Trends and Care Cascade Performance of Hypertension in Africa (1990–2019): A Dynamic Fractional Modelling Approach

Abstract

Hypertension is a major predictable contributor to cardiovascular morbidity and mortality in Africa, but the long-term trend analysis frequently uses descriptive or Gaussian-based procedures, ignoring the limited and autocorrelated character of proportionate data. The study aimed to conduct a rigorous longitudinal evaluation of age-standardised and crude hypertension prevalence and care cascade indicators between 1990 and 2019, and to address statistical shortcomings in previous research. In particular, the goals were to measure time changes, analyse sex differentials and assess the adequacy of the model with distributionally suitable measures. A dynamic model was used to analyse the results of a fractional time series, employing generalised least squares with an AR(1) correction and beta regression. Findings indicated a small but significant decrease in age-standardised prevalence among men (GLS 9 = -0.0050, p = 0.0405; Beta 9 = -0.0035, p = 0.003), and variability in women (interaction 9 = 0.0054, p = 0.001). There was a steady increase in diagnosis and treatment percentages, but the control improvement was inadequate and untreated extreme hypertension prevailed. The comparison between models showed that the beta regression (AIC = -341.51; pseudo R2 = 0.56) fit complexions better in comparison with the Gaussian methods. This research enhances the inferential validity and offers actionable evidence that supports the reinforcement of primary care systems, sex-sensitive policies, and long-term investment in the realisation of Sustainable Development Goal target 3.4 because it introduces temporal dependence and fractional modelling into a single framework.