MODELLING OF SOUTH AFRICAN CHILD HYPERTENSION: COMPARATIVE ANALYSIS OF THE FREQUENTIST QUANTILE REGRESSION APPROACHES

Abstract Background High blood pressure in children and adolescents raises the risk of hypertension in adults. This condition is referred to as child hypertension, or paediatric hypertension. Methods The frequentist quantile regression (FQR) approach is being used in this study to ascertain the risk factors of childhood hypertension in South Africa (SA). Data from the 2017–2018 Household Survey, Wave 5 of the South African National Income Dynamics Study (NIDS), was used in a retrospective analysis. A final data set of 1 812 was utilized for data analysis. An application of both the simplex and the Frisch-Newton frequentist quantile regression approaches were used to explore the risk factors of child hypertension in SA. Results In South African adolescent children, the prevalence of high blood pressure was 2.6% accountable to diastolic blood pressure (DBP) and 1.6% attributable to systolic blood pressure (SBP). By using only upper quantiles (\(\:\tau\:\in\:\left\{\text{0.75,0.95}\right\})\), body mass index, gender, pulse rate, exercise and cigarette consumption were found to be significant determinants of paediatric hypertension in SA. Conclusion The results of this investigation indicate that, according to the mean absolute error (MAE) and the root mean square error (RMSE), the BP models fitted using both the simplex and the Frisch-Newton frequentist quantile regression approaches revealed no difference between them, therefore, confirming the computational robustness of these frequentist quantile regression approaches.