This tool leverages longitudinal data from the Study of Women’s Health Across the Nation (SWAN) to deliver personalized 10-year cardiovascular risk estimates for midlife women. It integrates traditional risk factors—blood pressure, lipid profiles, smoking status, diabetes—with female-specific variables such as adverse pregnancy outcomes, menopausal status, and circulating sex hormone levels. A machine learning pipeline applies algorithmic feature selection and supervised model training, with cross-validation and benchmarking against established calculators to ensure robust calibration and discrimination. The prototype is deployed as an interactive R Shiny application, allowing users to input clinical and reproductive history details and receive individualized risk predictions generated by the trained model.
Description
This approach stands apart by addressing the systematic underestimation of cardiovascular risk in women inherent in existing tools. By incorporating reproductive health data alongside conventional metrics, the model captures risk signals unique to midlife women. Its machine learning–driven variable selection optimizes predictive power, and comparative performance testing demonstrates clear improvements over standard calculators. The user-friendly R Shiny interface bridges complex analytics and patient engagement, enabling accessible, evidence-based personalized risk assessment supported by NIH-funded SWAN cohort insights.
Applications
- Women's health clinic risk assessments
- Telehealth cardiovascular screenings
- Insurer risk stratification tools
- Employer wellness program integration
- Digital health app plugin
Advantages
- Enhanced predictive accuracy for midlife women through integration of female‐specific factors (adverse pregnancy outcomes, menopause status, sex hormone levels) alongside traditional CVD risk factors
- Personalized 10-year cardiovascular risk estimates generated via machine learning–optimized feature selection and supervised model training
- Improved calibration and discrimination compared to existing risk calculators, reducing underestimation of women’s CVD risk
- Interactive R Shiny interface for easy, real-time input of clinical and reproductive history and instant risk feedback
- Data-driven validation using longitudinal cohort data (SWAN), ensuring robustness and clinical relevance
- Supports early identification of high-risk individuals and informed, tailored prevention strategies
IP Status
Research Tool
