Methods: Eligible patients were fee-for-service Medicare beneficiaries 65 years and older hospitalized at an acute-care hospital for AMI between 2011 and 2014. Spending was defined as hospital-level risk-standardized payments associated with a 30-day episode of AMI care, stratified into low ( less than 25 percentile), average (greater than or equal to 25 to less than or equal to 75 percentile), and high (greater than 75 percentile) spending groups. The primary outcome was mortality within 30 days of admission. To examine the association between hospital-level 30-day spending and mortality, we fitted mixed-effects logistic regression models with random hospital intercepts to model 30-day mortality as a function of patients’ admitting hospital’s expenditure.
Results: We included 642,105 index hospitalizations for AMI at 2,319 acute-care hospitals. Hospitals with higher 30-day spending were larger, tended to be teaching hospitals, were more often located in an urban area, were more likely to have cardiac catheterization laboratories and cardiac surgery capability, and also had higher rates of coronary revascularization. Across hospital groups, median 30-day spending per beneficiary was $20,207 (low), $22,018 (average), and $24,174 (high). Higher hospital-level 30-day spending was associated with lower 30-day mortality even after adjustment for patient and hospital characteristics (adjusted odds ratio for additional $1000 spending, 0.989; 95% confidence interval 0.982-0.996, p=0.002). This relationship was not attenuated following additional adjustment for hospital-level revascularization rates.