Metodologia per lo studio dell'efficacia relativa di strutture sanitarie

In this paper we propose a methodology for the study of the relative efficacy of different health institutions under general conditions in which a) some outcomes are simultaneously connected with qualitative and quantitative observed indicators b) the errors' distribution is not normal; c) there is a lack of prior information for casual parameters that measure the relative effectiveness of the institutions. For this topic the Multivariate Sure Model expressed in a Multilevel way is used. The solutions are obtained by means of Bayesian inference methods. The relative effectiveness of some hospitals is studied by means of this methodology