Abstract
Background As screening programs in low- and middle-income countries (LMICs) often do not have the resources to screen the entire population, there is frequently a need to target such efforts to easily identifiable priority groups. This study aimed to determine (1) how hypertension prevalence in LMICs varies by age, sex, body mass index, and smoking status, and (2) the ability of different combinations of these variables to accurately predict hypertension. Methods and Results We analyzed individual-level, nationally representative data from 1 170 629 participants in 56 LMICs, of whom 220 636 (18.8%) had hypertension. Hypertension was defined as systolic blood pressure ≥140 mm Hg, diastolic blood pressure ≥90 mm Hg, or reporting to be taking blood pressure-lowering medication. The shape of the positive association of hypertension with age and body mass index varied across world regions. We used logistic regression and random forest models to compute the area under the receiver operating characteristic curve in each country for different combinations of age, body mass index, sex, and smoking status. The area under the receiver operating characteristic curve for the model with all 4 predictors ranged from 0.64 to 0.85 between countries, with a country-level mean of 0.76 across LMICs globally. The mean absolute increase in the area under the receiver operating characteristic curve from the model including only age to the model including all 4 predictors was 0.05. Conclusions Adding body mass index, sex, and smoking status to age led to only a minor increase in the ability to distinguish between adults with and without hypertension compared with using age alone. Hypertension screening programs in LMICs could use age as the primary variable to target their efforts.