Q and A with the Experts: Modelling COVID-19

Friday, April 3, 2020

With different countries, and different levels of government grappling with whether or not to publicly release their mathematical models and projections on the progression of the COVID-19 pandemic, we asked Professor Chris Bauch, an expert in mathematical and computer modelling of infectious disease outbreaks, to explain the basics to us.

What is a disease model/projection and how are they created?

 A disease model is a way of using mathematical equations or computer simulations to predict how many cases we can expect in the coming weeks or months.  They are created by writing down equations that represent the best current knowledge of how the disease spreads, the demographics and movement patterns of the host populations, the course of the infection in individuals, and how interventions like physical distancing reduce infection.  The equations are then solved to figure out how many cases we can expect in the future, under different scenarios for government actions, availability of health technologies, public reaction and many other variables.

 Are the models accurate?

 Models can be more or less accurate depending on the structure of the model, the quality of the data used to make the model, and how well the modeller has accounted for the impact of data uncertainties.  I think the best models from past epidemics and pandemics have been accurate enough to meaningfully inform public health policy.   One example includes the outbreak of foot-and-mouth disease in Great Britain in 2001.  Even though the models are not perfectly accurate, they are often the only tool we have to anticipate what might come.  The results of combining different models produced by different researchers are more reliable than the results of single models, generally speaking.  This is also a principle of modelling that has been used to improve weather forecasting models.  

 Should the public have access to the ‘worst-case scenario’ projections?

 Yes, I think so.  I don't believe in hiding information from the public.  But the public also needs to know how to interpret the projections.  They need to understand that there is always some uncertainty in model projections and that the worst-case scenarios will not necessarily play out.  Model projections will change depending on the steps we take to prevent the disease spread.  One of the reasons we make model projections is to point out what will happen if we don't do anything, or if we take the wrong actions, or if we wait too long.  In this way, they can stimulate action to prevent the worst-case scenarios predicted by the models.  This is how models have been used so far in the pandemic, successfully, I think.  By predicting what will happen if we do not do social distancing measures, they have stimulated action that will help prevent our healthcare systems from being overwhelmed. 

 The University of Waterloo has a number of experts available for comment on various aspects of the COVID-19 pandemic, click here to see the up-to-date list.

MEDIA CONTACT | Rebecca Elming
519-888-4567 x 30031 | @uwaterloonews | uwaterloo.ca/news

Attention broadcasters: Waterloo has facilities to provide broadcast quality audio and video feeds with a double-ender studio. Please contact us for more information.

Get these news items on your student portal

  1. 2023 (61)
    1. May (15)
    2. April (9)
    3. March (17)
    4. February (9)
    5. January (11)
  2. 2022 (127)
    1. December (7)
    2. November (12)
    3. October (8)
    4. September (8)
    5. August (4)
    6. July (4)
    7. June (13)
    8. May (23)
    9. April (10)
    10. March (15)
    11. February (13)
    12. January (10)
  3. 2021 (143)
    1. December (9)
    2. November (21)
    3. October (19)
    4. September (17)
    5. August (13)
    6. July (12)
    7. June (8)
    8. May (11)
    9. April (7)
    10. March (10)
    11. February (6)
    12. January (10)
  4. 2020 (143)
  5. 2019 (126)
  6. 2018 (99)
  7. 2017 (17)
  8. 2016 (17)
  9. 2015 (19)
  10. 2014 (24)
  11. 2013 (24)
  12. 2012 (18)