Effect of multi-criteria decision making method selection on health system capacity rankings in OECD countries

Abstract

The effective evaluation of healthcare systems is of vital importance for optimising resource allocation and improving public health. This study aims to conduct a comparative assessment of healthcare resource capacity across OECD countries and to examine how the selection of different Multi-Criteria Decision-Making (MCDM) methods affects country rankings in this process. In this cross-sectional study, 17 key health system indicators drawn from the World Bank database were used for 37 OECD countries. The criteria were objectively weighted using the CRITIC and Shannon Entropy methods. Subsequently, four different MCDM techniques (BSS, TOPSIS, EDAS and MAIRCA) were applied to rank the health system capacities of the countries. The individual rankings obtained from all methods were combined using the Borda count method to produce a final consensus ranking. The aggregated ranking using the Borda count method identified Hungary as the best-performing country. However, when comparing the results of individual methods, the MAIRCA method, in particular, differed significantly from the others. Furthermore, the chosen weighting technique also affected the results. The CRITIC method assigned the highest weight to the criterion “current health expenditure as a percentage of GDP”, while Shannon Entropy identified the criterion “local public health expenditure per capita” as the most important. Both the weighting scheme and the MCDM method selection critically affect the health system evaluation rankings. There was a very strong positive correlation between the CRITIC-TOPSIS and CRITIC-EDAS methods (ρ = 0.9708), while a strong negative correlation was found between CRITIC-TOPSIS and CRITIC-MAIRCA (ρ = -0.8542). The findings clearly demonstrate that health system evaluations are highly sensitive to both the chosen weighting scheme and the specific MCDM technique applied. These differences serve as a warning to policymakers and researchers not to rely on a single analytical model, and emphasise the necessity of using more robust, multi-method decision-support frameworks to conduct accurate and reliable health policy evaluations.