Assessing the progression of the COVID-19 pandemic in Canada using testing data and time-dependent reproduction numbers

OBJECTIVES: To compare a mathematical tool and time-dependent reproduction number (R(t)) estimates to assess the COVID-19 pandemic progression in a Canadian context. METHODS: Total number of reported cases were plotted against total number of tests for COVID-19 performed over time, with and without smoothing, for Canada and some Canadian provinces individually. Changes in curvature profile were identified as either convex or concave as indicators of pandemic acceleration or deceleration, respectively. R(t) estimates were calculated on an exponential growth rate. RESULTS: For Canada as a whole, the testing graphs had a slightly concave profile and a coincident decrease in R(t) estimates. Saskatchewan more recently had a convex profile with a gradual shift to a concave profile and also demonstrated a gradual decline in R(t) estimates. Curves and R(t) estimates for Alberta, British Columbia, Manitoba, Nova Scotia, Ontario and Quebec displayed a gradual shift towards concavity over time and an overall decrease in R(t) estimates, which is suggestive of a positive impact of public health interventions implemented federally and provincially. CONCLUSION: The present analyses compared a mathematical tool to R(t) estimates to ascertain the status of the pandemic in Canada. Caution should be taken when interpreting results due to factors such as varying testing protocols, available testing data unique to each province and limitations inherent to each method, which may generate different results using the two approaches. Analysis of testing data may complement metrics obtained from surveillance data to allow for a weight-of-evidence approach to assess the status of the COVID-19 pandemic.