Electronic Health Record-Based Prediction Model to Predict Primary Aldosteronism Risk in People With Hypertension
Authors' conclusion
This electronic health record-based prediction model achieved moderate discriminative performance, identifying primary aldosteronism in 29% of patients within the highest-risk decile, suggesting potential clinical utility for prioritizing screening efforts.
Study context
- Population
- People with hypertension screened for primary aldosteronism (PA) at an academic health system (April 2018 to January 2025)
- Participants
- 4,564 people
- Comparison
- Discovery vs Validation
- Discoveryn = 3,650
- Validationn = 914
- Outcome
- Discrimination of the PA prediction model (AUC) in the validation cohort
Methodological detailsEHR-based, machine learning, logistic regression model with 15 variables, discovery and validation cohorts, single academic health system
- EHR-based
- Machine learning
- Logistic regression model with 15 variables
- Discovery and validation cohorts
- Single academic health system
Performance
As reported in the abstract
Model discrimination for PA
| Test or model | AUC |
|---|---|
| Validation | 0.7 |
Validation
AUC0.7
- PA in top 10% highest-risk subgroup
- 29%
- Enrichment in top 10%
- 2.7-fold
- Recall in top 10%
- 27%
- Precision in top 10%
- 29%
- Recall in top 20%
- 41%
- Specificity in top 20%
- 82%
Further findings
PA prevalence
- Had PA
- 456 (10.0%)
Characteristics of PA vs non-PA patients
- PA patients older (years)
- 61.8 vs 57.9, P < .0001
- PA patients more likely Black
- 22.9% vs 8.1%, P < .0001
- PA patients higher systolic BP (mmHg)
- 148 vs 144, P = .001
- PA patients required more antihypertensive medications
- 3.8 vs 3.0, P < .0001
Model predictors
- Race
- P < 0.0001
- Hypokalemia
- P < 0.0001
- ≥4 antihypertensive medications
- P < 0.0001
- Obstructive sleep apnea
- P = 0.004
- Higher serum bicarbonate
- P < 0.001
- Reduced GFR
- P < 0.001
Original publication
Electronic Health Record-Based Prediction Model to Predict Primary Aldosteronism Risk in People With Hypertension. Endocr Pract. Published 17 December 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →