Opportunities and Challenges of Cardiovascular Disease Risk Prediction for Primary Prevention Using Machine Learning...

Area of Science:

• Cardiovascular Medicine
• Artificial Intelligence
• Health Informatics

Background:

• Cardiovascular disease (CVD) is a leading global cause of death.
• Machine learning (ML) offers superior predictive performance for CVD risk stratification compared to traditional methods.
• Integrating ML with Electronic Health Records (EHRs) allows for refined risk prediction using detailed patient data.

Purpose of the Study:

• To systematically review the literature on ML-driven models for long-term CVD risk prediction in primary prevention using EHR data.
• To evaluate the effectiveness and identify challenges associated with ML models in this context.

Main Methods:

• Systematic literature search in Medline and Embase (March 2024) for studies since 2010.
• Inclusion of systematic and narrative reviews evaluating ML models for CVD risk prediction in primary prevention using EHRs.
• Exclusion of studies on short-term prognostication or non-ML models.

Main Results:

• 22 studies met inclusion criteria; ML models demonstrate superiority in utilizing complex EHR data for precision cardiovascular risk assessment.
• Significant challenges include heterogeneous CVD outcome definitions, data incompleteness, and lack of external validation.
• Ethical and regulatory hurdles such as algorithmic opacity and equity concerns impede clinical integration.

Conclusions:

• ML-based CVD risk prediction holds transformative potential but faces methodological, technical, and regulatory barriers.
• Standardized validation, regulatory oversight, and interdisciplinary collaboration are crucial for clinical translation.
• A proposed standardized, transparent, and regulated EHR platform aims to improve model evaluation, reproducibility, and equitable implementation.