Coronavirus disease 2019 (COVID-19) has been the primary target of most epidemiologists and biostatisticians over the last two years, and dozens of papers of varying quality have been published on new methods of modeling the spread of the disease, the likelihood of an outbreak occurring, and the current threat posed by the disease. Researchers from Curtin University have formulated a new modeling framework to help identify the transmissibility of COVID-19 based on periods of varying case incidence.
Study: Estimating the transmissibility of SARS-CoV-2 during periods of high, low and zero case incidence. Image Credit: bob boz/Shutterstock
A preprint version of the study is available on the medRxiv* server while the article undergoes peer review.
The effective reproduction number depends on the number of contacts an infectious person makes and how likely that contact is to infect a second person. Many countries already offer advice on how to limit both of these factors, such as social distancing, handwashing, and avoiding crowded places.
The researchers have identified new techniques to estimate how effective these are. By observing changes in the rate of social contact and the probability of infection per contact and examining how these change the transmission rates of the virus, they aim to explore the effectiveness of behavioral changes in curbing the reproduction number.
To estimate the ability of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) to spread in a population over time, they used a novel semi-mechanistic model, pulling data from cases, population behaviors, and local health system effectiveness. Transmission is modeled separately for local to local transmission and cases imported from overseas. Local to local transmission is modeled using the average population-level trend driven by interventions that mostly target local to local transmission by policing, distancing behavior, and isolating infected individuals, as well as short-term fluctuations in local to local transmission.
This information allows the researchers to capture stochastic dynamics of transmission, including clusters of cases and periods of lower than expected transmission, and evaluate the ability of the virus to spread during these periods.
To estimate the transmission potential or TP, the scientists use three sub-models. Physical distancing behavior is split into two categories: macrodistancing, the reduction in the average rate of non-household contacts, and microdistancing, the reduction in transmission probability per non-household contact. Both are assessed through weekly surveys, the data of which can be used to infer temporal trends in behavior. The data on the number of days from symptom onset to case notification can be used to estimate the proportion of detected cases and the time taken to inform the infected.
The researchers focused on a period from March 2020 to January 2021 in Australia to demonstrate the value of their method. They estimated that transmission potential decreased substantially across the country through the second half of March 2020, to just below one, following an increase in macro and microdistancing behaviors. The local-to local transmission potential remained below one across April before steadily increasing in the summer months as restrictions were dismantled.
In New South Wales, a series of localized outbreaks between June and October 2020 were controlled with specific restrictions. The transmission potential hovered above one, suggesting these levels of controls were insufficient to protect the general population. In November 2020, South Australia suffered a cluster of more than 20 cases due to a breach of the mandatory quarantine order. Despite this, in previous days, case transmission remained low. The scientists estimated the transmission potential to be 1.71 on the day in question, suggesting that an epidemic could occur very rapidly once an outbreak is established.
This model could help inform public health policymakers, healthcare workers, and epidemiologists, allowing them to make the best decision available to them. Estimating the likelihood of the disease spreading allows the most effective social distancing policies to be identified, can help hospitals identify how many beds are likely to be occupied, and can provide further insight into the disease. This study should help provide insight into epidemic dynamics, account for variability in the types of contact made, and support situational assessment and planning for safe re-openings.
medRxiv publishes preliminary scientific reports that are not peer-reviewed and, therefore, should not be regarded as conclusive, guide clinical practice/health-related behavior, or treated as established information