Application of discriminant analysis for evaluating the status of coronary heart disease risk

Abstract

This dissertation examines some of the factors that increase the likelihood of developing Coronary Heart Disease (CHD) and shows the application of discriminant analysis for evaluating the status of CHD risk. The data for the study was collected from the Central Laboratory of Korle-Bu Teaching Hospital, Accra and mainly consists of a group of blood tests, recognized as good indicators of whether someone is likely to have a heart attack in the near future or not. The results of patients who are at high risk of developing Cl-ID and relatively healthier individuals with low CHD risk were analysed by means of descriptive statistics and the two-group discriminant function analysis. The study revealed that the observed individuals categorized into high-Cf-ID risk group on the average are ten years older than those categorized into low-CHD risk group. Analysis of the data also showed that the high-CUD risk group patients on the average have greater body mass and the level of lipids in the bloodstream is higher as compared to the low-Cf-ID risk group patients. The results of the discriminant function analysis also revealed that the levels of cholesterol, low density lipoprotein (LDL) in the bloodstream, and body mass index (BMI) contribute most to the development of high coronary heart disease risk, with age and fasting blood sugar (FBS) contributing the least. The collected data was also used to develop a discrminant functJbn, which best discriminates between the high-CHD risk individuals and low-CHD risk individuals. The developed function was successfully applied to classify the patients with unknown status of Cl-ID risk into low or high-Cl-ID risk groups.