KEYWORDS: machine learning, periprosthetic joint infections, revision total hip arthroplasty 

BACKGROUND: Periprosthetic joint infection (PJI) following total hip arthroplasty (THA) is associated with high morbidity
and a 25% 5 year mortality. 

OBJECTIVE: This is the first study to utilize machine learning (ML) robust oversampling techniques to
identify risk factors and predict postoperative mortality occurring during the same admission after revision THA for PJI.
These techniques potentially enable improved mortality risk prediction critical to identifying high-risk patients and
implementing preventative strategies.

METHODS: This study utilized the Nationwide Readmissions Database to identify adult
patients undergoing revision THA for PJI. ML classification models were developed to predict in-hospital mortality and
compared to logistic regression. Synthetic Minority Oversampling Technique (SMOTE) was applied to address data
imbalance by duplicating observations in the deceased cohort to improve classification accuracy. Model performance was
assessed using AUROC, Brier score, and F1 score, with SHapley Additive exPlanation (SHAP) analysis identifying key
predictors.

RESULTS: Among 19,099 patients undergoing revision THA for PJI, there was a 0.8% mortality rate (Figure 1). The
deceased cohort was older and had higher rates of congestive heart failure, coagulopathy, and renal failure.
Logistic regression (AUROC 0.855) and gradient boosting (AUROC 0.862) outperformed random forest (AUROC 0.765),
with gradient boosting demonstrating the highest F1 score (0.257). Following the application of SMOTE, all
models demonstrated improved AUROC, model calibration, and enhanced F1. SHAP analysis identified
fluid and electrolyte disorders, advancing age, and cardiac arrhythmia as key predictors of mortality. Elective admission had the greatest protective effect against mortality.

CONCLUSION: Fluid and electrolyte abnormalities were the strongest comorbid predictor of mortality. Given the established link between high-dose antibiotic-loaded bone cement spacers with AKI and subsequent electrolyte disturbances, optimizing electrolyte levels perioperatively may be a potential mortality-reducing intervention. Notably, elective admission was the strongest protective factor against mortality, likely due to its role in facilitating preoperative optimization.