Using Artificial Neural Networks in Diagnostics of Familial Combined Hyperlipidaemia Based on Levels of Certain Blood Parameters and Risk Assessment of Developing Cardiovascular Disease

Hyperlipidaemia is a condition that refers to an extremely high level of lipids in the blood. Cholesterol and triglyceride levels are highest in people with hereditary hyperlipidaemias, which are divided into five types. In this article we will discuss type IIb hyperlipidaemia. It is an inherited disorder that results in accelerated atherosclerotic early death, most commonly due to myocardial infarction. The aim of this study was to predict on the basis of blood lipid parameters, the risk of occurrence and development of cardiovascular disease, and to assist doctors in the proper diagnosis of this disease. For the purposes of this paper, artificial neural networks were used. Parameters used in the diagnosis of Familial hyperlipidaemia were: cholesterol (mmol/l), triglycerides (mmol/l), LDL (mmol/l), HDL (mmol/l), Apo B100 (mg/dl), systolic blood pressure and smoking status. The artificial neural network has an accuracy of 97.2%, reliably predicts the disease which means it can be used to help diagnose Familial hyperlipidaemia.