The digital divide among U.S. Hospitals delivering health services to populations with high social disparities.

Objective: The objective of this research is to examine the relationship between digital maturity and hospital care provided to patients who are living with high social disparities. Materials and methods: This cross-sectional observational study used digital maturity data from 2,247 acute-care hospitals in the United States, linked to CMS Cost Report data, CMS Hospital Service Area data, and ZIP-code-level Social Deprivation Index (SDI) data. EMRAM maturity was classified into three categories: Low=Stage 0, Medium=Stages 1-5, and High=Stages 6-7. The primary measure of social disparity was hospital-level weighted average SDI, dichotomized at the top 20%. Hospital-level administrative indicators included Medicaid caseload, charity-care burden, and bad-debt burden. Multinomial logistic regression was used as the primary model, adjusted for hospital characteristics and U.S. region. Results are reported as average differences in predicted probabilities of Low, Medium, and High EMRAM maturity. Pearson correlations and principal component analysis were used to assess convergence between SDI and hospital-level administrative proxy indicators. Results: Hospitals treating patient populations in communities with the top 20% of SDI were more likely to have Low EMRAM maturity and less likely to have High EMRAM maturity. In the adjusted multinomial logistic regression model, hospitals in the top SDI category were 5.9 percentage points more likely to have Low EMRAM maturity and 4.6 percentage points less likely to have High EMRAM maturity; the association with Medium EMRAM maturity was not statistically significant. Medicaid caseload and bad-debt burden were not significantly associated with EMRAM maturity. Charity-care burden was associated with EMRAM maturity differently by hospital ownership. Among nonprofit hospitals, top charity-care burden was associated with lower probability of Low EMRAM maturity and higher probabilities of Medium and High EMRAM maturity. Among proprietary hospitals, top charity-care burden was associated with lower probability of Low EMRAM maturity and higher probability of Medium EMRAM maturity, but not High EMRAM maturity. Among government hospitals, charity-care burden was not significantly associated with EMRAM maturity. Discussion and conclusions: Hospitals serving populations with the highest social disparities were more likely to have Low EMRAM maturity and less likely to have High EMRAM maturity, suggesting that patients in these communities may have reduced access to hospitals with advanced digital maturity that supports coordinated, data-driven, safe, and high-quality care. Hospital administrative measures, including Medicaid caseload, uncompensated care, charity care, and bad debt, did not measure the same construct as community-level social deprivation. Charity-care burden was not uniformly associated with digital maturity across all hospitals; rather, its association differed by ownership type, with the strongest pattern observed among nonprofit hospitals. These findings should be interpreted as observational associations from pre-pandemic data and not as evidence of causal relationships.