Earlier this week, WA Premier Mark McGowan said modelling indicated coronavirus would be spreading in Australia in April or May, and the peak of the epidemic would happen in August.
But where do these predictions come from, and how likely are they?
Around the world, including in Australia, pandemic forecasters are working frantically to predict how COVID-19 will play out.
Here’s what they can and can’t tell us, so far.
How to predict an unknown enemy
Researchers’ understanding of the new coronavirus, SARS-CoV-2, has exploded since it first emerged as a mystery illness in Wuhan, China, in December 2019.
We now have a better idea of how it spreads and who’s most at risk, and new research about the virus is being published almost daily. But there’s still so much we don’t know.
James McCaw, a University of Melbourne epidemiologist who is part of the team producing models of the potential pandemic for the Australian Government, says models are only as good as the information you can put into them.
At the moment, the lack of knowledge about the new coronavirus makes it very hard to nail down exactly what an outbreak in Australia will look like.
Two factors affect how a virus spreads: its generation time (the time between when one person becomes infected and when they infect someone else) and its reproduction number or R0, (loosely, the average number of people an infected person goes on to infect).
These influence what the most effective control measures might be, but researchers are still unpicking how these two factors work together in the COVID-19 outbreak.
“With the early epidemic data, it still remains compatible with a long generation time and a high R0, or a short generation time and low R0. And they influence how effective a control measure might be,” Professor McCaw said.
“We’re doing work on it, but we still can’t make any of those really crisp predictions.”
This means at this stage Professor McCaw can’t put figures on numbers of people who are likely to be infected, become seriously ill or die from the disease.
But even though the numbers are spongy at the moment, what we do know can still help us prepare.
What do we know?
Past pandemics and the emerging information about this specific virus all inform the plausible scenarios people like Professor McCaw are producing.
He says, based on the increase in cases in New South Wales and elsewhere, it seems we’re on the verge of sustained transmission — that is, when a virus is spreading throughout the community.
“Our models would suggest that if we are unable to definitively contain COVID-19 in Australia and the world, it will peak in the next two to four months,” he said.
How we react to an outbreak here affects when that peak of cases would come, and how high it is.
The worst-case scenario modelled by Professor McCaw and his colleagues has a very high peak of cases — meaning a large number of people would all be sick at the same time.
“[That scenario is] highly unrealistic because it’s the idea that we don’t change our behaviour, we don’t have any control response strategies. And of course, we will. And they should make a significant difference to how this disease spreads.”
Actions like introducing social distancing measures when the disease starts spreading in Australia will actually prolong the epidemic.
While this seems counterintuitive, it makes for a better outcome overall.
“It will extend the amount of time in which there is transmission activity within the population, but the overall magnitude of and burden on the health system will be far reduced. By slowing it down, the burden at any point in time will be lower and we will save lives.
“Fewer people overall [will get infected], and for those who are ill, however many that is, a lower burden on the health system will allow for better care for those people.”
The Government’s emergency response plan, which was put into effect last week, flags cancelling large events, closing childcare centres and locking down aged care homes as potential actions.
Professor McCaw said these large-scale, centralised control efforts do make a difference in the models.
But some of the most dramatic differences come from small actions people can take in their everyday lives: hand washing, practicing cough etiquette, staying away from others if you become sick.
“I’m always concerned that people will think that once it’s going, there’s nothing you can do about it, but that couldn’t be further from the truth,” he said.
“What the modelling says is that if everyone can make small changes to their daily routines, washing hands, not shaking hands, and being aware of when you might come into contact with infectious people, or if you were infectious, how you might reduce your level of transmission onwards … [that] can result in huge public health benefits.”
How long will a coronavirus pandemic last?
If the number of people searching online for a topic is anything to go by, one of the questions on many people’s minds at the moment is “how long will a pandemic last?”
Again, this isn’t something we can put a number or a date on yet. But Professor McCaw’s models hold some clues.
He said it was likely COVID-19 would become a permanent, seasonal disease in humans after this initial epidemic.
“Just as for [2009 swine flu], the virus will cause a large initial epidemic, perhaps followed by subsequent waves of infection, and then reduce to low levels,” he said.
“But it is unlikely to truly disappear, just like seasonal influenza doesn’t truly disappear each year.
“This is different to SARS — which we truly eliminated because we successfully controlled it before it could fully establish itself in the human population.”
Quick or precise: Choose one
Models for disease forecasting can range from very simple extrapolations of the current figures through to incredibly detailed models that simulate every single person in the population.
The simpler models are faster to produce than the detailed models, but they’re less accurate, and this is the line the Government is walking with its modelling at the moment.
And of course, these more detailed models are just as constrained by limited information about the virus itself.
Mikhail Prokopenko from the University of Sydney also models pandemics, using census and other data to essentially recreate every member of Australia’s population in digital form and see different ways a pandemic might play out.
His team’s models will show how the disease might spread around Australia geographically over time, what proportion of the population is likely to be infected and on what day of the epidemic the peak would come. But even preliminary results from his forecasts are still weeks away.
“There are so many moving parts, we need to take care in representing those,” Professor Prokopenko said.