Related Experiment Videos
Predicting Non-Home Discharge Following Primary Total Hip Arthroplasty Using an Artificial Neural Network to Identify
Campbell Dopke1, Theodore Quan1, Taylor Raffa1
1Department of Orthopaedic Surgery, George Washington University School of Medicine and Health Sciences, Washington, DC, USA.
The Iowa Orthopaedic Journal
|July 23, 2026
Summary
An artificial neural network model identified key factors predicting non-home discharge after total hip arthroplasty (THA). Age, operative time, hematocrit, functional status, and INR are crucial for optimizing patient care and reducing non-home discharges.
Area of Science:
- Orthopedic Surgery
- Health Services Research
- Artificial Intelligence in Medicine
Background:
- Total hip arthroplasty (THA) incidence is rising, increasing focus on cost containment and early postoperative discharge.
- Value-based care models necessitate understanding factors influencing discharge disposition to optimize patient outcomes and resource allocation.
- Identifying predictors of non-home discharge is crucial for developing targeted interventions and improving postoperative care pathways.
Purpose of the Study:
- To develop and apply an artificial neural network (ANN) model to identify key factors influencing discharge destination after primary THA.
- To determine both non-modifiable and modifiable patient characteristics associated with discharge to a non-home setting.
- To provide data-driven insights for preoperative patient optimization and reduce non-home discharges.
Main Methods:
- Utilized the National Surgical Quality Improvement Program (NSQIP) database for primary THA cases between 2016-2019.
- Analyzed demographic, comorbidity, preoperative, and intraoperative variables, inputting statistically significant factors (p<0.05) into an ANN model.
- Employed an Artificial Neural Network (ANN) model to predict discharge disposition with a focus on identifying significant predictive variables.
Main Results:
- Analyzed 124,691 primary THA patients; 15.5% were discharged to a non-home destination.
- The ANN model achieved a high predictive accuracy with a Receiver Operating Characteristic (ROC) Area Under the Curve (AUC) of 0.793.
- The top five predictors for non-home discharge were: patient age, operative time, preoperative hematocrit, functional status, and preoperative international normalized ratio (INR).
Conclusions:
- An ANN model effectively identified significant factors predicting non-home discharge following primary THA.
- Key predictive variables include age, operative time, preoperative hematocrit, functional status, and INR.
- These findings can guide clinicians in preoperative patient assessment and optimization to potentially reduce non-home discharges.