneity across subpopulations were assessed by including interactions between the average county-level mPINC score and key covariates, such as race/ethnicity and metro/nonmetro location (Supplementary Appendix, Section 4). We conducted several sensitivity analyses: using an arithmetic mean calculation of the mPINC score, a sample of hospitals with data across all survey waves, and focusing on the years 2017, 2019, and 2021 as these mPINC scores remain the same for each two-year period (Supplementary Appendix, Section 5).

Analyses were conducted in Stata 18.²⁹ The University of Connecticut Institutional Review Board (IRB) deemed the study exempt from IRB review (protocol #23-274-910).

Results

Table 1 summarizes maternal, infant, and county characteristics from 2017 to 2022, highlighting key trends. Breastfeeding initiation rates remained stable until a notable increase in 2022, coinciding with the infant formula crisis in February. Maternal age and education levels gradually increased, with a growing percentage of mothers holding college degrees and a decline in those with less than a high school education. The percentage of mothers who smoked and those participating in the WIC program steadily decreased. The racial and ethnic composition of mothers shifted, with the share of Hispanic mothers increasing and the proportion of White mothers slightly decreasing. BMI categories also showed a trend toward higher obesity rates. County characteristics, such as median household income and mPINC scores, had upward trends, with improvements in maternity care quality as indicated by the mPINC scores. Reflecting the geographic disparities mentioned in the introduction, breastfeeding initiation rates in the Delta and Appalachian regions are significantly lower than in other U.S. regions in our sample, at 69.31% and 76.54%, respectively, compared with the overall average of 85.95%. These disparities are mirrored in mPINC scores, with the Delta region scoring 73.60, the Appalachian region 74.76, and other regions 79.67. Both regions also exhibit notably lower median household incomes, higher child poverty rates, lower educational attainment, higher smoking rates, a lower percentage of mothers born outside the United States, and greater rurality—factors that are all important predictors of breastfeeding (Table 2).

Table 1.

Analytical Sample by Year of Birth, United States, 2017–2022

Description	Percentage or mean
Description	2017	2018	2019	2020	2021	2022
Breastfeeding initiation	84.42	84.48	84.66	84.49	84.13	86.10
Maternal characteristics
Age (years)	28.80	28.96	29.04	29.15	29.34	29.44
Final route and method of delivery
Spontaneous	64.39	64.52	64.68	64.40	64.15	64.18
Forceps	0.58	0.56	0.54	0.54	0.52	0.49
Vacuum	2.51	2.49	2.48	2.50	2.45	2.34
Cesarean	32.52	32.43	32.29	32.56	32.88	32.99
Marital status
Married	60.11	60.13	59.65	59.09	59.84	59.91
Prepregnancy maternal body mass index (BMI)
Underweight <18.5	3.32	3.13	2.99	2.78	2.64	2.66
Normal 18.5–24.9	43.20	41.96	40.79	39.64	38.75	37.87
Overweight 25.0–29.9	26.24	26.62	26.92	27.17	27.46	27.56
Obesity I 35.0–34.9	14.71	15.17	15.66	16.12	16.51	16.85
Obesity II 35.0–39.9	7.29	7.57	7.87	8.21	8.40	8.64
Extreme obesity III ≥40.0	5.25	5.55	5.77	6.07	6.24	6.41
Maternal level of education
Less than high school	12.89	12.37	11.94	11.55	10.75	10.86
High school diploma	25.28	25.60	26.04	26.49	26.03	26.46
Some college	28.47	28.07	27.76	27.23	26.74	26.15
College degree or more	33.36	33.96	34.26	34.74	36.48	36.53
Maternal race/ethnicity
Hispanic, may be of any race	21.16	21.30	21.76	22.15	22.29	23.65
American Indian or Alaska native	0.77	0.76	0.79	0.75	0.70	0.69
Asian	5.68	5.60	5.57	5.40	5.23	5.36
Black or African American	15.92	15.97	16.11	16.22	15.48	15.29
Multiracial	2.07	2.15	2.21	2.30	2.34	2.40
Native Hawaiian or Pacific Islander	0.21	0.22	0.22	0.22	0.22	0.23
White	54.19	54.01	53.33	52.95	53.74	52.38
Mother’s nativity
Born outside the United States, including territories	22.26	21.98	21.82	21.37	20.78	22.22
Payment source for delivery
Private insurance	50.26	50.66	51.16	51.43	52.54	52.20
Medicaid	42.88	42.38	42.41	42.45	41.37	41.65
Other^a	6.86	6.96	6.43	6.12	6.09	6.15
Maternal characteristics
Prior births now living
0	38.51	38.46	38.77	39.31	39.23	39.64
1	32.57	32.46	32.24	31.91	32.12	31.79
2	17.03	17.07	16.96	16.72	16.69	16.56
3	7.10	7.17	7.14	7.14	7.07	7.08
4 or more	4.79	4.83	4.89	4.93	4.89	4.93
Smoking status^b
Smoker	7.19	6.78	6.17	5.74	4.77	3.79
Women, Infants, and Children (WIC)^c
WIC participant	37.07	35.08	33.18	31.61	29.41	29.70
Infant Characteristics
Birth weight
≥2.5 kg	92.15	92.13	92.07	92.06	91.83	91.71
1.5–2.499 kg (low birth weight)	6.76	6.80	6.83	6.86	7.05	7.18
<1.5 kg (very low birth weight)	1.09	1.07	1.10	1.08	1.12	1.11
Obstetric estimate (maturity)
37 weeks and over	90.58	90.46	90.21	90.26	89.88	89.94
County characteristics
Median annual household income ($)	61,991	64,029	67,703	69,180	71,330	76,662
Average county mPINC score (points)^d	76.96	76.96	79.31	79.32	80.20	80.23
Percent of nonmetro counties	8.68	8.63	8.34	8.29	8.04	7.79
Poverty among individuals under 18 years	18.90	18.49	17.27	16.23	17.51	16.73
Number of live births (1,000s)	2,649	2,627	2,667	2,590	2,593	2,593

Missing or unknown values for each covariate were dropped when deriving percent distributions. Percentages may not add up to 100% due to rounding.

Include self-pay, Indian Health Service, CHAMPUS/TRICARE, and Other Government (Federal, State, Local).

Smoking 3 months before and during pregnancy.

WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.

mPINC (Maternity Practices in Infant Nutrition and Care) scores are weighted based on the number of births at each hospital, as reported in the mPINC survey.

Table 2.

Comparison of Analytical Sample Across the Delta Region, Appalachian Region, and the Rest of Our Sample, United States, 2017–2022

Description	Percentage or mean
Description	Rest	Delta region	Appalachian region
Breastfeeding initiation	85.95	69.31	76.54
Maternal characteristics
Age (years)	29.28	27.47	27.94
Final route and method of delivery
Spontaneous	64.49	61.23	64.63
Forceps	0.54	0.51	0.54
Vacuum	2.45	2.59	2.50
Cesarean	32.52	35.68	32.33
Marital status
Married	60.51	45.80	57.34
Maternal body mass index (BMI)
Underweight <18.5	2.88	3.16	3.32
Normal 18.5–24.9	40.78	35.35	37.85
Overweight 25.0–29.9	27.15	25.51	25.76
Obesity I 35.0–34.9	15.73	17.34	16.41
Obesity II 35.0–39.9	7.81	9.95	9.26
Extreme obesity III ≥40.0	5.65	8.68	7.40
Maternal level of education:
Less than high school	11.63	13.59	12.09
High school diploma	25.39	31.93	30.51
Some college	27.13	30.29	29.46
College degree or more	35.84	24.19	27.95
Maternal race/ethnicity:
Hispanic, may be of any race	23.66	7.82	8.91
American Indian or Alaska Native	0.81	0.19	0.14
Asian	5.90	1.77	2.06
Black or African American	15.29	38.55	12.07
More than one race	2.30	0.06	0.06
Native Hawaiian or Pacific Islander	0.24	1.32	1.95
White	51.80	50.30	74.82
Mother’s nativity
Born outside the United States, includes possessions	23.20	8.91	9.78
Payment source for delivery
Private insurance	52.00	39.42	49.00
Medicaid	41.30	57.51	46.20
Other^a	6.70	3.07	4.80
Maternal characteristics
Prior births now living
0	39.08	36.96	38.73
1	32.18	31.26	32.67
2	16.77	18.08	17.15
3	7.10	8.01	6.99
4 or more	4.88	5.68	4.46
Smoking status^b
Smoker	5.11	7.75	12.56
Women, Infants and Children (WIC)^c
WIC participant	31.91	35.08	33.18
Infant characteristics
Birth weight
≥2,5 kg	92.08	89.69	91.97
1.5–2.499 kg (low birth weight)	6.84	8.81	6.97
<1.5 kg (very low birth weight)	1.08	1.50	1.06
Obstetric estimate
37 weeks and over	90.31	88.06	90.11
County characteristics
Median income ($)	70,016	50,700	57,447
Average county mPINC score (points)^d	79.67	73.60	74.76
Nonmetro counties	6.57	22.01	23.37
Poverty for those under age 18	17.05	26.43	19.18
Number of live births (1,000s)	14,039	532	1,154

The counties included in the Delta and Appalachian regions are listed in Section 6 of the Appendix. The “Rest” column does not include these counties. In addition, some counties overlap between the Delta and Appalachian regions. Missing or unknown values for each covariate were dropped when deriving percent distributions. Percentages may not add up to 100% due to rounding.

Include self-pay, Indian Health Service, CHAMPUS/TRICARE, and Other Government (Federal, State, Local).

Smoking 3 months before and during pregnancy.

WIC, Special Supplemental Nutrition Program for Women, Infants, and Children.

mPINC (Maternity Practices in Infant Nutrition and Care) scores are weighted based on the number of births at each hospital, as reported in the mPINC survey.

Our results showed that for each additional point in the average county-level mPINC score, breastfeeding initiation rates increased by 0.10 percentage points (pp) (p < 0.001; Table 3, Column 1). To explore differences across racial and ethnic groups, an interaction term between the mPINC score and race/ethnicity was included (Table 3, Column 2). Among NHW mothers, the increase in breastfeeding initiation was 0.07 pp (p < 0.001), slightly below the overall average. In contrast, the effect was notably stronger for NHB mothers, with each point increase in the mPINC score corresponding to an additional 0.18 pp increase in breastfeeding (p < 0.001), totaling a 0.25 pp increase (95% confidence interval [CI], 0.24–0.25 pp; p < 0.001), more than three times the NHW rate. NHAI/AN mothers also experienced a stronger effect, totaling a 0.14 pp increase (95% CI, 0.12–0.16 pp; p < 0.001). For Hispanic mothers, each mPINC point was associated with a 0.09 pp increase in breastfeeding initiation (95% CI, 0.08–0.09 pp; p < 0.001).

Table 3.

Association Between County-Level mPINC Scores and Breastfeeding Initiation, Estimates from the Linear Probability Model Analysis, United States, 2017–2022

	(1)	(2)	(3)
Variables	[95% CI]	[95% CI]	[95% CI]
mPINC^d Score	0.097^a	0.066^a	0.079^a
	[0.095, 0.099]	[0.064, 0.068]	[0.077, 0.081]
Interaction terms
mPINCd Score × Race/Ethnicity
Non-Hispanic White (Ref)
Hispanic, may be of any race		0.019^a
		[0.014, 0.023]
American Indian or Alaska Native		0.072^a
		[0.052, 0.093]
Asian		−0.057^a
		[−0.064, −0.049]
Black or African American		0.183^a
		[0.177, 0.189]
Multiracial		−0.025^a
		[−0.037, −0.012]
Native Hawaiian or Pacific Islander		0.024
		[−0.011, 0.059]
mPINCd Score × County Type
Metro (Ref)
Nonmetro			0.091^a
			[0.086, 0.096]
Individual-level covariates	Yes	Yes	Yes
County covariates	Yes	Yes	Yes
State indicators (fixed effects)	Yes	Yes	Yes
Year indicators (fixed effects)	Yes	Yes	Yes
Number of observations	15,719,673	15,719,673	15,719,673

Robust standard errors. The coefficient represents a change in probability, expressed in percentage points.

Significant at p < 0.001determined by using a two-sided Wald test.

Significant at p < 0.01 determined by using a two-sided Wald test.

Significant at p < 0.05 determined by using a two-sided Wald test.

mPINC: Maternity Practices in Infant Nutrition and Care.

Conversely, NHA and non-Hispanic multiracial (NHMR) mothers showed slight declines compared with NHW mothers, with decreases of 0.06 pp (p < 0.001) and 0.03 pp (p < 0.001), respectively. After accounting for these differences, adjusted increases in breastfeeding initiation for NHA and NHMR mothers were 0.01 pp (95% CI, 0.00–0.02 pp; p < 0.001) and 0.04 pp (95% CI, 0.03–0.05 pp; p < 0.001), respectively. No significant difference was observed for non-Hispanic Native Hawaiian or Pacific Islander mothers compared with NHW mothers.

Finally, an interaction between the mPINC score and county type (metro versus nonmetro) demonstrated stronger increases in nonmetro counties (Table 3, Column 3). Each additional mPINC score point was associated with a 0.08 pp increase in metro areas (p < 0.001) and a 0.17 pp increase in nonmetro counties (95% CI, 0.17–0.18 pp; p < 0.001), nearly double of that observed in metro counties.

Each standard deviation increase in the mPINC score (14 points) is associated with a 1.36 pp rise in overall breastfeeding initiation. With 3,584,633 live births in U.S. hospitals in 2022, this equates to approximately 48,751 additional mothers initiating breastfeeding. However, this increase is not uniform across all population groups. The effect is strongest for populations with historically lower breastfeeding rates, particularly NHB mothers, who would experience a 3.49 pp increase, helping to narrow the existing gap. Similarly, NHAI/AN mothers would see a 1.92 pp increase, while mothers in nonmetro counties would experience a 2.38 pp gain. These findings suggest that improving mPINC scores could significantly reduce breastfeeding disparities among historically underserved populations.

Discussion

This study finds that improved hospital maternity care practices in support of breastfeeding are associated with a statistically significant increase in breastfeeding initiation for all infants. Specifically, newborns born in hospitals with higher mPINC scores, reflecting better breastfeeding support, tend to have better in-hospital breastfeeding outcomes. This aligns with previous research linking better maternity care practices to improved breastfeeding initiation, exclusivity, and duration.^6,12–14

The relationship between high quality maternity care and breastfeeding initiation is particularly strong among populations with historically lower breastfeeding rates, such as NHB and NHAI/AN mothers. For NHB mothers, prior studies indicate that targeted hospital quality improvement efforts, aimed at complying with the Ten Steps to Successful Breastfeeding, have helped reduce racial disparities.¹⁵ Similar effects for NHAI/AN mothers were underexplored. NHB mothers are twice as likely to have formula introduced in the hospital compared with White mothers.³⁰ Research from breastfeeding peer counselors notes that health care providers often neglect to discuss breastfeeding with African American women unless initiated by the mothers.³¹ International board-certified lactation consultants have reported that some providers assume women of color are less likely to breastfeed. NHB mothers also face barriers such as formula supplementation and limited access to lactation services, which can hinder breastfeeding success.³¹ Enhancing maternity care practices could help mitigate these challenges by addressing racial bias and improving access to support.

NHAI/AN mothers have faced structural barriers, including federal policies that displaced traditional breastfeeding practices, such as community-based lactation support, with less accessible Western approaches.³² Among Navajo mothers, common breastfeeding challenges include pain, latch issues, and concerns about milk supply, particularly for those breastfeeding for less than 6 months. Many expressed that lactation consultant support could have helped address these challenges.³³ Improved maternity practices may have reduced these barriers by offering better postdischarge breastfeeding support and facilitating referrals to community-based lactation providers. The smaller increase in breastfeeding initiation observed among NHA and NHMR mothers may be attributed to their higher baseline rates of breastfeeding initiation.

Finally, the stronger relationship between quality maternity practices and breastfeeding initiation in nonmetro areas highlights the critical importance of providing accessible, high-quality maternity care services in rural regions. Nonmetro areas closely align with the federal Office of Management and Budget’s designations of micropolitan counties (rural counties with an urban core of 10,000 to 50,000 residents) and noncore counties (rural counties without an urban core), and have experienced significant declines in maternity care availability. Between 2004 and 2014, 9% of micropolitan and noncore counties lost hospital maternity services, and an additional 45% had no maternity services to begin with.³⁴ Nationwide, over 400 maternity services closed between 2006 and 2020. Hospitals most affected by these closures are often those serving predominantly Black or Latino communities.³⁵ Insufficient maternity care disproportionately affects vulnerable populations in rural areas, who already face limited access to resources supporting positive maternal and infant health behaviors.³⁶ Our results suggest that when high-quality maternity practices are available, they are likely to address critical service gaps that positively impact breastfeeding initiation.

The mPINC survey is a valuable tool for hospitals to assess and monitor their breastfeeding support services. The CDC provides each participating hospital in the mPINC survey with a customized benchmark report, comparing its performance with similarly sized hospitals, those within the same regions, and all hospitals included in the survey. Hospitals can utilize this information to identify strengths and areas for improvement.³⁷ Given the evidence linking quality maternity care practices to better breastfeeding outcomes, public health initiatives and legislation aimed at improving hospital-based maternity care should remain a high priority.³⁸

This study has several key strengths. First, it utilizes national birth certificate data from 2017 to 2022, employing a standardized definition of hospital-based breastfeeding initiation. This approach allowed us to assess breastfeeding initiation among nearly all newborn infants from 47 states and the District of Columbia, covering approximately 70% of U.S. live births, while controlling for relevant individual and county-level covariates that may confound the analysis. Validation from other studies underscores the reliability of breastfeeding information recorded on birth certificates.³⁹ The national breastfeeding initiation trends observed in our data closely align with findings from the National Immunization Survey-Child (NIS-Child) (Supplementary Appendix, Section 7).² In addition, the study benefits from the mPINC survey, which provides comprehensive data on maternity care facilities with a notably high response rate, improving the understanding of the relationship between maternity care practices and breastfeeding initiation.

The primary limitation of this study is the inability to link mother−infant dyads to specific delivery hospitals as this information is not available in birth certificate data through the NVSS. This limitation prevents an accurate measurement of the intensity of treatment and introduces statistical noise, which may reduce the ability to detect significant effects. Other limitations include reliance on self-reports from hospital staff in the mPINC survey data, which may not fully represent all hospital practices accurately.²⁰ Although a standard protocol identifies key informants at each hospital, reliance on self-reporting may introduce bias. The mPINC survey is typically completed by managerial staff and physician leaders, although multiple staff members may contribute over time and could view and edit sections before final submission. This process could lead to differences in reporting between management and frontline staff. In addition, the lack of anonymity when completing the survey as a team may influence responses, potentially leading to biased or overstated reports on compliance with maternity care practices. This bias could make the observed effect of high-quality maternity care on breastfeeding initiation appear weaker than it truly is. Lastly, our study sample disproportionately excluded counties in nonmetro areas, as hospitals in these areas were less likely to respond to the survey. As a stronger relationship between mPINC scores and breastfeeding initiation was observed in nonmetro areas, this may lead to an underestimation of the total effect of better maternity practices on breastfeeding initiation.

Conclusions

Higher mPINC scores, reflecting better quality hospital maternity care practices, are associated with increased county-level breastfeeding initiation rates. This relationship is particularly strong among populations with historically lower breastfeeding rates, including NHB and NHAI/AN mothers, as well as in nonmetro counties. These findings emphasize the critical role of hospital-based maternity care in breastfeeding outcomes, suggesting that expanding access to supportive maternity care could help reduce breastfeeding disparities. They also underscore the need to address maternity care deserts, particularly in rural areas. Improving hospital practices may help overcome barriers to breastfeeding for underserved populations.

Footnotes

Acknowledgments

We would like to thank the teams at the Connecticut Department of Public Health and the Centers for Disease Control and Prevention for their valuable input and support. This study used mPINC data (2018–2022), as compiled from data provided through the Maternity Practices in Infant Nutrition and Care (mPINC) survey. The authors declare no conflicts of interest with respect to the research, authorship, or publication of this article. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the USDA and the State of Connecticut. No copyrighted material, surveys, instruments, or tools were used in the research described in this article.

Authors’ Contributions

L.S.E.: Conceptualization, methodology, formal analysis, investigation, and writing—original draft, writing—review and editing. T.A.: Conceptualization, methodology, investigation, supervision, funding acquisition, and writing—review and editing. All authors approved the final article as submitted and agree to be accountable for all aspects of the work.

Disclosure Statement

The authors have no conflicts of interest relevant to this article to disclose.

Funding Information

This project has been supported, in part by federal award number SLFRP0128, awarded to the State of Connecticut by the US Department of the Treasury, and by the USDA National Institute of Food and Agriculture, Hatch project No. 7007853.

Supplementary Material

References

Bartick

, Schwarz

, Green

, et al. Suboptimal breastfeeding in the United States: Maternal and pediatric health outcomes and costs. Matern Child Nutr, 2017; 13(1):e12366.

Centers for Disease Control and Prevention. NIS-Child Data Results [Internet]. Breastfeeding Data. 2024. [cited 2025 Feb 24]. Available from: https://www.cdc.gov/breastfeeding-data/survey/results.html

Chiang

, Li

, Anstey

, et al. Racial and ethnic disparities in breastfeeding initiation—United States, 2019. MMWR Morb Mortal Wkly Rep, 2021; 70(21):769–774.

Institute of Medicine. (US) Committee on Nutritional Status During Pregnancy and Lactation. Nutrition During Lactation [Internet]. National Academies Press (US); Washington (DC); 1991. [cited 2024 Jun 20]. Available from: http://www.ncbi.nlm.nih.gov/books/NBK235593/

Mieso

, Burrow

, Lam

. Beyond Statistics: Uncovering the roots of racial disparities in breastfeeding. Pediatrics, 2021; 147(5):e2020037887.

Barrera

, Beauregard

, Nelson

, et al. Association of maternity care practices and policies with in-hospital exclusive breastfeeding in the United States. Breastfeed Med, 2019; 14(4):243–248.

Allen

, Perrine

, Scanlon

. Breastfeeding supportive hospital practices in the US differ by county urbanization level. J Hum Lact, 2015; 31(3):440–443.

Lind

, Perrine

, Li

, et al.; Centers for Disease Control and Prevention (CDC). Racial disparities in access to maternity care practices that support breastfeeding – United States, 2011. MMWR Morb Mortal Wkly Rep, 2014; 63(33):725–728.

Wiener

, Wiener

. Breastfeeding Prevalence and Distribution in the USA and Appalachia by Rural and Urban Setting. RRH [Internet]. May 2, 2011. [cited 2025 Jan 29]. Available from: https://www.rrh.org.au/journal/article/1713

10.

Anstey

, Chen

, Elam-Evans

, et al. Racial and geographic differences in breastfeeding – United States, 2011–2015. MMWR Morb Mortal Wkly Rep, 2017; 66(27):723–727.

11.

Seiger

, Wasser

, Hutchinson

, et al. Barriers to providing lactation services and support to families in Appalachia: A mixed-methods study with lactation professionals and supporters. Am J Public Health, 2022; 112(S8):S797–S806.

12.

Nelson

, Perrine

, Freedman

, et al. Infant feeding-related maternity care practices and maternal report of breastfeeding outcomes. Birth, 2018; 45(4):424–431.

13.

Patterson

, Keuler

, Olson

. The effect of maternity practices on exclusive breastfeeding rates in U.S. hospitals. Matern Child Nutr, 2019; 15(1):e12670; doi: 10.1111/mcn.12670

14.

, Li

, Ashley

, et al. Associations of hospital staff training and policies with early breastfeeding practices. J Hum Lact, 2014; 30(1):88–96.

15.

Merewood

, Bugg

, Burnham

, et al. Addressing racial inequities in breastfeeding in the Southern United States. Pediatrics, 2019; 143(2):e20181897.

16.

Bookhart

, Anstey

, Kramer

, et al. Unpacking breastfeeding disparities: Baby-friendly hospital designation associated with reduced in-hospital exclusive breastfeeding disparity attributed to neighborhood poverty. Matern Child Health J, 2024; 28(8):1404–1412.

17.

National Center for Health Statistics. Centers for Disease Control and Prevention. U. S. Standard Certificate of Live Birth [Internet]. 2003. [cited 2023 Dec 6]. Available from: https://www.cdc.gov/nchs/data/dvs/birth11-03final-ACC.pdf

18.

National Center for Health Statistics. Vital Statistics Natality – All counties, (2016–2022), as compiled from data provided by the 57 vital statistics jurisdictions through the Vital Statistics Cooperative Program.

19.

Centers for Disease Control and Prevention. mPINCTM Survey Methods [Internet]. Breastfeeding Data. 2024. Available from: https://www.cdc.gov/breastfeeding-data/mpinc/methodology.html

20.

Nelson

, Grossniklaus

, Galuska

, et al. The mPINC survey: Impacting US maternity care practices. Matern Child Nutr, 2021; 17(1):e13092; doi: 10.1111/mcn.13092

21.

US Department of Agriculture. Rural-Urban Continuum Codes [Internet]. 2024. Available from: https://www.ers.usda.gov/data-products/rural-urban-continuum-codes/

22.

US Census Bureau. Small Area Income and Poverty Estimates (SAIPE) Program [Internet]. Available from: https://www.census.gov/programs-surveys/saipe.html

23.

Centers for Disease Control and Prevention. mPINC^TM Survey Scoring [Internet]. Breastfeeding Data. 2024. [cited 2025 Feb 24]. Available from: https://www.cdc.gov/breastfeeding-data/mpinc/scoring.html

24.

National Center for Health Statistics. User guide 2014, 2016-2022 Natality Public Use File [Internet]. 2024. [cited 2024 Apr 30]. Available from: https://www.cdc.gov/nchs/data_access/vitalstatsonline.htm

25.

National Center for Health Statistics. Centers for Disease Control and Prevention. NVSS – Facility Worksheets Guidebook – 40 [Internet]. 2022. [cited 2023 Dec 6]. Available from: https://www.cdc.gov/nchs/nvss/facility-worksheets-guide/40.htm

26.

National Center for Health Statistics (NCHS). NVSS – Revisions of the U.S. Standard Certificates and Reports [Internet]. 2003. [cited 2023 Dec 6]. Available from: https://www.cdc.gov/nchs/nvss/revisions-of-the-us-standard-certificates-and-reports.htm

27.

Standish

, Parker

. Social determinants of breastfeeding in the United States. Clin Ther, 2022; 44(2):186–192.

28.

Norton

, Dowd

. Log odds and the interpretation of logit models. Health Serv Res, 2018; 53(2):859–878.

29.

StataCorp. Stata Statistical Software: Release 18. StataCorp LLC: College Station, TX; 2023.

30.

McKinney

, Hahn-Holbrook

, Chase-Lansdale

, et al.; Community Child Health Research Network. Racial and Ethnic Differences in Breastfeeding. Pediatrics, 2016; 138(2). https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4960721/

31.

Robinson

, Fial

, Hanson

. Racism, bias, and discrimination as modifiable barriers to breastfeeding for African American Women: A scoping review of the literature. J Midwifery Womens Health, 2019; 64(6):734–742.

32.

Theobald

. Reproduction on the Reservation: Pregnancy, Childbirth, and Colonialism in the Long Twentieth Century. UNC Press Books; 2019; p. 289.

33.

Brown

, Talker

, Stoddard

, et al. Breastfeeding attitudes and practices in a rural Utah Navajo Community. Matern Child Health J, 2022; 26(2):397–406.

34.

Sonenberg

, Mason

. Maternity care deserts in the US. JAMA Health Forum, 2023; 4(1):e225541.

35.

Maternity Care Deserts Grow Across the US as Obstetric Units Shut Down [Internet]. PBS News; 2022. [cited 2024 Nov 18]. Available from: https://www.pbs.org/newshour/show/maternity-care-deserts-grow-across-the-us-as-obstetric-units-shut-down

36.

Hung

, Henning-Smith

, Casey

, et al. Access to obstetric services in rural counties still declining, with 9 percent losing services, 2004–14. Health Aff (Millwood), 2017; 36(9):1663–1671.

37.

Centers for Disease Control and Prevention. Maternity Practices in Infant Nutrition and Care (mPINC) Survey [Internet]. Breastfeeding Data. 2024. [cited 2025 Feb 24]. Available from: https://www.cdc.gov/breastfeeding-data/mpinc/index.html

38.

Pérez‐Escamilla

, Martinez

, Segura‐Pérez

. Impact of the Baby‐friendly Hospital Initiative on breastfeeding and child health outcomes: A systematic review. Matern Child Nutr, 2016; 12(3):402–417.

39.

Ahluwalia

, Helms

, Morrow

. Assessing the validity and reliability of three indicators self-reported on the pregnancy risk assessment monitoring system survey. Public Health Rep, 2013; 128(6):527–536.