<1> US hospital performance: A dynamic network analysis

Abstract: Healthcare is a critical and costly industry. In the U.S. a significant component of healthcare costs are expenses generated in hospitals. This paper reports the results of analyzing 607 U.S. hospitals between 2006-2009 using a dynamic network slack-based Data Envelopment Analysis (DEA) Model. We find accounting for the dynamic and network structure of the hospital lowers efficiency estimates. Further, hospitals are more efficient at providing hospital services compared to hotel services, but the efficiency of hospitals is not correlated with their size. Regarding the dynamic network slack-based DEA Model, we find slack-based approaches combine technical and allocative aspects of inefficiency and thus tend to have significantly lower efficiency levels than just radial technical efficiency measures. Further when applying an envelopment method like DEA, there are some benefits to averaging multiple years of data to remove variation and avoid estimating a frontier based on observations that might have significant noise in their measurement