Predicting ICU Readmissions

I developed a predictive healthcare model to identify patients at high risk of 30-day ICU readmission using electronic health record (EHR) data. By applying Python-based machine learning techniques, particularly XGBoost, the model analyzed patient history, vitals, and discharge conditions to generate risk scores. It achieved an accuracy of 84%, enabling clinical teams to take proactive measures within a 48-hour intervention window. As a result, ICU readmission rates were reduced by 23% in the first quarter after implementation.