THE FACTORS THAT CONTRIBUTE TO DIABETES MELLITUS IN MALAYSIA: ALTERNATIVE LINEAR REGRESSION MODEL APPROACH IN THE HEALTH FIELD INVOLVING DIABETES MELLITUS DATA
Abstract
ABSTRACT
Background: Diabetes mellitus (or diabetes) is a common disease that can cause of morbidity and mortality. Besides that, it is a serious deadly disease that making someone very weak and infirm. In Malaysia, the World Health Organization (WHO) has estimated the number of 0.94 millions of diabetics in 2000 will increase by 164% in the next 30 years which means the total number of diabetics is 2.48 millions in 2030. This study aimed to obtain significant factors associated with diabetes mellitus among patients in Malaysia.
Materials and Methods: The study also focused on efficiency model between multiple linear regression and alternative linear regression based on R-Square, adj R-Square, significant risk factors (p-value) and average width of the interval coefficients for each independent variable. Multiple linear regression and alternative linear regression analysis were used to identify risk factors contribute to diabetes mellitus among patients in Malaysia. The accepted level of significance was set below 0.05 (p<0.05) and all these methods are improved the programming language by using SAS 9.3 software.
Result: From the linear regression model, there is only one variable that contributes to diabetes mellitus among patients that is high factor (β = 12.82526, p < 0.0351). Whereas, by using alternative linear regression analysis, all independent variables such as mass index (β = -4.44754, p < 0.0001), total cholesterol (β = 0.06689, p < 0.0001), height (β = -1.98315, p < 0.0001), systolic blood pressure (β = 0.06941, p < 0.0001) and the weight (lbs) (β = 0.79864, p < 0.0001) are significant to diabetes mellitus. Average width of former multiple regression was found to be 61188.298 while using alternative linear regression model, the average width are 118.019.
Discussion and Conclusion: From this analysis, the most efficient method of obtained relationship between response and explanatory variable is alternative linear regression method compared to linear regression method.
Keywords: Diabetes Mellitus, Alternative Linear Regression Model and Multiple Linear Regressions