Dataset Information

Machine Learning Model Developed to Aid in Patient Selection for Outpatient Total Joint Arthroplasty.

ABSTRACT:

Background

Patient selection for outpatient total joint arthroplasty (TJA) is important for optimizing patient outcomes. This study develops machine learning models that may aid in patient selection for outpatient TJA based on medical comorbidities and demographic factors.

Methods

This study queried elective total knee arthroplasty (TKA) and total hip arthroplasty (THA) cases during 2010-2018 in the American College of Surgeons National Surgical Quality Improvement Program. Artificial neural network models predicted same-day discharge and length of stay (LOS) fewer than 2 days (short LOS). Multiple linear and logistic regression analyses were used to identify variables significantly associated with predicted outcomes.

Results

A total of 284,731 TKA cases and 153,053 THA cases met inclusion criteria. For TKA, prediction of short LOS had an area under the receiver operating characteristic curve (AUC) of 0.767 and accuracy of 84.1%; prediction of same-day discharge had an AUC of 0.802 and accuracy of 89.2%. For THA, prediction of short LOS had an AUC of 0.757 and accuracy of 70.6%; prediction of same-day discharge had an AUC of 0.814 and accuracy of 78.8%.

Conclusion

This study developed machine learning models for aiding patient selection for outpatient TJA, through accurately predicting short LOS or outpatient vs inpatient cases. As outpatient TJA expands, it will be important to optimize preoperative patient selection and effectively screen surgical candidates from a broader patient population. Incorporating models such as these into electronic medical records could aid in decision-making and resource planning in real time.

SUBMITTER: Lopez CD

PROVIDER: S-EPMC8666332 | biostudies-literature | 2022 Feb

REPOSITORIES: biostudies-literature

ACCESS DATA

Publications

Machine Learning Model Developed to Aid in Patient Selection for Outpatient Total Joint Arthroplasty.

Lopez Cesar D CD Ding Jessica J Trofa David P DP Cooper H John HJ Geller Jeffrey A JA Hickernell Thomas R TR

Arthroplasty today 20211208

<h4>Background</h4>Patient selection for outpatient total joint arthroplasty (TJA) is important for optimizing patient outcomes. This study develops machine learning models that may aid in patient selection for outpatient TJA based on medical comorbidities and demographic factors.<h4>Methods</h4>This study queried elective total knee arthroplasty (TKA) and total hip arthroplasty (THA) cases during 2010-2018 in the American College of Surgeons National Surgical Quality Improvement Program. Artifi ...[more]

PMID: 34917716

Similar Datasets

Project description:Venous thromboembolism (VTE) and major bleeding (MBE) are feared complications that are influenced by numerous host and surgical related factors. Using machine learning on contemporary data, our aim was to develop and validate a practical, easy-to-use algorithm to predict risk for VTE and MBE following total joint arthroplasty (TJA). This was a single institutional study of 35,963 primary and revision total hip (THA) and knee arthroplasty (TKA) patients operated between 2009 and 2020. Fifty-six variables related to demographics, comorbidities, operative factors as well as chemoprophylaxis were included in the analysis. The cohort was divided to training (70%) and test (30%) sets. Four machine learning models were developed for each of the outcomes assessed (VTE and MBE). Models were created for all VTE grouped together as well as for pulmonary emboli (PE) and deep vein thrombosis (DVT) individually to examine the need for distinct algorithms. For each outcome, the model that best performed using repeated cross validation was chosen for algorithm development, and predicted versus observed incidences were evaluated. Of the 35,963 patients included, 308 (0.86%) developed VTE (170 PE's, 176 DVT's) and 293 (0.81%) developed MBE. Separate models were created for PE and DVT as they were found to outperform the prediction of VTE. Gradient boosting trees had the highest performance for both PE (AUC-ROC 0.774 [SD 0.055]) and DVT (AUC-ROC 0.759 [SD 0.039]). For MBE, least absolute shrinkage and selection operator (Lasso) analysis had the highest AUC (AUC-ROC 0.803 [SD 0.035]). An algorithm that provides the probability for PE, DVT and MBE for each specific patient was created. All 3 algorithms had good discriminatory capability and cross-validation showed similar probabilities comparing predicted and observed failures indicating high accuracy of the model. We successfully developed and validated an easy-to-use algorithm that accurately predicts VTE and MBE following TJA. This tool can be used in every-day clinical decision making and patient counseling.

Project description:BackgroundRemoval of total joint arthroplasty from the inpatient-only list has created significant confusion regarding which patients qualify for an inpatient designation. The purpose of this study is to develop and validate a novel predictive tool for assessing who will be an outpatient vs inpatient after total knee arthroplasty (TKA).MethodsA cohort of Medicare patients undergoing primary TKA between January 2018 and September 2019 were retrospectively reviewed. Baseline demographics and patient characteristics were obtained, and their distributions for outpatient (less than 2 midnights) and inpatient stay were assessed. Subsequently, a XGBoost machine learning model was trained using 80% of the TKA patients, and the remaining 20% of patients were involved in testing the model's performance in terms of accuracy and the average area under the receive operating characteristic curve.ResultsEight hundred ninety-nine Medicare patients underwent TKA at our institution between January 2018 and September 2019. Of which, 625 patients had outpatients stays, and 274 qualified for inpatient designation. Significant associations were demonstrated between inpatient visits and the following factors: higher body mass index, increased age, better functional scores, multidimensional fatigue inventory, Charlson Comorbidity Index, American Society of Anesthesiologists score, female gender, cardiac history, and the Revised Cardiac Risk Index. The XGBoost model for predicting an inpatient or outpatient stay was 63.3% accurate, with area under the receive operating characteristic curve of 68.8%.ConclusionsUsing readily available key baseline characteristics, functional scores, and comorbidities, this machine-learning model accurately predicts the probability of an "outpatient" vs "inpatient" stay after TKA in the Medicare population. body mass index, age, VR12 functional scores, and multidimensional fatigue inventory scores had the highest influence on this predictive model.

Dataset Information

Machine Learning Model Developed to Aid in Patient Selection for Outpatient Total Joint Arthroplasty.

Background

Methods

Results

Conclusion

Publications

Machine Learning Model Developed to Aid in Patient Selection for Outpatient Total Joint Arthroplasty.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets