Impact of Policy Changes on Infant Feeding Decisions Among Low-Income Women Participating in the Special Supplemental Nutrition Program for Women, Infants, and Children

Abstract

Objectives. We present infant feeding data before and after the 2009 Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) food package change that supported and incentivized breastfeeding. We describe the key role of California WIC staff in supporting these policy changes.

Methods. We analyzed WIC data on more than 180 000 infants in Southern California. We employed the analysis of variance and Tukey (honestly significant difference) tests to compare issuance rates of postpartum and infant food packages before and after the changes. We used analysis of covariance to adjust for poverty status changes as a potential confounder.

Results. Issuance rates of the “fully breastfeeding” package at infant WIC enrollment increased by 86% with the package changes. Rates also increased significantly for 2- and 6-month-old infants. Issuance rates of packages that included formula decreased significantly. All outcomes remained highly significant in the adjusted model.

Conclusions. Policy changes, training of front-line WIC staff, and participant education influenced issuance rates of WIC food packages. In California, the issuance rates of packages that include formula have significantly decreased and the rate for those that include no formula has significantly increased.

The Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), funded by the US Department of Agriculture (USDA), provides nutritious food and nutrition and breastfeeding education to low-income pregnant and postpartum women, and children up to age 5 years. It is considered one of the premier public health nutrition programs in the United States; more than 9 million women, infants, and children receive WIC benefits,¹ with more than 1.4 million in California alone.² Nationwide, about half of all infants receive services from the WIC program.¹

Breastfeeding support is a significant priority of the WIC program and, as outlined in the 2011 Surgeon General’s Call to Action to Support Breastfeeding,³ is a nationally recognized primary prevention strategy for early childhood obesity. Numerous studies have shown breastfeeding to be associated with a 20% to 50% lower risk of overweight and obesity in children, depending on duration of breastfeeding and degree of supplementation.^4,5 At the same time, research has documented that WIC participation is associated with lower rates of breastfeeding when one compares WIC participants to nonparticipants.^6–12 Obesity disproportionately affects low-income individuals, and implementation of policies that increase breastfeeding rates among low-income families served by WIC may be an important strategy for combatting the obesity epidemic. Motivated by these assumptions, the 2006 Institute of Medicine Committee to Review the WIC Food Packages recommended significant policy changes to the WIC Program to incentivize and support breastfeeding.

The year 2009 marked a historic change to the WIC program by increasing support of breastfeeding and aligning the foods available through the WIC program, referred to as the WIC food package, with the 2005 Dietary Guidelines for Americans. On the basis of the Institute of Medicine recommendations, USDA made significant changes to the food packages for women and children to include fruits, vegetables, and whole grains, and to limit milk purchases to only lower-fat options for all women and all children older than 2 years.¹³ For postpartum women and infants, significant changes were made to the food packages to better incentivize and support breastfeeding: increasing the value of the WIC package for mothers who fully breastfeed, reducing the amount of formula for mothers who partially breastfeed, calibrating formula amounts for infants by age, and postponing complementary infant foods. In addition, California adopted the suggested federal policy of no routine issuance of infant formula to breastfeeding mothers in the first month postpartum.¹³ These breastfeeding-supportive changes were welcomed by the public health community, with hopes that the changes would increase breastfeeding rates among low-income mothers served by WIC. The objective of this study was to assess whether the key goals of the changes in the breastfeeding policies and food packages—to increase the issuance of the infant food package that does not include formula and decrease the issuance of the infant food packages that include formula—were achieved among a large population of WIC participants in California.

METHODS

The California WIC program implemented new WIC food packages on October 1, 2009. The breastfeeding policy changes included (1) an increase in the amount and variety of food benefits for fully breastfeeding mothers and infants receiving no formula from WIC, with the increase for breastfeeding mothers starting immediately postpartum and the increase for infants starting at age 6 months; (2) for those receiving formula, calibration of infant formula amounts to the age of the infant; and (3) adoption of the federal recommendation of no routine issuance of formula to breastfeeding mothers in the first 30 days postpartum.¹³

Study Site, Staff Training, and Participant Education

Public Health Foundation Enterprises (PHFE) WIC is the largest local agency WIC program in the country, serving more than 300 000 participants monthly in 60 sites in Los Angeles, Orange, and San Bernardino counties. This WIC program enrolls approximately 5000 infants each month with a racial/ethnic breakdown of approximately 84% Hispanic, 6% Black, 4% Asian, 4% White, and 2% other. In anticipation of the major changes in food packages for breastfeeding mothers and infants, PHFE began strategic and targeted staff training and participant education in April 2009, 6 months before the changes to the food packages. Staff training included new protocols and skill development, including a new Breastfeeding Assessment Interview tool that would be required for use with all breastfeeding mothers. Participant education focused on educating pregnant women of the impending food package changes, helping them understand the incentives for breastfeeding, and encouraging them to make their breastfeeding decisions early. This WIC program had a small breastfeeding peer counselor program at 7 of the 60 sites at the time of the policy change.

Variables and Analysis

All WIC administrative data in California are entered by WIC staff into the Integrated Statewide Information System (ISIS), which captures participant demographic data as well as food package information for each participant. Through a partnership with the California WIC Program, PHFE has received monthly downloads of its WIC participant data since December 2007.

Ideally, the outcome variable (breastfeeding behavior) would be captured through maternal interviews of breastfeeding behavior. Because data regarding the dose and frequency of breastfeeding are not collected this way, nor are they available in ISIS for all infants on the program, we used ISIS food package data as a proxy for breastfeeding rates. A recent study validated the use of infant food packages as a valid proxy for breastfeeding behavior.¹⁴ The current study included the rate of issuance by month of 3 WIC food packages for postpartum women and infants: (1) the “fully breastfeeding” package (receiving no formula from WIC), (2) the “combination breastfeeding” package (receiving some formula from WIC), and (3) the “formula only” package (receiving the maximum allowable amount of formula from WIC).

We examined issuance of these WIC food packages to the approximately 5000 infants enrolled in PHFE centers each month for the 3-year period from December 2007 to November 2010. We used rates of issuance of each package type at infant enrollment into WIC as a proxy for changes in breastfeeding initiation. We examined rates of issuance of the fully breastfeeding package to 2-month-old and 6-month-old infants as an indicator of changes in breastfeeding duration. Thus, for all analyses, we pooled observations of the approximately 5000 infants born each month, examining rates of issuance of the fully breastfeeding package to newborns, 2-month-old infants, and 6-month-old infants each month. These observations over the 3 years from 2007 to 2010 are not repeated observations of the same 5000 children, but each month we captured all enrolled newborns, 2-month-old infants, and 6-months-old infants issued a food package.

In addition to the outcome variable, we also captured demographic data over the study period. Because of the economic recession that took place during the course of the study, paired with findings that poverty status is often linked to lower rates of breastfeeding,^15–18 it was particularly important to document and control for demographic shifts during the study period. We extracted demographic variables from ISIS, which included number of infants enrolled each month, ethnic distribution of the population, family size, and poverty level—assessed as percentage of the population living below 100% of the federal poverty level.^19–21

Because of the potential impact of the staff training and breastfeeding education efforts that took place in the 6 months leading up to the policy change, we examined rates of issuance of the 3 food packages at 3 time periods: before any staff training or participant education on the new food packages (time 1: December 2007 to March 2009), during the period of training and education but before the change to the new food packages (time 2: April 2009 to September 2009), and after the change to the new food packages (time 3: October 2009 to November 2010).

We conducted data analysis with SAS version 9.2 (SAS Institute Inc, Cary, NC). We used the χ² test to compare categorical demographic variables over the 3 time periods. We used the analysis of variance (ANOVA) test to compare continuous variables and rates of fully breastfeeding, combination, and formula-only food packages at the 3 time periods. The first analysis examined rates of issuance of all 3 packages for infants when they were first enrolled into WIC (usually within the month following their birth). The second analysis examined rates of issuance of the fully breastfeeding package for infants aged 2 months and infants aged 6 months. To determine differences between these time periods for the rates of issuance we performed multiple comparison tests using Tukey’s Studentized Range honestly significant difference (HSD) test, set at P < .05. We used 1-way analysis of covariance (ANCOVA) to control for the 1 potential confounding variable found to change significantly over the study period: percentage of the population living below 100% of the federal poverty level.

RESULTS

Examination of demographic variables over the study period showed few changes over time. Enrollment levels continued to reflect the local birthrate, with about 69% of all births in the region served by WIC. Ethnicity and family size of the study population stayed constant through the study period. There was a significant increase over time in the number of families living below 100% of the federal poverty level (67.2% at time 1; 71.4% at time 2; 73.2% at time 3; P < .001).

The ANOVA test comparing the rates of issuance of each package type before the staff training and education started (time 1), during the period of staff training and participant education but before breastfeeding policy changes (time 2), and after the policy changes to the new food packages (time 3) showed significant differences for all 3 packages: fully breastfeeding (F = 230.5; df = 2,33; P < .001), combination feeding (F = 598.12; df = 2,33; P < .001), and formula only (F = 29.41; df = 2,33; P < .001). Rates of issuance of the packages at the 3 time points are shown in Table 1 and Figure 1. Rates of issuance of the fully breastfeeding package increased significantly with training and education alone and then again following the policy change. Conversely, rates of issuance of the combination package decreased significantly with education alone and then again after the policy change. Issuance of the formula-only package decreased significantly during the education period, and then went up slightly following the policy change but remained at significantly lower levels than before the education efforts began. These relationships remained significant after we controlled for percentage of the population living below 100% of the federal poverty level in the adjusted ANCOVA model (P < .001).

TABLE 1—

Rates of Special Supplemental Nutrition Program for Women, Infants, and Children Food Package Issuance Before Staff Training and Education (Time 1), During Training and Education but Before the Policy Change (Time 2), and After the Policy Change (Time 3)

Package Type	Time 1, %	Time 2, %	Time 3, %
Fully breastfeeding at enrollmenta	23.8	31.9	44.2
Combination breastfeeding at enrollmenta	46.1	42.6	28.6
Formula only at enrollmentb	30.1	25.4	27.2
Fully breastfeeding at 2 moa	12.7	15.8	19.9
Fully breastfeeding at 6 moc	8.5	9.3	13.9

^aAll changes (time 1 to time 2; time 2 to time 3; time 1 to time 3) significant at P < .001, according to Tukey’s Studentized Range (honestly significant difference) test.

^bChanges from time 1 to time 2 and time 1 to time 3 significant at P < .001; time 2 to time 3 significant at P < .05, according to Tukey’s Studentized Range (honestly significant difference) test.

^cChanges from time 1 to time 3 and time 2 to time 3 significant at P < .001, according to Tukey’s Studentized Range (honestly significant difference) test.

FIGURE 1—

Issuance rates of the fully breastfeeding, combination breastfeeding and formula feeding, and formula feeding–only packages at infant enrollment into the Special Supplemental Nutrition Program for Women, Infants, and Children in Southern California.

Note. Dashed vertical line indicates start of breastfeeding training and education (April 2009); solid vertical line indicates food package change (October 2009).

Figure 2 shows the rates of issuance of the fully breastfeeding package to infants at age 2 months and age 6 months. The ANOVA and Tukey HSD results (Table 1) illustrated significant increases in the rates of issuance of the fully breastfeeding package to 2-month-old infants both following the education alone and again following the policy change (F = 139.58; df = 2,33; P < .001). This relationship remained significant in the adjusted ANCOVA model (P < .001).

FIGURE 2—

Issuance rates of Special Supplemental Nutrition Program for Women, Infants, and Children fully breastfeeding packages to infants aged 2 and 6 months in Southern California.

Note. Dashed vertical line indicates start of breastfeeding training and education (April 2009); solid vertical line indicates food package change (October 2009).

Rates of issuance of the fully breastfeeding package to 6-month-old infants were not expected to change until 6 months following the food package change. In fact, they increased immediately after the policy change, strongly suggesting that staff training and participant education efforts before the policy change were instrumental in effecting these changes (F = 187.56; df = 2,33; P < .001). This relationship remained significant in the adjusted ANCOVA model (P < .001).

DISCUSSION

Findings from this study suggest that the recent WIC policy changes were associated with positive changes in the issuance of infant WIC food packages in California. Between the period before staff training and participant education began and after the change to the food packages (time 1 to time 3), the rate of issuance of the fully breastfeeding package increased by nearly 86%. Equally as important, rates of issuance of the formula-only package decreased by about 10%, suggesting that the policy changes did not have an unintended effect of increasing formula issuance. These changes happened despite evidence of increasing levels of poverty during the study period. Although we cannot disentangle the unique effects of the policy and food package changes from the instrumental supportive role played by WIC staff, our findings document that the breastfeeding policy changes and surrounding staff training and participant education had the intended effects of increasing issuance of the fully breastfeeding package and reducing issuance of packages that include formula.

These findings are critically important for at least 2 reasons. First, WIC participation has long been associated with reduced rates of breastfeeding.^6–12 This study suggests that when breastfeeding support and education in the WIC environment are supported by policy change and an enhanced food package, mothers take less formula from the WIC program and presumably breastfeed more. The incentivized breastfeeding package may have brought about a shift in the perception of the value of the fully breastfeeding package as equal to or greater than the value of infant formula. Future research should continue to examine the breastfeeding rates of WIC and non-WIC participants to establish whether these policy changes bring breastfeeding rates of WIC participants into alignment with those of nonparticipants.

Second, front-line WIC staff are instrumental in effecting changes in breastfeeding. The policy changes brought structural changes to the California WIC program that, although difficult to measure, clearly had an impact on staff. Staff trainings for breastfeeding support, though always an element of core WIC services, took on a vigorously renewed emphasis because of the new policy that supported and incentivized breastfeeding. This study showed that front-line staff with focused breastfeeding training can make a significant impact on thousands of women and infants. New tools enabled staff to communicate more effectively with participants, and the policy change helped buttress staff confidence in supporting breastfeeding. The expectation when this analysis began was that changes in issuance rates of the fully breastfeeding package would only come when the food package change took effect in October 2009 after the economic incentive was in place. In fact, the full economic impact of the new fully breastfeeding packages is only fully realized when infants turn 6 months, a very distal incentive for most low-income mothers. Thus, the observation that change began immediately after staff training and participant education, and months before the food packages changed, suggests that the role of the WIC staff was instrumental in the rapid success of the breastfeeding policy changes in California.

Much attention has been appropriately focused on the role of the breastfeeding peer counselor, and indeed the success of peer counselors in the WIC program in improving breastfeeding rates are important and extremely well documented.^22–30 This study did not examine the unique impact of peer counselors on participant breastfeeding rates because of the fact that peer counselors were available at a very small number of WIC sites through the course of this study. Research efforts that continue to highlight the beneficial role of peer counselors in the WIC setting are a similarly important policy focus as it is clear that one-on-one peer support is among the most successful means for increasing breastfeeding rates.

Limitations

Study limitations must be addressed. Although a randomized controlled trial is the ideal study design, federal guidelines required that the WIC food packages be changed all at one time within a state. Therefore, we chose a pre–post design to study the impact of the policy changes. Consideration of defining a control group of WIC sites and withholding staff training and participant education before the change was rejected as the food package change was not designed to be a study of the optimal way to improve breastfeeding at WIC, but was a public health strategy to enable all WIC mothers to breastfeed successfully. Insomuch as this was a study of the impact of a WIC policy change on WIC clients, we feel the pre–post design adequately captured the impact of the policy change.

This study took place in California, a region known for having higher breastfeeding rates than other regions in the United States³¹ and higher numbers of Mexican American mothers who are more likely to breastfeed.³² Although this study included more than 180 000 births across 3 years, findings may be limited to the largely Latina WIC population served by PHFE and may not be generalizable to all WIC participants across the nation. Research efforts to compare the impact of the policy change across regions are important, and findings from this study suggest that inclusion of measures that examine staff training and participant education strategies in the months leading up to the food package changes is important.

Finally, we used rates of issuance of each food package type at infant enrollment as a proxy for breastfeeding initiation, and examined rates of issuance of the fully breastfeeding package at 2 months and 6 months of age as indicators of breastfeeding duration. Our recent validation study of food package type as an indicator of feeding practices at home¹⁴ provided evidence that WIC mothers receiving the fully breastfeeding package are doing significant amounts of breastfeeding. Similarly, mothers receiving the full formula package are almost exclusively formula feeding. Thus, there is strong evidence that food package issuance data reliably capture feeding practices.

Conclusions

This study demonstrates that policy changes, paired with staff training and participant education, were associated with rates of issuance of WIC food packages. Nationally, the WIC program reaches more than 50% of all infants and has been criticized for undermining breastfeeding through the provision of infant formula. This study suggests that the recent changes to the WIC food packages are associated with reductions in the issuance of infant formula to WIC participants in California and increases in rates of breastfeeding.