An extended DEA model for hospital performance evaluation and improvement

This paper presents an extended data envelopment analysis (DEA) model which includes a new benchmark filtering measure to identify the decision making units (DMUs) which consistently have the best performance. Additionally, window analysis is utilized over multiple time periods in order to consider the time dependent nature of hospital data. The proposed approach, termed Panel-based Benchmarking, is applied to the data from previous research to demonstrate its effectiveness and benefits. Next, the proposed approach is applied to a real-world data set spanning a five year period supplied by the Iowa Hospital Association. Finally, the Malmquist method is used to compliment the proposed approach to verify temporal productivity performance of DMUs. The results indicate that the proposed model can generate benchmarks which consistently perform well over multiple time periods.