Using Data Envelopment to Estimate Hospital Efficiencies – A Teaching Case

Abstract

This teaching case has the intention to guide through the conduction of a meaningful Data Envelopment Analysis (DEA) in the healthcare sector. A data sample on German hospitals is provided and used throughout different tasks. Apart from the implementation of the DEA model itself, the study also covers areas of pre- and post-processing. As a result, the user of the case study is confronted with common pitfalls and learns to work with procedures, which have emerged as gold standards. The participant is encouraged to use methods for the detection of outliers and for the treatment of missing values to cover common issues in this field. The comparison of different DEA models enhances the understanding of the mechanics of DEA, especially the relevance of slacks for the analysis. In including quality data into the study, another essential feature for hospital analyses is addressed.

With the Helmsman DEA, an interesting, however, rather unfamiliar procedure is presented to achieve a meaningful inclusion of the quality data into the analysis. Using bootstrapping as a subsequent method completes the study. Finally, a recommendation for the grading of the tasks is given. The results and the source code to all implementations are provided.