O impacto da idade e da proximidade à morte nos custos assistencias de uma operadora de planos de saúde: um estudo sobre o red herring

Abstract

The present study aims to compare the variables “proximity to death” and “age” in determining of care costs. The techniques used were T-test of Means Differences, Wilcoxon Test, Quantile Regression and Classification and Regression Trees (CART). For this, administrative data from the assistance expenses of the beneficiary portfolio of a large medical cooperative of the State of Ceará were used, having as reference the biennium 2014/2015. With specific objectives, this research sought to compare the costs of the last twelve months of users who died with users who remained alive; to create a model able to explain the health costs from the age of the beneficiaries and their state of survival; and to create a model that allows explaining the health costs of the users who died, from their age and their proximity to death. The results obtained indicate that the variable proximity to death exerts a greater influence on the determination of the costs and that, since the beneficiary has died, there is a negative correlation between the months for death and the costs generated by him.