Leveraging artificial intelligence for cardiovascular risk: a primary care perspective.

Area of Science:

• Cardiology
• Artificial Intelligence
• Deep Learning

Background:

• Cardiovascular diseases (CVDs) are a leading global cause of death.
• Primary care physicians are crucial for early CVD detection and prevention.
• This study explores AI-driven cardiovascular risk (CVR) stratification in primary care.

Purpose of the Study:

• Investigate the efficacy of AI-based temporal deep learning (DL) models for CVR stratification.
• Support early detection and prevention of cardiovascular diseases in primary care settings.
• Enhance patient management and resource allocation through AI integration.

Main Methods:

• Implemented Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) temporal DL architectures.
• Utilized a synthetic patient cohort with demographic and clinical data (age, sex, BMI, BP, smoking, diabetes, hypertension).
• Trained models on sequential data to predict CVR categories and probabilities, benchmarking outputs against clinical recommendations.

Main Results:

• LSTM and GRU models successfully forecasted CVR across various time horizons.
• Model predictions were translated into clinically interpretable recommendations for follow-up and interventions.
• Demonstrated the capability of temporal DL to predict cardiovascular risk with actionable insights.

Conclusions:

• AI-driven CVR forecasting can enhance early intervention strategies in family medicine.
• Optimized patient management and resource allocation are potential benefits of integrating AI.
• Temporal DL shows promise for strengthening preventive care outcomes in primary care.