Remote Services, Benefits, and Program Retention Changes Among Young Children in WIC Authors: Erica L. Kenney, Matthew M. Lee, Rebecca S. Mozaffarian, Eric B. Rimm Published: 2026 ABSTRACT Key PointsQuestionDid introducing remote services and increasing the value of the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) benefits have associations with retention in WIC for children?FindingsIn this repeated cross-sectional study using 6 years of administrative data on 238 482 children participating in WIC in Massachusetts, retention in WIC had been declining for children over time. The introduction of a policy allowing remote WIC services was associated with a reversal of that trend, such that retention increased each month afterwards, but increasing the value of WIC benefits was not associated with further changes.MeaningThese findings suggest that making remote services available may help ensure more economically disadvantaged children benefit from WIC. Introduction Poverty threatens healthy development for infants and young children by hampering the ability of caregivers to afford and access enough healthful foods and preventive health services.1,2 To address this in the US, the Special Supplemental Nutrition for Women, Infants, and Children (WIC) program provides targeted nutritional and breastfeeding support, nutrition education, and referrals to health care services for over 6 million pregnant and postpartum women, infants, and children in households with low incomes. WIC is associated with reduced risk of food insecurity for children3,4 along with improved birth outcomes and cognitive development.5,6,7 Yet WIC is not leveraged to its full potential. While over 75% of eligible infants participate nationwide, participation drops precipitously with age; by age 4 years, only 25% of eligible children are enrolled.8 Several studies have explored reasons why participation declines with age. One is that the benefits become less valuable after the first year, when infant formula benefits stop and a toddler transitions into a WIC food package that, while nutritious, has a lower monetary value than the support for formula.9,10 Another is the complexity of participating in WIC, in terms of the administrative burden that comes with periodically recertifying eligibility in person (by verifying household income and other factors), the challenges participants may face associated with transportation, childcare, and time off in order to attend WIC clinic visits, and the relatively complicated nature of using food package benefits.9,10,11,12,13,14,15,16,17 With such challenges, many parents and/or guardians decide that continuing to navigate WIC is simply not worth it. Changes to WIC during the early stages of the COVID-19 pandemic to reduce person-to-person contact and address the economic crisis brought on by the pandemic serve as a repeated cross-sectional experiment of potential policy changes that could address these key issues of complexity and perceived value. First, Congress authorized waivers that allowed for remote WIC services and benefit administration in March 202018—implementing a policy that would, theoretically, reduce some of the administrative complexity that discourages retention. Additionally, in June 2021, Congress authorized increases in the value of the cash value benefit (CVB), which allows participant to buy produce—implementing a policy that quite literally increased the monetary value of WIC benefits. The value was initially increased substantially from an mean of $11 per beneficiary per month to $35 for all WIC recipients in June 2021, and a second change adjusted the benefit to $24 to $47 (depending on whether the beneficiary was a pregnant, postpartum, or breastfeeding parent or a child age 1 to 4 years) in October 2021. Preliminary evidence19,20,21 suggests the CVB increase contributed to better overall satisfaction with WIC, and that participants appreciated remote service delivery.22 However, it is unknown whether the CVB increase and introduction of remote services affected participant retention in WIC. In this study, we aimed to estimate whether these 2 policy changes were associated with the likelihood of postinfancy retention of children in WIC, hypothesizing that both policies would be associated with improved retention. Methods Study Design and Sample This repeated cross-sectional study used 6 years (January 2017 to December 2022) of monthly administrative data on child WIC participants from the Massachusetts Department of Public Health Nutrition Division, which oversees WIC in Massachusetts. The sample (eFigure 1 in Supplement 1) included all children aged 0 to 47 months who had been enrolled in WIC during infancy in Massachusetts during the study period, excluding children who lived in foster homes and those who left WIC early due to either death or becoming ineligible (ie, their household income rose above the eligibility threshold; they transferred out of state; or they were disqualified). This study was reviewed and approved by the Harvard School of Public Health institutional review board, which waived informed consent given the deidentified data. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline. Exposures We examined the implementation dates of 2 pandemic-era WIC policy changes as exposures (Figure 1). The first was March 2020, when WIC implemented a physical presence waiver and began issuing benefits and services remotely. The second was June 2021, when WIC increased the cash value of the CVB. Timeline of Policy Changes to the WIC and Other Relevant Nutrition Assistance Programs, 2020 to 2021 CVB indicates cash value benefit. P-EBT indicates pandemic electronic benefit transfer, SNAP indicates Supplemental Nutrition Assistance Program. WIC indicates Special Supplemental Nutrition Program for Women, Infants, and Children. Outcomes The primary outcome was monthly likelihood of retention of an eligible child in WIC, or the likelihood, in a given month, that a child younger than 4 years who had been certified during infancy and was still eligible was still certified in WIC, expressed as the monthly proportion of children who had been enrolled as newborns who still remained certified in WIC. The Massachusetts Department of Public Health Nutrition Division provided administrative WIC certification data from 2015 through 2022. For each month over this period, we calculated this population-level estimate of retention by dividing the number of children with an active WIC certification that month with a given age (in months) by the total number of children that had been certified as newborns who were calculated to be in that age group in that month. Retention for 24-month-olds in a given month, for example, represents the proportion of WIC-certified infants born 24 months prior, who at 24 months of age, were still actively certified in that month. When counting all those actively certified in a given month, this could include both individuals who had held an active certification for the entire time they are certified and individuals who might have at one point been terminated (eg, because they missed a required WIC appointment) and then reinstated (eg, because they belatedly renewed their certification). Because WIC participation declines substantially as a child grows older, we stratified analyses by age group (children aged 1 to 11 months, children aged 12 to 23 months, children aged 24 to 35 months, and children aged 36 to 47 months). For 12-month-olds, monthly estimates of retention begin in January 2016; for children aged 24 months, January 2017; and for children aged 36 months, January 2018. We excluded children aged 48 months and older from the analysis, given that their retention estimates would not have been available until January 2019 or later, which did not allow enough prepolicy time points for robust estimation. We restricted the analysis window to 2017 and onward, when retention estimates for the 3 youngest age groups were available. Our secondary outcome was retention with benefit usage (which is how the US Department of Agriculture defines participation in WIC), calculated similarly to the primary retention outcome, but including only children in households that were both actively certified and redeeming any WIC food package benefits that month. For this secondary outcome, we excluded data from June 2020 because of an administrative error in benefit redemption data. Covariates The administrative data included participants’ self-reported race or ethnicity (participants selected from Hispanic, non-Hispanic Asian, Non-Hispanic Black, Non-Hispanic White, or Non-Hispanic multiracial or other), participant sex (female or male), maternal highest level of education (less than high school or General Educational Development [GED], high school or GED, or college or above), and language spoken in the household (English, Spanish, or other language), all of which were reported to Massachusetts Department of Public Health Nutrition Division by the child’s parent or guardian. With regard to race or ethnicity, patients were provided with Non-Hispanic multiracial and other as selectable options with no additional specification. We stratified analyses by these variables to assess whether any changes in time trends associated with the policy introductions differed by these sociodemographic characteristics. We also explored whether other nutrition assistance program policy changes that could have impacted this population during the study period (Figure 1), including the introduction of online grocery shopping for the Supplemental Nutrition Assistance Program (SNAP) in June 2020, the introduction of pandemic electronic benefit transfer (P-EBT) benefits for children younger than 5 years old in March 2021, and an increase in SNAP benefit amounts related to an update to the Thrifty Food Plan in October 2021, were associated with changes in retention and thus possibly driving any observed changes in trends associated with the 2 main policy changes to WIC. Statistical Analysis We used descriptive statistics to summarize the distribution of key covariates of the certifications, by year of data. We then used an interrupted time series analytic approach with an autoregressive integrated moving averages (ARIMA) framework to test whether the introduction of each policy change (remote services in March 2020; CVB increase in June 2021) was associated with changes in time trends in WIC retention as well as our secondary outcome of retention plus benefit utilization. Interrupted time series (ITS) is a common method employed to evaluate the impact of public health policies and programs using time series data. ITS models assess whether the implementation of a policy (or other event) is associated with changes in time trends in a given outcome, relative to time trends before the event occurred. For each event studied, ITS models estimate whether there is an immediate shift in the average outcome (a level change) and whether there is a trend change in the outcome over time (a trend change). Standard assumptions of ITS models assuming independent errors may not hold when the series is particularly granular and when observations are autocorrelated over time. To account for autocorrelation in the retention outcome over time (eFigure 2 in Supplement 1), we used an ARIMA framework to fit our regression models with ARIMA errors.23 In brief, ARIMA models account for the correlation in outcomes over time in a time series (with potential adjustment for seasonality) so that impact of an intervention can be isolated from its temporal trend. For each age group, we fit an ARIMA ITS model with an intercept, an indicator for each of the policy changes (at March 2020 for remote services and June 2021 for the CVB increase), and running variables denoting time since implementation of each of the policies. We fit these models overall for each age group and by each age group-subgroup combination for the covariates described above. All analyses used R version 4.3.1 (R Project for Statistical Computing), with ARIMA models fit using maximum likelihood using the auto.arima function of the forecast package. Estimates and 95% CIs from ARIMA models are reported on the absolute percentage point (pp) scale. Statistical significance was set at an α of .05. To limit the likelihood of a type I error, we adjusted P-values within each set of analyses for multiple comparison using the Benjamini-Hochberg procedure for controlling the false discovery rate.24 Results The final sample included 616 958 certifications from 238 482 unique children participating in WIC; the sample was evenly distributed with respect to key demographic characteristics over the study period (Table 1). In the baseline year (n = 105 782), the mean (SD) age was 17.0 (14.0) months, half the sample was female (52 505 [50%]), and most children lived in households where the maternal highest level of educational attainment was high school or equivalent (69 992 [70%]). Most children (235 967 unique participants) resided in households that redeemed any WIC food package items. Table 1. Sociodemographic Characteristics of the WIC Certifications Sample From 2017 to 2022CharacteristicsParticipants by year of data, No. (%)2017 (n = 105 782)2018 (n = 98 835)2019 (n = 93 767)2020 (n = 102 990)2021 (n = 104,797)2022 (n = 110 787)Age at certification, mean (SD), mo17.02 (14.03)17.06 (14.01)17.01 (14.04)17.38 (13.85)17.65 (13.78)17.83 (14.01)Age at first certification, mean (SD), mo2.32 (6.74)2.20 (6.52)2.20 (6.56)1.96 (5.98)2.11 (6.40)2.73 (7.48)Sex Female52 505 (49.6)48 774 (49.3)46 326 (49.4)50 744 (49.3)51 579 (49.2)54 588 (49.3) Male53 277 (50.4)50 061 (50.7)47 441 (50.6)52 246 (50.7)53 218 (50.8)56 199 (50.7)Race or ethnicityb Hispanic46 788 (44.2)44 605 (45.1)42 669 (45.5)47 076 (45.7)48 337 (46.1)51 486 (46.5) Non-Hispanic Asian6167 (5.8)5548 (5.6)5236 (5.6)5235 (5.1)5085 (4.9)5159 (4.7) Non-Hispanic Black19 160 (18.1)18 392 (18.6)17 774 (19.0)19 214 (18.7)19 483 (18.6)20 690 (18.7) Non-Hispanic multiracial or othera2867 (2.7)2669 (2.7)2589 (2.8)2938 (2.9)3105 (3.0)3465 (3.1) Non-Hispanic White30 794 (29.1)27 618 (27.9)25 499 (27.2)28 527 (27.7)28 787 (27.5)29 987 (27.1)Maternal highest education Less than high school diploma or GED20 863 (19.7)19 281 (19.5)18 268 (19.5)19 715 (19.1)19 523 (18.6)20 562 (18.6) High school diploma or GED69 992 (66.2)65 066 (65.8)61 954 (66.1)69 277 (67.3)71 587 (68.3)75 933 (68.5) College and above9163 (8.7)8785 (8.9)8207 (8.8)8892 (8.6)9110 (8.7)9923 (9.0)Language spoken in household English66 682 (63.0)62 115 (62.8)58 785 (62.7)65 986 (64.1)65 545 (62.5)65 457 (59.1) Spanish22 211 (21.0)20 915 (21.2)19 587 (20.9)21 219 (20.6)21 869 (20.9)23 596 (21.3) Other language16 857 (15.9)15 792 (16.0)15 390 (16.4)15 781 (15.3)17 378 (16.6)21 730 (19.6)Redeemed any WIC food package item in the year105 226 (99.5)98 342 (99.5)93 261 (99.5)101 852 (98.9)102 989 (98.3)109 326 (98.7)Abbreviations: GED, General Educational Development; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children. a Participants are given multiracial and other as an option to select with no additional specification. Sociodemographic Characteristics of the WIC Certifications Sample From 2017 to 2022 Abbreviations: GED, General Educational Development; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children. Participants are given multiracial and other as an option to select with no additional specification. Across all age groups, from January 2017 to March 2020, the proportion of children retained in WIC each month was steadily declining by an estimated 0.006 pp (95% CI, −0.007 to −0.005) each month and reached a nadir in March 2020 (Figure 2), when the percentage of children being retained in each age group was 96% for children aged 1 to 11 months, 68% for children aged 12 to 23 months, 56% for children aged 24 to 35 months, and 48% for children aged 36 to 47 months. While the implementation of remote appointment and electronic benefits WIC waivers in March 2020 was not associated with an immediate change in retention for any age group, it was associated with changes in monthly trends in retention after March 2020 both overall and for different age groups (Table 2). The monthly trend in the percentage of children who had been certified as infants that were retained in WIC increased over the these 16 months by 1.04 pp per month (95% CI, 0.48 to 1.60 pp; P = .004) for children aged 11 to 23 months, 1.47 pp (95% CI, 1.06 to 1.87 pp; P < .001) for children aged 24 to 35 months, and 1.25 pp (95% CI, 0.79 to 1.72 pp; P < .001) for children aged 36 to 47 months. There were no significant changes in retention for infants, for whom retention stayed high at all times. In addition, we found no significant differences in retention trends by race or ethnicity, maternal education, or language (eTables 1-3 and eFigure 3 in Supplement 1). Meanwhile, there were no estimated changes in retention rates for any age group associated with the implementation of the increased CVB in June 2021. Tests for the impacts of other nutrition assistance policies that could have affected this population—online grocery shopping for SNAP participants, increased P-EBT access for young children, and increased SNAP benefit sizes—were also not significant (eTable 4 in Supplement 1). Line Chart of Proportion of Infant-Certified WIC Children Retained Across Study Period, by Age Group, Overall The first dashed line indicates March 2020—when the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) implemented remote benefits issuance and waived in-person appointments. The second dashed line indicates increases to the cash value benefit for fruits and vegetables in June 2021. Retention was calculated by dividing the number of infant-certified children in each age (in months) with an active certification by the total number of children of that age that had been certified as infants (regardless of whether they had an active or nonactive certification). Estimates were based on active, terminated, and reinstated certification administrative records from 2015 to 2022. By construction, this required differential burn-in periods by age, since active certifications prior to 2015 were structurally missing. The required burn-in period was equivalent to the age that retention was being estimated for (eg, for retention of children aged 12 months, the first 12 months of data [January 2015 to December 2015] had to be discarded; for children aged 24 months, the first 24 months of data [January 2015 to December 2016] had to be discarded; for retention of children aged 36 months, the first 36 months of data [January 2015 to December 2017] had to be discarded, and so on). Thus, retention curves for age groups are initiated at different points in calendar time. Table 2. Interrupted Time Series Model-Based Estimates of the Association Between COVID-19 WIC Policies and Changes to Program Retention, by Age GroupaAge group, moMarch 2020: WIC physical presence waiver and remote benefits issuanceJune 2021: WIC CVB for fruits and vegetables increaseEstimated level change (95% CI)P valuebEstimated slope change (95% CI)P valuebEstimated level change (95% CI)P valuebEstimated slope change (95% CI)P valuebAll ages combined−0.32 (−1.46 to 0.82).581.13 (0.71 to 1.55)<.001−0.37 (−1.52 to 0.77).58−0.42 (−0.84 to 0.005).050-11 −0.38 (−0.64 to −0.11).070.33 (0.09 to 0.57).080.12 (−0.15 to 0.39)>.99−0.22 (−0.47 to 0.04).7112-23−0.84 (−1.52 to −0.17).141.04 (0.48 to 1.60).0040.22 (−0.46 to 0.91)>.99−0.85 (−1.62 to −0.08).2824-35−0.44 (−1.22 to 0.35)>.991.47 (1.06 to 1.87)<.0010.64 (−0.10 to 1.38).71−0.63 (−1.08 to −0.17).0836-47−0.66 (−1.46 to 0.15).711.25 (0.79 to 1.72)<.001−0.27 (−1.23 to 0.69)>.99−0.30 (−0.77 to 0.17)>.99Abbreviations: CVB, cash value benefit; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children. a Models were specified using an autoregressive integrated moving averages framework to account for the correlation in retention outcomes over time. b P values adjusted for multiple comparisons using Benjamini-Hochberg control of the false discovery rate, and therefore are not always consistent with 95% CI. Interrupted Time Series Model-Based Estimates of the Association Between COVID-19 WIC Policies and Changes to Program Retention, by Age Groupa Abbreviations: CVB, cash value benefit; WIC, Special Supplemental Nutrition Program for Women, Infants, and Children. Models were specified using an autoregressive integrated moving averages framework to account for the correlation in retention outcomes over time. P values adjusted for multiple comparisons using Benjamini-Hochberg control of the false discovery rate, and therefore are not always consistent with 95% CI. For the secondary outcome of retention plus benefit usage, there were no statistically significant changes in retention associated with the March 2020 remote services waiver (eTable 5 in Supplement 1), either immediately or over time. However, there were 2.85 pp (95% CI, 1.56 to 4.11 pp; P < .001) and 3.03 pp (95% CI, 1.77 to 4.29 pp; P < .001) immediate increases in retention plus benefit usage levels in June 2021 for the age 12 to 23 months and age 24 to 35 months groups, respectively. Discussion In this study, we found that the implementation of waivers to allow for remote benefit issuance and appointments was associated with a turnaround in retention trends for WIC participants aged 0 to 3 years. Prior to March 2020, retention had been steadily declining month after month; after this time point, when the waivers were implemented and WIC was made easier to access, this trend reversed, with no discernible differences by race, ethnicity, maternal educational attainment, or language spoken at home. In secondary analyses, however, the introduction of remote services was not associated with changes in the proportions of children who were both retained in WIC and utilized food package benefits, suggesting that the waivers did not specifically impact participants’ utilization of the food benefits. Meanwhile, the increase in the cash value of WIC’s CVB later on, in June 2021, which was likely primarily beneficial for participants’ ability to purchase fruits and vegetables,19,20,21,25 was associated with small immediate increases in the proportion of children aged 12 to 23 months and 24 to 35 months retained and using WIC food benefits, then staying steady after those initial increases. Policy changes to other non-WIC food programs during the study period, meanwhile, were not associated with retention. These results suggest that WIC’s changes to allow for remote benefit issuance, certifications, and services may have helped reduce the likelihood of children leaving WIC early. While the changes were initially made to eliminate the risk of viral transmission, they also may have addressed one of the previous issues that prior research has identified as one barrier to maintaining WIC participation: the complexity of keeping up with required in-person activities, particularly for families with often limited access to childcare or transportation.9 Indeed, in a qualitative study conducted simultaneously with the present study, individuals who had been participating in WIC during the study period described how remote services removed these barriers to participation.26 These waivers, however, may not have addressed barriers to food benefit utilization, given that the introduction of remote services was not associated with increased retention plus benefit usage; it is possible that this was due to the unique context of the COVID-19 pandemic, when supply chain disruptions, understocked stores, and reluctance to shop in person may have made full benefit utilization difficult (as was also suggested in the qualitative study26). It is also possible, however, that separate interventions beyond increasing retention overall, perhaps more directly addressing the benefits themselves, are needed to ensure that benefits are fully utilized. Our results are similar to previous research that has evaluated other efforts to simplify the administrative burden of WIC and other social safety net programs.27,28 For example, transitioning from paper vouchers for WIC benefits to easier-to-use electronic benefit transfer cards in a given state (a transition that had already occurred in Massachusetts, in 2014) was associated with a 9% increase in participation.29 A recent study also found the numbers of participants in WIC to be higher during the COVID-19 pandemic in states that allowed for remote benefit reloading compared with states that required participants to reload benefits in person.30 Limitations This study has limitations. While a strength of this study is the use of complete retention data from an administrative dataset, meaning we were not reliant on self-reported participation status, we did exclude WIC children who enrolled in WIC after birth from analyses to accurately calculate population-level retention estimates each month. The study therefore may not be generalizable to all WIC-participating children, and may also not be generalizable beyond Massachusetts. While the use of multiple years of pre–COVID-19 data allowed us to robustly estimate prepolicy change trends in retention, providing increased assurance that the changes observed were not due to random error, the lack of a comparison group is a limitation. However, because all WIC participants were exposed, a comparison group was not possible. The observed retention changes could have been due to another event or policy change or factors related to the pandemic itself; however, the other policy changes that we could directly evaluate had no association with retention. Meanwhile, while a suite of other economic assistance policies for which WIC households may also have been eligible were introduced in March 2020 (precluding us from adjusting for their introduction, since it was the same time point), including increases in the cash value of SNAP benefits and similar administrative flexibilities to WIC, we would have expected these policies to result in reduced retention in WIC, given that they would have made WIC participation less economically imperative. Conversely, while it could be argued that the turn-around in WIC retention trends could have been due to increased economic anxiety, one would have expected perhaps a more immediate shift in retention if this were the case; we would have also expected to see fewer households excluded because of increased income, which we did not observe. Similarly, because the CVB increase in June 2021 occurred while the remote services waivers were still in place, we cannot determine whether the CVB increase itself would have been associated with increased retention plus benefit usage in the absence of remote services. Future research should further evaluate the impact of these policies in more controlled settings. Conclusions In this repeated cross-sectional experimental study, the implementation of remote services was associated with increased likelihood of child participants in WIC staying enrolled in this beneficial program. Policy efforts to streamline WIC services and make remote services permanent, such as the More Options to Develop and Enhance Remote Nutrition in WIC Act of 2025,31 should be prioritized.