Impact of nurse home visiting on take-up of social safety net programs in a Medicaid population

Abstract

 

Childhood poverty can affect health and development across the life course. Access to social safety net programs may alleviate poverty-related hardships like food insecurity among low-income families, yet many eligible households do not enroll. We used a randomized controlled trial (n = 5670) to evaluate the impact of the Nurse–Family Partnership (NFP) home visiting program during pregnancy and the first 2 years after delivery on take-up of social programs including the Special Supplemental Nutrition Program for Women, Infants, and Children (WIC) and Supplemental Nutrition Assistance Program (SNAP). The NFP services were funded as part of a Medicaid Section 1915(b) waiver in South Carolina. We found that NFP participants were more likely to be enrolled in SNAP or WIC during pregnancy (87.8% vs 86.0%) and were enrolled in SNAP for 0.6 months longer in the first 2 years after delivery than control participants. Nurse home visiting moderately increased take-up of social safety net programs in pregnancy and the first years of life, even in a context with already high rates of participation. This study contributes important evidence on the effectiveness of Medicaid-funded initiatives for addressing social needs of low-income families.

Clinical Trial Registration

ClinicalTrials.gov; ID NCT03360539 (https://clinicaltrials.gov/study/NCT03360539).

Introduction

In the United States, 16% of all children aged 18 years and younger live in poverty, one of the highest rates among comparable countries.^1,2 Childhood poverty, particularly during the “sensitive period” of the first years of life, can affect physical, socio-emotional, and cognitive development and have long-term consequences on health and social outcomes across the life course.^3,4 Increasing access to social safety net programs for families with low income is critical for reducing poverty, alleviating food insecurity, and supporting children's healthy development.⁵ A growing body of research has shown that participation in safety net programs like the Supplemental Nutrition Assistance Program (SNAP), a federal program that provides nutrition benefits to families with incomes below 130% of the federal poverty level (FPL), and the Supplemental Nutrition Program for Women, Infants, and Children (WIC), a federal program that provides nutrition benefits and other resources to pregnant and postpartum women and children under 5 with incomes below 185% FPL, have substantial health, educational, and economic benefits for children, with the earlier the intervention, the larger the impacts.^6-11 Evidence on the impact of Temporary Assistance for Needy Families (TANF), a joint federal and state program that provides temporary cash benefits to families with lower income who have children, is weaker due to low reach, but some studies have found benefits in reducing child maltreatment.^12,13

However, millions of families who are eligible for social safety net programs do not participate.¹⁴ Reasons for incomplete take-up are complex, including lack of awareness, social stigma, language barriers, and administrative burden and complexity.^15,16 These barriers impede participation and increase program churn, whereby eligible families repeatedly exit and re-enter programs, reducing their effectiveness.^15,17 Only about half of eligible people receive WIC benefits.¹⁸ In a third of US states, participation in SNAP among eligible people is less than 75%.¹⁹ Because of the complexity and fragmentation of the US safety net system, individuals who participate in 1 program like Medicaid still face barriers to participation in other programs for which they are eligible.¹⁶ Moreover, the populations that are most critical to reach with social programs may be the least likely to receive them.¹⁵ Yet, there is limited research on policy approaches or interventions to facilitate take-up of social safety net programs among eligible families.^20-22

Home visiting programs, which are designed to support pregnant women and children experiencing economic hardship, may be well positioned to improve take-up of safety net programs among eligible families. Linking families to needed services and community resources, also called service coordination, is a central component of evidence-based home visiting models.^23,24 Home visitors can increase families’ awareness of available programs, assist in completing onerous paperwork and documentation, provide instrumental support such as contacting agencies and setting up appointments, reduce stigma, and provide encouragement and emotional support.²⁴

While connecting families to social programs is a stated goal of most home visiting programs, there is little evidence on the effectiveness of home visiting on take-up of safety net programs. Historically, some home visiting programs considered reduced use of social programs across the maternal life course as a positive indicator of economic self-sufficiency.^25,26 Randomized evaluations of the Nurse–Family Partnership (NFP) have examined its impact on reductions in receipt of welfare; while no effects were found at ages 2 or 4 years, evaluations at ages 6, 9, and 18 years found fewer months of TANF and SNAP receipt.^26-30 More recently, a randomized trial of Healthy Families Oregon found that participants in the program were more likely to be enrolled in TANF and to have received more days of SNAP over the first 2 years.³¹ Yet, considering the well-established benefits of programs like SNAP on child well-being, much more evidence is needed on the extent to which home visiting programs can increase take-up of social safety net programs in the first years of life. In particular, there is little evidence of impact on take-up during pregnancy, a critical period for fetal development.^7,32 Few studies have measured the use of services or resources using administrative records instead of retrospective self-report.³¹ In addition, little is known as to whether home visitors can help reduce program churn.³³ Finally, evidence is needed to identify whether any positive effects on take-up are sustained among the most disadvantaged families who face greater social vulnerabilities, such as poor mental health, and who may have the greatest barriers in accessing services.

In 2016, the state of South Carolina began to offer NFP home visiting services as part of a Medicaid Section 1915(b) waiver. Philanthropic funding supported the scale-up of the NFP program through a “Pay for Success” model, which makes payments back to funders contingent on the effects of the program on predetermined health outcomes evaluated in a randomized trial (see Appendix Section 1 for complete details).³⁴

The objective of this study was to determine the effect of an intensive nurse home visiting program on take-up and use of social safety net programs, including WIC, SNAP, and TANF, in pregnancy and the first 2 years after delivery. Our hypotheses were that nurse home visiting would increase take-up of SNAP and WIC during pregnancy and increase duration of use and decrease churn of SNAP and TANF in the first 2 years postpartum.

Data and methods

Trial design

We conducted an individually randomized controlled trial of the NFP home visiting program in South Carolina. Participants were individually randomized to either a treatment group that was offered access to NFP services or to a control group. The primary outcomes of the trial were a composite of adverse birth outcomes, interbirth intervals of less than 21 months, and major injury or concern for abuse or neglect in child's first 24 months.³⁴ Previous work found no impact of the intervention on birth outcomes or on outcomes related to the use of prenatal care, but did find that the intervention reduced the use of the emergency department in the first 12 months.^35-37 This study is an analysis of preregistered secondary outcomes related to take-up and use of social safety net programs. This study follows CONSORT (Consolidated Standards of Reporting Trials) guidelines for reporting randomized trials and was approved by Harvard T.H. Chan School of Public Health Institutional Review Board (IRB15-2939). Complete details of the study design were published in the study protocol.³⁴

Trial enrollment took place from April 1, 2016, through March 17, 2020 (enrollment was stopped due to the COVID-19 pandemic). Individuals were referred to NFP by way of a variety of sources, including their local health care providers, schools, Medicaid, or self-referral, through 1 of 9 NFP implementing sites in government agencies and hospital systems throughout South Carolina.³⁴ The NFP staff were encouraged to enroll people from low-income ZIP codes. Trained NFP staff assessed potential participants’ eligibility. Eligible participants provided written informed consent and completed a short baseline survey. Participants were then randomly assigned on-the-spot in a 2:1 treatment-to-control ratio using computer-assisted software.

Eligible participants were nulliparous pregnant women at 28 weeks of gestation or less who were eligible for Medicaid during pregnancy and resided in an NFP-served catchment area. These criteria reflect NFP's general eligibility for pregnant women to access their programs. Individuals who were 14 and younger, incarcerated, or in a lockdown facility were excluded, due to the challenges of including younger children and incarcerated individuals in research. Participants randomized to the treatment arm were offered NFP services throughout pregnancy and up to 24 months after delivery. NFP services consist of educational sessions, health screenings, and life-course goal setting conducted by a nurse during biweekly or monthly home visits. Nurses assess client needs and preferences and may provide referrals and follow-up support (eg, encouragement, information, or hands-on assistance) in accessing social safety net services. Nurses flexibly support clients with additional or fewer visits as necessary. After March 23, 2020, nearly all visits were conducted via telehealth. Services were provided in languages other than English using bilingual nurses for Spanish, where available, or interpreter services. Participants randomized to the control arm received usual care in South Carolina, which included access to all other community and medical services. All study participants were provided with a list of available community resources at baseline.³⁵ All participants were also eligible to receive a postpartum infant home visit within 6 weeks of delivery, which is covered by the South Carolina Medicaid program.

Data

We matched study participants to vital records and Department of Social Services records via a probabilistic match based on name, race, social security number, birth date, and Medicaid ID. The analytical sample was restricted to participants with an “index” live birth in matched vital records within 120 days of the expected delivery date reported on the baseline survey. To assess receipt of NFP services, we matched participants to intervention programmatic data.

Outcomes

Receipt of WIC was determined from the vital record. In the Department of Social Services records, we extracted dates of enrollment and exit from SNAP and TANF services by month. Based on these records and the date of delivery, we calculated the following preregistered outcomes: whether or not a participant received SNAP or WIC during pregnancy, the number of months the participant received SNAP or TANF between pregnancy and 2 years postpartum, and SNAP or TANF benefit churn between pregnancy and 2 years postpartum. Churn was defined as having received a program at any time during the year and having experienced at least 1 break in participation of 4 months or less that started or ended during that year.³³ We also examined further exploratory outcomes including separately examining WIC, SNAP, and TANF receipt and the number of months of receipt in the first 1 and 2 years postpartum. Details on time periods for outcomes are provided in Appendix Table S1.

Statistical analysis

Analysis was performed between January 11, 2024, and July 2, 2024. Per our pre-analysis plan and protocol, we used an intent-to-treat approach and ordinary least-squares linear regression models to compare outcomes between the intervention group and control group.^34,38 We estimated unadjusted models and models adjusted for baseline covariates. Covariates used in the adjusted models included participant's age, race and ethnicity, gestational age at study enrollment, relationship with the father of the child, education, employment, use of social services, housing stability, health care utilization, health behaviors, and physical and mental health status, measured during the baseline survey (definitions provided in Appendix Section 2). Tests were 2-sided. Statistical significance was defined as P ≤ .05.³⁴ We used dummy variables to account for missing baseline covariates in regression models, which is appropriate in randomized trials when covariates are balanced between treatment and control.³⁹ We also examined outcomes in the prespecified subgroup of participants facing social vulnerabilities, which was defined as individuals who were younger than 19 years old, had not finished high school, or had challenges with mental health (defined by a Patient Health Questionnaire-2 [PHQ-2] score of ≥3 at baseline or reported receiving mental health treatment in the year before enrollment). We use the term “social vulnerabilities,” recognizing that vulnerabilities are not inherent characteristics but rather the result of wider structural causes.⁴⁰ As robustness checks, we examined results for the subgroup of participants who did not report any use of WIC, SNAP, or TANF at enrollment, and, as many social programs implemented changes to eligibility and enrollment procedures during the COVID-19 pandemic, we also examined results for the subgroup of participants who were unaffected by COVID-19 (details provided in Appendix Table S1). Analyses were conducted using Stata, version 14.2 (StataCorp LLC).

Limitations

This study has limitations. First, some participants were not matched to administrative data sources; however, this attrition was balanced across treatment and control. Second, we did not have access to income data to confirm eligibility of participants for means-tested programs; while all participants in the trial were eligible for pregnancy Medicaid (income ≤199% FPL), income eligibility levels in South Carolina for WIC (185% FPL), SNAP (130% FPL), and TANF (30% FPL) are more stringent. Third, while NFP is an established home visiting program that operates in 40 states,⁴¹ results may not be generalizable to other states with different population characteristics and safety net program eligibility and administrative contexts. Fourth, while the use of administrative data allows for objective measures of social service participation, errors in data may lead to mismeasurement of outcomes. Finally, the COVID-19 pandemic and restrictions resulted in most home visits moving from in-person to telehealth, as well as changes in benefits of social programs (SNAP, for example, issued emergency allotments to supplement benefits⁴²). Our results were robust to the sample of participants who were unaffected by COVID-19.

Results

There were 5670 participants enrolled in the trial and 4932 participants (3295 in the home visiting group and 1637 in the usual care group) were matched to administrative records (Appendix Figure S1). Match rates were not different between treatment and control groups (Appendix Table S2). Baseline characteristics were balanced across treatment and control groups (Table 1). Most participants were non-Hispanic Black (55.2%) or non-Hispanic White (34.7%), while 5.8% were Hispanic, 3.1% were more than 1 race, and 1.2% were Asian, Indigenous, Native Hawaiian, or Pacific Islander. Among participants, 22.3% had less than a high school diploma and 35.4% had only a high school diploma or equivalent. At baseline, 52.8% of participants were receiving WIC and 25.1% were receiving SNAP, while less than 1% were receiving TANF. Over one-third reported experiencing food insecurity (36.5%) and two-thirds reported high stress (66.0%).

Participation in the home visiting program was high during pregnancy but waned in the first year and second year postpartum (Table 2). While 98.3% of participants received at least 1 completed nurse home visit during pregnancy, 76.6% received a visit during the first year postpartum, and 52.1% received a visit in the second year postpartum. In-person visits were most common during pregnancy, while telehealth visits, which tended to be shorter, became more common in the postpartum years, largely due to the restrictions imposed by the COVID-19 pandemic. Across pregnancy, first year postpartum, and second year postpartum, visit duration decreased, with an average length of 67, 58, and 49 minutes, respectively. Among treatment participants with at least 1 completed nurse home visit, 32% received a referral for WIC and 18% received a referral for SNAP in the pregnancy period. Referrals decreased in the first year postpartum, with 19.2% receiving a referral for WIC and 13.7% receiving a referral for SNAP in the first year postpartum and 8.9% receiving a referral each for WIC and SNAP in the second year postpartum. TANF referrals remained low, reflecting restrictive eligibility for the South Carolina TANF program. Among intervention and control participants, the percentage receiving a Medicaid-reimbursed home visit was low, with 6.3% of intervention and 11.5% of control participants (Appendix Table S3).

Figure 1 shows the program effect estimates for both the full sample and the subgroup of participants facing social vulnerabilities. For the full sample, compared with usual care, intervention participants were more likely to be enrolled in SNAP or WIC during pregnancy (adjusted between-group difference: 2.1 percentage points [pp]; 95% CI: 0.3 to 4.0 pp) (Figure 1). Intervention participants were also more likely to be enrolled in SNAP in the first year postpartum (3.1 pp; 95% CI: 0.4 to 5.7 pp) and were enrolled for 0.6 months longer (95% CI: 0.1 to 1.1 months) in the first 2 years postpartum than usual-care participants. There was no difference in outcomes of SNAP receipt in the second year postpartum, or TANF receipt or SNAP or TANF churn in the first 2 years postpartum. Results for additional exploratory outcomes are shown in Appendix Table S4. Effect sizes were similar for the subgroup of participants facing social vulnerabilities, although less precise due to the smaller sample size (Figure 1). Results were also similar for the group of participants who were not receiving social services at enrollment (Appendix Table S5) and those unaffected by the COVID-19 pandemic (Appendix Tables S6 and 7).

Discussion

This study investigated the impact of the NFP home visiting program on take-up of social safety net programs. We found that a nurse home visiting program moderately improved take-up and use of social safety net programs including SNAP and WIC in pregnancy and in the first 2 years after delivery by 2%–5%. The impact of the program on take-up for the subgroup of participants facing social vulnerabilities was similar to that for the full sample. These effects are similar to a home visiting intervention in Oregon that increased receipt of SNAP by 2.8% in the first 2 years after delivery, and slightly lower than recently reviewed interventions that increased SNAP and WIC participation by 5%–8%.^31,43,44

In our sample of Medicaid-eligible pregnant women, participation in social programs in the usual-care group was fairly high (86% participated in SNAP or WIC during pregnancy and 59% participated in SNAP in the first year postpartum). Larger structural determinants beyond the reach of home visitors, such as income eligibility levels and sanctions, may have hampered home visitors’ efforts to further increase participation.⁴⁵ In addition, family ambivalence, perceived stigma, and lack of time have been identified as family-level barriers to successfully connecting to services.³² Previous research has found that incomplete take-up of social programs is influenced by multiple factors at individual, interpersonal, and institutional levels.²⁰ Additionally, in this trial, the NFP program was scaled to include a broader population, which, while demographically similar to the usual NFP population, was found to be more engaged in medical and social care.³⁶ Further research is needed to better understand the challenges of service coordination within home visiting and the effectiveness of home visitors in connecting clients to services across different programs.²⁴

The largest program impacts we found were for increased WIC and SNAP participation during pregnancy and the first year postpartum, whereas the effects subsided by the second year postpartum. This may be due to changing client needs or eligibility over time. It may also be due to waning participation in the home visiting program over time, where fewer participants received home visits in their second year after birth. Identifying strategies to decrease participant attrition in home visiting programs is critical to supporting child development in the early years—for example, by reducing housing instability, which has been identified as a primary reason for home visiting attrition.⁴⁶

Our study is situated in an innovative policy context in which home visiting services were offered under a Medicaid Section 1915(b) waiver. Since 2022, the Centers for Medicare and Medicaid Services has authorized a variety of demonstration projects to address social determinants of health under the health-related social needs (HRSN) framework, including housing and nutrition supports, case management, and transportation.⁴⁷ Evaluation of these innovative models is critical to understanding the most cost-effective ways to identify and address social needs, coordinate social safety net programs, and promote child health and development.⁴⁸ Our study contributes to the evidence base on the effectiveness of innovative models in meeting the social needs of pregnant Medicaid enrollees and their families. More evidence is needed on what types of community-based care coordination programs are most effective in linking families with low-income to safety net programs.

In addition to Medicaid's interest in connecting families with the safety net, clinical practices are increasingly seeking to screen and address the social determinants of health in obstetric and pediatric settings.^7,49,50 However, there is limited evidence of effectiveness of clinical interventions on take-up of social programs. An intervention comparing screening of pediatric patients in emergency, inpatient, and outpatient settings found that only a small proportion of families with food insecurity were ultimately connected to resources at any of the settings.⁵¹ A pilot program to identify families with food insecurity and enroll eligible families in SNAP in a pediatric clinic found that only 8% of eligible patients completed the enrollment process.⁵² Similarly, a primary care intervention to increase WIC enrollment that included extensive provider educational activities did not sustainably change screening, counseling, or referrals.⁵³ Challenges faced by pediatricians in clinical interventions include time constraints, uncertainty as to how to handle positive screens, and patient stigma. Developing collaborative relationships between clinicians and evidence-based home visting programs may address some of these challenges.⁵⁴ Further research on the effectiveness of clinical interventions in addressing patient's social needs is needed, including interventions that partner physician practices with home visiting programs are needed.

Conclusion

Overall, we found that a nurse home visiting program for pregnant Medicaid enrollees moderately increased take-up of social safety net programs in pregnancy and the first 2 years after delivery. Future research could explore how other innovative models of care can be used to improve take-up rates and address health-related social needs of eligible families.

Acknowledgments

The Abdul Latif Jameel Poverty Action Lab led the implementation of the trial. The authors thank current and former leadership and staff at South Carolina Department of Health and Human Services, NFP nurse home visitors, and NFP leadership and NFP staff. This paper was presented at the 2025 Population Association of America Research Meeting.

Supplementary material

Supplementary material is available at Health Affairs Scholar online.

Funding

Supported by the Children’s Trust of South Carolina, Arnold Ventures, Duke Endowment, BlueCross BlueShield of South Carolina Foundation, and J-PAL North America’s US Health Care Delivery Initiative.

Role of funder

Researchers at J-PAL were involved in the design and conduct of the study, as well as data collection, management, analysis, and interpretation of the data, and review of the manuscript. No other funder had a role in the design and conduct of the study; collection, management, analysis, and interpretation of data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Data sharing statement

The data use agreements for this study do not permit sharing individual-level linked data publicly.

Notes

Author notes