Análise custo efetividade para suporte à decisão em saúde: uma aplicação à nefrologia utilizando árvore de decisão

Abstract

This study aims to improve health decision-making by analyzing the cost-effectiveness of treatments for Chronic Kidney Disease (CKD) in the Hospital Information System (SIH) of the Unified Health System (SUS), using a tree-based approach of decision. The research included a systematic review using the PRISMA methodology, cost-effectiveness analysis and the construction of a decision tree model. The literature review revealed advances in the application of machine learning, combined with economic theories, to support healthcare decisions. This review served as the basis for the development of a method for building the decision tree model presented in this dissertation. CKD is responsible for approximately 1.2 million deaths annually worldwide, causing high costs and a significant reduction in patients’ quality of life. Early diagnosis is crucial to prevent and delay its effects. It is estimated that the error rate in diagnoses varies between 10 and 15%, highlighting the importance of healthcare decision support technologies to optimize resources and improve patients’ quality of life. The cost-effectiveness analysis indicates that the treatment of CKD has an average effective cost of R$1,964.84, while kidney transplantation from a deceased donor has the highest effective cost, R$44,269.83. The estimated decision tree demonstrates an accuracy of 90%, revealing that age group, especially if the patient is 65 years old or older, is the most influential attribute on treatment results.