Cardiovascular diseases (CVD) in women and minority groups are not studied enough, not recognized early, not diagnosed well and not treated properly. CVD, including heart failure, heart attack and stroke, is a major global health problem and the leading cause of death among women. CVD killed nearly 18 million people in 2019, and it is estimated that this figure will rise to 22 million by 2030.
Given the insufficient research, recognition, diagnosis and treatment of women at risk of or with CVD, the aim of our project, CARDIA, is to address a gap in CVD prevention. We will develop a new AI algorithm to personalize CVD prevention interventions based on a variety of datasets including the UK Biobank dataset, using an ML methods such as naive Bayes, support vector machines, random forests and neural networks among others. The algorithm will be able to classify patients into different risk categories. Ensemble models, another machine learning method, will also be developed to improve predictive accuracy for capturing different risk factors (e.g., mental health symptoms) more prevalent in women. Additionally, explainable methods (SHAP, LIME, PIMP or Explainable Boosting Machine (EBM)) will be integrated to understand the reasoning behind the algorithm’s predictions. The outcome of our project will be a tool for primary care providers and patients to predict and prevent CVD.
Therefore, CARDIA could: i) address gender disparities in the existing literature, ii) provide technology for the prediction and prevention of CVD based on signs of CVD and iii) impact other prediction and prevention strategies to open new avenues for discovery across various fields.