Volume 43, Issue 4, October 2024, Pages 1023–1035
Keffa Denis Diomande1, Seydou Sangare2, N’Guessan Behou Gérard3, and Kone Tiémoman4
1 Virtual University of (VUCI), Abidjan, Côte d’Ivoi, Côte d’Ivoire
2 Virtual University of (VUCI), Abidjan, Côte d’Ivoi, Côte d’Ivoire
3 Virtual University of , Research and Digital Expertise Unit, Abidjan, Côte d’Ivoi, Côte d’Ivoire
4 Virtual University of , Research and Digital Expertise Unit, Abidjan, Côte d’Ivoi, Côte d’Ivoire
Original language: English
Copyright © 2024 ISSR Journals. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
The phenomenon of leaving against medical advice remains a significant issue in public reference institutions in Côte d’Ivoire. Thus, one out of twelve adult patients hospitalized in the Orthopedics – Traumatology department of the Treichville University Hospital often interrupts their treatment in favor of traditional Bone-Setters or other destinations. However, despite recent advances in machine learning, it is still challenging to predict what type of destination these absconding patients will choose. Therefore, this article first aims to sequentially establish two datasets based on medical records: one original and the other after feature selection. Then, based on these datasets, this research involved four supervised machine learning models (Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), and Gradient Boosting (GB)). The results obtained from performance metrics during testing, after five cross-validations, show that Random Forest is the most robust model for both datasets. Finally, a second analysis indicates that the Random Forest built on the original dataset remains the best model overall, with an AUC-ROC of 96%, an accuracy of 86%, a precision of 84%, a recall of 100%, and an F1-Score of 91%. These results suggest that this model offers hope for early and accurate prediction of the destination the absconding patient will opt for, thus positively impacting their care.
Author Keywords: Cross validation, datasets, destinations, Discharge Against Medical Advice, Orthopedics – Traumatology, supervised machine learning.
Keffa Denis Diomande1, Seydou Sangare2, N’Guessan Behou Gérard3, and Kone Tiémoman4
1 Virtual University of (VUCI), Abidjan, Côte d’Ivoi, Côte d’Ivoire
2 Virtual University of (VUCI), Abidjan, Côte d’Ivoi, Côte d’Ivoire
3 Virtual University of , Research and Digital Expertise Unit, Abidjan, Côte d’Ivoi, Côte d’Ivoire
4 Virtual University of , Research and Digital Expertise Unit, Abidjan, Côte d’Ivoi, Côte d’Ivoire
Original language: English
Copyright © 2024 ISSR Journals. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
Abstract
The phenomenon of leaving against medical advice remains a significant issue in public reference institutions in Côte d’Ivoire. Thus, one out of twelve adult patients hospitalized in the Orthopedics – Traumatology department of the Treichville University Hospital often interrupts their treatment in favor of traditional Bone-Setters or other destinations. However, despite recent advances in machine learning, it is still challenging to predict what type of destination these absconding patients will choose. Therefore, this article first aims to sequentially establish two datasets based on medical records: one original and the other after feature selection. Then, based on these datasets, this research involved four supervised machine learning models (Logistic Regression (LR), Decision Tree (DT), Random Forest (RF), and Gradient Boosting (GB)). The results obtained from performance metrics during testing, after five cross-validations, show that Random Forest is the most robust model for both datasets. Finally, a second analysis indicates that the Random Forest built on the original dataset remains the best model overall, with an AUC-ROC of 96%, an accuracy of 86%, a precision of 84%, a recall of 100%, and an F1-Score of 91%. These results suggest that this model offers hope for early and accurate prediction of the destination the absconding patient will opt for, thus positively impacting their care.
Author Keywords: Cross validation, datasets, destinations, Discharge Against Medical Advice, Orthopedics – Traumatology, supervised machine learning.
How to Cite this Article
Keffa Denis Diomande, Seydou Sangare, N’Guessan Behou Gérard, and Kone Tiémoman, “Binary Classification Model of Adult Patients Deserting the Orthopedic Traumatology Department of a Reference Hospital: A Machine Learning Approach to Strengthen Traditional Medicine,” International Journal of Innovation and Applied Studies, vol. 43, no. 4, pp. 1023–1035, October 2024.
