SNAP Linked to Lower Health Care Use in Adults on Medicaid/Medicare

Tackling food insecurity through enrollment in the Supplemental Nutrition Assistance Program (SNAP) led to substantial drops in health care use and costs among adults 65 years of age or older dually enrolled in Medicare and Medicaid, researchers found.

Food insecurity, an issue that plagued over 35 million Americans in 2019 and as many as 54 million during the Covid-19 pandemic, is associated with poor health outcomes as individuals are forced to eat worse diets, balance the cost of food against other living essentials, and bear the psychological burden of uncertain or insufficient access to food. And these negative effects also impose a considerable financial burden in the form of high use of acute care services and increased health care costs, Seth A. Berkowitz, MD, MPH, of the University of North Carolina at Chapel Hill, and colleagues explained in Annals of Internal Medicine.

SNAP, “the nation’s largest direct effort to fight food insecurity,” has been shown to increase food access; however, many individuals who are eligible for the program do not participate, including older adults on Medicare and Medicaid who are especially likely to benefit from SNAP.

In order to assess the association between enrollment in SNAP and health care use and costs, Berkowitz and colleagues made use of what they called a “unique circumstance… As part of a state program to increase SNAP enrollment, persons dually eligible for Medicare and Medicaid received outreach for SNAP enrollment. This allowed for previously unavailable linkages among data sets related to SNAP outreach, SNAP participation, and health care use and cost.”

Using this data, Berkowitz and colleagues found that “SNAP enrollment was associated with fewer inpatient admissions, emergency department visits, and long-term care admissions over approximately 22 months of follow-up. Enrollment in SNAP was also associated with approximately $2,360 lower annual Medicaid spending per person. Further, these findings were similar across several analytic approaches that make different methodological assumptions. Despite this, enrollment in SNAP was low overall, suggesting that there is substantial room for improvement with regard to persons accessing benefits for which they are eligible.”

For their incident user retrospective cohort study, Berkowitz and colleagues used data from the outreach records of Benefits Data Trust and North Carolina Medicaid claims from September 2016-July 2020 to identify older adults age ≥65 years in North Carolina who were dually enrolled in Medicare and Medicaid but were not initially enrolled in SNAP.

They evaluated three outcomes related to health care use—inpatient hospital admissions, emergency department visits, and long-term care admission—and two related to health care expenditures—the sum of claims paid by North Carolina Medicaid and total allowable expenditures (the highest amount Medicaid could have paid had the person not also had Medicare coverage).

The final analysis consisted of 115,868 individuals (mean age 74.2 years; 67.4% women; 34.5% non-Hispanic Black), of whom only 5,093 (4.4%) enrolled in SNAP during the study period. Mean follow-up time was roughly 22 months.

“In outcome regression analyses, SNAP enrollment was associated with fewer inpatient hospitalizations (−24.6 [95% CI, −40.6 to −8.7]), emergency department visits (−192.7 [CI, −231.1 to −154.4]), and long-term care admissions (−65.2 [CI, −77.5 to −52.9]) per 1,000 person-years as well as fewer dollars in Medicaid payments per person per year (−$2,360 [CI, −$2,649 to −$2,071]).”

The study authors noted that these findings suggest the recently announced U.S. Department of Agriculture revision to SNAP benefit levels, which “will tend to increase their value substantially,” has the potential to make a major mark on public health. But this benefit cannot take their full effect unless the public health system implements efforts to increase SNAP participation. And, while coordination between agencies in state governments may help, they argued that it will also be important to reassess “the administrative burden of means-tested programs.

“Even if barriers to enrollment can be overcome with outreach, these burdens may not need to be present in the first place,” they wrote. “Lengthening recertification periods and streamlining social assistance by moving from several programs with similar eligibility criteria (for example, SNAP, the Housing Choice Voucher Program, and the Low Income Home Energy Assistance Program) to a single, more comprehensive program deserves consideration, as do more universal social assistance programs that focus on categorical eligibility rather than means testing. Future research should explore the health effects of alternate approaches to social assistance and investigate how soon benefits may accrue.”

In an editorial accompanying the study, Craig Gundersen, PhD, of Baylor University in Waco, Texas, wrote that the reduction of health care costs identified in this study “is astounding and exceeds the average annual SNAP benefit levels for seniors of $1,488. Not only do the findings of this study provide further evidence that ’SNAP is medicine,’ but Berkowitz and colleagues also show that enrolling Medicare beneficiaries in SNAP actually saves the government money.”

In light of this, Gundersen agreed that the low rates of SNAP enrollment among this population is a major concern, and he noted that one possible solution to this issue is to make enrollment automatic, “reconstructing SNAP as a universal basic income. In a modification of a standard universal basic income, if all households with incomes under 400% of the poverty line (approximately $100,000 for a family of 4) received about 125% of the current maximum SNAP benefit, there would be an estimated 98% decline in food insecurity in the United States, at a cost of $564 billion. Although this is not an inexpensive proposal, any comprehensive cost–benefit calculation should account for the potential reductions in health care costs that Berkowitz and colleagues observed.”

Gundersen also shot down the idea proposed by some experts that the U.S. turn away from focusing on food insecurity and instead focus on “nutrition security.” Such a plan, he argued, is problematic from a research perspective, “insofar as there is no accepted measure for this in contrast to food security, which has had a well-established metric for more than 25 years.” In addition, the practicality of such a plan is questionable, as there is already evidence that addressing food security improves nutrient intakes.

Perhaps most importantly, Gundersen argued, is that putting the emphasis on nutrition would lead to “calls that would lead to changes in the structure of SNAP such that it no longer allows recipients dignity and autonomy in their food procurement. As a consequence, there would be declines in SNAP participation and increases in food insecurity. Given the benefits documented in Berkowitz and colleagues’ study, we must resist the efforts of those who wish to decimate SNAP and, instead, enhance it further so that the program flourishes.”

In regard to the economic benefits of expanded SNAP access, Berkowitz and colleagues added that it is important to “distinguish between studying changes in health care use as indicators of health and viewing SNAP as a program to produce a ’return on investment’ by reducing health care costs. We view SNAP as a program that provides critical nutrition and income support to millions of Americans, rather than a cost containment strategy for the health care system. Given the clear connection between income and health, programs, like SNAP, that provide nutrition and income support to persons made vulnerable by the political economy are key tools for advancing health equity.”

Study limitations included a lack of access to Medicare data; an inability to assess how these results will generalize to other age groups and those outside of North Carolina; and the potential for bias caused by residual confounding.

1. SNAP enrollment was associated with fewer inpatient admissions, emergency department visits, and long-term care admissions over approximately 22 months of follow-up, as well as approximately $2,360 lower annual Medicaid spending per person.

2. Despite benefits in health care use and spending, only 4.4% of eligible individuals enrolled in SNAP during the study period, suggesting the need for a concerted effort to improve program enrollment among older U.S. adults.

John McKenna, Associate Editor, BreakingMED™

Cat ID: 151

Topic ID: 88,151,282,585,464,730,142,255,94,151,60,925