Heterogeneous Methodology to Support the Early Diagnosis of Gestational Diabetes

Abstract

Gestational diabetes mellitus (GDM) is a public health problem. Along with changes in eating habits, increased purchasing power, and climate change, among others, the number of women with gestational diabetes complicated by pregnancy is increasing. GDM generates problems for the mother and for the baby. Therefore, early diagnosis is important to indicate adequate medical follow-up and treatment in a timely manner. In this context, we present a hybrid methodology of a specialized system structured in the Bayesian networks, the multicriteria approach of decision support, and artificial intelligence. In such a methodology, input parameters are proposed in order to support the early diagnosis of GDM, based on the symptoms of diseases that manifest in concomitance or that develop due to the favorable environment caused by the evolution of undiagnosed diabetes. The diseases and symptoms studied were extracted from the medical literature. The diseases were weighted using the Bayesian networks, based on data from the Health Maintenance Organization with coverage in 11 Brazilian states. The weights of the symptoms were tabulated according to the analysis of medical specialists, organized by the multicriteria methodology, applying multiattribute utility theory (MAUT) methods, in particular, MACBETH, by using the Hiview computational tool. Finally, the information was structured in the knowledge base of a specialist system, made in Expert SINTA software.