Improving hospital efficiency is an emerging area of interest among policy makers in the new global economy's healthcare system. To ensure accurate efficiency analyses, we consider the nonhomogeneous input/output characteristics of various hospital departments, particularly the Department of Medicine, Department of Surgery, and Department of Other Specialist Medicine. These departments employ co-inputs to produce nonhomogeneous outputs. Specifically, we employ data envelopment analysis to evaluate the efficiency of 15 veterans hospitals in Taiwan. Empirical results show that the performance of the Department of Surgery has higher quality than that of the Department of Medicine and Department of Other Specialist Medicine. In addition, we include another data science technique, namely, impulse response function analysis. The findings indicate that "the New Hospital Accreditation" introduced in 2007 and revised in 2011 improved the efficiency of all departments in their respective first year of introductions. By contrast, the efficiencies of the Department of Surgery and Department of Other Specialist Medicine immediately decreased in the second year after the introductions.
Keywords: Data envelopment analysis; Hospital accreditation; Hospital efficiency; Impulse response function; Nonhomogeneous departments.