Project Summary/Abstract
This project will examine the effects of added Supplemental Nutrition Assistance Program (SNAP) benefits
on healthcare outcomes and disparities by race/ethnicity and disability. SNAP has been found to alleviate food
insecurity (albeit not fully), and is disproportionately utilized by racial/ethnic minorities and people with
disabilities, who have greater sensitivities to food insecurity (given “thinner margins of health” stemming from
exposures to other inequities). SNAP policy changes could therefore affect healthcare outcomes for
racial/ethnic minorities and people with disabilities in different ways than for their counterparts. Further, these
groups experience healthcare disparities (e.g., in disease prevention and management, unfavorable clinical
events, and expenditures) that are likely affected by food insecurity: individuals exposed to food insecurity miss
needed care and medications, use emergency and inpatient care, and require expenditures all at higher rates
than unexposed individuals, and effects are elevated among people with chronic conditions.
We will use both national Medicaid data and a novel set of linked claims, public health, and administrative
SNAP data from Massachusetts to exploit a staggered set of natural experiments created by the Families First
Coronavirus Response Act: (i) April 2020: ~60% of SNAP households were given variable amounts of added
monthly benefits (up to over $1,000/month, with an average of $161/unit/month over the first year), (ii) May
2021: those with no/small boosts were given up to $95/month in added benefits, and (iii) since April 2021: 18
states have ended the new benefits, which will end for remaining states in March 2023. In MA, SNAP-linked
data will allow us to use these changes to model dose-responses and treatment removal precisely. T-MSIS
Medicaid data and state variability will broaden insights and improve the generalizability of our findings.
As these SNAP changes occurred, multiple contemporaneous disruptions to healthcare, public health,
social welfare policies, and economic conditions were occurring as well. To address this issue, we will leverage
the significant heterogeneity in the geography and timing of the disruptions, in addition to the robust efforts that
have taken place during the pandemic to provide detailed, publically available data on these disruptions. We
will use confirmatory factor analysis, machine learning techniques, and an expert panel to guide our
understanding of and adjustment for these factors (Aim 1). We will then use growth curve models to measure
the effects of each SNAP change on healthcare outcomes (i.e., preventive use, unfavorable events, and
expenditures) (Aim 2) and on healthcare disparities based on race/ethnicity and disability status (Aim 3) in MA
and the US. Changes to SNAP’s funding and administration are routinely debated. This study will inform these
debates by offering rigorous, actionable evidence on the potential public health impact of these policy choices.