Dr. Omar E Cornejo is a faculty member in the School of Biological Sciences at Washington State University. His research focuses on evolutionary genomics and populations genetics of hosts and pathogens. Dr. Cornejo has teamed up with a group of scientists from the US, Australia and Sweden to develop model-guided assessments of how different mitigation strategies might influence the spread of COVID-19. These epidemiological models track susceptible (S) individuals, exposed individuals (E), Infectious individuals that can transmit the disease (I), and recovered individuals (R). Similar models have been used with diseases like influenza or other coronavirus, but they have been updated to address issues related specially to the COVID-19 disease that is produced by the infection with the SARS-CoV-2 virus.
Dr. Cornejo has developed a version of the model to assess the relative importance of considering quarantining low-risk and high-risk groups on the progress of the COVID-19 epidemic. These models allowed the research team to first ask how different efforts to mitigate (via social distance) would impact the spread of the disease. Their analysis suggests that if only high-risk individuals are required to maintain quarantine, then the total number of infected individuals will be minimally affected. If only low risk individuals are required to quarantine this would be more effective but not enough to control the epidemic. The control of the epidemic is only possible if we all individuals socially distance.
The following set of analyses let the team analyze how much reduction in the transmission rate (contact rate times the probability of contagion) would be needed to effectively reduce the total percent of infected individuals in a population. Their models demonstrate that an aggressive implementation of measures early during the epidemic could have greatly reduced time of the epidemic and to relax social distancing requirements. Unfortunately, this did not happen, and the models predict that the transmission needs to be reduced to at least 65% of the current rate to bring the epidemic to a halt.However, if the measures used to reduce transmission are relaxed too soon, then we will likely see spikes in the infection rates. If we take a more cautious approach and the quarantine measures are relaxed in a progressive manner, then it is more likely that we will be able to delay the return of the disease to a time when vaccines and more effective treatment has been.
Questions regarding how these models translate into actual social networking and contact with others is a pressing question. The majority of mechanistic models presented so far in the growing COVID-19 literature, including those discussed above, are poorly suited to capture changes in transmission or contact rate that occur heterogeneously across a population’s contact network. It is also not easy to translate changes in the “mean rate of transmission” to measures that public health system can implement. Therefore, in collaboration with Ryan McGee, PhD student from University of Washington, the team has developed an explicit network model of our SEIR models (see https://github.com/oeco28/SEIR-MODEL-COVID19 and https://github.com/ryansmcgee/seirsplus). This network-based framework allows them to assign parameters and other properties (e.g., age group, risk level, willingness to comply with an intervention) and direct interventions (e.g., testing) to individuals based on their state, connectedness in the network, or other factors. Their network model also allows to study population structures (e.g., households, schools, workplaces, communities) that have an important impact on transmission patterns. In addition, network-based interventions, such as social distancing, case isolation, and contact tracing, can be implemented directly with corresponding real-time modifications to the contact network. This makes changes in epidemic outcomes (e.g., total infections, reproduction number) easily interpretable in terms of numbers and patterns of contacts among individuals, something that is not feasible with mean-field models. Currently these last generation of network models are being used to explore specific questions about mitigation and how the social structure of different places can affect the spread of the epidemic and our ability to control it. Specifically, the team is testing if there will be differences in the spread of the epidemic will be great in societies or communities that have a larger number of multigenerational households where low-risk and high-risk individuals are in frequent contact. They are also exploring what are the limits of testing/contact tracing and isolating as strategies as the general lockdown policies are relaxed. Preliminary results suggest that tracing could potentially help but have limited effect if not done properly.
Other research groups are encouraged to access the code and explore scenarios on their own. Dr. Cornejo and his collaborators are continuously providing support to the Departments of Health of Idaho and Washington and are in constant contact with other groups (Marc Lipsitch at Harvard School of Public Health, Joshua Weitz at Georgia Tech). More up to date information about our advances are frequently posted on Twitter via @ocornejoPopGen, @evokerr, @CT_Bergstrom and @RS_McGee.