On-call pandemic response

Sleeping with his mobile phone next to his bed during the pandemic has become the norm for Professor George Milne, who has been woken up early by the Health Department of WA asking him to run worst-case scenarios through UWA’s COVID-19 virtual world computer disease model for Perth.

These scenarios included the introduction of infectious cases due to the sudden arrival of an infected cruise ship, failure to quarantine and a spike when social distancing is wound back and borders reopened.

"One scenario we analysed was if a plane-load of people landed in Perth airport and half of them were infectious, they were not quarantined and they all managed to “escape” into the community. How would WA cope with that?"

Professor George Milne

“These are experiments you just cannot run in the real world.”

Other requests have been to analyse worst-case scenarios to calculate the demand for ICU beds and ventilators.

The turnaround time for running new scenarios through the UWA computer model is as little as 12 hours — so if his mobile rings at 7am, by 7pm his team has simulated what would have happened in Perth and produced the new epidemic curves and reports.

Engaged by the Health Department of WA to assist the COVID-19 healthcare demand team, Professor Milne’s research group evaluated the effectiveness of alternative social distancing strategies to deal with the scenario using a city model the size of Newcastle in New South Wales, with results multiplied by 8 to be scaled up for Perth’s population of 2.2 million.

“Using this, we can very quickly look at scenarios in advance and we can work out what measures are needed to contain spread of the virus,” Professor Milne said.

“For example, if we were to get a runaway outbreak as is occurring in Victoria now, we are able to model that scenario and determine how long it would take to control, and what resources would need to be put in place.

“We are able to take current case data from the Health Department, input it to the computer model and then run the model to predict where we would be, in terms of daily cases, in a week, in two weeks, a month, six months and 12 months. Then we do it with an adjustment of social distancing measures in place and see what the difference is. This allows us to generate quite different epidemic curves and raw numbers on how many people we predict would be infected in WA, including those who are asymptomatic.

“With these numbers, we can calculate hospitalisation rates and from that ICU demand and deaths as well.

“At times during the pandemic, our UWA team was put under pressure but we rose to the challenge and we were fortunate that we had the expertise and the technology to make a contribution.”

This story is part of the series Computer modelling COVID-19 spread critical to flattening the curve.