A Bayesian Network Model to Establish a Digital Twin Architecture for Superior Falls Risk Prediction

Implementing digital twin technology as an abstraction from the reality of physical systems for real-time evaluation of trends and outcomes has dominated the manufacturing sector, but its importance in healthcare cannot be overemphasized. We develop a digital twin architecture using a Bayesian network approach for establishing the predisposition of patients to falls risk on admission. This framework provides decision support for seamless management of inpatients\u27 falls risk acuity by interfacing patients\u27 clinical, demographic, and psychosocial attributes with a Bayesian network for pattern classification to predict falls risks levels. By using the Synthetic Minority Oversampling Technique (SMOTE) to upsize the minority falls risk class, it was possible to enhance the prediction accuracy of falls risk vulnerability after a case study with real hospital data. Our results show that the chances of optimizing patients\u27 experience with the digital twin framework are very high because of the enhanced management of falls risk on admissions