Development and Internal Validation of a Multivariable Clinical Risk Score for Adult Hypertension among Non-Pregnant Women: A Biomarker Analysis of the 2024/25 EDHS from Ethiopia

Abstract Background Sub-Saharan Africa faces a growing hypertension epidemic. In Ethiopia, limited clinical infrastructure leaves most cases undiagnosed as standard risk models require unavailable laboratory profiles. This study developed, internally validated and translated a non-laboratory clinical risk score for hypertension among non-pregnant Ethiopian women using 2024/25 Ethiopian Demographic and Health Survey biomarker data. Methods Analyzing data from 12,850 non-pregnant women aged 15 to 49 years (80% derivation; 20% validation cohorts) hypertension was defined by elevated blood pressure or medication use. Multivariable logistic regression identified predictors converted into a point-based scoring system via Framingham scaling. Model performance followed TRIPOD guidelines using 1000 bootstrap resamples assessing discrimination via the area under the receiver operating characteristic curve, calibration via the Brier score and net benefit via Decision Curve Analysis. Results Weighted hypertension prevalence was 15.8%. Seven independent predictors were retained: age, body mass index, urban residence, household wealth, tobacco use, high parity and family history of non-communicable diseases (score range: 0 to 18). The score demonstrated excellent discrimination in both derivation (area under the receiver operating characteristic curve: 0.824) and validation cohorts (area under the receiver operating characteristic curve: 0.818) alongside strong calibration (Brier score: 0.091). At a 6-point cutoff sensitivity was 81.4% and specificity was 78.2%. Decision Curve Analysis confirmed net clinical benefit. Conclusions This validated non-laboratory clinical risk score provides a practical screening tool for Health Extension Workers and primary care nurses to effectively triage high-risk women and reduce Ethiopia's cardiovascular burden.