Did Maryland’s All-Payer Global Budget Revenue Model Change Hospital Margins and Revenue per Discharge? A Difference-in-Differences Analysis of Maryland and Massachusetts, 2010–2023

Vollständiger Abstract

Worum geht es in dieser Arbeit?

Background Maryland implemented an all-payer hospital global budget revenue (GBR) model in 2014 to decouple hospital revenue from volume and improve revenue predictability. This study evaluated whether Maryland’s 2014 GBR implementation was associated with differential changes in hospital total margins, inpatient revenue per discharge, and COVID-period margin volatility relative to Massachusetts. Methods A matched hospital-year panel for Maryland and Massachusetts spanning 2010–2023 (2,261 matched observations) was constructed by merging Medicare cost report-based financial measures with provider characteristics using CMS certification identifiers. Two-way fixed-effects difference-in-differences models with hospital and year fixed effects and hospital-clustered standard errors estimated post-2014 changes in outcomes in Maryland relative to Massachusetts. Identification was assessed using an event-study specification and a placebo test (fake treatment in 2012, restricted to the pre-period). Financial stability under pandemic shock was evaluated using a COVID-period differential effect on margin volatility, defined as the absolute year-over-year change in total margin. An additional restricted sensitivity analysis for 2010–2019 adjusted for available case-mix and payer-mix covariates. Results GBR implementation was not associated with a statistically detectable change in total margins (β = 0.021, p = 0.232). In contrast, GBR was associated with higher log inpatient revenue per discharge in Maryland relative to Massachusetts after 2014 (β = 0.287, p = 0.003), consistent with an approximate 33% increase in revenue per discharge (exp[0.287]−1). However, in the restricted 2010–2019 sensitivity analysis adjusting for available case-mix and payer-mix covariates, this association was attenuated and no longer statistically significant (β = 0.030, p = 0.513). Margin volatility during COVID-19 did not differ significantly between states (β = −0.018, p = 0.161). The placebo test showed no spurious effect in the pre-period (β = −0.008, p = 0.801), supporting the validity of the design. Conclusions Maryland’s all-payer global budgets were associated with higher revenue per discharge in the primary model, but with no detectable change in total margins or in differential margin volatility during COVID-19. The revenue-per-discharge finding was sensitive to adjustment for available case-mix and payer-mix covariates in a restricted sensitivity analysis, supporting a cautious interpretation. These findings are consistent with GBR altering revenue intensity without clear evidence of improved profitability or shock-absorbing effects over 2010–2023.