Abstract
Modeling death rates has been suggested as a potential method to screen hospitals and identify superior and substandard providers. This article begins with a review of one hospital death rate study and focuses upon its findings and limitations. It also explores the inherent limitations in the use of large data sets to assess quality of care.