Abstract

Coronary artery disease (CAD) remains the leading cause of morbidity and mortality globally. While traditional 
risk prediction models rely on clinical, biochemical, and demographic factors, they often omit the contribution of 
genetic variation and social determinants of health (SDOH). Advances in whole-genome sequencing (WGS) have 
enabled population-scale assessment of polygenic risk, while SDOH capture environmental and socio-economic 
influences on disease development. This study integrates WGS-derived polygenic hazard scores with SDOH data 
to improve CAD risk prediction, assess model interpretability, and evaluate real-world performance across diverse 
populations. Using data from the UK Biobank and independent cohorts, models combining genomic and social risk 
factors demonstrated superior predictive performance and improved calibration compared with models using 
either data type alone. However, disparities in predictive accuracy persist across populations, highlighting 
challenges in equity and access. Implementation considerations, including infrastructure, governance, patient and 
clinician engagement, and ethical frameworks, are critical for translating these integrative approaches into clinical 
practice. Our findings underscore the potential of integrated socio-genomic models to enhance precision medicine 
while emphasizing the need for careful attention to fairness, transparency, and real-world applicability.