Toward a health situation index with uncertain availability of experts: The case of the countries associated to the Pan American Health Organization (PAHO)

In literature, several health situation index (HSI) have been developed from different perspective. These HSI require to compute many variables and generally consider the experts opinion to provide a tangible and consensus result due to the different realities under study, e.g, the different countries associated to the Pan American Health Organization (PAHO). However, the question about its practical implementation remains open under a more real scenario where the availability of experts is not necessarily guaranteed. This work proposes a HSI for the countries associated to the PAHO with uncertain availability of experts. We consider the variables in the official yearbooks published by PAHO, during the years 2005-2017, and apply a principal component analysis (PCA) to select a representative subset of them. This information is used for the design of a survey based on Majority Judgment method, which is to be answered by experts not necessarily available. Under this scenario, the behavior of those unavailable experts is modeled by an specific artificial neural network (ANN) based on a set of historical information. The obtained results allow to compute the proposed health situation index and to obtain a health situation raking for the countries associated to the PAHO.