Main Article Content
Abstract
The integration of Artificial Intelligence (AI) into healthcare, particularly for predictive analytics, offers transformative potential in shifting from reactive to proactive care models. However, implementation is challenged by fragmented healthcare data, interoperability gaps, data quality issues, and limited clinical integration. Current healthcare systems are predominantly reactive, focusing on acute interventions rather than prevention, which limits effective chronic disease management. AI-powered predictive analytics can address this by analyzing historical and real-time data to forecast disease progression, identify at-risk patients, and enable timely interventions reducing costs, emergency visits, and hospital readmissions. Yet, barriers such as inadequate infrastructure, clinician resistance, and concerns over trust and explainability persist. The proposed system adopts a modular, service-oriented architecture supporting hybrid deployment, integrating data ingestion, preprocessing, feature engineering, model training, and prediction services with clinician and administrator dashboards. Functional requirements include patient risk prediction and operational forecasting, while non-functional requirements emphasize performance, scalability, reliability, and compliance. Stakeholder analysis identifies clinicians, administrators, and other healthcare actors, each with specific needs for usability, integration, and transparency.