Fuzzy analysis of internment factors for pediatric burned

摘要

The decision between hospital internment or clinical treatment of pediatric burned represents a delicate and complex task for medical experts that are unprovided of computational tools to help in this decision process. The social context of child family, the available hospital structure, and mainly the clinical picture of the burned child are important factors to be evaluated in this decision context. Important criteria such as: extent, depth, location of the burn, child's age, and social aspects must be weighted by the expert to reach a decision. The uncertainty derived from imprecise linguistic terms such as moderate injury or great extension are the natural reasoning mechanism in this area. There are no precise boundaries separating these linguistic values. The medical reasoning based on these imprecise values leads to the conclusion about the injury classification (the diagnostic) and about the correct decision to be made: interning the child or prescribing an ambulatory treatment. An expert system that translates the uncertain medical knowledge in a set of fuzzy rules is the proposal of this research. The model was designed as an interactive and flexible tool that allows the representation of distinct scenarios through modifications of parameters such as weights of interment factors and type or parameters of membership functions.