Polymerase chain reaction (PCR) tests are mostly recognized as a primary standard to identify a patient with infectious diseases like COVID-19. Rapid Antigen Testing (RAT) kits are alternatives that enable easy and quick assessment while their accuracy is yet to be certified when compared to PCR Tests. Literature shows that there is a time-dependency on returning negative result by using the RAT. There is an unidentified period for each patient even when they exercise frequent rapid testing. Longer unidentified period leads to higher risk to infect others. Risk models are therefore essential for one to assess the potential impacts. We adopt the probabilistic framework to describe the likelihood of a patient returning a negative result by RAT tests, together with the chance for a patient...[ Read more ]

Polymerase chain reaction (PCR) tests are mostly recognized as a primary standard to identify a patient with infectious diseases like COVID-19. Rapid Antigen Testing (RAT) kits are alternatives that enable easy and quick assessment while their accuracy is yet to be certified when compared to PCR Tests. Literature shows that there is a time-dependency on returning negative result by using the RAT. There is an unidentified period for each patient even when they exercise frequent rapid testing. Longer unidentified period leads to higher risk to infect others. Risk models are therefore essential for one to assess the potential impacts. We adopt the probabilistic framework to describe the likelihood of a patient returning a negative result by RAT tests, together with the chance for a patient to infect another healthy individual. Intuitively, if a patient is very likely to be infectious, it should also has a higher chance to be identified by tests due to the high viral load. By our model, we would like to evaluate the trade-off between the probability of infection and the probability of accurate testing.

We collected individual-level data from COVID-19 in Hong Kong to quantify the time-dependent impacts for not identifying a patient. For given uncertainties, we propose an linear scheduling model to minimize the infection risk by utilizing the of using RAT. We aim to apply the model in contextual setting such that the decision makers can decide for testing frequency and operation strategies to guarantee the level of safety.