Efficiency in university hospitals: A genetic optimized semi-parametric production function

Abstract

This paper investigates the social-welfare efficiency drivers of public university hospitals in Brazil by focusing on how the surrounding social welfare conditions may affect their performance. A novel Genetic Envelopment Analysis (GEA) approach is developed here to this end. Subsequently, LASSO regression is applied to filter the impact of social-welfare related variables –on efficiency scores. Results indicate that beds, number of employees and number of doctors are the influential factors in determining the efficiency level, while the operating scales are not relevant to the productivity level. We further find that there is a degree of difference related to the efficiency level among the hospitals in the sample. Finally, our results show that GEA estimates present higher discrimination and dispersion compared to DEA, SFA and TOPSIS, also GEA provides the most reliable and accurate results. In the second stage analysis, we find that female population ratio and high school ratio significantly affect the efficiency level in a negative manner, while the urban population ratio has a significant and positive impact. Based on these results, we provide important policy implications.

Keywords

• Productivity and competitiveness
• Genetic envelopment analysis (GEA)
• LASSO regression
• Public health care
• University hospitals