AI-Driven Clinical Decision Support Systems Transforming Healthcare Delivery and Efficiency

Abstract

Cardiovascular disease (CVD) continues to be one of the leading causes of global mortality, responsible for an estimated 17.9 million deaths each year while exerting significant pressures on health systems. The key to reducing the burden of disease and improving patient outcomes is early accurate risk prediction. Conventional clinical assessment approaches often find it difficult to analyze multi-layered interplays of multiple biological, demographic, and lifestyle-related factors. This study introduces an AI-based Clinical Decision Support System (AI-CDSS) for cardiovascular disease risk prediction through the integration of machine learning (ML) models and explainable artificial intelligence (XAI). Patients and methods: The framework was generated using a clinical dataset of 70,000 patients with 18 predictive features. The dataset was preprocessed with missing value treatment, categorical encoding, feature selection, and normalization using standard scaling. Five ML algorithms (Logistic Regression (LR), Support Vector Machine (SVM), Random Forest (RF), Extreme Gradient Boosting (XG Boost), and Artificial Neural Network (ANN)) were assessed using an 80:20 train-test split with 5-fold cross-validation. Model performance accuracy, precision, recall, F1-score, and ROC-AUC. The experimental section showed that the highest values of predictive performance, with an accuracy of 94.2%, precision of 93.8%, recall of 94.5%, F1-score (F-Measure) of 94.1% and CRM or area under ROC curve (AUC) were obtained by the XG Boost model for cardiovascular prediction SHAP analysis identified age, systolic blood pressure, cholesterol level (total Chol<154mg/dL), BMI and smoking status as the top 5 most important predictive features. We present an AI-CDSS framework that uses a combination of machine learning and explainable AI methods to offer better prediction performance while maintaining clinical interpretability. The results pointed out that AI-based decision support systems improve early identification of cardiovascular risk, assist in personalized therapy planning, and facilitate healthcare delivery.