Association for Public Policy Analysis & Management

Background: Food-related hardships are associated with poor health outcomes among children. The Supplemental Nutrition Assistance Program (SNAP), formerly the Food Stamp Program, has been found to reduce food hardships and is associated with improved health outcomes. However, these relationships between food hardships, SNAP, and health outcomes have not been explored among children with special health care needs (SHCN). Children with SHCN face elevated levels of an array of economic hardships and disproportionately experience poor health. Given the heightened health and economic vulnerabilities of this population, we hypothesized that children with SHCN might me more responsive to the beneficial effects of SNAP receipt.

Sample: We analyzed pooled data from the 2016 and 2017 iterations of the National Survey of Children’s Health, which includes a validated measure of SHCN status. Starting in 2016, the survey also included a measure of food insufficiency, defined as being unable to afford appropriate and/or enough food at any time over the past year (SNAP receipt was similarly measured as any receipt over the previous year). Following previous studies, we limited the sample to children from households with income below 200% of the federal poverty level. This approach increases sensitivity in detecting those potentially eligible for SNAP by allowing for reporting errors and for fluctuations in income and assets over the previous year.

Measures: We examined two binary outcomes: the presence of any food insufficiency, and whether the child’s health was excellent (versus very good, good, fair, or poor). We chose this dichotomization for health because fair/poor health is rarely reported in children and is less strongly correlated with negative healthcare outcomes than non-excellent health status.

Analysis: Differences by SHCN status in the effects of SNAP on the two outcomes were assessed using bivariate probit models that adjusted for covariates. These models were utilized to address self-selection into SNAP, and as in previous studies we used SNAP administrative policies that varied by state as instruments. Average treatment effects were calculated for interpretability.

Results: Among low-income households, children with (n = 4843) versus without (n = 14,061) SHCN were more likely to experience food insufficiency (62% versus 47%) and to receive SNAP (41% versus 30%), but they were less likely to have excellent health status (30% versus 67%). The relative reductions in the probability of experiencing food insufficiency that was associated with receiving SNAP benefits were similar among children with and without SCHN: from 69% to 50% (27% reduction) and from 53% to 34% (36% reduction), respectively (p = 0.14). However, SNAP benefits were associated with a larger relative increase in the probability of experiencing excellent health status among children with versus without SHCN: from 28% to 34% (22% increase) versus from 65% to 71% (10% increase), respectively (p = 0.02).

Conclusions: We found SNAP receipt to be associated with similar reductions in food insufficiency regardless of SHCN status. However, SNAP receipt was associated with larger increases in excellent health status for children with SHCN. Children with SHCN may be more sensitive to changes in food insufficiency status.