Association Between 30-Day Episode Payments and Acute Myocardial Infarction Outcomes Among Medicare Beneficiaries

Abstract

Background: Recent policy efforts have focused on improving the value of acute myocardial infarction (AMI) care. Medicare payment programs, for example, increasingly evaluate hospital performance based on spending, as determined by payments made to institutions and providers, and outcome measures for a longitudinal episode of AMI care. Little is known about the relationship between total 30-day payments-both in the inpatient and immediate postdischarge timeframe-and outcomes after an admission for AMI.

Methods and results: Using Medicare claims data, we identified Medicare fee-for-service beneficiaries ≥65 years of age who were hospitalized at an acute-care hospital for AMI between July 1, 2011, and June 30, 2014, and examined the association between hospital-level 30-day payments for an episode of AMI care and patient 30-day mortality using mixed regression models with a logit link function and random hospital intercepts. Our cohort included 642 105 index hospitalizations for AMI at 2319 acute-care hospitals. Overall mean 30-day episode payments per beneficiary were $22 128 (SD, $1750). The observed 30-day mortality rate was 12.9%. Higher 30-day payments were associated with lower 30-day mortality after adjustment for patient characteristics and comorbidities (adjusted odds ratio for additional $1000 payments, 0.986; 95% confidence interval, 0.979-0.992; P<0.001). Additional adjustment for potential mediating factors, including hospital characteristics, coronary revascularization rates, and discharge disposition, did not significantly attenuate the relationship (adjusted odds ratio for additional $1000 payments, 0.987; 95% confidence interval, 0.980-0.994; P<0.001).

Conclusions: Higher hospital-level 30-day payments-both inpatient and in multiple settings after discharge-for AMI care were associated with lower 30-day mortality among beneficiaries. This may have implications for payment programs that incent reduction in payments without considering value.

Keywords: Medicare; health expenditures; mortality; myocardial infarction; percutaneous coronary intervention.

Figure 1. Geographic Variation of County-Level 30-day Episode Payments for Acute Myocardial Infarction Care

Geographic areas in white reflect locations which did not have hospitals that were included in our study cohort.

Figure 2

Association Between Hospital-Level 30-Day Episode Payments Following a Hospitalization for Acute Myocardial Infarction and Patient-level 30-Day Mortality

Figure 3. Unadjusted and Multivariate Adjusted 30-day Patient Mortality Rates for Every $1000 Increase in Hospital-Level 30-day Episode Payments

Model A - Unadjusted Model B - Adjusted for patient characteristics and comorbidities Model C - Adjusted for patient characteristics and comorbidities as well as hospital characteristics Model D - Adjusted for patient characteristics and comorbidities, hospital characteristics, and hospital-lcvcl 30-day revascularization rates Model E - Adjusted for patient characteristics and comorbidities, hospital characteristics, hospital-lcvcl 30-day revascularization rates, and SNF and homccarc utilization rates