Nova modelagem para um sistema de apoio à decisão para enfermagem baseada em evidências : o SADEBE versão 2.0

Abstract

Decision-making models help you evaluate different options and choose the best solution to solve a problem or achieve a goal. They are effective tools to model uncertainties and improve the quality of decisions. Then a decision model was developed - the Decision Support System for Evidence-Based Nursing (SADEBE) version 1.0, whose pillars for the development of the system were: Evidence-Based Practice (EBP) and Case-Based Reasoning (CBR). However, despite providing good results, the previous version of SADEBE had some limitations, namely: a smaller database, some domain dependencies and external infrastructures. Thus, a new version was proposed to increase the database and improve the decision-making model, in order to select the best conducts for the care of patients in the Immediate Postoperative Period (IPO) of abdominal surgery in the Intensive Care Unit (ICU). Among the methods tested, the modeling that best suited the problem were the Hidden Naive Bayes (HNB) and Random Forest (RF) models, considered satisfactory for this purpose. With the expansion of the study database and using another model for decision making, it became possible to solve all the problems and surpass the old version in the performance of its activities.