Non-homogeneity in the efficiency evaluation of health systems.

BACKGROUND: An international comparison of health system performance is a popular tool of health policy analysis. However, the efficiency evaluation of health systems is a practical example of an international comparison in which non-homogeneity is expected. The objective of this paper is to evaluate the efficiency of health systems by models in which a degree of non-homogeneity among countries is considered. METHODS: We study the problem of non-homogeneity of health systems in the theoretical framework of the data envelopment analysis (DEA), which is a popular method of efficiency evaluation with hundreds of applications from various fields. DEA assume the homogeneity of production units and the homogeneity of the environment in which the production units operate. Hence, we compiled a summary of 14 recommendations on how to deal with the non-homogeneity in the DEA models. The analysed sample includes 38 OECD member countries. The data are from the year 2019. RESULTS: As an example, we evaluated the health system efficiency of the Czech Republic. We used the DEA models with the neighbourhood measure of distance and the constraint limiting the comparison of countries with different levels of economic development. The health system inputs were the numbers of physicians, nurses, and hospital beds. In the production of the intermediate outputs (doctor consultations, inpatient care discharges), the Czech Republic should look at Poland, Slovakia and Slovenia. In the production of health outcomes (life expectancy), the peer countries are France, Italy and Switzerland. CONCLUSIONS: The results of the DEA analysis are only indicative because no single analytical method can determine whether a health system is better or worse than others. We need to combine different methods, and DEA is one of them. We consider DEA as an exploratory method, not a method providing definitive answers.