Abstract: Based on the undesired output-super-efficiency SBM model, the efficiency of medical and health services in 31 provinces, autonomous regions and municipalities in China from 2009 to 2018 is measured, and the regional gap and dynamic evolution of health and medical efficiency is explored by using the Dagum Gini coefficient and Kernel density estimation methods. The spatial spillover effect of its influencing factors is also exhibited through the spatial Dubin model. The results can be exhibited as follows: (1) The efficiency of medical and health services in China and the four major regions of the East, the Central, the West and the Northeast are gradually increasing. The proportion of high-efficiency provinces is gradually increasing, and the proportion of low-efficiency provinces is gradually decreasing. (2) A "declining" trend is presented in the overall gap in the efficiency of China’s medical and health services. The intra-regional gap between the central and western regions is gradually narrowed, while the intra-regional gap between the eastern and northeastern regions has presented a widening trend. The gaps in the efficiency of medical and health services in the four regions are all shrinking. Part of the overlap in the process of medical development and reform is the main source of the overall gap in the efficiency of medical and health services. There is basically no polarization in the efficiency of medical and health services in the eastern and northeastern regions, while polarization exists in the central and western regions for most years. (3) A significant and positive spatial spillover effect is exhibited in the Chinese medical and health service efficiency. Per capita disposable income and the proportion of technical personnel have a significant negative direct effect on the efficiency of medical and health services. Furthermore, the proportion of graded hospitals behaves a significant negative spillover effect. The positive direct effects and spillover effects of population density and old-age dependency ratio are both very significant.

Key words: medical efficiency; undesired output-super-SBM; Dagum Gini coefficient; kernel density estimation; spatial Durbin mode