Abstract

Hypertension “also referred to as high blood pressure” is a chronic medical condition that remains a critical public health concern. Recent studies have indicated that hypertension weakens the body’s immunity, making it susceptible to others dreadful diseases, such as coronary heart diseases, cardiovascular diseases, stroke, diabetes and the recent COVID 19 outbreak. Globally, efforts have been made to mitigate the frequent succumbing of patients to hypertension, however the occurrences of this disease continue to increase. In Namibia, a statistical tool was developed to identify potential factors associated with blood pressure. Numerous studies used a variety class of regression methods, including (Linear models (LM), Generalized linear model (GLM), Generalized additive models (GAM) and Mixed effects to predict high blood pressure. Some of these methods do not reflect the true values as they are not flexible enough, lack stability, do not conform to their assumptions, weak interpretability and restricted to model various distributions. The aim of this study was to apply alternative class of regression model Generalized additive models for location scale and shape (GAMLSS) model, to analyze the distributive occurrences of high blood pressure in Namibia. The GAMLSS model offers a flexible approach as it allows flexible modelling of the mean, variance, skewness and kurtosis. The results show that for the median parameter, high blood pressure increased with age, BMI, glucose (SBP only) and haemoglobin (DBP only) levels. In relation to the latter finding, the frequency in high blood pressure patients was higher among middle class male adults residing in urban areas, with no level of education, and with a higher alcohol consumption rate. On the other hand, the findings presented a negative relationship between high blood pressure and weight, as well as height, the BMI and DBP observed in the scale parameter. With the shape parameter, high blood pressure was associated no attainment of education (SBP), increased BMI and alcohol intake (DBP). The centile estimation indicated that 25% of males aged 42 – 64 and 25% of females aged 47 – 64 years were hypertensive. Moreover, 10 % of males aged 35–41years and females aged 35- 46 years were hypertensive. The study showed that the GAMLSS was the ideal model, which responded to variables from a wide range of distribution, comparing in comparison to other statistical models. The study further illustrated that the GAMLSS was more flexible than other regression models.

More details

School	: School of Natural and Applied Sciences
Issued Date	: 2021

Download full document