Mediated Relationships between Supplemental Nutrition Assistance Program Benefits, Food Hardships, Health Status, and Emergency Department Use Among Low-Income Children with and without Special Health Care Needs
Abstract
Research Objective
The Supplemental Nutrition Assistance Program (SNAP) has been found to reduce food hardships, which are associated with poor health and health care outcomes among children. Evidence that SNAP improves adult health and healthcare outcomes is growing, but studies including children have been limited. The need to assess these potential relationships among children with special health care needs (SHCN) is particularly pronounced, given their heightened social and health care complexity. We therefore aimed to (i) examine relationships between SNAP, food hardships, health status, and health care use among children, and (ii) examine whether SHCN status modified any of these relationships.
Study Design
To estimate the relationship between SNAP and food hardships, we followed prior studies in using a bivariate probit model with state-level SNAP administrative policies as instruments (to account for disproportionate self-selection into SNAP among households with high food hardships). We estimated this model within a structural equation modeling framework to simultaneously estimate downstream associations with health status and health care use. Household food insufficiency (FI) and excellent health status (versus very good, good, fair, or poor; a robust dichotomization among children) were modeled as mediators of the relationship between any SNAP receipt and any emergency department (ED) use over the prior year. We interacted SHCN status with SNAP to assess modification, and we adjusted for sociodemographic and environmental variables (e.g., exposure to smoking). Associations were estimated as the combination of direct and indirect effects. Predicted probabilities were calculated for interpretability.
Population Studied
Using pooled 2016, 2017, and 2018 National Survey of Children's Health data (nationally representative), we examined 17,791 children in households with income below 150% of the federal poverty level (a cutoff used in prior studies), 4710 of whom had SHCN based on a validated instrument.
Principal Findings
SNAP was associated with: decreased FI likelihood, increased excellent health status likelihood, and decreased ED use likelihood; SHCN status was associated with an increase in the magnitude of each of these relationships. All associations had p-values ≤0.001. See table:
| FI | Excellent Health Status | ED Use | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Total | No SHCN | SHCN | Total | No SHCN | SHCN | Total | No SHCN | SHCN | |
| No SNAP | 65% | 63% | 73% | 46% | 54% | 22% | 35% | 29% | 53% |
| SNAP | 32% | 31% | 35% | 66% | 74% | 45% | 21% | 16% | 33% |
| Δ | −33 | −32 | −37 | +20 | +20 | +22 | −14 | −13 | −19 |
| ΔΔ | −5 | +3 | −7 | ||||||
|
Δ: difference in percentage points (pp), SNAP vs. no SNAP ΔΔ: difference-in-difference in pp; (SNAP vs. no SNAP if SHCN) vs. (SNAP vs. no SNAP if no SHCN) |
|||||||||
Conclusions
SNAP was associated with significant beneficial effects regarding a child's likelihood of experiencing FI, excellent health status, and ED use, with even more beneficial effects found for children with SHCN.
Implications for Policy or Practice
Though an incomplete solution, SNAP may improve child health and healthcare outcomes, particularly for children with elevated vulnerabilities. Greater investments in food hardship relief by healthcare systems (e.g., SNAP enrollment assistance, food Rx programs) and policymakers (e.g., increased SNAP benefits, reduced enrollment barriers) may pay substantial dividends in improved outcomes for these populations and the health systems providing them care.
Primary Funding Source
National Institutes of Health.
