Related Experiment Video
Updated: Nov 24, 2025
07:33
In Vitro Application of a Wireless Sensor in Flexion-Extension Gap Balance of Unicompartmental Knee Arthroplasty
Published on: May 5, 2023
909
Development of a Machine Learning Algorithm to Predict Nonroutine Discharge Following Unicompartmental Knee
Yining Lu1, Zain M Khazi2, Avinesh Agarwalla3
1Department of Orthopedic Surgery and Sports Medicine, Mayo Clinic, Rochester, MI.
The Journal of Arthroplasty
|December 28, 2020
Summary
Machine learning accurately predicts nonhome discharge after knee replacement surgery. This tool helps clinicians optimize patient care and reduce hospital stays by identifying at-risk individuals.
Area of Science:
- Orthopedic Surgery
- Health Informatics
- Machine Learning
Background:
- Optimizing discharge destination after unicompartmental knee arthroplasty (UKA) is crucial for patient outcomes and healthcare costs.
- Predicting nonhome discharge is essential for efficient resource allocation and patient management.
Purpose of the Study:
- To develop a machine learning algorithm for predicting nonhome discharge in patients undergoing UKA.
- To identify key factors influencing discharge destination after UKA.
Main Methods:
- Retrospective analysis of a national surgical outcomes database (2015-2019).
- Development and evaluation of five machine learning algorithms to predict nonroutine discharge.
- Performance assessment using discrimination, calibration, and decision curve analysis.
Main Results:
- Out of 7275 UKA patients, 3.6% had nonroutine discharge.
- Key predictors included: hospital stay, hematocrit, BMI, sodium, ASA classification, gender, and functional status.
- An extreme boosted model achieved an AUC of 0.875 and was integrated into a web application.
Conclusions:
- The developed model can aid clinician decision-making for elective UKA patients.
- Identifies patients needing early insurance authorization for nonmodifiable risks.
- Suggests prehabilitation for modifiable risk factors to improve discharge outcomes.
