AI-Augmented Predictive Analytics Model for Early Risk Stratification in Cardiovascular Disorders

Abstract

Cardiovascular diseases (CVDs) are the number one cause of death all over the world, which underscores the necessity of proper and early risk stratification to initiate clinical intervention. Conventional risk assessment instruments, including the Framingham Risk Score and ASCVD Risk Estimator, utilize a small number of clinical variables and linear assumptions, which could decrease the predictive capability in complicated patient groups. The purpose of the research is to create an AI-enhanced predictive analytics framework of early risk stratification of cardiovascular diseases to enhance the prediction accuracy and aid clinical decision-making. A structured clinical dataset made up of demographic and physiological characteristics was implemented in a supervised machine learning framework using a structured clinical dataset. Preprocessing of the data involved processing of missing values, normalization and feature selection. The data was separated into training and testing data and the Random Forest and Gradient Boosting models were trained and optimized. Core predictive measures were used to assess model performance, with the following served as core measures Areas Under the Receiver Operating Characteristic Curve (AUC-ROC), accuracy, precision, recall (sensitivity), specificity, and F1 score. The proposed model had a good predictive performance with AUC-ROC of 0.91, accuracy of 88.3, sensitivity of 90.2, specificity of 85.6, and an F1 score of 0.89, better than the traditional models. The AI predictive model has a great advancement in the early cardiovascular risk stratification, which is more accurate and with better clinical utility. The given approach can help to facilitate proactive healthcare intervention and maximize patient outcomes.