Benchmarking the Health Sector in Germany – An Application of Data Envelopment Analysis

At present, a first round of hospital benchmarking as required by German law on health care reform takes place. After extensive discussions between hospitals and insurance companies, which are jointly responsible to deliver benchmarking results, a method with some peculiar characteristics was chosen. In this paper it is argued that the deficiencies of said method could be overcome by using Data Envelopment Analysis (DEA). The reasons that make DEA an advisable tool for policy decisions within the context of relative performance evaluation in the health care sector are discussed. In order to illustrate the potential of nonparametric frontier estimation for hospital benchmarking in Germany, a comparison of hospitals, which provide the same basic clinical care, is carried out. Controlling for differences in the case mix and for possible heterogeneity of the services which hospitals provide, substantial productivity differences can be detected. Beyond simply identifying inefficient providers DEA leads to additional insight about the reasons of inefficiency and to useful management implications.