Sistema de apoio à decisão para a prática da enfermagem baseada em evidências em Unidade de Terapia Intensiva

• Ano de defesa: 2022
• Autor(a) principal: Rosenstock, Karelline Izaltemberg Vasconcelos
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: por
• Instituição de defesa: Universidade Federal da Paraíba; Brasil; Ciências Exatas e da Saúde; Programa de Pós-Graduação em Modelos de Decisão e Saúde; UFPB
• Palavras-chave em Português: Enfermagem baseada em evidências.; Avaliação de processos (cuidados de saúde); Redes Bayesianas; Evidence-based nursing; Process assessment (health care); Bayesian networks; CNPQ::CIENCIAS DA SAUDE::SAUDE COLETIVA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/22788

In the field of nursing research, there is an increase in interest in seeking scientific evidence aimed at solving complex problems in care practice. This model is called Evidence-Based Nursing, its practice takes into account the levels of scientific evidence presented in publications. Due to the amount and complexity of information in the health area, there is a need to produce methods for evaluating scientific evidence and decision support models as tools to aid in the practice of evidence-based nursing, in order to expose different strategies and their respective consequences in terms of risks and benefits for a given clinical issue. Therefore, this study aims to develop a computer-based decision-making model for use in evidence-based nursing practice. It is an exploratory scientific and experimental research, prescriptive with a quantitative approach. The research site was the Intensive Care Unit of the University Hospital Lauro Wanderley, where 77 cases of patients were selected with transcripts of the records of the first 24 hours of admission from the Patient Admission Form, Nursing History, Clinical Evolution and prescriptions on the day of admission by intensive care nurses about the care provided. To define the decision model, the Waika to Environment for Knowledge Analysis (WEKA) software was used, using the Hidden Naive Bayes (HNB) classifier as it presents the best results. In the present study, the HNB is composed of three characteristics of the case bank: identification/vital signs of patients, nursing diagnoses and nursing interventions. These characteristics were used to store the cases in the database and it is through them that interventions for each new case will be sought in this database. This study developed from a new decision support model for evidence-based nursing practice involving Bayesian Networks and a methodological approach to extract information from scientific evidence, constituting a management and decision support tool for nurses called SADEBE. It was observed that there are still no conceptual models that abstract specifics of Evidence-Based Practice (EBP) to be reused and instantiated in different applications and areas of knowledge with the integration of research evidence with contextual information. In this context of urgent adoption of measures to minimize the gap between scientific advances and care practice, this study contributes to achieving excellence in evidence-based nursing practice, through the use of instruments, methods and valid procedures for the obtaining reliable scientific evidence. Furthermore, the validation of a decision model based on inferences, cases and evidence contributes to overcoming barriers to the implementation of research results in practice.