COVID-19 Ch 11: Modeling

The eleventh episode of our Anatomy of a Pandemic series has arrived, and just in time. Have you found yourself trying to sift through headlines claiming “this model predicts that” and “that model predicts this”, but you’re not sure where the truth really lies? Then this episode is for you. With the help of Dr. Mike Famulare from the Institute for Disease Modeling (interview recorded April 29, 2020), we walk through the basics of mathematical modeling of infectious disease, explore some of the current projections for this pandemic, and discuss some guidelines for evaluating these headline-making models. As always, we wrap up the episode by discussing the top five things we learned from our expert. To help you get a better idea of the topics covered in this episode, we’ve listed the questions below: What is a math model and what are some of the goals of mathematical modeling? So talking specifically now about infectious disease models, can you walk us through what the basic components are of an infectious disease model, like an SIR model? Where do you get the data that you use to estimate the parameters in an SIR model - what is based on actual data and what has to be estimated? Infectious disease outbreaks often have a curve-like shape, with the number of infected individuals on the y-axis and time on the x-axis. Can you explain why infectious disease epidemics tend to follow a curve? Can you talk us through some of the assumptions that you have to make when you're constructing one of these models and how that kind of relates to the uncertainty inherent within models? How might that uncertainty affect interpretation? What are some examples of the various ways we use infectious disease models in public health policy? Can you talk about how models might be used at various stages of a pandemic to guide public health measures? How might our use of models early on in a pandemic be different from the middle of one? Speaking specifically about COVID-19 now, can you talk about what a basic model for this pandemic might look like?  Are models for COVID-19 using only lab-confirmed cases of the disease or clinical-confirmed cases as well? Looking back on these earlier models of COVID-19, what can we take away from the performance of these models? Is there any agreement among models as to what policies might be the best in terms of keeping cases and deaths as low as possible?  For those of us who have no background in mathematical or statistical modeling, are there guidelines that we should use to evaluate these models or compare them? What should we (as in the general public) be taking away from these models? Are there any positive changes you hope to see come out of this pandemic, either as a member of the community or as a math modeler? For a deeper dive into the wonderful world of infectious disease models, we recommend checking out this recent video from Robin Thompson, PhD of Oxford Mathematics titled “How do mathematicians model infectious disease outbreaks?” The video was posted on April 8, 2020.