Abstract 204: Impact of Global Budgets on Hospital Cardiovascular Care in Maryland

Abstract

Background: Global budgets (GB), an innovative method of hospital payment, have been considered effective in containing expenditures. In January 2014, the state of Maryland introduced Global Budget Revenues (GBR), transitioning from per case to capitated, population-based payments (PBP). Under this model, hospitals are provided yearly funds, thus being financially incentivized to improve efficiency of services. Prices are set by the Health Services Cost Review Commission (HSCRC), which collects data and monitors hospitals’ quality of care. Potential adverse incentives of GB would include underprovision of necessary services or avoiding sicker, high cost patients (cream skimming). This study aimed to evaluate the effects of GBR on the care of hospitalized patients with important CV conditions. Methods: We analyzed HSCRC inpatient claim databases containing anonymized information on statewide hospital admissions between fiscal years 2013 and 2018. Using ICD codes, we identified patients with a principal diagnosis of heart failure (CHF), acute myocardial infarction (AMI), and acute ischemic stroke (AIS) from a pool of 1,959,237 inpatient admissions. Outcome measures were: hospitalizations, length of stay (LOS), inpatient PCI and CABG procedure rates, casemix adjusted 30-day readmission rates (CARR), risk standardized in-hospital mortality, and hospitalization charges. Trends in outcome measures before and after GBR implementation were compared using segmented regression analysis, with p values indicating trend change. Results: With introduction of GBR, CHF and AMI hospitalization trends did not change significantly, whereas AIS admissions stabilized downward (p<0.0001). There was an increase in mean LOS for CHF (p=0.059). CARR followed pre-policy declining trends, with more consistent improvements noted for the AMI cohort (p=0.005). In-hospital mortality continued to decrease, although at slower rates for CHF and AMI cohorts (p=0.03; p=0.019). CABG procedure rates declined significantly (p<0.0001). Mean hospitalization charges increased for all three CV conditions, faster for CHF (p=0.03) and AIS (p<0.0001), slower for AMI (p=0.0003). Conclusions: Adoption of GBR in Maryland had no harmful effects on in-hospital outcomes and quality measures. There was a modest improvement in potentially avoidable hospitalizations and mildly reduced utilization offset by significant rise in charges. Some observations could result from care shifts to the outpatient setting. Quality of CV care was not affected by GBR suggesting that efficiency improvements might have occurred in other areas of the healthcare system. Additionally, combining GB with pay for performance programs and assiduous quality monitoring might have mitigated adverse incentives. As GBR transitions to Maryland Total Cost of Care, long-term effects of PBP on CV outcomes will require further investigation.