A Web-Based Perinatal Decision Support System Framework Using a Knowledge-Based-Approach to Estimate Clinical Outcomes: Neonatal Mortality and Preterm Birth in Twins Pregnancies.

摘要

There are two main contributions to knowledge presented in this thesis: (1) an improved method for predicting neonatal mortality and preterm birth in twin pregnancies, and (2) a framework to build a web-based perinatal decision support system (PEDSS) using a knowledge based approach.;The PEDSS includes three main components: the knowledge-base, a workflow engine, and a mechanism to communicate results. This tool provides prediction results within seconds and assists clinicians in the decision making process.;Earlier identification of clinical outcomes may lead to more efficient allocation of resources. Thus, two novel prediction models using Decision Trees(DT) and Hybrid Artificial Neural Network(ANN) were evaluated. The DT prediction model had the highest performance outcome for predicting neonatal mortality (sensitivity=62.24%, specificity=99.95%, Positive Predictive Value (PPV)=72.34%, Negative Predictive Value (NPV)=99.92%) using information available within 10 minutes after birth, and preterm birth in twin pregnancies (sensitivity=79.32%, specificity=91.97%, PPV=66.85%, NPV=95.66%) before 22 weeks gestation.