Yusuf Yesil, MD, MIS 
Up to 20 percent of people with high blood pressure actually suffer from primary aldosteronism. Yet clinicians rarely screen for this hormonal condition. Patients spend years cycling through ineffective medications while their cardiovascular risk escalates.
The bottleneck is not a lack of treatments. It is a failure of detection.
The Data-Mining Solution
An AI model trained on 30 years of routine electronic health records offers a way out of this diagnostic blind spot. Using an XGBoost architecture, the tool analyzed data from over 225,000 adults and successfully flagged over 90 percent of cases.
This is not about replacing clinical judgment. It is about automating the search for subtle patterns that human doctors easily miss in a standard 15-minute consultation.
By scanning existing patient histories, the algorithm highlights who needs specialized hormone testing. This bypasses the need for expensive new diagnostic infrastructure.
The Shift to Precision
The clinical implications are massive. Instead of treating hypertension as a generic symptom, clinicians can target the root endocrine cause. This matches a broader shift in cardiovascular care, where tools like HyperScore are mapping multi-organ damage from high blood pressure.
However, integrating these models into busy hospital workflows remains a major hurdle. Algorithms only work if doctors actually receive, understand, and trust the alerts. If health systems cannot bridge the gap between software prediction and clinical action, these high-performing models will simply sit silent in the background.
