Abstract
Abstract: Applying a log-logistic accelerated failure time mixed effects model to a sample of 95,504 in-hospital patients with acute myocardial infarction (AMI) between 2005 and 2010 in the United States, we measured the relative contribution of hospitals (vs. patients) in explaining in-hospital AMI mortality. Before adjusting for age, race, income, 29 comorbidities of AMI patients, and primary payer, hospital characteristics explained 19.93% of the variance in AMI in-hospital mortality. After controlling for these, variance explained declined by 5.65%, to 14.28%. These findings have implications for policymakers in assessing hospitals' "responsibility" for AMI patient mortality, for hospitals in allocating resources toward improving AMI patient care, and for medical intermediaries in making liability judgments and payment allocations to hospitals.