Improving productivity using government data: The case of US Centers for Medicare & Medicaid's ‘Nursing Home Compare’

Abstract

The US Government’s Centers for Medicare & Medicaid Services (CMS) rates more than 15,000 nursing homes nationwide using a five-star scale. The outcomes are disseminated in various ways including a user friendly and informative web page. The ratings are generated using publicly available data. One objective of this work is to explore and extract these data in a replicable manner and to reveal how the government uses them to generate the star ratings. Another objective is to compare these ratings with classifications obtained with frontier analysis, a generalization of data envelopment analysis (DEA), using the same data and attributes. Frontier analysis can be made to generate results that closely parallel those from CMS. Frontier Analysis offers concrete benefits and advantages – derived mainly from its basis on linear programming – such as identification of peer performers, benchmarking, simplified sensitivity/scenario analyses, establishing star distributions, and incorporating management directives. Frontier Analysis provides transparency, simplicity, objectivity, and modeling flexibility. This work makes the case to governments to use quantitative methods such as frontier analysis to replace current highly specialized, complex, and esoteric practices while still attaining the objectives of effectively summarizing large amounts of data and information, simplifying the consumer’s decision-making process, and spotlighting excellence.

Keywords:

• Productivity
• data envelopment analysis (DEA)
• frontier analysis (FA)
• government data
• ratings

Acknowledgements

We wish to acknowledge Dr. Edward Mortimore, Technical Director, Survey and Certification Group CMS, who provided us with information, explanations, and clarifications about how CMS generates NHC ratings through personal communications. We also acknowledge the helpful comments from the four referees.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 The appendix collects a list of abbreviations and acronyms used in this article.

2 As of July 2019, these measurements come from actual records. See special note after the concluding remarks.

3 Instructions to download these files are in an appendix.