Hospital Readmission Risk Prediction and Care-Management Targeting

The nurses can make 12 follow-up calls a day and 60 patients go home. Who gets the call, what is that worth compared with calling at random, and is the list fair by age, race and gender? This project turns a 30-day readmission risk model into a daily call list with a reason and a talking point for each patient, a capacity planner (what one more nurse is worth) and a fairness check of who is called. Prediction is the means, not the headline: an adapted LACE index and logistic regression are compared with LightGBM on the public UCI Diabetes 130-US Hospitals data, every split by patient. Capacity, call effect and costs are stated assumptions; the project is a method demonstration, not a clinical decision tool.

Category: Healthcare Management

Master of Business Administration Project MBA06 from ProTechSC Academy, with graphical abstract, video demonstration and project guidance.