Cases of COVID-19 have gone undetected, likely causing future waves. The aim of our research is to develop mathematical and statistical tools for the early detection of future BC COVID-19 waves, and to evaluate control strategies for a future wave. A key component is the estimation of unreported cases and the probability of transmission in high-risk subgroups (such as the elderly and homeless). Our mathematical model will determine disease spread and testing policies interactions. We will then identify early detection strategies for future waves. To track the patterns of individual behaviours and evaluate intervention strategies, we will develop a computer simulation model. With other provinces facing the same problems, our tools can be applied to the national pandemic.
The most exciting development from this project is a model to estimate hidden (e.g. asymptomatic or untested and unconfirmed) COVID-19 cases based on public case, recovery, death count data. We applied this model to the Northern Health Region for the first 30 weeks of the pandemic and found that:
A second development is our theoretical investigation into the effect of contact tracing through a novel mathematical model. The reproduction number (the average number of new infections coming from a single COVID-19 positive case) decreases as contact tracing efforts increase. This is because cases that are found through contact tracing will often isolate. However, when there is a large transmission rate, contact tracing alone may not control the infection. Thus, additional public health measures to decrease transmission are required for contact tracing to be effective. Further, increased testing rate increases the effectiveness of contact tracing.
The Island Health team has made great strides studying outcomes of control measures through agent-based mathematical models to simulate real world scenarios. The team has also been working on profiling contact events between COVID-19 cases and other patients or healthcare workers while accessing or providing service. This provided another method of contact tracing within Island Health service locations.
Our methods and tools are in place for the next pandemic (including open source R code available online). Dr. Cowen was interviewed by the CBC in November 2021, making the public aware of the issue of hidden case counts in the Northern Health Authority region of B.C. This interview came just as case counts began to increase in the region with the Delta wave.
The work done to estimate hidden populations can be done with new variants, different regions, and across the country. The tools that have been developed for Island Health specifically in relation to pandemic response, can and have been adapted to provide actionable insights into other public health crises (i.e. opioid) and pressing operational challenges (e.g. pressure on emergency departments).
Our research is part of an ongoing program to improve population models, and we are continuing to apply our methods to a broad range of problems. We have continued our research on developing statistical models for count data and have drafted a manuscript for a Canada-wide model of COVID-19. Similarly, we are just finishing up drafting an article on the mathematical models that consider contact tracing. These include:
PhD student Matthew Parker has plans to present the Canada-wide modelling work at several conferences this summer and Dr. Cowen has been invited to present the work at the International Chinese Statistical Association Canadian Chapter in Banff.
MSc student Viet Dao is developing a statistical model based on Island Health lab testing data to provide another method of estimating hidden cases. Dr. Cowen is in the process of hiring a post-doc funded by NSERC EIDM and a UVic Aspirations 2030 Fellowship to continue this work, deploying it to all of BC. Dr. Cowen has hired a second post-doc funded by the Canadian Statistical Sciences Institute to develop models using Island Health’s data to estimate the homeless population and study the number of COVID-19 cases within this population.