Ethnic Differences in Breastfeeding Duration among WIC-Eligible Families

Women's Health Issues 21-5 (2011) 374–382

www.whijournal.com

Original article

Racial/Ethnic Differences in Breastfeeding Duration among WIC-Eligible Families Patrice Johnelle Sparks, PhD * Department of Demography and Organization Studies, University of Texas at San Antonio, San Antonio, Texas Article history: Received 18 June 2010; Received in revised form 8 March 2011; Accepted 9 March 2011

a b s t r a c t Purpose: This research documented racial/ethnic differences in breastfeeding duration among mothers from seven diverse racial/ethnic groups in rural and urban areas of the United States that initiated breastfeeding among income and categorically eligible WIC participants. Methods: Using data from the Longitudinal 9-Month–Preschool Restricted-Use data file of the Early Childhood Longitudinal Study–Birth Cohort, this research first assessed racial/ethnic differences in breastfeeding initiation and duration, maternal and child health characteristics, social service usage, and sociodemographic characteristics. Next, breastfeeding survivorship and Cox proportional hazards models were estimated to assess potential racial/ethnic disparities in breastfeeding duration once these control variables were accounted for in multiple variable models. Findings: Breastfeeding initiation rates and breastfeeding durations of 6 months were lower among WIC-eligible mothers compared with all mothers. WIC-eligible, foreign-born Mexican-Origin Hispanic (FBMOH) mothers were most likely to breastfeed for 6 months. Breastfeeding duration rates dropped quickly after 4 months of duration among WIC-eligible mothers that initiated. Two crossover patterns in breastfeeding durations were noted among 1) FBMOH and non-Hispanic Black mothers and 2) Asian and Native American mothers. A FBMOH breastfeeding duration advantage was noted compared with non-Hispanic White mothers once all control variables were included in the Cox proportional hazard models. No other racial/ethnic disparities in breastfeeding duration were noted. Conclusion: More attention to educational programs and broad forms of support as part of WIC are needed to help reach the breastfeeding duration goals of Healthy People 2010 and continued support of the Loving Support Peer Counseling Program may serve as an ideal policy for local WIC offices. Copyright Ó 2011 by the Jacobs Institute of Women’s Health. Published by Elsevier Inc.

Introduction The health benefits of breastfeeding for infants and mothers have been well documented (Ahluwalia, Morrow, & Hsia, 2005; American Academy of Pediatrics 2005; Bartick & Reinhold, 2010; Bartick, Stuebe, Shealy, Walker, & Grummer-Strawn, 2009; James & Lessen, 2009; Keister, Roberts, & Werner, 2008; Ogbuanu et al., 2009). Breastfeeding promotion efforts by the American Academy of Pediatrics Section on Breastfeeding highlight these benefits as part of their recommendation for

Supported by a Tenure-Track Research Award (TRAC) from the University of Texas at San Antonio. * Correspondence to: Patrice Johnelle Sparks, PhD, Department of Demography and Organization Studies, University of Texas at San Antonio, 501 West Durango, MNT 2.270K, San Antonio, Texas 78207. Phone: (210) 458-3141; fax: (210) 458-3164. E-mail address: [email protected]

6 months of exclusive breastfeeding and the encouragement of continued breastfeeding during the first year of the infant’s life (American Academy of Pediatrics, 2005). Although breastfeeding rates in the early postpartum period have increased and remained steady over the past few years in the United States (Ryan, Zhou, & Acosta, 2002), only 20% of women still breastfeed at 12 months (Thulier & Mercer, 2009). The rates of breastfeeding exclusivity at 6 months are lower and miss the Healthy People 2010 goal of 17%, with approximately 13.6% of women exclusively breastfeeding of the almost 74% of women who initiate (Centers for Disease Control and Prevention, 2009). Even with improvements in breastfeeding initiation and short-term exclusivity, minority and low-income women remain at a disadvantage for continuing breastfeeding once they start or meeting these recommendations (Forste & Hoffmann, 2008; Forste, Weiss, & Lippincott, 2001; Lee, Elo, McCollum, K& Culhane, 2009). Participation in the Supplemental Nutrition Program for Women, Infants, and Children (WIC) may provide one outlet to

1049-3867/$ - see front matter Copyright Ó 2011 by the Jacobs Institute of Women’s Health. Published by Elsevier Inc. doi:10.1016/j.whi.2011.03.002

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

promote and increase breastfeeding behaviors among lowincome mothers. However, the research evidence is mixed on the association between WIC participation and breastfeeding initiation and duration, with some studies noting a breastfeeding advantage, whereas others note a breastfeeding disadvantage among WIC participants (Beal, Kuhlthau, & Perrin, 2003; Bunik, Krebs, Beaty, McClatchey, & Olds, 2009; Chatterji & Brooks-Gunn, 2004; Hurley, Black, Papas, & Quigg, 2008; Jiang, Foster, & Gibson-Davis, 2010; Sparks 2010b). These differences may in part be due to self-selection for both WIC participation and breastfeeding behaviors among mothers (Besharov & Germanis, 2000; DeBate & Pyle, 2004; Lee & Mackey-Bilaver, 2007; Schwartz, Popkin, Tognetti, & Zohoori, 1995). Lee et al. (2009) present a comprehensive conceptual framework using social network theory and ecological theories of parenting that offers an interesting approach to examine WIC usage and breastfeeding. The social environment provides an important formal set of social networks that may provide breastfeeding support for mothers beyond her connections with family and friends. One of WIC’s central goals is to provide breastfeeding promotion and support to participants, and recent data indicate that the percentage of breastfeeding participants has increased (U.S. Department of Agriculture et al., 2010a, 2010b). Additionally, WIC benefits are tailored to women who breastfeed for longer durations; therefore, support from this program could serve as one means to increase breastfeeding durations in the United States (Jiang et al., 2010). By considering both formal and informal social networks, as well as sociodemographic characteristics, a more complete pattern of breastfeeding duration can be assessed with policy implications for formal programs like WIC. The purpose of this research was to assess racial/ethnic differences in breastfeeding duration among WIC-eligible participants using seven diverse racial/ethnic groups in rural and urban areas of the United States. More specifically, this research documents monthly breastfeeding survivorship for these different racial/ethnic groups then provides a multivariate analysis of breastfeeding duration, controlling for many sociodemographic characteristics. These findings contribute to the research literature and policy debate by clearly detailing breastfeeding patterns for groups often overlooked in the literature, including several Hispanic groups and Native Americans. If low-income minority women are less likely to initiate and continue breastfeeding, it is important to identify potential barriers to sustained periods of breastfeeding among women who initiate and focus on programs that could implement new policies and programmatic goals. Further breastfeeding programs must be developed and implemented that can meet the needs of these vulnerable women and their infants, and WIC serves as one outlet to reach this population. Methods Data Source Data for this analysis were taken from the Early Childhood Longitudinal StudydBirth Cohort (ECLS-B) Longitudinal 9-MonthdPreschool Restricted-Use data file (U.S. Department of Education, 2008). The ECLS-B follows a nationally representative probability sample of 14,000 children born in the United States during 2001, with specific sampling strategies to include a diverse set of socioeconomic and racial/ethnic groups in the population (U.S. Department of Education, 2005). Responses to the various instruments provide a comprehensive survey data source on child

375

health, developmental, and educational outcomes, while also detailing sociodemographic characteristics, social service use, and family dynamics. For this analysis, information contained on the child’s birth certificate and responses from the parental interview, conducted in person at the respondent’s home using a standardized questionnaire instrument when the sample child was 9 months and 2 years old were used. As a longitudinal study, the ECLS-B follows the same sampled children from birth through kindergarten, with approximately 10,550 children included in the first wave of data collection. Data for this analysis were obtained by permission and approval of the Institute of Education Sciences Data Security Office of the U.S. Department of Education, National Center for Education Statistics. Because of the sensitive nature of the information contained in the data, the author obtained a restricted data license allowing the use of the data source for this research purpose. All protocols specified in the restricted data license agreement were followed to protect the confidentiality of respondents. Sample Population The primary objective of this research was to examine breastfeeding duration, ranging from 1 to 12 months, among mothers who initiated breastfeeding and met income or categorical eligibility requirements to receive WIC benefits. More specifically, sampled children were excluded if their mother did not initiate breastfeeding and if the child’s family income was above 185% of the federally designated poverty threshold at the time of the 9-month parental interview or the mother was not receiving benefits from Medicaid, food stamps, or Temporary Assistance for Needy Families. These restrictions to the full sample gave an approximate sample size of 3,550 children for this analysis. Per NCES restricted data license agreements, all sample sizes reported here are rounded to the nearest 50 to protect respondent confidentiality. Breastfeeding Duration Mothers were first asked if they had initiated breastfeeding when the sampled child was born. Responses from this question were used to provide a description of the proportion of mothers that initiated breastfeeding. Next, mothers indicating that they had ever breastfed their child were asked how many months they had breastfeed the child. Possible answers to this question ranged from 0 (indicating less than 1 month) to 26 months. Responses from the 9-month and 2-year waves of data collection were used to ensure that all mothers indicating they had initiated breastfeeding had a value for breastfeeding duration in months. Therefore, the dependent variable for this analysis was the length of time in months the mother breastfed, with appropriate censoring indicators if the mother had not stopped breastfeeding at 12 months (right censored cases). Covariates Maternal race/ethnicity Three separate variables taken from the 9-month parental interview and the child’s birth certificate were used to specify if the mother was 1) Hispanic or non-Hispanic; 2) U.S. or foreign born; and 3) the mother’s race (self-identified). This classification scheme led to seven categories of race/ethnicity, including nonHispanic Whites (NHW), foreign-born Mexican-origin Hispanics

376

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

(FBMOH), U.S.-born Mexican-origin Hispanics (USMOH), nonHispanic Blacks (NHB), Native Americans, Asians (of all ethnicities), and other Hispanics, including Puerto Ricans and Cubans. Sociodemographic characteristics Low birthweight status (yes, no) was determined from the child’s birth certificate, and children weighing less than 2,500 grams at birth were identified as low birthweight. WIC usage (yes, no) was assessed by responses to the question, “Did you or your child receive benefits from WIC, that is the Special Supplemental Nutrition Program for Women, Infants, and Children, in the past 12 months?” Mothers were asked to give a self-assessment of their own health status, and a dummy variable was created to indicate poor maternal health status if the mother reported her health as fair or poor. Additionally, a variable was created based on the mother’s report of her previous live birth history to indicate first birth status with the sampled child (yes, no). Maternal age in years at the time of the child’s birth was included on the child’s birth certificate (<20, 20–34, or
35 years of age). The mother’s highest level of education completed was also available on the child’s birth certificate (
including family poverty status, maternal educational level, age of child when the mother returned to work, marital status, and urban residence. To deal with item missing data, predictive mean matching was used for each of the independent variables described, which replaces missing values with a new value based on responses or values on other items (Bodner, 2008). Imputation of missing values for the dependent variable was taken from the second wave of the data collection if information was missing at wave 1. Mean values for imputed values did not differ significantly from the original items. All protocols as specified in the restricted data license held by the author were followed to ensure respondent confidentiality. This research also received expedited institutional review board approval from the author’s university. Results Significant differences in the distribution of all variables based on maternal race/ethnicity were found for all covariates in the analysis, except for first birth status, including all breastfeeding variables (Table 1). Among all mothers, Asian, FBMOH, and other Hispanic mothers were most likely to initiate breastfeeding. Among WIC-eligible mothers, FBMOH mothers were most likely to breastfeed for 6 months, whereas NHB mothers were least likely to breastfeed for 6 months. USMOH mothers were least likely to breastfeed for 6 months among the Hispanic racial/ethnic groups. More than one quarter of WIC-eligible FBMOH mothers who initiated breastfeeding were still breastfeeding at 6 months. Overall, WIC usage was very high among mothers in this sample, but Asian mothers were least likely to make use of the program if they were eligible, whereas FBMOH and NHB mothers used WIC at the highest rates. The majority of mothers were married; however, Native American, other Hispanic, and FBMOH mothers were most likely to be cohabitating, and NHB and USMOH mothers were most likely to be single. Clear differences were observed by race/ethnicity for maternal educational levels, with NHW and Asians mothers most likely to have some college or more education, and FBMOH mothers most likely to have not finished high school. Most mothers did not reenter the workforce after their child was born, but mothers returning to work did so after about 3 months across all of the racial/ethnic groups. Breastfeeding survivorship by maternal race/ethnicity is shown in Figure 1 and detailed in Table 2. Differences in breastfeeding survivorship were noted as early as 3 months when comparing the racial/ethnic groups with the highest and lowest proportion of WIC-eligible mothers still breastfeeding, where appropriately 63% of FBMOH mothers were still breastfeeding while only 44% of USMOH mothers were breastfeeding among those that initiated breastfeeding (Table 2). After 4 months, fewer than half of NHW, USMOH, NHB, Native American, Asian, and other Hispanic mothers who initiated breastfeeding were still breastfeeding. This may be due to the high percentage of mothers returning to work once the child was 4 months of age, as noted in the bivariate analysis in Table 1. A crossover effect was observed in breastfeeding survivorship among USMOH and NHB mothers at 5 months, indicating that initially USMOH mothers quit breastfeeding at a faster rate than NHB women; however, USMOH women breastfeed at a more constant rate between 6 and 12 months. A similar type of crossover pattern was observed between Native American and Asian mothers, but the crossover occurred earlier, around 4 months. At 12 months, breastfeeding survivorship was highest

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

377

Table 1 Weighted Percentages of Breastfeeding Behaviors and Sociodemographic Characteristics by Maternal Race/Ethnicity, Early Childhood Longitudinal StudydBirth Cohort (ECLS-B), Wave 1dChild 9 Months Old Non-Hispanic Whites

Foreign-Born Mexican-Origin Hispanic

U.S.-Born Mexican-Origin Hispanics

Non-Hispanic Blacks

Breastfeeding behaviors among all mothers (n w 10,550) n (rounded to nearest 50) 5,000 750 400 1,800 71.29 79.77 65.57 49.27 Mother initiated breastfeedingy y 16.16 21.13 10.10 5.88 Mother breastfed for 6 months Breastfeeding and sociodemographic characteristics among WIC-eligible mothers initiating breastfeeding (n w 3,550) n (rounded to nearest 50) 1,300 500 200 650 14.29 28.50 11.37 9.98 Mother breastfed for 6 monthsy 7.72 5.94 4.92 11.05 Infant low birthweighty 74.46 88.99 85.43 88.42 Mother or child used WIC benefits in past 12 monthsy Parent self-rated poor health* 8.92 13.78 7.83 9.11 First birth 38.32 32.37 36.52 39.44 31.65 55.08 43.30 53.88 Family lives below poverty thresholdy Mother’s age at birth, yrs* <20 10.36 8.77 18.54 12.62 20–34 79.47 80.12 76.40 76.83
35 10.17 11.11 5.06 10.55 y Maternal education Less than high school 22.32 71.50 37.61 30.28 High school diploma 38.12 17.99 43.21 38.79 Some college or more 39.56 10.51 19.18 30.93 Month after birth mother returned to worky Mother not in labor force, months 44.30 63.54 43.96 37.76 1 9.09 3.71 7.43 6.99 2 17.17 7.44 18.94 13.96 3 9.60 7.13 7.39 13.48 4 4.80 5.44 6.57 8.07 5 2.32 1.70 1.77 5.86 6 4.28 4.37 5.74 5.61 >6 8.44 6.67 8.20 8.27 y Mother’s martial status Married 69.22 62.15 59.91 57.62 Cohabitating 18.19 30.41 21.47 14.12 Single 9.08 5.30 17.33 24.96 Widowed or divorced 3.51 2.14 1.29 3.30 72.03 95.49 95.67 95.46 Urban residencey

Native American

Asian

Other Hispanic

550 64.23 11.37

1,500 80.37 17.14

550 77.10 14.54

250 15.70 5.41 88.72

400 15.83 7.54 68.04

250 16.31 6.93 79.71

12.28 39.56 45.27

8.38 42.87 33.44

5.70 40.53 39.38

17.07 74.26 8.67

7.45 72.68 19.87

9.34 77.21 13.45

35.59 35.34 29.07

24.74 30.87 44.39

46.97 31.83 21.20

41.58 8.44 13.24 13.94 5.64 3.46 3.69 10.01

55.75 6.06 12.17 8.81 3.92 3.31 4.18 5.80

40.03 7.29 21.42 16.73 4.03 0.87 3.37 6.26

48.55 37.48 11.29 2.68 71.85

77.88 11.54 8.95 1.63 98.31

54.74 31.32 10.25 3.69 97.94

Weight: W1R0. 2 * p  .01 for differences across racial/ethnic and residence groups, on the basis of c tests for equal distributions. y p  .001 for differences across racial/ethnic and residence groups, on the basis of c2 tests for equal distributions.

among FBMOH, Native American, NHW, and Asian WIC-eligible mothers. Breastfeeding survivorship at 12 months was highest among FBMOH mothers and lowest among NHB mothers. Few differences in breastfeeding durations were noted in the Cox proportional hazards model when only maternal race/ ethnicity was included in the model (model 1, Table 3). Among WIC-eligible mothers, FBMOH mothers were more likely to continue breastfeeding over the period compared with NHW mothers, whereas NHB mothers were less likely to continue breastfeeding compared with NHW mothers. Breastfeeding discontinuation risks were higher among women giving birth to a low birthweight infant, WIC recipients, mothers reporting poor and fair self-rated health, and younger mothers (model 2, Table 3). A protective effect was still observed for FBMOH mothers and continued breastfeeding with the inclusion of these variables in Model 2. Mothers with less than a high school education or a high school diploma were more likely to discontinue breastfeeding over the 12-month period, compared with mothers with some college education or more (model 3, Table 3). Mothers returning to work 1 to 3 months after the child was born were more likely to discontinue breastfeeding than mothers who did not reenter the labor force after controlling for all other variables in the model. Single mothers were more likely to discontinue breastfeeding than married mothers. No

differences in breastfeeding durations were noted between cohabitating, widowed or divorced, and married mothers. WIC-eligible mothers living in urban areas were more likely to continue breastfeeding than their counterparts in rural areas. The only significant racial/ethnic disparity in breastfeeding duration was noted between FBMOH and NHW mothers in the full model; the association noted shows a breastfeeding duration advantage for FBMOH mothers even when controls for maternal educational levels, return to work, marital status, and residential location were included in the model. Although the results are not presented herein, interaction terms for maternal race/ethnicity and WIC participation were included in a final model with all covariates included in model 3. A marginally significant association was noted for other Hispanic mothers using WIC (hazard ratio ¼ 0.71; 95% confidence interval ¼ 0.51–0.99). More specifically, other Hispanic mothers using WIC were more likely to continue breastfeeding than NHW mothers using WIC. All other model coefficients remained the same as those presented in model 3 in Table 3. Discussion With increased attention to the health, developmental, and economic benefits of breastfeeding (Bartick & Reinhold, 2010), it

378

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

Figure 1. Breastfeeding survivorship by maternal race/ethnicity among WICeligible mothers.

is not surprising that breastfeeding rates have increased or remained fairly stable over the past few years. However, results from this research indicate that breastfeeding initiation and continuation rates do not meet the goals of Healthy People 2010 (Centers for Disease Control and Prevention, 2009; U.S. Department of Health and Human Services, 2000). Further results presented here from a diverse set of WIC-eligible families indicate wide variation in breastfeeding rates across racial/ethnic groups. If breastfeeding initiation, continuation, and exclusivity rates are to increase in the near future, racial/ ethnic minority and low-income mothers must be given the tools and resources to be successful at navigating decisions with breastfeeding. From an ecological theoretical perspective, this involves situating individual women and their decision-making process within their local environment and giving them appropriate social and institutional support to obtain education assistance from programs like WIC for successful breastfeeding transitions at birth (DeBate & Pyle, 2004; Hovell, Wahlgren, & Gehrman, 2002; Lee et al., 2009; McLeroy, Bibeau, Steckler, & Glanz, 1988).

Several sociodemographic factors have been identified in the literature to influence breastfeeding duration, including lowincome status (Ahluwalia, Morrow, Hsia, & Grummer-Strawn, 2003; Ahluwalia et al., 2005; Barton, 2001; Hurley et al., 2008; Ryan, Zhou, & Gaston, 2004), maternal race/ethnicity (Ahluwalia et al., 2003; Chin, Myers, & Magnus, 2008; Forste et al., 2001; Hurley et al., 2008; Li, Fridinger, & GrummerStrawn, 2004; Li & Grummer-Strawn, 2002), immigrant status (Celi, Rich-Edwards, Richardson, Kleinman, & Gillman, 2005; Gibson-Davis & Brooks-Gunn, 2006; Gibson, Diaz, Mainous, & Geesey, 2005), maternal educational levels (Chin et al., 2008; Li, Darling, Maurice, Barker, & Grummer-Strawn, 2005), maternal age (Scott & Binns, 1999; van Rossem et al., 2009), marital status (Kiernan & Pickett 2006; Lee et al., 2009; Thulier & Mercer, 2009), employment status (Fein, Mandal, & Roe, 2008; Fein & Roe, 1998; Guendelman et al., 2009), WIC participation (Chatterji & Brooks-Gunn, 2004; Jacknowitz, Novillo, & Tiehen, 2007), rural residential location (Flower et al., 2008; Ryan et al., 2004; Sparks, 2010b), and low birthweight status of the infant (Smith, Durkin, Hinton, Bellinger, & Kuhn, 2003). Results from this study support other work that finds low-income, minority women, and WIC participants are less likely to breastfeed (Lee et al., 2009) and continue breastfeeding if they initiate (Chatterji & Brooks-Gunn, 2004). Low-income status may work through several mechanisms to influence breastfeeding cessation. Research shows that low-income mothers may have to return to work more quickly after a birth owing to less flexibility in certain employment sectors to accommodate continued breastfeeding if breastfeeding is initiated (Guendelman et al., 2009; Hill, Arnett, & Mauk, 2008; Raisler, 2000). Further, limited social support from a partner or other family members may make breastfeeding more difficult for women from more deprived backgrounds (Mitra, Khoury, Hinton, & Carothers, 2004). WIC usage also poses potential problems with breastfeeding continuation among low-income mothers owing to the availability of formula supplements (Chatterji & Brooks-Gunn, 2004; McCann, Baydar, & Williams, 2007; Racine, Frick, Guthrie, & Strobino, 2009), and results presented here indicate that WIC participants were more likely to discontinue breastfeeding over the 12-month period. However, the benefits a mother receives from WIC depends on whether or not she breastfeeds, and mothers and children receive the benefits for a shorter amount of time if the mother elects to use formula (Jiang et al., 2010). Bivariate tests show that WIC-eligible FBMOH mothers use WIC at the highest rates and the greatest proportion

Table 2 Kaplan–Meier Survival Estimates for Breastfeeding Duration Among WIC-Eligible Mothers Who Initiated Breastfeeding Based on Maternal Race/Ethnicity, Early Childhood Longitudinal StudydBirth Cohort, Wave 1: Child 9 Months Old (n w 3,550) Month

Non-Hispanic White

Foreign-Born Mexican Origin Hispanic

US-Born Mexican Origin Hispanic

Non-Hispanic Black

Native American

Asian

Other Hispanic

0 1 2 3 4 5 6 7 8 9 10 11 12

1.0000 0.9457 0.6404 0.4828 0.3856 0.3145 0.2777 0.2181 0.1898 0.1676 0.1431 0.1301 0.1156

1.0000 0.9312 0.7387 0.6307 0.5305 0.4735 0.4342 0.3635 0.3065 0.2790 0.2436 0.2200 0.1984

1.0000 0.9570 0.6129 0.4409 0.2796 0.2151 0.1881 0.1452 0.1237 0.1129 0.1075 0.0914 0.0860

1.0000 0.9624 0.6729 0.5039 0.3224 0.2285 0.1831 0.1221 0.1033 0.0923 0.0673 0.0626 0.0548

1.0000 0.9174 0.6696 0.5609 0.4087 0.3261 0.3087 0.2435 0.2217 0.1783 0.1652 0.1522 0.1391

1.0000 0.9229 0.6427 0.5244 0.4319 0.3470 0.3316 0.2751 0.2417 0.2031 0.1671 0.1465 0.1285

1.0000 0.9389 0.6511 0.4892 0.3813 0.3489 0.3166 0.2482 0.2230 0.1871 0.1619 0.1367 0.1151

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

379

Table 3 Cox Proportional Hazards Models of Breastfeeding Duration Among WIC-Eligible Mothers, Early Childhood Longitudinal StudydBirth Cohort (n w 3,550)

Maternal race/ethnicity Non-Hispanic White (reference) Foreign-Born Mexican-origin Hispanic U.S.-Born Mexican-origin Hispanic Non-Hispanic Black Native American Asian Other Hispanics Low birthweight Mother or child used WIC in past 12 months Parent self-rated poor health First birth Mother’s age at birth, yrs <20 20–34 (reference)
35 Family lives below poverty threshold Maternal education Less than high school High school diploma Some college or more (reference) Month after birth mother returned to work Mother not in labor force (reference) 1 2 3 4 5 6 >6 Mother’s martial status Married (reference) Cohabitating Single Widowed or divorced Urban residence Wald Chi-square 2-loglikelihood Degrees of freedom

Model 1 HR (95% CI)

Model 2 HR (95% CI)

Model 3 HR (95% CI)

1.00 0.73 1.16 1.16 0.93 0.93 0.93

1.00 0.71 1.13 1.09 0.92 1.00 0.93 1.26 1.26 1.16 1.06

1.00 0.70 1.11 1.08 0.89 1.06 0.90 1.30 1.19 1.14 1.07

(0.65–0.82)* (0.99–1.36) (1.06–1.28)y (0.80–1.08) (0.82–1.04) (0.82–1.06)

(0.64–0.80)* (0.96–1.32) (0.99–1.21) (0.79–1.07) (0.88–1.12) (0.81–1.06) (1.16–1.37)* (1.15–1.39)* (1.03–1.31)z (0.98–1.14)

1.25 (1.11–1.40)* 1.00 0.92 (0.82–1.02)

(0.62–0.80)* (0.94–1.31) (0.98–1.20) (0.77–1.04) (0.93–1.20) (0.79–1.04) (1.19–1.41)* (1.08–1.32)* (1.01–1.28)z (0.98–1.15)

1.14 (1.00–1.29)z 1.00 0.93 (0.93–1.04) 1.06 (0.98–1.14) 1.21 (1.09–1.34)* 1.22 (1.11–1.33)* 1.00 1.00 1.29 1.15 1.18 1.08 0.96 1.07 0.94

60.50* 47,221.46 6

157.44* 47,127.09 12

(1.12–1.49)* (1.03–1.29)y (1.05–1.34)y (0.93–1.25) (0.78–1.20) (0.90–1.26) (0.82–1.07)

1.00 1.09 (0.99–1.20) 1.13 (1.00–1.28)z 1.14 (0.93–1.39) 0.90 (0.81–0.99)z 221.52* 47,063.82 26

Abbreviations: CI, confidence intervals; HR, hazard ratios. * p  .001. y p  .01. z p  .05.

of these mothers also live in poverty, yet FBMOH mothers have the highest rates of breastfeeding initiation and 6-month breastfeeding and witness a breastfeeding duration advantage compared with WIC-eligible NHW mothers in the multivariate models. More research is needed to see how targeted educational programs and other components of the WIC program may support breastfeeding among a diverse group of mothers who decide to breastfeed during the prenatal period. Beyond women using WIC, health programs and policies at the local level must support employers to ensure women have adequate time and space to pump breast milk once they return to work (DeBate & Pyle, 2004). Other associations between the variables highlighted and breastfeeding duration are not constant across racial/ethnic groups in this analysis. The sociodemographic profiles of NHB, FBMOH, USMOH, and other Hispanic mothers are very similar (Table 1). However, distinct patterns are noted in breastfeeding durations between these groups (Figure 1 and Table 2), with a consistent FBMOH breastfeeding advantage despite their very low socioeconomic positions relative to the other racial/ethnic groups. The higher sustained breastfeeding durations of FBMOH women and the crossover pattern observed between NHB and

USMOH mothers speaks more broadly to the literature that recognizes the roles of cultural beliefs and traditions in determining breastfeeding behaviors among Hispanic women more generally (Ceti, Rich-Edwards, Richardson, Kleinman, & Gillman, 2005; Gill, 2009; Gorman, Madlensky, Jackson, Ganiats, & Boies, 2007). The higher continued breastfeeding rates observed among FBMOH women compared with USMOH or other Hispanic women support other research that finds lowered rates of breastfeeding initiation and continuation among more acculturated Hispanic subgroups (Gibson-Davis & Brooks-Gunn, 2006; Kimbro, Lynch, & McLanahan, 2008; Singh, Kogan, & Dee, 2007). Health programs would benefit from knowing how formal and informal social networks among this ethnic group may translate into successful breastfeeding practices. More research is needed to understand the duration crossover observed between Asian and Native American mothers, which is a new contribution from this research. Limitations Some discussion of limitations to this research is warranted. First, the data source used for this analysis is nationally

380

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

representative of all births occurring in the United States in 2001, but the analytic sample included only children living in families that met income or categorical requirements to receive WIC benefits. With that being said, there was specific oversampling for low-income families, which helped to ensure a large enough sample size across the variables used and appropriate adjustments where made to the analysis to correct standard errors estimates and other statistics for hypothesis testing. Still, the ECLS-B sample was not originally designed to study this specific subpopulation. Second, with the use of any secondary data source, and longitudinal data in particular, item missing data becomes important for consideration. All variables from this analysis came from one wave of data collection or the child’s birth certificate; however, information from the second wave of data collection was used to impute missing values on the dependent variable for breastfeeding duration if it was missing at wave one. Fewer than 3% of all cases were missing on any one variable, and in most cases, the value was much smaller; predictive mean matching, as detailed, was used to account for item missing data. Further, all missing cases for breastfeeding duration in wave one were available in wave two. Predictive mean matching was also used for this variable to compare the two imputation methods, but the mean and median was not substantially different between the two methods and the information reported by the mother at wave two was used to impute values. Third, initial tests were conducted to see if each of the racial/ethnic groups, except for Native Americans, could be classified as native or foreign born. Because of small sample sizes with the number of women discontinuing breastfeeding in any one month, this analysis was not possible. However, future data collection strategies would benefit from designing a sample that considered this characteristic because the U.S. population is becoming more diverse as a result of immigration. Last, self-selection bias associated with both WIC participation and breastfeeding behaviors is always a problem when using observational data (Jiang et al., 2010; Schwartz et al., 1995; Sparks, 2010a). Implications for Policy and Practice A contribution of this study is the documentation of breastfeeding survivorship among a diverse set of WIC-eligible mothers from seven racial/ethnic groups in both rural and urban areas of the United States, including Native American mothers. For most mothers, breastfeeding rates drop quickly after 3 months among mothers who initiate, and this is likely because of the number of mothers returning to work between 1 and 3 months after the child is born. Work-related responsibilities might play an important role in determining breastfeeding duration for low-income mothers (Gielen, Faden, O’Campo, Brown, & Paige, 1991; Guendelman et al., 2009), as noted in both the bivariate and multivariate results. The new Patient Protection and Affordable Care Act, which requires employers to provide break time in privately designated locations for new mothers to express milk, may provide some mothers, particularly those who must reenter the workforce soon after their child is born, with the opportunity to continue breastfeeding and gain wider acceptance for breastfeeding in public (U.S. Department of Labor, 2010). However, employers with fewer than 50 employees are not subject to this act, and this could disproportionately impact low-income and minority mothers working for smaller businesses. Other mechanisms, such as the child’s health, WIC participation, support from a spouse, knowledge and skills from some college education or more, and residential location seem to

account for differences in breastfeeding duration among this sample of WIC-eligible mothers. Further, only FBMOH mothers come close to meeting the recommendations of Healthy People 2010 once health behaviors and sociodemographic characteristics are accounted for in the multivariate models relative to NHW mothers. Taken together, policy recommendations must consider the multifaceted determinants of breastfeeding behaviors using something akin to the behavioral ecological model for interpretation (DeBate & Pyle, 2004; Hovell et al., 2002; Lee et al., 2009; McLeroy et al., 1988). First results presented here indicate that the vast majority of WIC-eligible mothers initiate breastfeeding; however, many of these mothers discontinue breastfeeding some after their child is born. Several perceived barriers identified by new mothers in qualitative studies, such as economic considerations, maintaining female autonomy and independence, receiving adequate support from family and friends, experiencing limited support from health care professionals, and fatigue, may explain the sharp discontinuation patterns for some mothers (Flower et al., 2008; Guttman & Zimmerman, 2000; McFadden & Toole, 2006; Schmied & Lupton, 2001). Larger societal attitudes toward breastfeeding, particularly breastfeeding in public or in the workplace, may present different barriers to breastfeeding initiation and duration for low-income and minority women (Acker, 2009; Guttman & Zimmerman, 2000; Li, Rock, & Grummer-Strawn, 2007). More broadly, the results presented herein speak to the need to identify appropriate counseling and educational programs, as well as support mechanisms, to meet the needs of many low-income, minority women. Additionally, the strong WIC participation levels noted herein make WIC an ideal starting point for addressing many of these concerns. Second, the Loving Support Peer Counseling Program provided in WIC offices throughout the United States may offer one established program to address the concerns of low-income mothers about breastfeeding. This program provides WIC offices with flexibility in delivering services, but eight activities are targeted to promote breastfeeding among participants, including 1) providing breastpumps, 2) training WIC staff in breastfeeding promotion activities, 3) conducting media campaigns and providing educational materials, 4) supporting other counseling activities, 5) hosting classes and support groups for WIC partici pants, 6) making lactation consultants available to all participants, 7) offering training for lactation consultant certification, and 8) supplying a telephone hotline to address questions and concerns from participants (U.S. Department of Agriculture et al., 2010a). The majority of programs participate in the first three items, but more work could be done to promote all of these activities, particularly items 7 and 8. More targeted educational and support programs available during WIC sessions (Beal et al., 2003; Chatterji & Brooks-Gunn, 2004) based on the Loving Support Peer Counseling Program could offer low-income mothers the formal social support necessary to navigate breastfeeding and other social and family demands. Lactation consultants, a service generally available to higher income mothers (Guttman & Zimmerman, 2000), could provide lowincome women of all races/ethnicities with the guidance and support to continue breastfeeding if they encounter problems after they have initiated breastfeeding (Keister et al., 2008). Additional funding for this program was authorized in 2010 and speaks to the comprehensive approach needed to meet the breastfeeding needs of a diverse set of women with many competing economic, social, and personal decisions regarding parenting.

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

Last, women must feel comfortable with their feeding choices, whether that is to breastfeed or formula feed. Many low-income women face substantial barriers or pressures to breastfeed in their immediate environment, even if they have the knowledge about the benefits of breastfeeding (Schmied & Lupton, 2001). Culturally sensitive programs must acknowledge the many competing interests when making the decision to initiate or continue breastfeeding, and strong social and formal support networks must respect and provide assistance for the feeding decisions of all mothers. Acknowledgments The author thanks Corey S. Sparks, PhD, for assistance with graphics for this manuscript. References Acker, M. (2009). Breast is bestdBut not everywhere: Ambivalent sexism and attitudes toward private and public breastfeeding. Sex Roles, 61, 476–490. Ahluwalia, I., Morrow, B., Hsia, J., & Grummer-Strawn, L. (2003). Who is breastfeeding? Recent trends from the pregnancy risk assessment and monitoring system. Journal of Pediatrics, 142, 486–491. Ahluwalia, I. B., Morrow, B., & Hsia, J. (2005). Why do women stop breastfeeding? Findings from the pregnancy risk assessment and monitoring system. Pediatrics, 116, 1408–1412. American Academy of Pediatrics. (2005). Policy statement. Breastfeeding and the use of human milk. Pediatrics, 115, 496–506. Bartick, M., & Reinhold, A. (2010). The burden of suboptimal breastfeeding in the United States: A pediatric cost analysis. Pediatrics, 125, e1048–e1056. Bartick, M., Stuebe, A., Shealy, K. R., Walker, M., & Grummer-Strawn, L. M. (2009). Closing the quality gap: Promoting evidence-based breastfeeding care in the hospital. Pediatrics, 124, E793–E802. Barton, S. J. (2001). Infant feeding practices of low-income rural mothers. The American Journal of Maternal/Child Nursing, 26, 93–97. Beal, A. C., Kuhlthau, K., & Perrin, J. M. (2003). Breastfeeding advice given to African American and white women by physicians and WIC counselors. Public Health Reports, 118, 368–376. Besharov, D. J., & Germanis, P. (2000). Evaluating WIC. Evaluation Review, 24, 123–190. Bodner, T. E. (2008). What improves with increased missing data imputations? Structural Equation Modeling: A Multidisciplinary Journal, 15, 651–674. Bunik, M., Krebs, N. F., Beaty, B., McClatchey, M., & Olds, D. L. (2009). Breastfeeding and WIC Enrollment in the Nurse Family Partnership Program. Breastfeeding Medicine, 4, 146–150. Celi, A. C., Rich-Edwards, J. W., Richardson, M. K., Kleinman, K. P., & Gillman, M. W. (2005). Immigration, race/ethnicity, and social and economic factors as predictors of breastfeeding initiation. Archives of Pediatrics & Adolescent Medicine, 159, 255–260. Centers for Disease Control and Prevention. (2009). Breastfeeding report card: United States, 2009. Atlanta: U.S. Department of Health and Human Services. Ceti, A. C., Rich-Edwards, J. W., Richardson, M. K., Kleinman, K. P., & Gillman, M. W. (2005). Immigration, race/ethnicity, and social and economic factors as predictors of breastfeeding initiation. Archives of Pediatrics & Adolescent Medicine, 159, 255–260. Chatterji, P., & Brooks-Gunn, J. (2004). WIC participation, breastfeeding practices, and well-child care among unmarried, low-income mothers. American Journal of Public Health, 94, 1324–1327. Chin, A. C., Myers, L., & Magnus, J. H. (2008). Race, education, and breastfeeding initiation in Louisiana, 2000-2004. Journal of Human Lactation, 24, 175–185. DeBate, R. D., & Pyle, G. F. (2004). The behavioral ecological model: A framework for early WIC participation. American Journal of Health Studies, 19, 138–147. Fein, S. B., Mandal, B., & Roe, B. E. (2008). Success of strategies for combining employment and breastfeeding. Pediatrics, 122, S56–S62. Fein, S. B., & Roe, B. (1998). The effect of work status on initiation and duration of breast-feeding. American Journal of Public Health, 88, 1042–1046. Flower, K. B., Willoughby, M., Cadigan, R. J., Perrin, E. M., Randolph, G., & The Family Life Project Investigative Team. (2008). Understanding breastfeeding initiation and continuation in rural communities: A combined qualitative/ quantitative approach. Maternal and Child Health Journal, 12, 402–414. Forste, R., & Hoffmann, J. P. (2008). Are US mothers meeting the Healthy People 2010 breastfeeding targets for initiation, duration, and exclusivity? The 2003 and 2004 National Immunization Surveys. Journal of Human Lactation, 24, 278–288. Forste, R., Weiss, J., & Lippincott, E. (2001). The decision to breastfeed in the United States: Does race matter? Pediatrics, 108, 291–296.

381

Gibson, M. V., Diaz, V. A., Mainous, A. G., & Geesey, M. E. (2005). Prevalence of breastfeeding and acculturation in Hispanics: Results from NHANES 1999-2000 study. Birth: Issues in Perinatal Care, 32, 93–98. Gibson-Davis, C. M., & Brooks-Gunn, J. (2006). Couples’ immigration status and ethnicity as determinants of breastfeeding. American Journal of Public Health, 96, 641–646. Gielen, A. C., Faden, R. R., O’Campo, P., Brown, C. H., & Paige, D. M. (1991). Maternal employment during the early postpartum period: Effects on initiation and continuation of breast-feeding. Pediatrics, 87, 298–305. Gill, S. L. (2009). Breastfeeding by Hispanic women. Journal of Obstetric Gynecologic and Neonatal Nursing, 38, 244–252. Gorman, J. R., Madlensky, L., Jackson, D. J., Ganiats, T. G., & Boies, E. (2007). Early postpartum breastfeeding and acculturation among Hispanic women. BirthIssues in Perinatal Care, 34, 308–315. Guendelman, S., Kosa, J. L., Pearl, M., Graham, S., Goodman, J., & Kharrazi, M. (2009). Juggling work and breastfeeding: Effects of Maternity leave and occupational characteristics. Pediatrics, 123, E38–E46. Guttman, N., & Zimmerman, D. R. (2000). Low-income mothers’ views on breastfeeding. Social Science & Medicine, 50, 1457–1473. Hill, G. J., Arnett, D. B., & Mauk, E. (2008). Breast-feeding intentions among low-income pregnant and lactating women. American Journal of Health Behavior, 32, 125–136. Hovell, M. F., Wahlgren, D. R., & Gehrman, C. A. (2002). The behavioral ecological model: Integrating public health and behavioral science. In R. J. DiClemente, R. A. Crosby, & M. C. Kegler (Eds.), Emerging theories in health promotion practice and research (pp. 347–385). San Francisco: Jossey-Bass. Hurley, K. M., Black, M. M., Papas, M. A., & Quigg, A. M. (2008). Variation in breastfeeding behaviours, perceptions, and experiences by race/ethnicity among a low-income statewide sample of Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) participants in the United States. Maternal and Child Nutrition, 4, 95–105. Jacknowitz, A., Novillo, D., & Tiehen, L. (2007). Special supplemental nutrition program for women, infants, and children and infant feeding practices. Pediatrics, 119, 281–289. James, D. C. S., & Lessen, R. (2009). Position of the American Dietetic Association: Promoting and supporting breastfeeding. Journal of the American Dietetic Association, 109, 1926–1942. Jiang, M., Foster, E. M., & Gibson-Davis, C. M. (2010). The effect of WIC on breastfeeding: A new look at an established relationship. Children and Youth Services Review, 32, 264–273. Keister, D., Roberts, K. T., & Werner, S. L. (2008). Strategies for breastfeeding success. American Family Physician, 78, 225–232. Kiernan, K., & Pickett, K. E. (2006). Marital status disparities in maternal smoking during pregnancy, breastfeeding and maternal depression. Social Science & Medicine, 63, 335–346. Kimbro, R. T., Lynch, S. M., & McLanahan, S. (2008). The influence of acculturation on breastfeeding initiation and duration for Mexican-Americans. Population Research and Policy Review, 27, 183–199. Lee, B. J., & Mackey-Bilaver, L. (2007). Effects of WIC and Food Stamp Program participation on child outcomes. Children and Youth Services Review, 29, 501–517. Lee, H. J., Elo, I. T., McCollum, K. F., & Culhane, J. F. (2009). Racial/ethnic differences in breastfeeding initiation and duration among low-income inner-city mothers. Social Science Quarterly, 90, 1251–1271. Li, R., Fridinger, F., & Grummer-Strawn, L. (2004). Racial/ethnic disparities in public opinion about breastfeeding: The 1999–2000 healthstyles surveys in the United States. Advances in Experimental Medicine and Biology, 554, 287–291. Li, R. W., Darling, N., Maurice, E., Barker, L., & Grummer-Strawn, L. M. (2005). Breastfeeding rates in the United States by characteristics of the child, mother, or family: The 2002 National Immunization Survey. Pediatrics, 115, E31–E37. Li, R. W., & Grummer-Strawn, L. (2002). Racial and ethnic disparities in breastfeeding among United States infants: Third National Health and Nutrition Examination Survey, 1988–1994. Birth – Issues in Perinatal Care, 29, 251–257. Li, R. W., Rock, V. J., & Grummer-Strawn, L. (2007). Changes in public attitudes toward breastfeeding in the United States, 1999-2003. Journal of the American Dietetic Association, 107, 122–127. McCann, M. F., Baydar, N., & Williams, R. L. (2007). Breastfeeding attitudes and reported problems in a national sample of WIC participants. Journal of Human Lactation, 23, 314–324. McFadden, A., & Toole, G. (2006). Exploring women’s views of breastfeeding: A focus group study within an area with high levels of socio-economic deprivation. Maternal and Child Nutrition, 2, 156–168. McLeroy, K. R., Bibeau, D., Steckler, A., & Glanz, K. (1988). An ecological perspective on health promotion programs. Health Education Quarterly, 15, 351–377. Mitra, A. K., Khoury, A. J., Hinton, A. W., & Carothers, C. (2004). Predictors of breastfeeding intention among low-income women. Maternal and Child Health Journal, 8, 65–70. Ogbuanu, C. A., Probst, J., Laditka, S. B., Liu, J. H., Baek, J., & Glover, S. (2009). Reasons why women do not initiate breastfeeding: A southeastern state study. Women’s Health Issues, 19, 268–278.

382

P.J. Sparks / Women's Health Issues 21-5 (2011) 374–382

Racine, E. F., Frick, K., Guthrie, J. F., & Strobino, D. (2009). Individual net-benefit maximization: A model for understanding breastfeeding cessation among low-income women. Maternal and Child Health Journal, 13, 241–249. Raisler, J. (2000). Against the odds: Breastfeeding experiences of low income mothers. Journal of Midwifery & Womens Health, 45, 253–263. Ryan, A. S., Zhou, W. J., & Acosta, A. (2002). Breastfeeding continues to increase into the new millennium. Pediatrics, 110, 1103–1109. Ryan, A. S., Zhou, W. J., & Gaston, M. H. (2004). Regional and sociodemographic variation of breastfeeding in the United States, 2002. Clinical Pediatrics, 43, 815–824. Schmied, V., & Lupton, D. (2001). Blurring the boundaries: Breastfeeding and maternal subjectivity. Sociology of Health & Illness, 23, 234–250. Schwartz, J. B., Popkin, B. M., Tognetti, J., & Zohoori, N. (1995). Does WIC participation improve breast-feeding practices? American Journal of Public Health, 85, 729–731. Scott, J. A., & Binns, C. W. (1999). Factors associated with the initiation and duration of breastfeeding: A review of the literature. Breastfeeding Review, 7, 5–16. Singh, G. K., Kogan, M. D., & Dee, D. L. (2007). Nativity/immigrant status, race/ ethnicity, and socioeconomic determinants of breastfeeding initiation and duration in the United States, 2003. Pediatrics, 119, S38–S46. Smith, M. M., Durkin, M., Hinton, V. J., Bellinger, D., & Kuhn, L. (2003). Initiation of breastfeeding among mothers of very low birth weight infants. Pediatrics, 111, 1337–1342. Sparks, P. J. (2010a). Childhood morbidities among income- and categoricallyeligible WIC program participants and non-participants. Journal of Children and Poverty, 16, 47–66. Sparks, P. J. (2010b). Rural-urban differences in breastfeeding initiation in the United States. Journal of Human Lactation, 26, 118–129. Thulier, D., & Mercer, J. (2009). Variables associated with breastfeeding duration. JOGNN-Journal of Obstetric Gynecologic and Neonatal Nursing, 38, 259–268.

U.S. Department of Agriculture, Food and Nutrition Service, and Office of Research and Analysis. (2010a). WIC Breastfeeding peer counseling study, final implementation report. Alexandria, VA: Author. U.S. Department of Agriculture, Food and Nutrition Service, and Office of Research and Analysis. (2010b). WIC participant and program characteristics 2008. Alexandria, VA: Author. U.S. Department of Education. (2008). Early childhood longitudinal study, birth cohort (ECLS-B), longitudinal 9-month-preschool restricted-use data file (NCES 2008-024). Washington, DC: National Center for Education Statistics. U.S. Department of Education (Ed.). (2005). User’s manual for the ECLS-B ninemonth private data file and electronic code book. Washington, DC: Author. U.S. Department of Health and Human Services. (2000). Healthy people 2010: Understanding and improving health (2nd ed). Washington, DC: U.S. Government Printing Office. U.S. Department of Labor. (2010). Fact sheet #73: Break time for nursing mothers under the FLSA. Washington, DC: U.S. Department of Labor, Wage and Hour Division. van Rossem, L., Oenema, A., Steegers, E. A. P., Moll, H. A., Jaddoe, V. W. V., Hofman, A., et al. (2009). Are starting and continuing breastfeeding related to educational background? The Generation R Study. Pediatrics, 123, E1017–E1027.

Author Descriptions P. Johnelle Sparks, PhD, is an assistant professor in the Department of Demography at the University of Texas at San Antonio. Her research interests include maternal and child health, racial/ethnic and income disparities in child health, spatial inequality, and research methods.