If you turn on the news right now, chances are you’ll probably hear someone talking about models. And they are important. They’re a crucial tool in helping us visualise the path of COVID-19, meaning we can start to make predictions about which interventions might work and which might not.
So what actually are they?
Models are representations of the real world. Think about a cardboard box. We know three measurements:
- l (length),
- w (width), and
- h (height)
And the formula for the volume of a cuboid is:
Volume = l x w x h
So we have a (very simple) mathematical model of the space in that box. Of course, the model is not the same as the real thing. We didn’t think about the thickness of the cardboard, or any other subtleties. But this sort of mathematical representation is essentially what models are. They’re used to help scientists describe and explain things they can’t experience directly.
Models and pandemics
Now let’s turn to epidemiology (the study of how often diseases occur in different groups of people and why). Using mathematical and computational models of the world, epidemiologists can begin to predict how diseases may progress in a given population. And – crucially – the effect that different interventions may have on the spread. With the right models, we can start to answer questions like “how big will the outbreak be?”, “how will it develop?”, and “how could we control it?”.
These models, and the statistical tools that underpin them, are highly complex. In fact, to most of us, they’re something of a black box. Many sources of data are used in mathematical modelling, with some forms of model requiring vastly more data than others. And in reality, many models aren’t complex enough to capture the multifarious subtleties of disease outbreaks. Some are based upon assumptions which could turn out to be wrong. Others produce vastly different pictures. But they’re still helpful.
This is because modelling, when done well, allows us to test a variety of possible strategies in computer simulations before applying them in reality. Scientists spend significant amounts of time building, testing, comparing, and revising models. And more time still interpreting and discussing their implications.
“All models are wrong, but some are useful.”
Insights drawn from models can be used to inform real-life decisions, including government policy. The reason we’re all social distancing is because models predict that this would massively reduce the spread of COVID-19.
Even so, there’s a lot that models don’t capture. They cannot anticipate the development of an antiviral treatment that reduces the need for hospital beds, for example. But that’s the nature of modelling. They are always complex, often flawed, and sometimes useful.
This means that models get argued about. A lot. And particularly when they have significant implications, as with models of COVID-19. But understanding what models are for, and their limitations, makes them a vital tool in our response to pandemics.