Critical aspects of the usage of statistical classification models

Abstract

Statistical classification models are used every day in health care as evaluation systems allowing building homogeneous groups to help in prognostics, diagnostics, therapy choosing and any other situation. The usefulness of them is not only evident in the medical surroundings, but many of them tribute directly to the performance and policies improvement in the Public Health area. As part of the health technologies, they require a continuous assessment in the different environments they are performed. On the other hand, using them mechanically, without a critical mentality, could bring about more risks than benefits, so their predictive or explanatory power does not justify the usage of their outcomes in an unquestioned way.