A new mathematical model, called Distancing-SEIRD, shows that continuing social distancing at the current levels in Ontario and Quebec for six months (until mid-September), could save about 100,000 lives, according to University of Waterloo researchers
The machine learning-enabled model was recently the subject of a study that is pending peer review.
“I think we need to wait to ease the social distancing restrictions,” said Mohammad Kohandel, a professor in Waterloo’s Department of Applied Mathematics. “People are tired of the social distancing restrictions, but they need to be patient as we need to continue this a little bit longer and then start to remove things gradually as that seems to be the best approach. To keep the number of patients who need the treatments below the number of resources available, we need to go four to six months with social distancing from the time it started.”
The researchers modified a standard mathematical model that is used in epidemiology to simulate the spread of infections, called an SEIRD model. They used a machine-learning algorithm to feed the model data collated by the Johns Hopkins University Center for Systems Science and Engineering collected from 184 countries between January 22, 2020, and April 13, 2020. The data includes daily counts of confirmed COVID-19 cases, deaths, and daily counts of people who have recovered from the disease.
The researchers also used the cellphone tracking data that was collected by Google for North America but explored the results for Canada to find the ranges of people who were adhering to social distancing. The data showed that approximately 60 per cent of people are practicing social distancing in Ontario, and 70 per cent in Quebec.
The Distancing-SEIRD model then predicted that the total number of deaths over six months in Ontario with no social distancing would be as high as over 100,000 people. In contrast, a more strict case of social distancing, with 60 per cent adherence, would result in a drastic reduction in the death toll – reducing fatalities by more than 50 per cent over the first six months.
With Quebec having experienced higher numbers of COVID-19 cases, the model indicates that strict social distancing measures could save tens of thousands of lives over six months, compared to no social distancing.
Our model parameters are all fixed assuming that everything remains the same,” said Michelle Przedborski, a Research Assistant Professor in Waterloo’s Faculty of Mathematics. “The model shows that if we have a strict social distancing policy and then we stop social distancing all at once the tendency is that we don’t actually flatten the curve, we just shift the infection peak. And effectively, we risk having the same peak numbers of infections as we would if we hadn’t social distanced.”
The mathematicians are currently working on an extension of the model to include the effects of other factors, such as sex and age. They are keen to work on the model further with public health officials to aid in the making of policy decisions.
The study, Mathematical modeling of COVID-19 containment strategies with considerations for limited medical resources, authored by Kohandel, Przedborski, and Waterloo’s PhD candidates Brydon Eastman and Cameron Meaney, has been submitted for publication.
