Modelling the transmission of the omicron variant of SARS-CoV-2 in Ontario using inter and intra-county population mobility

Liwei Yang^{*1, 2} (blairyeung), Yuhao Yang¹ (martinyang0416), Muxin Tian ¹ (realtmxi)

¹ Department of computer science, University of Toronto, 40 St George St, Toronto, M5S 2E4, Canada

² Department of Cell & Systems Biology, University of Toronto, 25 Harbord St, Toronto, M5S 3G5, Canada

^* This is the corresponding author.

Figure 1: The convolutional model.

Input of the model

Model parameters not subject to fitting:

Google population mobility index in Ontario

Figure 2: Google population mobility in Ontario.

Assumed scenarios:

1. Constant (use the moving average of year 2022)
2. Return to baseline (restore to baseline, four more assumptions)
3. Time-series forecast

Seasonality of the transmission activity in Ontario

Figure 3: Seasonality of SARS-CoV-2 transmission activity

Assumed scenarios:

1. 10% sensonality
2. 20% sensonality
3. 40% sensonality

Vaccination and vaccine effectiveness

Figure 4: Vaccination status in Ontario. The actual data we used are specified to public health unit.

• We used the phu-specific data to estimate the vaccine fraction of different ages bands.

• We used the provincial dose administration data to estimate the vaccination level.

• We then combined the two data to estimate the number of vaccine immunity for the counties.

Assumed scenarios:

1. Future vaccination performed as the ending speed.
2. No future vaccination
3. 100 - 500 boost dose per day per 100,000 poppulation

Figure 5: Vaccine effectiveness against clinical infection.

Assumed scenarios:

Unweighted average of multiple articles.

Commutating matrix

Figure 6: Commuting matrix of Ontario, censused in 2015.

This is an ongoing project

The model can run now! However, there are still some features not implemented yet and some calibrations to be done.

TODOs:

Data to be collected

• Average delay from COVID-19 onset to hospitalization for the omicron variant of SARS-CoV-2. Preferably age-specific.
• Average delay from COVID-19 onset to admission of ICU for the omicron variant of SARS-CoV-2. Preferably age-specific.
• Average delay from COVID-19 onset to death for the Omicron variant of SARS-CoV-2. Preferably age-specific.
• Specific dose administration count. (The current one contains 1, 2, 3 doses only. We may want to include 4, 5 dose.
• Vaccine effectiveness against general infection (estimated prior distribution using posterior distribution probability).
• Vaccine effectiveness against symptomatic disease, given infection (conditional prior distribution).
• Vaccine effectiveness against hospitalization, given infection (conditional prior distribution).
• Vaccine effectiveness against deaths, given infection (conditional prior distribution).

Features to be implemented

• Forecasting the population mobility using previous Google mobility.
• Use the synthesized commutation matrix to compute the inflow and outflow mobility of inter-county population.
• Use the inter-county population flow to estimate the inter-county flow of infected and immunized individuals.

Model fitting

The model need to be fitted to the previous year's transmisison of COVID-19 before it is used for forecast.

Currently working on the calibration and optimization of the model.

Completed features:

Vaccine effectiveness

We estiamted the vaccine effectiveness using the vaccination effectiveness derived Andrews et al., Gold, CDC.

Population mobility & seasonality

The model can now accept and use population mobility & seasonality ot estimate

Calibration and vaccine effectiveness

We are currently working on the calibration of the model.