Multi-criteria decision-aiding for public hospitals: The role of interactions among pairs of access and quality criteria

Abstract

More robust and resilient health systems must safeguard populations and economies in search of a more sustainable future. Their complexity, alongside the crises they have been facing, postulates the need for appropriate performance assessment mechanisms to improve the access and quality of their services. Given the existence of multiple criteria regarding these dimensions, this work proposes a collaborative multi-criteria decision-aiding framework to categorise the performance of public hospitals in terms of the access and quality of their services according to a five-star rating. Besides, although the literature is keen on assuming criteria independence, the fact is that there are several potential benefits of considering interactions between criteria. However, there is not an unmitigated, nor consensual, understanding of this concept yet, nor applications to the health sector. Therefore, our framework adopts an extension of the Electre Tri-nC method to incorporate interactions between criteria. In the end, from a sample of 26 hospitals in Portugal between 2018 and 2021, the results show that the majority are assigned to “Poor” (“Two stars”) and “Average” (“Three stars”) categories over the considered 4-year period - something that increased over time. Ultimately, assuming criteria dependence reveals more realistic results in comparison to a criteria independence assumption. Furthermore, the number of doctors and nurses is found not to influence the performance categories attributed to each hospital. In the end, several robustness analyses prove the framework’s credibility.