READMISSION

Hospital readmissions create challenges for healthcare systems, increasing costs and putting pressure on resources. This project introduces a machine learning-based system designed to predict patient readmissions, helping medical personnel and hospitals take early action to improve patient care and manage resources more effectively. By analyzing electronic health record (EHR) data, the model assesses a patient’s risk and provides explanations for key factors influencing the prediction. The system was trained on a dataset containing patient details such as age, medical history, lab results, and past hospital visits. It was developed using Python for machine learning, Express.js for the backend, and TypeScript with React for the frontend, ensuring smooth data processing and an easy-to-use interface. Strong security features like authentication, encryption, and error handling were added to protect patient information. The result shows that the model was able to achieve 63.87% accuracy, with recall scores of 72% and 55% in different areas. These results highlight the model’s ability to predict readmissions while also showing areas where improvements can be madethrough better data processing and tuning. By using predictive analytics, this system helps healthcare professionals make informed decisions, reduce avoidable readmissions, and improve hospital efficiency. This project demonstrates how AI-powered solutions can transform healthcare by enabling proactive patient management and better decision-making.