Binary Logistic Regression Modeling for Characterisation of Hypertension and Some Non-Hereditary Risk Factors

High blood pressure is a serious concern for public health, which is a crucial area of study as it is a key transmitting agent for coronary artery diseases and other complications. The population of hypertensive individuals is on the increase due to a number of factors, and the rate of prevalence of this morbidity and its terminal effect on humanity is alarming. This study aimed to identify some preventable and controllable risk factors of hypertension. Secondary data (n = 310) were collected from the departments of endocrinology and cardiology of the Teaching Hospital at the University of Ilorin (UITH), Kwara State. A binary logistic regression model was fitted on factors like age, gender, and smoking. Results showed that age, gender, alcohol consumption, height, BMI, and hours of daily rest are positive predictive factors for hypertension, where gender, working status, and BMI are statistically significant response variables (p-value < 0.05). Also, the odds of developing hypertension with respect to gender, working status, and BMI are 4.25, 0.55, and 7.09, respectively, when other predictor variables are held constant. The receiver operating characteristics, ROC, which measures the sensitivity and specificity of the model (AUC = 0.7141), indicated the probability that the model is more likely to assign a higher probability to a positive case (hypertension) compared to a negative case. In conclusion, the developed model, being one of the recent studies to examine the predictive power of a model, can be adopted for better precisions of the explanatory risk factor variables for hypertension and ultimately help to reduce its prevalence.