Predicting Patient Readmission Risk
Executive Summary
The solution combines time-based ADT (Admission, Discharge, and Transfer) data with provider data and machine learning to predict patient readmission risk, reducing manual effort and enabling consistent, data-driven assessments. Its core analytical approach uses supervised learning with a neural-network model to assess readmission risk across multiple time horizons. In addition to predicting 90-day readmission risk, the solution provides insights for 30-day and 60-day periods, enabling a more nuanced, multi-horizon assessment rather than relying on a single binary outcome.
About Our Client
Client Name: Confidential
Industry: Healthcare
Location: USA
Technologies
Programming Language: Python
Framework: Keras / TensorFlow
Learning Type: Supervised Learning
Model: Neural Network
Data Format: CSV
