Structural Discrimination in Nonprofit Hospital Community Benefit Spending.

Importance: Nonprofit hospitals receive substantial tax exemptions to provide a community benefit. However, little is known about the distribution of community benefit spending (CBS) across US communities with varying degrees of social vulnerability beyond the hospital's immediate geographic area. Objective: To assess associations of CBS per capita with community-level characteristics and social determinants of health. Design, Setting, and Participants: This cross-sectional study used Internal Revenue Service Series 990 Tax Forms from 2018 to 2023, to create a dataset of CBS for nonprofit hospitals in the US. Facility-level CBS allocation to counties was based on inpatient utilization to more accurately reflect a hospital's community. Data were analyzed from January to December 2024. Exposures: County-level race and ethnicity characteristics and socioeconomic factors, including educational attainment, proportion living below 138% of the federal poverty level (FPL), and the Social Vulnerability Index (SVI) score. Main Outcomes and Measures: The primary outcome was total CBS per capita. Generalized linear regression models with a γ log-link function were used to assess the association of CBS per capita with community-level social determinants of health characteristics. Results: A total of 2465 nonprofit hospitals across 3140 US counties were included. Allocation of CBS varied significantly across communities, with the counties in the highest quintile receiving a mean (SD) of $540 ($250) per capita compared with counties in the lowest quintile with $22 ($16) per capita. Communities in the highest quintile of CBS had a higher proportion of White residents, while communities in the lowest quintile had a higher proportion of residents who were non-Hispanic Black or Hispanic, had lower educational attainment, and were living with incomes below 138% of the FPL. For every 1% proportional increase in non-Hispanic Black or Hispanic residents in a community, there was 1.61% (95% CI, 1.38%-1.84%) and 0.88% (95% CI, 0.63%-1.14%) less CBS per capita, respectively. In addition, there was less allocation of CBS per capita among counties with a greater proportion of people with low educational attainment, greater levels of poverty, or higher SVI scores. These results were consistent before and during the COVID-19 pandemic. Conclusions and Relevance: This cross-sectional study found that nonprofit hospitals' CBS was regressively allocated across US communities, with more socially vulnerable or racially and ethnically minoritized communities receiving less benefit than more affluent, non-Hispanic White communities, suggesting that the nonprofit tax system may be structurally discriminatory and contributing to health disparities.