Medicare Hospital Outcome Measures and Unrelated Socioeconomic and Clinical Factors

Abstract

Prior studies generally find non-existing or weakly negative (heart failure) associations between risk adjusted readmission and mortality rates at the hospital level. We examined all Hospital Referral Regions (HRR), and found persisting modest positive associations for acute myocardial infarction and stronger negative associations for heart failure at the HRR level. Subgroup analyses suggest differences in local characteristics of beneficiaries are associated with differences in RSRR and RSMR. It is clear that community infrastructure – the psychosocial context in which a discharged patient operates – should be an important predictor of post-discharge events. Yet such community differences are not currently well-addressed in CMS’ hospital readmission and mortality models. Given the importance, use for accountability, and basis for substantial financial penalties of 30 day risk standardized outcome measures, it is vital that CMS hospital performance measures continue to be scrutinized.

Keywords: Medicare