Sistema de apoio à decisão clínica para manejo e encaminhamento de pacientes com nódulos de tireoide na atenção primária

• Ano de defesa: 2023
• Autor(a) principal: Farias, Fernanda Lima de Vasconcellos
• Tipo de documento: Dissertação
• 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: Glândula tireoide - Nódulo; Tireoide - População brasileira; Atenção primária a saúde; Thyroid nodule; Clinical decision support system; Primary care; CNPQ::CIENCIAS DA SAUDE::SAUDE COLETIVA
• Link de acesso: https://repositorio.ufpb.br/jspui/handle/123456789/30152

Thyroid nodules are a frequent condition in Brazilian population and can be diagnosed through physical exam or imaging tests. Most of them are benign, usually not requiring additional diagnostic or therapeutic interventions. Traditionally, the endocrinologist is the specialist in thyroid pathologies, as well as in other diseases thar are even more prevalent and have greater morbidity and mortality, such and diabetes mellitus and obesity. Considering the low number of endocrinologists in Brazil and the possibility of monitoring patients with thyroid nodules in primary care, we propose a clinical decision support system, with the goal of differentiating more complex cases that require specialized evaluation from those with low complexity that can be managed by the non-specialist physician, suggesting in the latter situation the appropriate management. The knowledge-based system was validated through the comparison between its proposed decisions and a reference decision, taken by two endocrinologists, based on the analysis of clinical cases. We observed strong agreement order for the system to be validated, it is necessary to compare the agreement of its decision with another decision considered as reference, and we chose in this project to compare with the decision taken by two endocrinologists, based on the analysis of clinical cases. We observed a strong agreement between decisions, as the Kappa coefficient was 0.8687 with CI 95% (0.7575 – 0.9800), the overall accuracy was 93.5%, with 94.1% sensibility and 93.0% specificity.