Abstract 20: Socioeconomic Status and Hospital Variation in 30-Day AMI Episode-of-Care Payments

Objective: We examined how payments for a 30-day episode-of-care following AMI differ for hospitals with higher and lower proportions of dual-eligible AMI patients. Background: With the implementation of bundled payment programs, an increasing number of providers are being reimbursed for episodes-of-care as opposed to fee-for-service. Providers may have concerns about how their patient mix impacts this change in reimbursement. For example, low SES patients may require greater healthcare resources due to higher disease burden or lack of social support thereby increasing the costs of care for an episode. Methods: We performed a national hospital-level study of Medicare fee-for-service beneficiaries 65 or older hospitalized with AMI (principal discharge diagnosis of ICD-9 codes 410.xx excluding 410.x2) in 2008. To create comparable 30-day episode-of-care payments across hospitals, we standardized payments by removing or averaging Medicare-specific geographic and policy payment adjustments associated with each care setting. For each hospital, we calculated a risk-standardized payment (RSP) using hierarchical generalized linear models. We risk-adjusted payments for patient age, history of cardiac procedures, and select comorbidities in the 12 months prior to and including the index admission. We limited our analyses to hospitals with 25 or more AMI index admissions and compared RSPs for hospitals with the highest (>30%) and lowest ( < 10%) proportion of dual-eligible AMI patients. We also compared the proportion of unadjusted payments attributed to index and post-acute care for these two groups. Data were analyzed using two-sample t-tests (alpha=0.05). Results: Our final sample included 1,864 hospitals. 268 had >30% dual-eligible AMI patients and 531 had < 10%. Hospitals with the highest proportion of dual-eligible patients had a significantly lower (p < 0.0001) RSP (mean = $19,962) than hospitals with the smallest proportion of dual-eligible patients (mean RSP = $20,455). In addition, hospitals with the highest proportion of dual-eligible patients spent a significantly lower (p < 0.0001) proportion of their overall payments during the index admission (mean = 74.5%) than hospitals with the lowest proportion of dual-eligible patients (mean = 77.4%). Conclusions: As a group, hospitals that cared for the highest proportion of dual-eligible patients had a lower RSP for a 30-day episode-of-care for AMI, spent a lower proportion of 30-day episode-of-care payments during the index admission, but a higher proportion during post-acute care. Given current practice patterns, this finding may suggest that bundled payments for AMI episodes-of-care generally would not penalize hospitals that provide for high proportions of low SES patients.