Hospital Global Budgets: Theory and Evidence of Intended and Unintended Outcomes in Maryland's Global Budget Program

Policymakers and payers are increasingly adopting prospective and value-based payment models, in an effort to align the financial incentives of providers with the health outcomes of their patients.  Most new payment models in the US have focused on payments for discrete episodes of care (e.g., bundled payments in Medicare) or for specific populations (e.g., Medicare’s Value-Based Payment Modifier), as an alternative to traditional fee-for-service reimbursement.  Unique among large-scale payment demonstrations, Maryland recently placed all of the state’s acute care hospitals under global budgets, prospectively setting the revenues a hospital can receive for all inpatient, outpatient, and emergency department services.  The state implemented this global budget program in two phases: first to rural hospitals by 2011, and then statewide in 2014.

Maryland’s program aims to improve care quality and coordination beyond the hospital setting, by making hospitals financially accountable for preventable hospitalizations and readmissions in the markets they serve.  Anecdotal evidence suggests that the state’s hospitals have responded by implementing primary care-focused population health management programs.  In addition, a recent descriptive study indicates that these efforts were successful in reducing admissions and readmissions during the first year of the program’s statewide implementation.  However, economists have noted that several features of prospective payment could have negative welfare implications for patients and providers, which have not been studied in the context of Maryland’s program.  For example, to increase profits, hospitals could selectively admit lower-cost patients (“cream-skimming”), channel volume to competing hospitals, or steer their patients to hospital-owned ambulatory care facilities whose revenues are not globally budgeted. 

This paper assesses the effects of Maryland’s first wave of hospital global budget contracts on this range of intended and unintended outcomes.  Using Medicare claims, we study differential trends in hospital admissions, readmissions, and spending, among fee-for-service Medicare beneficiaries who live in markets served by Maryland’s rural hospitals, relative to beneficiaries in propensity score matched control markets. We examine total inpatient utilization and disease-specific admissions, which globally budgeted hospitals are likely to target for reductions.  Correspondingly, we study patient flows to primary care providers, hospital-owned outpatient and ambulatory care practices, and to non-globally budgeted hospitals, to investigate mechanisms – both intended and unintended – by which hospitals could have achieved the admissions changes that we observe.  Our evaluation is currently underway; the results are forthcoming.

Unlike previous evaluations of Maryland’s global budget program, our identification strategy makes use of a richer set of control hospitals from rural markets throughout the US.  We also study a broad range of outcomes over a four-year evaluation period, to assess how the program’s effects evolved over time.  Our results will provide important insights for payment reform initiatives beyond Maryland, including efforts by CMS to introduce global budgets for other rural hospitals in the US.