Improving food security and diet quality during pregnancy is important, given the far-reaching effects of maternal nutrition during this critical window of the life course on the health of both women and their children. In the short term, poor maternal nutrition has been associated with adverse fetal development, resulting in low birthweight and preterm birth.(Gete et al., 2020; Lechtig et al., 1975; Saenger et al., 2007) Also, it can lead to maternal complications, such as gestational-diabetes-mellitus (GDM) and gestational-weight-gain (GWG) outside the recommended range.(Diemert et al., 2016; McIntyre et al., 2019; Silva-Zolezzi et al., 2017) These adverse maternal health conditions can contribute to the long-term development of chronic diseases.(Christian & Stewart, 2010; Wu et al., 2004) For example, pregnant women with GDM are more likely to develop type 2 diabetes,(Kim et al., 2002) and children with low birth weight have increased heart disease and premature mortality in later life.(Negrato & Gomes, 2013) Due to disparities in upstream socioeconomic factors, US women in historically marginalized groups (e.g., Black and Hispanic women, those with lower education) are more likely to experience poor maternal nutrition, increasing the risk of adverse maternal health and having low birthweight infants.(Headen et al., 2015) This intergenerational transmission of inequity may contribute to the persistent health disparities in the U.S.(Aizer & Currie, 2014) Consequently, interventions and policies that improve food security and nutrition during pregnancy are essential to prevent the intergenerational perpetuation of health disparities.

The Special Supplemental Nutrition Program for Women, Infant, and Children (WIC), initiated in the 1970s, is among the largest US safety net programs aimed at improving nutrition for low-income pregnant and postpartum women and children younger than 5.(Oliveira, 2009) The program provides nutrition supports (including food packages and breastfeeding education) to approximately seven million women and children every month.( United States Department of Agriculture, 2023) In 2009, the WIC food package was revised for the first time to match the current dietary guidelines based on recommendations of the Institute of Medicine (now known as the National Academy of Medicine).(Institute of Medicine, 2006) The food package was modified to include more whole grains and less saturated fat (e.g., reduced-fat milk).(Institute of Medicine, 2006) For the first time, women also received a cash value voucher specifically for fruits and vegetables, valued at $8 for pregnant women and $10 for breastfeeding women.(US Department of Agriculture Food and Nutrition Service, 2014) The revised food packages were designed to be cost-neutral for WIC, meaning increases in expenditures on healthier food had to be counterbalanced by reductions in expenditures elsewhere (e.g., juice, whole milk).(Hanson & Oliveira, 2009) For example, in fiscal year 2010 right after the new package was implemented, nearly 10% of the total federal costs of WIC food (or $614 million) were attributable to newly introduced spending on fruits and vegetables. This offset other spending on presumably less healthful food; consequently, overall the revision led to changes in food families brought home.(Vericker et al., 2016) Empirical work has confirmed improved dietary quality among WIC participants.(Hamad, Batra, et al., 2019; Oh et al., 2016; Schultz et al., 2015)

However, few studies have examined whether this resulted in improved maternal or infant health. One study found reduced maternal preeclampsia and increased likelihood of recommended GWG and appropriate-for-gestational-age birthweight in California.(Hamad, Collin, et al., 2019) Another study using national data found reductions in more-than-recommended GWG for women.(Pulvera et al., 2022) Few studies have examined the effects on perinatal outcomes in a large multi-state sample. To fill this gap, we utilized national birth certificate data and conducted a rigorous quasi-experimental analysis to examine the effects of the 2009 WIC revision on perinatal health (infant birthweight, GDM, and GWG), including subgroup analyses to examine differential effects among historically marginalized groups. Additionally, we explored whether the revision affected program participation among pregnant women. While the improved package could encourage greater participation, program spending was capped and the revision was intended to be cost-neutral, which may have limited health impacts. These questions are particularly informative for ongoing policymaking in the context of recent revisions during the COVID-19 pandemic to further expand the value of the WIC fruit and vegetable vouchers.

This study is among the first to examine the effects of the 2009 WIC revision on perinatal health. We used a rigorous quasi-experimental analysis and impute WIC eligibility as exposure. We found limited evidence that the revision had large positive effects on infant and maternal health. Instead, we found small reductions in infant birthweight and SGA. Although these effects size are small (ranging from −0.2% to 0.1%) and thus may not be clinically meaningful at the individual level, they may represent meaningful (albeit modest) changes at the population level.(Guyatt et al., 2002) The effect sizes are slightly smaller to those found in recent studies examining the effect of the WIC revision on perinatal health, which used different study samples and methods.(Hamad, Collin, et al., 2019; Pulvera et al., 2022) The subgroup analyses showed inconsistent associations for GDM and GWG by race/ethnicity, nativity, and education.

In terms of mechanisms for the observed effects, previous work has found the revision was associated with increased consumption of whole grains and decreased consumption of white bread and juice.(Oh et al., 2016; Schultz et al., 2015) Results about the impact on fruit and vegetable consumption were less consistent, with national or state studies showing positive effects while local studies showing little effects.(Zhang et al., 2020) A handful of studies have also found increased availability (i.e., supply) of healthy food and beverages in authorized WIC stores.(Schultz et al., 2015) In terms of changing program participation as a mechanism, we did not find that the WIC revision was associated with WIC participation by pregnant women. This is consistent with the USDA reports about how new spending on the food package worked around implementation of the new package. Improvements to the food package, such as the fruit and vegetable cash voucher, might attract more eligible women to participate in the program, or might encourage existing participants to redeem a larger share of their benefits. However, the 2009 revision was intended to be cost-neutral, and states’ WIC funding was not increased in the short run with the revision. Thus, spending on the cash value voucher had to be offset by reduced state funding on other food. The fact that the revision was mostly focused on encouraging more healthful food consumption and not increasing spending on food overall may explain some of our contradictory and null findings. For example, it may have prevented some participants from purchasing the foods they prefer. This suggests less substantive impacts would be expected, compared to more recent pandemic-era expansions to the size of the cash value voucher that were not revenue neutral. Thus, results need to be interpreted cautiously compared to predictions of increasing spending in the program.

Only a few studies have evaluated the effects of the package revision on more downstream outcomes such as maternal and infant health. Comparing findings across these studies face challenges due to variations in samples, study timeframes, and study design:

Our findings regarding birthweight-related outcomes largely mirror those of a study evaluating the effects of the 2009 WIC revision using 2006–2012 California birth certificate data.(Hamad, Collin, et al., 2019) Specifically, both studies found reduced birthweights and SGA and increased AGA. While the prior study also found reductions in LGA, suggesting that the reductions in birthweight might be explained by fewer LGA infants, our multi-state study did not find similar results for LGA. Another study using multi-state data from the CDC’s Pregnancy Risk Assessment and Monitoring Survey (PRAMS) did not find changes in SGA or AGA after the 2009 revision.(Pulvera et al., 2022)

For GDM, while we observed a null effect of the 2009 revision on GDM in the overall sample, we found differential effects across racial/ethnic, nativity, and education groups. Specially, we noted increased GDM among foreign and Hispanic women, decreased among US-born women and those without a college degree. The disparity raises questions about the underlying mechanisms driving these differences. Factors such as genetic, lifestyle, social-cultural, and other environmental factors may contribute,(Chen et al., 2019) yet further research is necessary to elucidate the complex pathways linking the WIC revision to GDM. Our findings differ slightly from those of the California study, which also found a null effect on GDM in the overall population but reduced effects among Black, Hispanic, Asian, and other racial/ethnic groups. However, their results of GDM should be interpreted cautiously due to potential violations of the parallel trends assumption. Another study using PRAMS data also identified similar violation and therefore did not access the effects on GDM.

Regarding GWG, both of the other studies reported reductions in more-than-recommended GWG for mothers.(Hamad, Collin, et al., 2019; Pulvera et al., 2022) This finding is significant considering the link between excessive GWG and increased risk of pregnancy complications, adverse birth outcomes, as well as increased obesity and chronic disease later in life for both mothers and children.(Baran et al., 2020; Bodnar et al., 2010; Simko et al., 2019) In our study, we similarly observed reductions in GWG. However, because of the lack of pre-pregnancy body mass index measurements in birth certificate data, we were unable to determine whether the reductions in GWG were primarily driven by reduced more-than-recommended GWG.

Several factors may potentially explain the lack of strong positive findings. First, research has shown that mothers’ nutritional status just before conception and during early pregnancy can influence pregnancy outcomes, such as low birthweight and SGA.(Ramakrishnan et al., 2012) However, the majority of women enroll in WIC during their second trimester, which means they miss the important early intervention stage. Second, there may be spillover of WIC benefits to other family members.(Bitler et al., 2022; Robinson, 2013; Ver Ploeg, 2009) If pregnant women passed food benefits to other family members, maternal health and infant health may not have improved as much as expected from the increased fruit and vegetable redemption. We investigated this by running the same analysis among single first-time mothers. We observed slightly larger coefficients indicating a reduced probability of SGA and GDM (Supplement Table S8); however, these findings are not statistically significant. Third, it is also possible that increased fruit and vegetable acquisitions and reduced other WIC food acquisition were not sufficiently large enough nutritional changes in magnitude to affect the birth outcomes we study. Additionally, our findings of a small reduction in infant birthweight and GWG may result from decreased overall caloric intake, an important factor influencing GWG and infant birthweight, due to increased consumption of fruit and vegetable and reduced consumption of other food.

There are several limitations of this study. First, our measure of predicted WIC eligibility may lead to some misclassification that could attenuates results towards the null if classical in nature, even while reducing issues related to self-selection; nevertheless, linkage of administrative data on WIC participation is infeasible at this time but can be explored in the future. Second, data limitations (i.e., lack of income data) prevented us from using alternative approaches, such as comparing WIC-eligible women with “near-eligible” women, who would be more comparable. The violation of the parallel trends assumption in the staggered design prevented us from leveraging the geographical and temporal variation in policy adoption. Third, we were only able to include data from 27 states, as the key outcome, GDM, was unavailable in other states adopting the new birth certificate after 2009. Since WIC program administration, food package design, and the characteristics of pregnant women vary by state, our results might not generalize to other states or nationwide. Fourth, although the largest advantage of the birth certificate data is its comprehensiveness and generalizability, there are known data quality issues with the reporting of GDM and other health outcomes on birth certificates,(Haghighat et al., 2016; Reichman & Schwartz-Soicher, 2007) and thus may be subject to measurement error. Finally, our analysis of program participation was limited by the aggregated nature of the data and the lack of specific details (e.g., on pregnancy status), but unfortunately more granular data is not available at this time.

Our study contributes to the growing body of evidence examining the effects of the 2009 WIC revision on maternal and infant health. While most studies found the revision improved diary quality for some program participants, we found limited evidence of large positive effects on short-term infant and maternal health. This underscores the importance of continued research in this area to enhance our understanding of the potential benefits and limitations of the 2009 revision—including its cost-neutrality—and to inform policy and programmatic efforts aimed at improving maternal and infant health among vulnerable populations. It may be that larger improvements—such as the $25-$45 fruit and vegetable vouchers implemented during the COVID-19 pandemic—may be more impactful.(United States Department of Agriculture, 2022)