Performance evaluation of nonhomogeneous hospitals : the case of Hong Kong hospitals

Li, Yongjun and Lei, Xiyang and Morton, Alec (2019) Performance evaluation of nonhomogeneous hospitals : the case of Hong Kong hospitals. Health Care Management Science, 22 (2). pp. 215-228. ISSN 1572-9389

[img]
Preview
Text (Li-etal-HCMS2018-Performance-evaluation-of-nonhomogeneous-hospitals)
Li_etal_HCMS2018_Performance_evaluation_of_nonhomogeneous_hospitals.pdf
Accepted Author Manuscript

Download (2MB)| Preview

    Abstract

    Throughout the world, hospitals are under increasing pressure to become more efficient. Efficiency analysis tools can play a role in giving policymakers insight into which units are less efficient and why. Many researchers have studied efficiencies of hospitals using data envelopment analysis (DEA) as an efficiency analysis tool. However, in the existing literature on DEA-based performance evaluation, a standard assumption of the constant returns to scale (CRS) or the variable returns to scale (VRS) DEA models is that decision-making units (DMUs) use a similar mix of inputs to produce a similar set of outputs. In fact, hospitals with different primary goals supply different services and provide different outputs. That is, hospitals are nonhomogeneous and the standard assumption of the DEA model is not applicable to the performance evaluation of nonhomogeneous hospitals. This paper considers the nonhomogeneity among hospitals in the performance evaluation and takes hospitals in Hong Kong as a case study. An extension of Cook et al. (2013) [1] based on the VRS assumption is developed to evaluated nonhomogeneous hospitals' efficiencies since inputs of hospitals vary greatly. Following the philosophy of Cook et al. (2013) [1], hospitals are divided into homogeneous groups and the product process of each hospital is divided into subunits. The performance of hospitals is measured on the basis of subunits. The proposed approach can be applied to measure the performance of other nonhomogeneous entities that exhibit variable return to scale.