AUTHOR=Liu Bo , Ma Yijiang , Zhou Chunxiao , Wang Zhijie , Zhang Qiang 

TITLE=A novel predictive model of hospital stay for Total Knee Arthroplasty patients

JOURNAL=Frontiers in Surgery

VOLUME=Volume 9 - 2022

YEAR=2023

URL=https://www.frontiersin.org/journals/surgery/articles/10.3389/fsurg.2022.807467

DOI=10.3389/fsurg.2022.807467

ISSN=2296-875X

ABSTRACT=Abstract
Objective: 
The aim was to explore the main risk factors and develop a predictive nomogram of hospital stay in an undergoing Total knee arthroplasty (TKA) patient.
Methods: 
We used 2622 patients undergoing TKA in Singapore in a retrospective cohort study. Hospital extension was defined based on the 75% quartiles (Q3) of hospital stay. We randomly divided all patients into two groups in a 7:3 ratio: the training and validation groups. We performed the univariate analysis in the training group, in which variables with p-values < 0.05 were included and then subjected to multivariate analysis. The Multivariable logistic regression analysis was applied to build a predicting nomogram by using the variables that the p-values < 0.01. To evaluate the prediction ability of the model, we calculated the C-index.
Furthermore, the ROC, Calibration, and DCA curves were drawn to assess the model. Finally, we verified the accuracy of the model using the validation group. Similarly, the C-index. Furthermore, the ROC curve, the Calibration curve, and the DCA curve were applied to evaluate the model in the validation group.
Results: 
Finally, 2266 patients were included in this study. The 75% quartiles (Q3) of hospital stay was six days .457(20.17%) patients were recognized as hospital extensions. There were 1588 patients in the training group and 678 patients in the validation group. Age, Hb, D.M., Operation Duration, Procedure Description, Day of Operation, Repeat Operation, and Blood Transfusion were used to build the prediction model. The C-index was 0.680 (95% CI: 0.734 - 0.626) in the training group and 0.710 (95% CI: 0.742 - 0.678) for the validation set . The calibration curve and DCA indicated the hospital stay extension model's good performance in training and validation groups.
Conclusion:
To make better use of these factors and identify patients' risk factors early, the medical team can plan a patient's rehabilitation path as a whole. Its advantages lie in better resource allocation, maximizing medical resources, improving the functional recovery of patients, and reducing the overall cost of hospital stay and surgery. We also hope this study can help clinicians in the future.