Uses and opportunities for machine learning in hypertension research

Area of Science:

• Cardiology
• Medical Informatics
• Artificial Intelligence

Background:

• Artificial intelligence (AI) offers valuable insights for hypertension specialists.
• Significant AI applications are emerging, validated on extensive datasets.

Purpose of the Study:

• To review AI applications for predicting clinical outcomes in hypertension using big data.
• To showcase AI's role in identifying cardiovascular events and hypertension risk.

Main Methods:

• Review of four distinct AI applications utilizing large datasets.
• Methods include deep learning, support vector machines (SVM), neural networks, and machine learning algorithms.
• Wearable biosensors and photoplethysmography were employed for risk assessment.

Main Results:

• AI models predicted cardiovascular events with 56%-71% accuracy and 68%-71% sensitivity/specificity.
• Machine learning achieved 51% sensitivity and 99% specificity in hypertension classification (AUC 87%).
• Wearable devices identified hypertension risk with >80% sensitivity and >90% positive predictive value.

Conclusions:

• AI methods are being actively applied within the field of hypertension.
• These examples highlight AI's capability in predicting clinical outcomes and assessing risk.
• AI demonstrates significant potential for improving hypertension management and patient care.