Artificial intelligence for precision viral surveillance of emerging infectious disease (EID): Data-driven digital twin metaverse-envisioned study

Precision containment strategies incorporating artificial intelligence (AI)-driven dynamic viral shedding models are pivotal for the effective control of emerging infectious diseases (EIDs). Among the foundational applications within the Metaverse, digital twins-which integrate physical and virtual environments via augmented reality (AR) and mixed reality (MR)-offer a promising solution. Leveraging IoT-like laboratory-based viral shedding data together with demographic and clinical features, this study aims to develop a data-driven digital twin model for precision viral surveillance to monitor EIDs and to provide an immersive framework for evaluating the effectiveness of contact tracing, isolation, and quarantine protocols within the Metaverse. We proposed a digital twin thread architecture, comprising a temporal data pipeline designed to support multiple twin functionalities. The process began with the development of the physical twin, which incorporated dynamic cycle threshold (Ct) data from serial RT-PCR tests-serving as IoT-like laboratory inputs along with the associated demographic and clinical data. The underlying parameters of infectious disease dynamics were learned through Markov-based statistical machine learning, applied to these time-series data. A virtual avatar representing these digital threads-a virtual thread cohort-was rendered in virtual reality (VR). Analytic twins, enhanced via AR, overlaid virtual data onto the physical twin to bridge observed and inferred states. Subsequently, decision twins, implemented through MR, were utilized to assess the effectiveness of immersive, precision-guided interventions such as contact tracing, isolation, and quarantine. This framework was applied to COVID-19 outbreaks caused by the Alpha and Omicron variants of concern (VOCs) in Changhua, Taiwan, using viral shedding data. A showcase case study on precision contact tracing during the Alpha VOC outbreak was presented. A noise-driven privacy protection method was implemented for addressing the concern of patient confidentiality. From the physical twin data of 269 confirmed Alpha VOC cases, a virtual thread cohort of 1,000,000 simulated cases was generated. Analytic twins, enabled by AR, synthesized data from both physical observations and virtual predictions, capturing real-time dynamics that were otherwise unobservable. Using this framework, the initial Alpha VOC cluster was analyzed to derive key transmission indicators. Decision twins identified optimal Ct-guided contact tracing windows: for individuals with Ct values between 18 and 25, retrospective tracing for 7 days achieved 30 % effectiveness, 13 days yielded 60 %, and 24 days reached 90 %. For Omicron VOC, the effectiveness of quarantine among vaccinated individuals (with booster) reached 77 % after 3 days and 94 % after 7 days, compared to 39 % and 76 % in unboosted individuals, respectively. The utility of precision contact tracing within the Metaverse was validated by the Alpha VOC outbreak showcase study along with the presentation of a noise-driven approach for data privacy protection and data security. This Ct-guided, data-driven digital twin model demonstrates a novel approach to EID containment, highlighting the potential of the Metaverse as a convergence of physical and cyber domains. Our findings illustrate the applicability and scalability of digital twin frameworks in precision public health and underscore their broader implications for future healthcare innovations taking data security and privacy protection into account.