Table 3.

Studies evaluating the ability of machine learning algorithms to predict healthcare utilization of  TSA

Author, year	Outcomes measured	Follow up	Most accurate algorithm	Mean absolute error	Outcomes
Biron et al., (2020) [26]	Length of stay	N/A	N/A	N/A	Machine learning may be used to predict whether individuals had a one-day LOS or shorter following TSA
Karnuta et al., (2020) [28]	Length of stay, discharge disposition, and inpatient charges	1 Month	N/A	Accuracy in Chronic/degenerative conditions -Total Cost: 76.5% -Length of Stay: 91.8% -Disposition (home): 73.1% Accuracy in Acute/traumatic conditions -Total Cost: 70.3% -Length of Stay: 79.1% -Disposition (home): 72%	For both chronic/degenerative and acute/traumatic shoulder arthroplasty, artificial neural networks displayed medium to high accuracy and reliability in predicting inpatient cost, LOS, and discharge disposition
Lopez et al., (2021) [33]	Non-home discharge and 30-day postoperative complication rates	1 Month	Both had similar accuracy, but the artificial Neural Network had better discriminative ability	Accuracy for Artificial Neural Network Non-home discharge Accuracy: -Boosted Decision Tree: 90.3% -Artificial Neural Network: 89.9%	Machine learning has the capacity to reliably predict non-home discharge following elective TSA
Lopez et al., (2022) [34]	Prolonged operative time and 30-day postoperative complication rates	1 Month	Artificial Neural Network	Accuracy for Artificial Neural Network Prolonged operative time Accuracy: -Boosted Decision Tree: 85.6% -Artificial Neural Network: 84.7%	Machine learning models can predict which patients are more likely to require longer TSA operations