Challenges to Implementation of the 2009 Food Package Changes

Implementation of the 2009 food packages involved challenges at the federal, state, and local agency levels, as well as for vendors and manufacturers. A summary of the key challenges and actions taken to resolve these challenges is presented in Table 2-1 and described in detail in the phase I report (NASEM, 2016). For the most part, these challenges have since been overcome, but they remain relevant to consider in future rounds of food package revisions.

TABLE 2-1

Challenges to Implementation of the 2009 Food Package Revisions.

At the state agency level, notable challenges included production and distribution of the 1-pound loaf of bread, dealing with the “dangling quart” of milk, changes to container sizes available in the marketplace (e.g., juice and peanut butter), and implementation of the CVV (a description of the CVV is provided in Chapter 1). But along with these challenges, as noted in Chapter 1, a key benefit of the 2009 food package changes, finalized in 2014, was the ability for states to tailor packages (USDA/FNS, 2007a). Although the additional options led to some inconsistencies in specific foods available from state to state, it enabled state administrators to make decisions that maximize the suitability of the foods for their own population and that also contain costs. For example, the Final Rule allows children ages 12 to 24 months to receive fat-reduced milks if overweight or obesity is a concern (USDA/FNS, 2014a). As of 2015, 72 percent of WIC state agencies had adopted this option, covering 60 percent of WIC participants (USDA/FNS, 2015A).

At the vendor level, despite challenges to ensuring WIC foods were available, most evidence suggests that the food package revisions were beneficial for vendors, increasing both sales and profitability for the items offered in the revised food package (Andreyeva et al., 2011) and increasing sales of newly eligible food items to non-WIC customers (Gittelsohn et al., 2012). However, the revised food packages were designed to be cost-neutral to WIC. Thus, while sales apparently increased from WIC foods for some items (reduced-fat milk, whole grains, fruits, and vegetables), sales likely decreased for others (whole milk, juice) (Andreyeva et al., 2013a; Andreyeva and Luedicke, 2013b, 2014). Although some vendors reported difficulty finding and maintaining suppliers for some foods, available evidence finds that prices for these items did not increase, suggesting that vendors adjusted their supply quantities without incurring increased costs (Zenk et al., 2014).

At the local level, results from multiple studies have documented the effect of WIC nutrition education on participant knowledge, attitudes, and behavior change (USDA/ERS, 2007; Kavanagh et al., 2008; Ritchie et al., 2010; USDA/FNS, 2010; Sullivan et al., 2011; Whaley et al., 2012a,b; Hildebrand et al., 2014; Isbell et al., 2014; USDA/FNS, 2016c). A few of these studies have examined the effects of nutrition education on participant adaptation to the 2009 food package changes in particular. In California, Ritchie et al. (2010) demonstrated that nutrition education alone led to increased consumption of low-fat milk and whole grains even before the 2009 changes to the WIC food packages. Following the change, consumption of these foods increased further (Whaley et al., 2012b). A study of the effect of the 2009 breastfeeding food package change on rates of breastfeeding demonstrated significant increases in exclusive breastfeeding in the 6 months prior to the policy change when staff training and participant education focused on the upcoming policy changes (Whaley et al., 2012a). Similar changes were not evident in other states where staff training and participant education specific to the breastfeeding food package changes were not a focus prior to the food package change. Together, these studies suggest that achieving the intended health impact of the WIC food packages and 2009 revisions is facilitated by staff training and participant education.

Finally, the electronic benefit transfer (EBT) system is designed to enhance efficiency across the program by limiting purchases to only those foods authorized by the program. However, the linked databases that code “WIC-approved” foods must be updated continually in response to changes in the marketplace, which poses an administrative burden. USDA-FNS is in the process of developing a nationally representative Universal Product Code (UPC) database in collaboration with states, which should help to alleviate some of this burden. Because WIC benefits are grouped by EBT systems at the household rather than individual level, an additional benefit of the EBT system is the flexibility it allows in food acquisition when more than one family member is a WIC participant. Another vendor-level benefit of the EBT system was demonstrated by Phillips et al. (2014), who reported that EBT implementation both improved the ability of vendors to track inventory and stabilized inventory because participants were able to make purchases throughout the month instead of during a single visit. In addition, vendor reimbursement occurred more quickly.

WIC Participation Rates

Although the national WIC caseload increased between 2006 and 2010 (see Figure 2-1), reaching a peak of approximately 9 million in 2010, participation declined thereafter to approximately 8 million by 2014 and continued to decline into 2015 (USDA/ERS, 2015a). The year 2014 marked the fourth consecutive year—and only the fourth year in the program's history—that participation for women, infants, and children all fell (see Figure 2-2). In parallel, between fiscal years (FYs) 2013 and 2014, overall expenditures across USDA nutrition assistance programs decreased 5 percent and participation in the Supplemental Nutrition Assistance Program (SNAP) and the National School Lunch Program decreased by 2 and 1 percent, respectively. However, during the same period, participation in the School Breakfast Program and the number of meals served in the Child and Adult Care Food Program both increased by 2 percent (USDA/ERS, 2015b).

FIGURE 2-1

Annual number of participants in the WIC program constructed from monthly averages of participants, FY2004–2015. NOTES: FY2014 is the latest complete data. Data for FY2015 may be incomplete. SOURCE: USDA/FNS, 2016b.

FIGURE 2-2

Average annual WIC participation by participant category, 2004–2014. NOTE: No participation data were available for 2005, 2007, or 2009. SOURCES: USDA/FNS, 2007b, 2010, 2015b.

The reasons for the overall decline in WIC participation since 2010 are unclear, but the decline may be at least partially attributable to decreasing U.S. birthrates, as well as to the nation's improving economic health. From 2007 to 2009, the United States experienced an economic downturn that may have caused more individuals to have incomes low enough to ensure eligibility for WIC. During the recession, Medicaid, SNAP, and Temporary Assistance for Needy Families (TANF), experienced increases in participation and received increased funding through the American Recovery and Reinvestment Act of 2009 (KFF, 2009, 2015; CBO, 2012; EOPUS, 2014). Participants in these programs who meet WIC age and physiological state criteria are automatically eligible for WIC. Not only has the country been experiencing a still-incomplete recovery from the recession that began in 2008, but other recent changes in Medicaid, SNAP, and TANF may have affected WIC eligibility and participation as well. Additionally, between October 1 and 16, 2013, the federal government experienced a shutdown, which resulted in a gap in funding for the WIC program at the beginning of the fiscal year. Although most states maintained WIC services, some offered modified services; program recovery was slow in some states, lasting up to a year.

In addition to these demographic and economic changes that may have influenced WIC participation between 2010 and 2015, the committee considered whether the food package changes initially implemented in 2009 may have contributed as well. To do this, the committee used state-level data on participation and the number eligible for WIC from 2006–2012 (USDA/FNS, 2011a, 2013, 2014b, 2015b; Bitler and Hoynes, 2013). The analysis considered the effects of national trends, time invariant state factors, the date of implementation of the new food package, unemployment rate, births per capita, and participation in TANF/SNAP/Unemployment Insurance (UI). Details of the estimation method are discussed in Appendix F. The results suggest no significant difference between participation before versus after implementation of the new food packages. The estimated effect was not statistically significant, and it was small in magnitude.

Demographics of the WIC-Participating Population

WIC Participation by Age and Physiological State

The proportion of WIC participants varies among women by physiologic state and among children by age, but this proportion has remained relatively stable since 2006 (see Table 2-2). Data from 2012 indicate that 63 percent of individuals eligible for WIC nationwide participated in the program (USDA/FNS, 2015b). This includes 85 percent of eligible infants, 53 percent of eligible children, 71 percent of eligible pregnant women, and 77 percent of eligible postpartum women (USDA/FNS, 2015b). The majority of women who participate in WIC are pregnant. Of the remaining women, approximately half are breastfeeding and half are postpartum but not breastfeeding. More than half of infants are fully formula-fed. Of the remaining infants, approximately half are fully and half are partially breastfed. The majority of WIC participants are children (53 percent in 2014) (USDA/FNS, 2015c). Many eligible children discontinue participation after 1 year of age (USDA/FNS, 2011a), although the specific reasons for this behavior are unknown.

TABLE 2-2

Participation of Women and Children in WIC by Physiological State or Feeding Mode, FY2005–2006 and 2015–2016.

Racial and Ethnic Composition of the WIC-Participating Population

WIC serves a population with a diverse racial and ethnic composition, and this composition, based on major race and ethnicity categories, has not changed more than 3 percent between 2006 and 2014 (USDA/FNS, 2007a, 2015c) (see Figures 2-3a and 2-3b). Of note in the figures below, the percentage of WIC participants who reported being white increased between 2006 and 2014 by a few percentage points (see Figure 2-3a), yet the percentage of Hispanic/Latino participants remained the same despite an increase in proportions of Hispanic people in the U.S. population (see Figure 2-3b) (USCB, 2011). There was also a decrease in the percentage of WIC participants who reported being American Indian and Alaskan Native between 2006 and 2014. The U.S. population has increased 9 percent since 2005, with the greatest contributions to population growth from immigration, temporary and permanent residency, and other population shifts (DHS, 2014). According to the U.S. Census Bureau, the majority of growth in the U.S. population from 2000 to 2010 resulted from an increase in Hispanic and Asian populations (USCB, 2011). The 2010 American Community Survey found that 92 percent of the U.S. Hispanic population comprises 10 subgroups, with the top three being Mexican, Puerto Rican, and Cuban (Motel and Patten, 2012). Although there are no available national data on racial and ethnic subgroups in the WIC-participating population, these subgroup proportions likely reflect a national trend.

FIGURE 2-3a

Distribution of race of WIC participants, 2006 and 2014. SOURCES: USDA/FNS, 2007b, 2015c.

FIGURE 2-3b

Distribution of ethnicity of WIC participants, 2006 and 2014. SOURCES: USDA/FNS, 2007b, 2015c.

Food Expenditures of WIC Participants

The committee was tasked with planning and implementing an analysis of food expenditures for the WIC-participating population using national data. Data from the Food Acquisition and Purchase Survey (FoodAPS) (USDA/ERS, 2015c) were used in phase I to compare household expenditures on total food, food at home, and food away from home among three groups: (1) WIC-participating households, (2) WIC-eligible, nonparticipating households (i.e., households with incomes less than or equal to 185 percent of the poverty-to-income ratio and with a pregnant woman or child less than 5 years old), and (3) WIC-ineligible, nonparticipating households (i.e., households with incomes greater than 185 percent of the poverty-to-income ratio) with a pregnant woman or children less than 5 years of age (NASEM, 2016).¹

These data also allow for examination of detailed expenditures by WIC-participating households for specific WIC food categories. As described in further detail in Chapter 10 of the phase I report (NASEM, 2016), FoodAPS is a nationally representative, stratified sampling of 4,826 households from April 2012 through January 2013. Data on food expenditures were collected over a 7-day period, with all household members asked to track and report their food purchases or acquisitions during the survey week. Information on each “event” (i.e., purchase or acquisition) included whether the food was purchased or acquired at home or away from home, total expenditure, and source of payment (i.e., specifically whether the event included use of payment with a WIC voucher and, if so, the amount of payment allotted to WIC items). Although in most cases it was not possible to know exactly which foods were obtained with the WIC food voucher (because more than one type of tender may have been reported), by examining individual food codes, purchased items could be categorized as WIC food items. Data were weighted using the household weights provided in the FoodAPS documentation.² Mean food expenditures were compared between WIC-participating households and both types of non-WIC-participating households using t-tests and the Pearson chi-square statistic (Rao and Scott, 1984).³

The analysis of the FoodAPS data provides evidence on the total food expenditures and expenditures on food groups for WIC-participating households to assess the relative contribution of the WIC food packages to their food expenditures. Overall, as shown in Table 2-3 (i.e., updated from Table 10-2 in the phase I report [NASEM, 2016]), the average weekly expenditures for food at home (FAH) for WIC-participating households was $124.20/week, similar to weekly expenditures of WIC-eligible, nonparticipating households ($113.20/week). WIC-participating households had marginally greater (p = 0.10) expenditures on food away from home (FAFH) expenditures ($60.59/week) than those of WIC-eligible, nonparticipating households ($47.06/week). The average value of WIC expenditures for these WIC-participating households was $10.75/week. Among all WIC-participating households (i.e., both those that made WIC purchases during the week and those that did not), this represents 8.8 percent of total FAH expenditures. Among the nearly one-third of WIC-participating households (32.3 percent) that made a purchase with a WIC benefit during the week, WIC expenditures represented 24.3 percent of reported FAH expenditures for that week.

TABLE 2-3

Summary Statistics for Weekly Food Expenditures for WIC-Participating and Other Households.

The data also allowed comparison of expenditures for specific WIC food items among the three groups of households. As shown in Table 2-4, the overall average level of expenditures on identified WIC food items ($21.55/week) was significantly larger for WIC-participating households than for eligible non-WIC-participating households ($14.82/week, p = 0.01), and was similar to the average expenditures of the non-WIC-participating households with income greater than 185 percent of the poverty income ratio ($21.31/week). The share of WIC food items in FAH expenditures was higher for WIC-participating households than for the other households (18 percent overall for WIC-participating households, and 13 percent and 13.5 percent of total FAH expenditures for the low-income and higher-income non-WIC-participating households, respectively; p = 0.01). Among households with WIC expenditures during the week (WIC paid >0), expenditures on WIC food items averaged 29 percent of total FAH expenditures. Because WIC-participating households also used other resources to acquire WIC food items, the average value of WIC food items acquired was greater than the value of reported WIC expenditures in the week.⁴

TABLE 2-4

Weekly Household Expenditures: Total Expenditures, Expenditures When WIC Is Used (WIC Paid >0) and When WIC Is Not Used (WIC Paid = 0).

Also shown in Table 2-4, for most of the identified food items, WIC-participating households spent more than other households. For example, WIC-participating households spent a weekly average of $1.42 on eggs, compared to $0.72 for WIC-eligible, nonparticipating households and $1.00 for non-WIC-participating higher-income households; and WIC-participating households spent $3.98 per week on milk, compared to $3.05 per week for eligible, non-WIC-participating households and $3.88 per week for non-WIC-participating higher- income households. Weekly expenditures on breakfast cereals were close in value across the three groups. WIC-participating households spent less on vegetables and fruits ($1.79/week) than did the other higher-income households ($2.60/week).⁵

The FoodAPS data show how expenditures on WIC food items were made in the “shopping trip” (event) among the household groups. At the level of purchase and acquisition events, as shown in Table 2-5, the average value of WIC expenditures as a purchase event was $21.74. Expenditures for most WIC food items were larger when WIC vouchers were used at the shopping event (WIC Paid >0) than otherwise (WIC Paid = 0). Compared to when WIC vouchers were used during an event (WIC paid >0) ($15.68/event),⁶ average spending on identified WIC food items was less when WIC vouchers were not used (WIC paid = 0) ($3.58/event). The same pattern holds for most individual food items. For example, when WIC vouchers were used (WIC paid >0), WIC-participating households spent an average $3.43 on milk during a single shopping trip, compared to, on average, $0.59 when vouchers were not used (WIC paid = 0).⁷ Among the low income households, WIC-participating households had higher expenditures for infant formula, eggs, and juice in shopping trips (shopping events) compared with other low income households.

TABLE 2-5

Food Expenditures Events During Week: Total for Event, Amount WIC Paid (WIC Paid >0) and Other Paid (WIC Paid = 0).

The analysis of expenditure patterns among the three groups of households shows that the expenditures of WIC-participating households differ from other households in several ways. In comparison to other eligible households with income less than 185 percent of the poverty-to-income threshold, WIC-participating households spent more on total food expenditures and more on FAFH. Comparable households with higher-income spent more than WIC-participating households both in total and for FAH and FAFH. WIC-participating households spent more on identified WIC food items in a week than did comparable households with income less than 185 percent of the poverty-to-income ratio. The amount spent on identified WIC food items was comparable to the spending level of the higher-income, non-WIC-participating households.

Breastfeeding Trends in the WIC-Participating Population Compared to the U.S. Population

Healthy People 2020's goals for breastfeeding are presented in Table 2-6 (HHS, 2015). In 2011, the U.S. Surgeon General called for action to support these goals, recommending that families, communities, health care centers, and employment sites provide the support necessary for women to initiate and continue breastfeeding (HHS, 2011). To assess progress toward reaching these goals, the U.S. Centers for Disease Control and Prevention (CDC) estimates breastfeeding prevalence across the United States every 2 years, using data from the U.S. National Immunization Survey. The most recent estimates reflect statistics for children born in 2013 and represent survey data from 2014 and 2015 (see Table 2-6). As shown in the table, although the national goal for initiation of breastfeeding has nearly been achieved, goals for duration of breastfeeding have been more challenging to meet (CDC, 2016a). This may result, in part, from differences in breastfeeding behavior related to racial and ethnic groups, maternal education and age, and WIC participation (CDC, 2010).

TABLE 2-6

Healthy People 2020 Breastfeeding Objectives Compared to 2014–2015 Proportion (%) of Children Who Were Breastfed at Various Ages.

In addition to varying by income, the proportion of women that breastfeed varies among racial and ethnic groups. In 2009 and 2013, for example, the prevalence of breastfeeding at 6 months was consistently lowest for non-Hispanic African Americans (33 and 39 percent of infants, respectively) and non-Hispanic American Indian/Alaska Native (39 and 41 percent, respectively) and highest for non-Hispanic Asian (86 and 84 percent, respectively) (see Table 2-7). Studies of breastfeeding prevalence in the WIC-participating population show similar variation, with fewer African-American women initiating and sustaining breastfeeding compared to other racial and ethnic groups (Hurley et al., 2008; Marshall et al., 2013; McKinney et al., 2016). The underlying reasons for racial and ethnic differences in breastfeeding prevalence are not well understood at this time, but differences among racial/ethnic groups in ever breastfeeding are now relatively small (from 79 to 99 percent of the proportion of non-Hispanic whites who ever breastfed) (see Figure 2-4). There are much larger differences among racial/ethnic groups in continued breastfeeding at 6 months (from 36 to 111 percent of the proportion of non-Hispanic whites who continued to breastfeed at 6 months) (see Figure 2-4).

TABLE 2-7

Prevalence of Ever Breastfeeding and Breastfeeding at 6-Months Postpartum by Race: 2009 and 2013.

FIGURE 2-4

Breastfeeding prevalence by racial/ethnic group expressed as a percentage of values for non-Hispanic whites: United States, 2013. SOURCE: HHS, 2015.

In the WIC-participating population, the most recent WIC Participant and Program Characteristics (PC2014) report indicated that, in 2014, 70 percent of infants (served by agencies that provided data) were ever breastfed (USDA/FNS, 2015c). Although other measures of breastfeeding prevalence were available to the committee, the longest time-series for which all-infant and WIC-infant prevalence of breastfeeding could be compared was at 6 months of age (see Figure 2-5) (1971 through 2013 for U.S. infants; 1978–2013 for WIC infants) (Ryan et al., 2002; CDC, 2015). In 2013, all-infant 6-month breastfeeding prevalence was 49 percent, while the WIC-infant estimate was 39 percent. Although a lower proportion of WIC infants were breastfed than those in the general population, the prevalence of breastfeeding in both groups has been increasing since the late 1970s.

FIGURE 2-5

Six-month breastfeeding prevalence: 1971–2013. NOTES: Data exclusively for WIC eligible, nonparticipants were not available. Therefore, the comparison of all infants to WIC infants is an underestimate of the difference of interest, namely WIC (more...)

Six-month breastfeeding prevalence in the U.S. population has consistently tracked with income. As shown in Table 2-8, from 2008 to 2013, 6-month breastfeeding prevalence ranged from as low as 33 to 38 percent for women under 100 percent of the poverty level to as high as 60 to 70 percent in women at 600 percent or more of the poverty level (CDC, 2016b). Although breastfeeding increased during this time period for women in all income levels, increases between 2008 and 2013 were highest among women with income above 200 percent of the poverty level. Other available data indicate that between 2004 and 2008, breastfeeding prevalence was lower for WIC-participating women compared to WIC-eligible, nonparticipating women, with substantial differences across racial/ethnic groups (CDC, 2010).

TABLE 2-8

6-Month Breastfeeding Prevalence (%) in the United States by Income from 2008 to 2013.

Promotion of Breastfeeding in the WIC Program

A number of breastfeeding promotion and support strategies have been in place as part of Healthy People 2020 that may have helped to increase the prevalence of breastfeeding in both WIC-participating and non-WIC-participating populations. These include strategies to increase the proportion of employers that have worksite lactation support programs, reduce the proportion of breastfed newborns that receive formula supplementation within the first 2 days of life, and increase the proportion of live births that occur in facilities that provide recommended care for lactating mothers and their infants (HHS, 2015). The Baby-Friendly Hospital Initiative has been assisting hospitals with meeting these goals (IOM, 2011b; WHO, 2012).

Additionally, since the food package revisions implemented in 2009, there has been a concerted effort within WIC to increase the proportion of women who breastfeed. The committee's phase I review of breastfeeding (see Chapter 7 in NASEM, 2016) found that the 2009 changes to the food package were associated with small improvements in breastfeeding initiation (USDA/FNS, 2011a; Whaley et al., 2012a; Chiasson et al., 2013; Langellier et al., 2014; Joyce and Reeder, 2015). However, it was not possible to determine whether these improvements resulted from the food package changes themselves, the enhanced breastfeeding promotion and support activities begun at about the same time, or both.

As of 1996, states were no longer required to conduct an annual evaluation of breastfeeding promotion and support activities (P.L. 104-193, Personal Responsibility and Work Opportunity Reconciliation Act of 1996). As a result, other than in Whaley et al. (2012a), little information exists as to which specific breastfeeding promotion and support activities within WIC have been the most successful. Implementation of the USDA's Loving Support^© program is one way that WIC breastfeeding activities are supported, but implementation varies depending upon available state-level resources.

USDA-FNS Guidance Related to Provision of Infant Formula

Effects of the 2009 Food Package Changes on the Choice to Breastfeed

The food packages implemented in 2009 were designed to encourage exclusive breastfeeding by allowing very limited issuance of infant formula to breastfeeding women in the first 30 days postpartum. States may allow the issuance of one can of formula in the infant's first month of life, but they are not permitted to create food packages that routinely issue formula to breastfeeding women during this period (USDA/FNS, 2014a). Although the intention of this policy was to support breastfeeding in the immediate postpartum period, evidence indicates the policy had an unintended consequence because women had to choose between two options: (1) receive no formula from WIC, or (2) receive the maximum formula allowance provided to nonbreastfeeding mothers. In some states, a third option to receive 1 can of formula is available in the first month. The results of the largest study of the impact of the food package changes on breastfeeding rates showed that there were no changes in breastfeeding initiation that could be attributed to the revised food packages (USDA/FNS, 2011b). Although issuance of the full breastfeeding package (the first option noted above) increased in the first month, there was also an increase in issuance of the full-formula package (the second option noted above) (USDA/FNS, 2011b). The results of this study also indicated that although issuance of the full breastfeeding package was higher in the first month after implementation of the 2009 food packages, the transition of infants out of this package and into a partially breastfeeding or fully formula-feeding package occurred at a significantly faster rate compared to before implementation of the 2009 food packages. Although there is some evidence that the allowance of one can of formula in the first month helps support breastfeeding, data are not available on either the impact of a one-can policy or its breadth of implementation.

Barriers and Incentives to Breastfeeding in the WIC-Participating Population

The committee's phase I report (NASEM, 2016, Chapter 7) covered barriers, motivators and incentives for breastfeeding as well as the effect of the WIC breastfeeding food package on breastfeeding promotion, initiation, and duration among both WIC and low-income populations. Identified barriers to breastfeeding were numerous and included social norms, cultural factors, social structures, employment,¹⁰ and biomedical factors (see Table 7-4 in NASEM, 2016), although the influence of each specific barrier on breastfeeding in the WIC-participating population could not be determined.

There is some evidence that breastfeeding promotion and support provided through the WIC program improves breastfeeding initiation (Olson et al., 2010; Whaley et al, 2012a; Haider et al., 2014; Hildebrand et al., 2014) and duration (Bonuck et al., 2005; Olson et al., 2010; Pugh et al., 2010; Chapman et al., 2013; Haider et al., 2014; Reeder et al., 2014). Data are less convincing for the effects of promotion and support on exclusivity of breastfeeding (Anderson et al., 2005, 2007; Bonuck et al., 2005; Hayes et al., 2008; Meehan et al., 2008; Hopkinson and Konefal Gallagher, 2009; Petrova et al., 2009; Sandy et al., 2009; Bunik et al., 2010; Olson et al., 2010; Pugh et al., 2010; Kandiah, 2011; Chapman et al., 2013; Haider et al., 2014; Hildebrand et al., 2014; Howell et al., 2014; Reeder et al., 2014). WIC participants may perceive that the program delivers conflicting messages by supporting breastfeeding while also distributing infant formula at no cost to participants (Holmes et al., 2009).

WIC participation has been associated with a lower proportion of women who initiate breastfeeding and shorter durations of exclusive and any breastfeeding compared to women not participating in WIC (Li et al., 2005; Hendricks et al., 2006; Ryan and Zhou, 2006; Jacknowitz and Tiehen, 2007; Flower et al., 2008; Bunik et al., 2009; Ziol-Guest and Hernandez, 2010; Jensen, 2012; Mao et al., 2012; Marshall et al., 2013; Ma et al., 2014; Bullinger and Gurley-Calvez, 2016; Gregory et al., 2016). However, these findings are dependent upon methodology applied and treatment of selection bias.¹¹ Using variation in food prices as an instrumental variable (as a means of controlling for selection bias), it was reported that WIC participation decreases the proportion of women that exclusively breastfeed by nearly 50 percent (Bullinger and Gurley-Calvez, 2016). Yet, Gregory et al. (2016), using a propensity-matching method, reported that WIC participation was not associated with a lower likelihood of breastfeeding at 3 months and that WIC participants and nonparticipants differed with respect to baseline factors that predict breastfeeding. Both studies used the Infant Feeding Practices II dataset to develop their samples. Evidence on the effect of timing of entry into WIC on these outcomes is not conclusive (Joyce et al., 2008; Ziol-Guest and Hernandez, 2010; Tenfelde et al., 2011; Langellier et al., 2012; Ma and Magnus, 2012; Jacobson et al., 2015; Metallinos-Katsaras et al., 2015).

Dietary Patterns of Americans Have Changed

There have been several changes in consumption of food groups by the U.S. population since the 2006 report. Most notably, after decades of increases, mean energy intake decreased significantly between 2003–2004 and 2009–2010 for the U.S. population overall (Ford and Dietz, 2013). Additionally, between 2007–2008 and 2011–2012 (periods immediately before and after the 2009 food package changes), in women 20 years and older, consumption of whole grains increased 34 percent, consumption of seafood low in omega-3 fatty acids increased by 26 percent, and consumption of nuts and seeds increased by 28 percent (see Table 2-11). In contrast, consumption of soy products decreased by 30 percent. Between 2007–2008 and 2011–2012, in children ages 2–5 years, consumption of seafood high in omega-3 doubled, yogurt consumption increased by 83 percent, and consumption of whole grains increased by 46 percent (see Table 2-12). The committee's analysis (see Chapter 4) indicated that, despite these increases, intakes of whole grains and seafood are nonetheless below recommended amounts for both WIC-eligible women and children. Economic Research Service (ERS) data indicate that per capita wheat product consumption (whole grain and refined) has declined since 2000 (USDA/ERS, 2016a), a trend which could be related to consumer interests in lowering carbohydrate intake or in gluten-free products. Although comparison of 2007–2008 and 2011–2012 is provided for discussion, the direction and magnitude of change varies depending upon the survey years compared.

TABLE 2-11

Trends in Food Consumption from Selected Food Groups: Mean Intakes for U.S Women, 20 Years and Older, NHANES 2005–2012.

TABLE 2-12

Trends in Food Consumption from Selected Food Groups: Mean Intakes for U.S. Children, 2 to 5 Years of Age, NHANES 2005–2012.

Expansion of the Variety of Foods Available in the Marketplace Since 2006

According to an ERS report on trends in new food products in the United States (USDA/ERS, 2014a), from 2006 to 2010, the share of new products in the categories of fruit and vegetables, dairy products, pasta and rice, and infant food increased while the share of new snack product and beverage introductions decreased. The variety of available fresh fruit has been robust over the past decade, possibly spurred by a greater ethnic diversity of consumers (Bentley and Pérez, 2015). On the other hand, the availability of 100% fruit juice has fallen steadily since its peak in 1977. Availability per capita of 100% orange juice, which accounts for half of the availability all fruit juice in the United States, decreased 54 percent between 1977 and 2012 (Bentley and Pérez, 2015).

The introduction of new-product claims gives some indication of changes in the food supply that result from changes in consumer preferences. In 2010, “no gluten” ranked among the top 10 claims for the first time, with nearly triple the number of products making this claim compared to 2006 (USDA/ERS, 2014a). “Low or no fat” product claims also moved to the top 10 claims for the first time since 2006. From 2006 to 2010, launches of new products with “whole grain” claims doubled worldwide (Oldways Whole Grains Council, 2016).

The Food Supply and the WIC Program

Several recent changes in the U.S. food supply with implications for the revised food packages can be attributed to the influence of WIC itself. For example, since the Child Nutrition and WIC Reauthorization Act of 2004, which included provisions for infant formula manufacturers to bid for state WIC contracts, manufacturers have been bidding for formulas supplemented with the fatty acids docosahexaenoic acid (DHA) and arachidonic acid (ARA).¹² When bidding began, these products were relatively new additions to formula options in the marketplace (USDA/ERS, 2011a). At present, nearly all infant formulas sold in the United States are supplemented with DHA and ARA, an indication of WIC's significant influence on the infant formula market (USDA/ERS, 2011a). Another example of WIC's influence on the U.S. food supply is anecdotal evidence suggesting that increased availability of 1-pound loaves of whole wheat bread can be attributed to the addition of this item to the WIC food packages in 2009 (Oliveira and Frazao, 2015).

Changes in Dietary Guidance

The 2006 IOM review of WIC food packages drew on the 2005 Dietary Guidelines for Americans (DGA) (USDA/HHS, 2005). The DGA are updated every 5 years, and since the 2006 review, the 2010 and the 2015–2020 DGA have been issued. As described in the statement of task, the committee was charged with aligning WIC food packages for individuals ages 2 years and older with the most current DGA. Recommendations for infants and children less than 2 years of age draw on the recommendations of the American Academy of Pediatrics (AAP) and other authoritative groups. Changes in dietary guidance since 2005 relevant to the WIC food packages are summarized below.

In addition to these changes, the Dietary Guidelines Advisory Committee reported strong evidence that the seafood industry has rapidly expanded to meet demand and that fisheries are increasingly employing sustainable management strategies to avoid long-term collapse (USDA/HHS, 2015). This is important given WIC's broad reach and recommendations included herein related to fish in the food packages.

Changes to Recommended Food Group Intakes

This review presents an opportunity to improve the alignment of the food packages after two changes in the DGA and after introduction of the 2009 food package changes. Compared to the 2005 DGA (see Table 2-13), the 2010 DGA reorganized the vegetable food group into five subgroups. The recommended food intakes increased for “red-orange vegetables,” “starchy vegetables,” and “beans and peas.” The recommended quantities of “dark-green vegetables” and “other vegetables” decreased. There were no changes in recommended intakes of total fruit, grains, protein foods, or oils. Recommended intakes of dairy foods were slightly increased for two levels of caloric intake. Compared to the 2010 DGA, the 2015–2020 DGA included no changes to the recommended amounts from each of the major food groups or food subgroups, except for small changes to the subgroups of protein foods. In summary, the major change in the DGA since the 2005 edition relevant to this review is the reformulation of the vegetable subgroups. This reformulation has already resulted in the committee's recommendation that WIC participants should be allowed to acquire white potatoes with their CVV (IOM, 2015).

TABLE 2-13

USDA Food Intake Patterns for Kcal Levels of Interest: Comparison of 2005 and 2015–2020 DGA.

Shortfall Nutrients and Nutrients of Public Health Concern

The 2015–2020 DGA identified nine nutrients (vitamin A, vitamin D, vitamin E, vitamin C, folate, calcium, magnesium, fiber, and potassium) as “shortfall” nutrients, that is, nutrients that are underconsumed relative to Dietary Reference Intake recommendations (see Table 2-14 for a comparison across the 2005, 2010, and 2015–2020 DGA). Calcium, vitamin D, fiber, and potassium were further classified as nutrients of public health concern because their underconsumption has been linked to adverse health outcomes. For adolescent and premenopausal females, iron was also identified as a shortfall nutrient because of risk of iron-deficiency. Compared to the 2010 DGA, the 2015–2020 DGA no longer identified choline and vitamin K in adults, phosphorus in children, and vitamin B12 in adults over 50 as shortfall nutrients. In the 2010 DGA, folate was not identified as a shortfall nutrient, rather a nutrient of concern for women capable of becoming pregnant; the newer guidelines continue to recommend that women of reproductive age supplement a diet rich in vegetables, fruits, and grains with foods enriched with folic acid or that include folic acid supplements. The 2015–2020 DGA still consider iron a nutrient of public health concern.

TABLE 2-14

Shortfall Nutrients and Nutrients of Public Health Concern Noted in the Dietary Guidelines for Americans: 2005, 2010, and 2015–2020.

Cultural Variation in Satisfaction with the 2009 Food Package Changes

Multiple studies have documented moderate to high satisfaction with the 2009 changes in the WIC food packages (Gleason and Pooler, 2011; Whaley et al., 2012b; Ishdorj and Capps, 2013; Bertmann et al., 2014; Ritchie et al., 2014). However, evidence also indicates cultural variation in participants' satisfaction with certain types or amounts of food items. Black et al. (2009) conducted interviews and focus groups with WIC participants and caregivers throughout Maryland to assess perceptions of the proposed food package changes and examine differences in food preferences by race and ethnicity. Although food preferences appeared to be similar between non-Hispanic black and non-Hispanic white participants, Hispanic respondents were more likely than non-Hispanic respondents to prefer beans compared to peanut butter and to express dislike for frozen and canned vegetables. In a statewide survey of nearly 3,000 mostly (79 percent) Hispanic WIC participants and caregivers in California, Ritchie et al. (2014) reported that nearly all (91 percent) respondents were satisfied with the new food items introduced (vegetables and fruits, whole grains, and lower-fat milk). Additionally, a higher proportion of primarily Spanish speakers (compared to primarily English speakers) were satisfied with vouchers for whole grains, vouchers for lower-fat milk, and the amount of juice; and a higher proportion of primarily English speakers (compared to primarily Spanish speakers) were satisfied only with the amount of milk and not with the amounts of other foods. Actual consumption of foods was difficult to ascertain. In a quasi-experimental study in northern Illinois, Zenk et al. (2012) found that the availability of fresh fruit and vegetables at authorized WIC retailers increased significantly after the 2009 package changes for African-American culturally specific vegetables and fruits, but not for Latino culturally specific vegetables and fruits.¹⁴

Redemption of WIC Foods

In addition to published studies of the acceptability of the 2009 food packages, the committee also examined state data on foods actually purchased (i.e., from states that have implemented the EBT system). A summary of redemption data gathered from a sample of states at various time points between 2012 and 2016 is presented in Table 2-15 (see the table footnotes for time ranges). Available data indicate that redemption was variable for most food categories, but consistently high for infant formula and consistently low for jarred infant meats. For foods provided in more than one food package, it was not possible to determine if redemption was different among food packages. There were not enough data to determine whether states with high concentrations of particular racial and ethnic groups have higher redemption for specific items (e.g., tortillas or beans are common Hispanic staples). Nonetheless, the committee considered possible regional differences in food and substitution preferences in its deliberations.

TABLE 2-15

Redemption of WIC Foods, 2013–2016.

Barriers to WIC Participation and Redemption

Key barriers to WIC participation and redemption of WIC foods are summarized in Box 2-1, and details of the committee's review are available in Appendix G, Table G-1. Although the quantitative evidence identified by the committee was sufficient to support only an association, not a causal relationship, between these barriers and either participation or redemption, possibilities for maximizing both participation and redemption include: streamlining the registration process (Gilbert et al., 2014), enhancing customer service and reducing wait times for participants (Christie et al., 2006), informing participants of local vendors (Gleason et al., 2014), ensuring culturally-appropriate nutrition messaging (Phillips et al., 2014), enhancing the perceived value of packages (Gleason and Pooler, 2011), and examining the impact of minimum stocking requirements on food availability (Gleason et al., 2011). As the EBT system is implemented nationwide (by 2020), it may remove several of the identified barriers (Phillips et al., 2014).

BOX 2-1

Summary of Literature Findings on Barriers to WIC Participation and Redemption.

The WIC Participant Shopping Experience

Although studies are limited, qualitative work among WIC programs nationwide suggests that the participant shopping experience can be another key barrier to participation and redemption of WIC foods. Prior to the food package revisions, a survey administered to parents and caretakers of WIC participants in New York State found that issues with food procurement (store policies, food availability) and the WIC food packages (adequacy, satisfaction with the items) were barriers to participation (Woelfel et al., 2004).¹⁵ Several small studies were carried out after implementation of the 2009 food package changes to evaluate the perception and use of the CVV by WIC participants (Christie et al., 2006; Gleason and Pooler, 2011; Najjar, 2013; Bertmann et al., 2014; Gleason et al., 2014). Bertmann et al. (2014) reported that CVVs were inconsistently redeemed in Arizona. They identified the following barriers to redemption: participants' perception of annoyance or anger expressed by cashier or other shoppers; cashiers' lack of training; fluctuation in enforcement of WIC redemption rules from store to store and week to week; and feelings of embarrassment or judgment when using the CVV. The authors cautioned, however, that their findings might not be generalizable to other WIC-participating populations. In a Wisconsin study of WIC participant CVV redemption patterns, Gleason and Pooler (2011) reported positive responses overall to the package changes, but with differences in non-use and maximum use of the CVV among some WIC subpopulations. Some participants described a level of discomfort with having to do math in the store, which the researchers hypothesized may be enough to deter use of the benefits. Additionally, the authors identified several vendor-level challenges, including difficulties in maintaining fresh foods (particularly in smaller stores), anticipating client demand, and having the correct sizes available. Najjar (2013) reported that helpful vendors and both vendor and participant understanding about the use of the CVV can positively affect the WIC shopping experience.

The effect of allowing split tender for CVV purchases (using a different payment method for the amount over the CVV benefit) on redemption has yet to be comprehensively evaluated. Inasmuch as WIC provides vouchers for all other foods based on quantity, not value, WIC participants may pay less attention to food prices when redeeming their vouchers (Hoynes and Schanzenbach, 2015). The CVV, however, is a cash benefit, and purchasing power may vary regionally. In a study of 26 metropolitan market areas, Leibtag and Kumcu (USDA/ERS, 2011b) found that the 20 most commonly purchased vegetables and fruits cost 30 to 70 percent more in the highest-priced compared to the lowest-priced market areas.