Christian Klemt1, Venkatsaiakhil Tirumala1, Yasamin Habibi1, Anirudh Buddhiraju1, Tony Lin-Wei Chen1, Young-Min Kwon2. 1. Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA. 2. Bioengineering Laboratory, Department of Orthopaedic Surgery, Massachusetts General Hospital, Harvard Medical School, 55 Fruit Street, Boston, MA, 02114, USA. ymkwon@mgh.harvard.edu.
Abstract
BACKGROUND: A reliable predictive tool to predict unplanned readmissions has the potential to lower readmission rates through targeted pre-operative counseling and intervention with respect to modifiable risk factors. This study aimed to develop and internally validate machine learning models for the prediction of 90-day unplanned readmissions following total knee arthroplasty. METHODS: A total of 10,021 consecutive patients underwent total knee arthroplasty. Patient charts were manually reviewed to identify patient demographics and surgical variables that may be associated with 90-day unplanned hospital readmissions. Four machine learning algorithms (artificial neural networks, support vector machine, k-nearest neighbor, and elastic-net penalized logistic regression) were developed to predict 90-day unplanned readmissions following total knee arthroplasty and these models were evaluated using ROC AUC statistics as well as calibration and decision curve analysis. RESULTS: Within the study cohort, 644 patients (6.4%) were readmitted within 90 days. The factors most significantly associated with 90-day unplanned hospital readmissions included drug abuse, surgical operative time, and American Society of Anaesthesiologist Physical Status (ASA) score. The machine learning models all achieved excellent performance across discrimination (AUC > 0.82), calibration, and decision curve analysis. CONCLUSION: This study developed four machine learning models for the prediction of 90-day unplanned hospital readmissions in patients following total knee arthroplasty. The strongest predictors for unplanned hospital readmissions were drug abuse, surgical operative time, and ASA score. The study findings show excellent model performance across all four models, highlighting the potential of these models for the identification of high-risk patients prior to surgery for whom coordinated care efforts may decrease the risk of subsequent hospital readmission. LEVEL OF EVIDENCE: Level III, case-control retrospective analysis.
BACKGROUND: A reliable predictive tool to predict unplanned readmissions has the potential to lower readmission rates through targeted pre-operative counseling and intervention with respect to modifiable risk factors. This study aimed to develop and internally validate machine learning models for the prediction of 90-day unplanned readmissions following total knee arthroplasty. METHODS: A total of 10,021 consecutive patients underwent total knee arthroplasty. Patient charts were manually reviewed to identify patient demographics and surgical variables that may be associated with 90-day unplanned hospital readmissions. Four machine learning algorithms (artificial neural networks, support vector machine, k-nearest neighbor, and elastic-net penalized logistic regression) were developed to predict 90-day unplanned readmissions following total knee arthroplasty and these models were evaluated using ROC AUC statistics as well as calibration and decision curve analysis. RESULTS: Within the study cohort, 644 patients (6.4%) were readmitted within 90 days. The factors most significantly associated with 90-day unplanned hospital readmissions included drug abuse, surgical operative time, and American Society of Anaesthesiologist Physical Status (ASA) score. The machine learning models all achieved excellent performance across discrimination (AUC > 0.82), calibration, and decision curve analysis. CONCLUSION: This study developed four machine learning models for the prediction of 90-day unplanned hospital readmissions in patients following total knee arthroplasty. The strongest predictors for unplanned hospital readmissions were drug abuse, surgical operative time, and ASA score. The study findings show excellent model performance across all four models, highlighting the potential of these models for the identification of high-risk patients prior to surgery for whom coordinated care efforts may decrease the risk of subsequent hospital readmission. LEVEL OF EVIDENCE: Level III, case-control retrospective analysis.
Authors: Richard Iorio; Andrew J Clair; Ifeoma A Inneh; James D Slover; Joseph A Bosco; Joseph D Zuckerman Journal: J Arthroplasty Date: 2015-09-09 Impact factor: 4.757
Authors: Heather S Haeberle; James M Helm; Sergio M Navarro; Jaret M Karnuta; Jonathan L Schaffer; John J Callaghan; Michael A Mont; Atul F Kamath; Viktor E Krebs; Prem N Ramkumar Journal: J Arthroplasty Date: 2019-06-11 Impact factor: 4.757
Authors: Daniel E Goltz; Sean P Ryan; Thomas J Hopkins; Claire B Howell; David E Attarian; Michael P Bolognesi; Thorsten M Seyler Journal: J Bone Joint Surg Am Date: 2019-03-20 Impact factor: 5.284
Authors: Adam I Edelstein; Mary J Kwasny; Linda I Suleiman; Rishi H Khakhkhar; Michael A Moore; Matthew D Beal; David W Manning Journal: J Arthroplasty Date: 2015-05-27 Impact factor: 4.757
Authors: Daniel E Goltz; Billy T Baumgartner; Cary S Politzer; Marcus DiLallo; Michael P Bolognesi; Thorsten M Seyler Journal: J Arthroplasty Date: 2017-08-18 Impact factor: 4.757
Authors: Prem N Ramkumar; George F Muschler; Kurt P Spindler; Joshua D Harris; Patrick C McCulloch; Michael A Mont Journal: J Arthroplasty Date: 2016-11-17 Impact factor: 4.757