ARTICLE |
a Departments of Pediatrics
d Emergency Medicine, Boston University School of Medicine, Boston, Massachusetts
b Department of Epidemiology, Boston University School of Public Health, Boston, Massachusetts
c Breastfeeding Center, Boston Medical Center, Boston, Massachusetts
| ABSTRACT |
|---|
|
|
|---|
METHODS. Massachusetts Community Health Information Profile, an online public health database that was created by the Massachusetts Department of Public Health, includes breastfeeding initiation data that are obtained from the electronic birth certificate, which we used to compare breastfeeding rates among preterm and term infants. Birth-linked demographics and data that also were accessed were maternal age, race/ethnicity, birthplace, and health insurance (public or private) as an indicator of socioeconomic status and infant's gestational age. We assessed the association between breastfeeding initiation and maternal birthplace, as well as race/ethnicity and the other potential confounders, using logistic regression.
RESULTS. There were 80624 births in Massachusetts in 2002, and 8.2% (6611) of newborns had a gestational age <37 weeks. The state's overall breastfeeding initiation rate was 74.6%. We excluded records of mothers who were younger than 15 years and older than 39 years, nonsingleton births, infants with a gestational age <24 weeks and >42 weeks, and records with missing data. Of the total births in Massachusetts, 67884 (84%) met inclusion criteria for this study. Breastfeeding initiation rates were lowest among preterm infants of the youngest gestational ages. Breastfeeding initiation was 76.8% among term infants born at 37 to 42 weeks, 70.1% among infants born at 32 to 36 weeks, and 62.9% among infants born at 24 to 31 weeks. In univariate analysis, among preterm infants, a lower proportion of US-born black, Asian, and Hispanic mothers initiated breastfeeding than US-born white mothers; nonUS-born black and nonUS-born Hispanic mothers had the highest breastfeeding initiation rates. Among term infants, US-born black mothers had the lowest initiation rates, and nonUS-born black and nonUS-born Hispanic mothers had the highest. In multivariate logistic regression, however, after controlling for mother's age, race, birthplace, and insurance, US-born white mothers were least likely to breastfeed either term or preterm infants when compared with any other racial/ethnic group, including US-born black mothers. The likelihood that nonUS-born Hispanic mothers would breastfeed was almost 8 times greater than that for US-born white mothers for a preterm infant and almost 10 times greater for a term infant. In multivariate logistic regression analysis stratified by gestational age for both preterm and term infants, older mothers and mothers with private health insurance were most likely to breastfeed.
CONCLUSIONS. In Massachusetts, preterm infants were less likely to receive breast milk than term infants, and the likelihood of receiving breast milk was lowest among the youngest preterm infants. In multivariate logistic regression, mothers who were born outside the United States were more likely than US-born mothers to breastfeed either term or preterm infants in all racial and ethnic groups. In an unexpected finding, US-born white mothers were less likely to breastfeed term or preterm infants than US-born black mothers or mothers of any other racial or ethnic group.
Key Words: breastfeeding premature preterm infants maternal birthplace Massachusetts
Abbreviations: CDCCenters for Disease Control and Prevention MassCHIPMassachusetts Community Health Information Profile GAgestational age ORodds ratio AORadjusted odds ratio CIconfidence interval
Breastfeeding is the optimal form of infant nutrition.1 Among premature infants, formula feeding increases the risk for necrotizing enterocolitis,2 delayed brainstem maturation,3 decreased scoring on cognitive and developmental tests,47 sepsis,8 and delayed visual development.9, 10 With this in mind, many interventions are designed to increase breast milk consumption among preterm infants.1116 However, breastfeeding rates among US preterm infants are not available; the national surveys17, 18 that report annual breastfeeding rates do not report gestational age. Although data are collected in states where gestational age and breastfeeding are recorded on the birth certificate, published work is restricted to a few small studies.1921 One study reviewed 151 charts of infants who were
34 weeks' gestational age and admitted to the NICU in a rural Illinois hospital and found a breastfeeding initiation rate of 49.7%, with no significant association with maternal race/ethnicity.19 A study by Geraghty et al20 examined breastfeeding rates among term and preterm infants, but the research was restricted to multiple births. In addition, although the Centers for Disease Control and Prevention (CDC) Pregnancy Risk Assessment Monitoring System collects both breastfeeding rates information22, 23 and data on gestational age,24 we have not been able to identify any Pregnancy Risk Assessment Monitoring Systembased studies that examined both together.
National surveys calculate breastfeeding rates among term infants according to maternal race/ethnicity, but maternal birthplace is not recorded. Maternal birthplace likely is important because breastfeeding is the cultural norm in the countries of origin for many nonUS-born US residents. The goals of this study were to compare breastfeeding initiation rates among preterm and term infants in Massachusetts in 2002 and to investigate the role of maternal birthplace and race/ethnicity on breastfeeding initiation.
| METHODS |
|---|
|
|
|---|
Birth-linked demographics and data also accessed through MassCHIP were maternal age (categorized as 1519, 2029, or 3039 years), race/ethnicity (categorized as white/non-Hispanic, black/non-Hispanic, Hispanic, or Asian/Pacific Islander/non-Hispanic), maternal birthplace (categorized as US-born [50 states; Washington, DC; and Puerto Rico] or nonUS-born), health insurance (public or private) as an indicator of socioeconomic status, and infant's gestational age (GA; categorized as 2431, 3236, or 3742 weeks). Race/ethnicity and birthplace were recoded into a single variable for analysis. For creation of a data set with broad applicability, certain exclusions were made. We excluded mothers who were younger than 15 years (73) and older than 39 years (3085), nonsingleton births (3951), births to Native Americans (167), and infants with a GA <24 weeks (182) and >42 weeks (21). Native Americans were excluded because the number was too small to analyze, and infants at extreme ends of the gestational age range were excluded because of potential medical complications that might affect breastfeeding. We used health insurance as a marker for socioeconomic status; therefore, we excluded insurance types that did not clarify income (non-Medicaid/Medicare government insurance [2090] and self-pay [471]). An additional 4064 records were excluded because 1 of the following variables was missing: maternal age (176), insurance status (1918), race (1530), or birthplace (93) or infant's GA (347). A total of 12740 births (fewer than the sum of the individual categories, because some exclusions fell into >1 category) were excluded, leaving 67884 births for analysis.
Descriptive analysis included frequencies. We first assessed breastfeeding initiation rates for preterm (2436 weeks' GA) and term infants, overall, and stratified by maternal factors (age, race/ethnicity/birthplace, and insurance). We then assessed the association between GA (preterm 2431 weeks, preterm 3236 weeks, and term), first on a univariate basis and then adjusted for all other variables in the model. Finally, we used univariate and multivariate logistic regression to assess the role of maternal factors in predicting breastfeeding initiation. Data were analyzed using Stata/SE 8.2 for Windows (Stata Corp, College Station, TX).
| RESULTS |
|---|
|
|
|---|
3742 weeks), 70.1% among infants who were born at 32 to 36 weeks, and 62.9% among infants who were born at 24 to 31 weeks. Demographic information and breastfeeding initiation rates by gestational age are presented in Table 1.
|
|
|
|
| DISCUSSION |
|---|
|
|
|---|
Low breastfeeding rates among premature infants are of particular concern given the importance of breast milk for these infants.27, 9, 10 However, these rates are not unexpected and have been reported in smaller studies.1921 Breastfeeding a preterm infant is more complex than breastfeeding a term infant, requiring additional knowledge, support, and equipment such as a breast pump. Because less affluent mothers are less likely to be able to afford a breast pump26 and may have additional practical problems, such as transportation to the NICU, low breastfeeding rates among poorer mothers of preterm infants might be anticipated.
One might expect, however, that considerable proportions of immigrant mothers would be affected by similar issues as poor mothers, such as lack of access to breast pumps. We found that nonUS-born mothers had significantly higher breastfeeding initiation rates than US-born mothers for preterm as well as for term infants. For example, nonUS-born black mothers were 3.1 times more likely to initiate breastfeeding in their preterm infants than US-born white mothers, and nonUS-born Hispanic mothers were 7.6 times more likely.
Breastfeeding rates among black mothers of preterm infants are particularly noteworthy because black infants are almost twice as likely to be born preterm as white infants27 (a trend that was apparent in our study [Table 1]). In addition to our findings on nonUS-born black mothers and infants, we note that, among both preterm and term infants, US-born black mothers had the lowest breastfeeding initiation rates, but after controlling for maternal factors and gestational age, US-born black mothers had higher breastfeeding initiation rates than US-born white mothers. Black mothers are reported consistently at the national level as having the lowest breastfeeding rates.18, 28 However, these data do not consider birthplace and are not adjusted for confounding.
Our results regarding maternal birthplace expand on those published in a recent, smaller study, which analyzed data on 4207 mothers and found that, looking at all ethnicities combined, US-born mothers had an 85% reduction in the odds of initiating breastfeeding (OR: 0.150; P < .01) and a 66% reduction in the odds of breastfeeding at 6 months (OR: 0.344; P < .01) when compared with nonUS-born mothers. That study also found that each additional year of living in the United States was associated with a 4% decrease in the odds of initiating breastfeeding (OR: 0.96; P < .01) and a 3% decrease in the odds of breastfeeding at 6 months (OR: 0.97; P < .05). Given these findings, we suggest that national surveys, which currently report breastfeeding rates only on the basis of race, need consistently to consider maternal place of birth.
Because rates of breastfeeding are so relatively high in certain nonUS-born groups (Table 3), we examined the makeup of nonUS-born Massachusetts residents. According to the 2000 US census, 772983 (12%) Massachusetts residents were born outside the United States. Within this group, the largest proportions were from Europe (32.2%), Latin America (30.0%; comprising Caribbean [14.5%], South America [9.3%], and Central America [6.5%]), and Asia (26.1%).25
One possible limitation of the study is whether exclusions led to outcomes that were not representative of the entire state. Exclusions were made to limit outlying cases that might affect breastfeeding initiation, such as extremely preterm infants with high mortality (<24 weeks). We also excluded multiple births, which are common in preterm deliveries, but with the limited data available, any reasons for variation in breastfeeding rates within this group would be difficult to control for. It also is a limitation of this study that we do not have information on factors that are known to affect breastfeeding initiation, such as substance abuse or maternal illness. These variables may be responsible in part for the low rates among preterm infants, rather than prematurity alone.
The main limitation of the study, however, lies on our reliance on breastfeeding rate data from the electronic birth certificate. In this regard, we are reassured by the knowledge that the 2002 breastfeeding initiation rate obtained from the Massachusetts birth certificate (75%)28 is comparable to the 2002 initiation rate for Massachusetts described by the 2 prime sources of breastfeeding rate data nationally: the Ross Mothers' Survey (73%)18 and the CDC National Immunization Survey (74%),28 which use different data collection methods. Generally, birth certificate data are widely used in research, and published studies that have used such data include reports from the CDC's Morbidity and Mortality Weekly.17, 20, 2935 Other recently published studies specifically used Massachusetts birth certificate data.36, 37
Hypothetically, the wording of the question ("Are you breastfeeding, or do you intend to breastfeed?") may lead to misleading answers, from mothers of preterm infants in particular, because barriers to breastfeeding a preterm infant may be too difficult for mothers to overcome, even if they state intent to breastfeed. However, our study found a significantly lower rate of breastfeeding/"intent to breastfeed" among mothers of preterm infants than among mothers of term infants. If, nonetheless, more mothers of preterm than term infants are stating intent than are actually breastfeeding, then our study is underestimating the gap between breastfeeding rates of preterm and term infants. Exaggeration of breastfeeding in mothers of preterm infants specifically would inflate the preterm breastfeeding rate and bring it closer to the term rate. Therefore, we do not believe that our results exaggerate the difference between breastfeeding initiation rates in term and preterm infants.
| CONCLUSIONS |
|---|
|
|
|---|
| FOOTNOTES |
|---|
Address correspondence to Anne Merewood, MPH, IBCLC, Division of General Pediatrics, Maternity Building, 4th Floor, 91 E Concord St, Boston Medical Center, Boston, MA 02118. E-mail: anne.merewood{at}bmc.org
The authors have indicated they have no financial relationships relevant to this article to disclose.
| REFERENCES |
|---|
|
|
|---|
This article has been cited by other articles:
![]() |
S. S. Hawkins, K. Lamb, T. J Cole, C. Law, and the Millennium Cohort Study Child Health Group Influence of moving to the UK on maternal health behaviours: prospective cohort study BMJ, May 10, 2008; 336(7652): 1052 - 1055. [Abstract] [Full Text] [PDF] |
||||
![]() |
A. Merewood Breastfeeding: Promotion of a Low-tech Lifesaver NeoReviews, July 1, 2007; 8(7): e296 - e300. [Abstract] [Full Text] [PDF] |
||||
![]() |
R. Perez-Escamilla and P. Putnik The Role of Acculturation in Nutrition, Lifestyle, and Incidence of Type 2 Diabetes among Latinos J. Nutr., April 1, 2007; 137(4): 860 - 870. [Abstract] [Full Text] [PDF] |
||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||