An Analytical Framework for TJR Readmission Prediction and Cost-Effective Intervention

Area of Science:

• Health Economics
• Medical Informatics
• Machine Learning in Healthcare

Background:

• Hospital readmissions after total joint replacement (TJR) pose significant financial and clinical burdens.
• Predicting TJR readmission risk is challenging due to data complexities and class imbalance.

Purpose of the Study:

• To develop an analytical framework for assessing the cost-effectiveness of interventions aimed at reducing TJR readmissions.
• To create a machine learning model for predicting individual TJR patient readmission risk within 90 days post-discharge.

Main Methods:

• Utilized data sampling and boosting techniques to address class imbalance in readmission prediction.
• Developed an ensemble of models for robust risk prediction.
• Integrated a cost analysis framework considering misclassification costs for decision support.

Main Results:

• Successfully developed a predictive model for TJR readmission risk.
• Demonstrated a cost-effective intervention selection framework through a community hospital case study.
• The framework effectively identifies strategies to reduce readmissions.

Conclusions:

• The proposed analytical framework provides a data-driven approach to manage TJR readmissions.
• Machine learning and cost-effectiveness analysis can optimize intervention strategies for healthcare cost reduction.
• This approach supports informed decision-making for hospitals aiming to lower readmission rates.