Modeling of Risk for Diabetes Mellitus and Hypertension Using Bi-response Probit Regression

摘要

One of statistical analysis used to find relation among categorical response variable either categorical or continue predictor variables is logistic regression analysis. Logistic regression has two link functions as logit and probit link. Assumption applied on bi-response probit regression model is that both of response variables are connected. Diabetes mellitus and hypertension are a related disease. They can occur at the same time and are known as comorbidity diseases, i.e. diseases that may exist on the same patients. Hypertension sufferers have chances to become diabetic patients. Moreover, hypertension may be possessed by diabetic patients. Therefore, the aim of this study was to design a model of risk for diabetes mellitus and hypertension occurence simultaneously using bi-response probit regression approach and identify significant factors influencing diabetes mellitus and hypertension. Based on the result of chi-square test to find the relation between diabetes mellitus and hypertension, Pearson Chi-Square scored 15.009 while p-value 0.000. It means that there was a closed relation on both diabetes mellitus and hypertension so it can be concluded that both of response variables is dependent. Furthermore, based on the smallest AIC's score was obtained the best bi-response probit regression model with significant factors influencing diabetes mellitus and hypertension occurrences were Body Mass Index (BMI), systolic blood pressure, and diastolic blood pressure.