Inteligência artificial para predição de desfecho e recursos necessários ao paciente submetido a cirurgia

Abstract

The assistance for patients in need of surgery involves multiple variables and is susceptible to various outcomes. Principles of Artificial Intelligence, represented here by Bayesian Networks, are tools that allow determining cause-and-effect relationships in scenarios of uncertainty with multiple variables. The study proposed here aimed to construct and test a Bayesian Network with variables related to patient care in surgery. Information regarding surgical production was used, totaling 895 procedures. The data from these procedures enabled the learning of the constructed Bayesian Network. The construction of the Bayesian Network was performed, allowing the determination of predicted outcome probabilities. The use of Bayesian Networks showed potential for aiding in medical decision-making and, due to its ability to predict the resources needed for the patient, it can assist in the managerial and administrative decisions of the surgical department.