Abstract
The applications of digital twin (DT) technology in healthcare have shown potential to transform patient care. This paper presents a simulation-based proof-of-concept patient digital twin (PDT) for monitoring chronic diseases, namely, hypertension, diabetes, and respiratory conditions, utilizing publicly available datasets to demonstrate the potential for remote patient support. The PDT can simulate remote monitoring of a patient's health status by integrating physiological data streams and medical records. The PDT framework was demonstrated through simulation-based experiments using publicly available datasets, enabling proof-of-concept evaluation of patient monitoring and the drug-class prediction pipeline. However, the existing PDTs for patient remote monitoring need to be shaped with sufficient information to provide a comprehensive analysis. This limitation can be addressed by incorporating the patient's medical records along with real-time vitals in the PDT, allowing the patient's health status to be monitored and analyzed more effectively. Through it, the patient's health status can be monitored and analyzed effectively. We designed and simulated a PDT for chronic disease patients in a simulation-based setting to demonstrate the concept. We emulated the collection of patient vitals using public datasets to represent sensors attached to the patient's body (physical twin) and accessed simulated medical records. By real-time simulation of the patient data, a DT of the patient (i.e., virtual replica) is created. We performed cloud-based real-time simulation so that the DT can be accessed remotely and continuously observed in a simulated environment to demonstrate the feasibility of early abnormality detection and timely intervention. The drug-class prediction module achieved 96.53% accuracy on a rule-driven synthetic dataset, demonstrating the effectiveness of the system pipeline as a preliminary validation step before proceeding to clinical evaluation.