DEVELOPMENT OF A CARDIOVASCULAR DISEASE DIAGNOSIS AND MANAGEMENT SYSTEM USING PREDICTIVE MODELLING

Early diagnosis of cardiovascular disease (CVD) is crucial in reducing death and optimizing patient outcome, necessitating advanced diagnostic facilities. Delayed diagnosis leads to extremely complex situations, increasing the cost of health care and reducing quality of life. This project builds a model for CVD diagnosis and control using machine learning models—Logistic Regression, Decision Trees, and Random Forest—to optimize accuracy in diagnosis. By analyzing patient data like clinical history, symptoms, and test results like electrocardiograms (ECGs) and blood markers, the system identifies patterns that are indicative of CVD like coronary artery disease and heart failure. The system applies predictive modeling to risk stratification to enable early interventions and individualized treatment plans. Models are trained on diverse datasets across institutions with rigorous data preprocessing like normalization and outlier removal to maintain data quality. Feature selection and cross-validation are employed to enhance model generalizability, while supervised learning algorithms with hyperparameters tuned enable optimal performance on validation sets. Early findings prove that machine learning models are better than traditional approaches. The Random Forest model for heart failure was predicted with an AUC-ROC value of 0.92 (compared to 0.80 with traditional methods), and Logistic Regression achieved 95% sensitivity and 88% specificity for coronary artery disease diagnosis. These improvements enable earlier, more accurate diagnosis and saving lives. Combining these technologies in healthcare practice would transform CVD management by offering real-time risk assessment tools and data-driven insights, enhancing active, personalized treatment. Overall, this project completes the requisite gap in improved CVD diagnosis by offering a scalable solution that can reduce mortality, enhance patient quality of life, and optimize healthcare resource utilization. Real-world validation and integration within electronic health records should be the subject of future research.