Prevalence and determinants of hypertension among adults of reproductive age in Tanzania: analysis of a cross-sectional Demographic and Health Survey

Abstract

Objective: We aimed to estimate prevalence and identify determinants of hypertension in adults aged 15-49 years in Tanzania.

Design: We analysed cross-sectional survey data from the 2022 Tanzania Demographic and Health Survey and Malaria Indicator Survey conducted between February and July 2022. Descriptive statistical analysis, logistic regression, machine learning and geospatial methods were used to estimate prevalence and determine determinants of hypertension.

Setting: Tanzania.

Participants: A total of 13 385 participants aged 15-49 years were included in the analysis.

Primary outcome: The primary outcome variable was hypertension, defined as either systolic blood pressure (BP)≥140 and/or diastolic BP≥90 mm Hg or under anti-hypertensive drugs.

Results: The prevalence of hypertension among adults of reproductive age was 11% (95% CIs 10.09 to 11.56) in Tanzania, varying significantly across risk factors. Prevalence was high in people aged 40-49 (22.11%, 95% CI 20.07 to 24.29) and obese (23.69%, 95% CI 20.67 to 27.00). The mean prevalence of hypertension was also high in the southern, eastern, western, southern highlands, north-west and north-eastern part of the country, correlating with the spatial distribution of older age (30-49) and higher body mass index (BMI) (≥25). Individuals aged 40-49 had nearly six times (adjusted OR (AOR): 5.68, 95% CI 4.10 to 7.83) the odds of hypertension relative to those aged 15-19. Obese individuals had higher odds (AOR: 2.88, 95% CI 2.01 to 4.13) compared with overweight individuals (AOR: 1.93, 95% CI 1.36 to 2.74). Machine learning results showed age and BMI as the most important determinants of hypertension and that significant interactions between risk factors exist.

Conclusion: The prevalence of hypertension varied across risk factors and the strongest determinants of hypertension in adults of reproductive age were age and BMI.

Keywords: Demography; Geographical mapping; Hypertension; Machine Learning; Prevalence.

Figure 1. Spatial distribution of observed prevalence of hypertension in Tanzania, 2022. Each circle represents a sampled location (cluster).

Figure 2. Variable importance plot of risk factors by their categories.

Figure 3. SHAP summary plots of risk factor importance. (A) Ranking of risk factors based on mean absolute Shapley values, and (B) SHAP value and direction of influence. SHAP, SHapley Additive exPlanation.

Figure 4. Dependency plot to show interaction between risk factors. Each risk factor is coloured by the values of a strongly interacting risk factor. SHAP, SHapley Additive exPlanation.

Figure 5. Continuous surface maps of (A) estimated hypertension prevalence, (B) upper and (C) lower confidence intervals, (D) proportion of individuals aged 30–49 years and (E) proportion of overweight and obese individuals.