Metaverse health care has increasingly gained attention.1,2 While it has been applied to various medical fields such as radiology, oncology, training, and others using the artificial technology of augmented virtual reality (AR) and extended virtual reality (XR), the digital twin (DT) is one of the major elements that links physical space with virtual space for the service layer of interest, such as health care.3 The integration of metaverse technologies in healthcare, especially through digital twins, enhances medical services, including diagnostics, therapy, and patient engagement in areas like cancer treatment, mental health, and chronic disease management.4 The applications of DT to personal health care have been fully reviewed elsewhere, from basic medical research at different levels of the body to clinical research and population-based health care studies.5
Driven by the COVID-19 pandemic, DT has been deemed to have great potential for real-time monitoring, big data analytics, and making timely decisions to offer preventive measures for early detection of epidemics and disease. In the absence of DT, it is very difficult to have a prompt response to the outbreak of an emerging infectious disease (EID), particularly when the underlying pathogen has a high potential for mutations. It's hard to keep EID like the coronavirus disease 2019 (COVID-19) pandemic under control because of new variants of concern (VOCs) of the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) that appear quickly, starting with Alpha and passing through Beta, Gamma, Delta, and finally Omicron. Each variant or subvariant manifests various extents of transmissibility, duration of infectiousness, disease severity, and the escaped immune response transcended from their corresponding mutation points. For instance, the N501Y mutation in Alpha and Omicron accounts for high transmissibility, and the E484K mutation in Beta, Delta, and the E484A mutation in Omicron is related to an escaped immune response.6
To cope with the thorny issue of rapid mutations of SARS-CoV-2 variants, it is indispensable to have precise and real time surveillance of the spread of SARS-CoV-2 infection. The recent study has proposed the use of viral shedding measured by cycle threshold (Ct) through the reverse transcriptase-polymerase chain reaction (RT-PCR) test, as a gradient relationship has been shown between the Ct value and transmissibility and the severity of the disease for patients infected with SARS-CoV-2.7,8 It has also already been used as an indicator for evaluating the efficacy of anti-viral therapy.9,10 The median time for a high Ct value was used to judge whether the VOCs or variants of interests (VOIs) would lead to the new wave of the pandemic or affect the duration of hospitalization.11
However, there are several obstacles to making use of the Ct value for precision surveillance of the outbreak of EID at the individual level. First, the EID, like SARS-Cov-2 infection, has been infectious during pre-symptomatic (non-persistent asymptomatic) or asymptomatic periods, and Ct is not often measured until the presence of a symptom. Information on the digital instances of Ct (the same individual) before the symptom is often unavailable unless a precision contact tracing investigation is undertaken to identify the suspected infected contact. Second, data on the digital instances Ct value are often incomplete, even when RT-PCR is offered for the suspected infected contacts during the quarantine and isolation period. Third, precision surveillance of EID at the individual level is often faced with how to provide a precise duration of quarantine and isolation, which is often influenced by the status of vaccination, when the prevalence of infections such as Omicron (B.1.529) is rampant. These three obstacles, together with the dynamic properties of Ct, render the avatar of the real world, as noted in the previous review article of DT, hardly available when precision surveillance strategies need to be pursued.5
The Ct value also plays an important role in precision surveillance for containing epidemics of the disease at the population level. The dynamic changes in the mean Ct value at the population level were used to estimate the Rt in the specific area7. In addition to the disease progression, the Ct-value can reflect the timeline of infection for affected patients. However, the digital instances of the Ct-value of each COVID-19 patient varied from individual to individual. It is therefore interesting to derive a series of unbiased parameters of the kinetic epidemic trajectory at the population level, such as generation time.
DT Application to Ct-guided Surveillance for Covid-19 Epidemic
Taiwan has maintained a well-contained epidemic of COVID-19 until the first large-scale community-acquired outbreak occurred in May 2021, with 600–700 daily cases until the end of July 2021. The immediate response and contact tracing have been conducted by the local government in order to curb the rampant transmission. By doing this, we can get multiple RT-PCR readings, which let us figure out how the Ct level changed after confirmation and how that changed level relates to the first signs of a disease. We developed a novel Ct-guided surveillance model to solve the potential issues of the traditional contact tracing method, which may involve time- and effort-consuming quarantining of all possible contacts within 14 days prior to a given case being identified. It enables one to have precision quarantine and isolation for suspected infected persons when contacting infectives.
We hereby developed a Markov process to delineate the dynamic changes of Ct-value for patients affected by COVID-19 before and after symptoms occurred for estimating transition rates between states, trained by data from a community-acquired outbreak of SARS-CoV-2 Alpha and Omicron VOC variants in Changhua, Taiwan. “In Silico” simulation, the machine learning algorithm was programmed to generate outcomes based on the digital instances of trained parameters to predict the time of infectiousness in the pre-symptomatic phase and the time to symptoms of each given case in order to answer the following questions:
“How many days before is required for enrolling persons suspicious of infection through close contact with infectives of Alpha VOCs given information on viral shedding level upon conformation? This is related to the precision contact tracing investigation for the outbreak of the emerging infectious disease.
“How many days are required to quarantine the suspected infected persons through close contact with infectives of Omicron VOCs given information on viral shedding level upon conformation with and without booster vaccination”? This pertains to precision surveillance of quarantine and isolation for pre-symptomatic cases and symptomatic cases to forestall the spread of the emerging infectious disease.
At the population level, such a method can be used to estimate the generation time and series interval and also to compute the reproductive number for assessing the spread of outbreaks of emerging infectious disease with the use of a modeling strategy.