What can Australia learn from New Zealand’s coronavirus modelling?

Following some spectacular question-dodging by Prime Minister Scott Morrison, deputy chief medical officer Paul Kelly announced modelling on Australian coronavirus cases would be “unlocked” later this week. 

New Zealand’s modelling has already been released, with one report thanking “our Australian colleagues for their valuable work in providing modelling data on numbers of infected cases for different scenarios”. 

But researchers say while estimated case rates match other modelling, New Zealand’s modelling is preliminary at best and excludes a number of necessary variables to be applied to Australia.

In their “plan for” scenario, 65% of New Zealand’s population would be infected, with 34% of the population developing symptoms. Deaths would range from between 12,600 to 33,600 people.

Tom Kompas, a foundation director of the Australian Centre for Biosecurity and Environmental Economics and chief investigator in the Centre of Excellence for Biosecurity Risk Analysis, told Crikey the modelling needed more data and variables. 

“It uses a basic supervision risk and intensity model,” he said. “As more data comes in, you can make it more elaborate.”

But where it currently stands, Kompas said, lacks depth. 

“The way in which you validate parameters and infectious diseases change results if you don’t have the data,” he said. 

Frustrated with the secrecy from the government, Kompas is currently working with a team to develop a model applicable to Australia. 

“Ours is a bit more comprehensive,” he said. It uses publicly-available data and is set to be released within the coming days.  

Dr Mahmoud Elkhodr, information communication technology lecturer at Central Queensland University has previously used artificial intelliegence (AI) predictive analysis to model coronavirus spread. He told Crikey it was difficult to apply New Zealand’s modelling to Australia. 

“Is it applicable? Impossible to tell … there are a lot of variables involved. Some were not even considered in the report.”

Transmission sources, he said, were a key driver in Australian cases, with those arriving from overseas isolated. Border controls are mentioned in one New Zealand model, though it focuses on travel restrictions over quarantine.

Similarly, there’s little modelling on what effect introducing new medicines or vaccines, and when, would have on the ICU rate. Australia’s rate of testing is significantly higher than New Zealand’s too, which would also effect the model, Elkhodr said.

“The next two weeks are very crucial in determining where we are heading. We will either follow the UK curve and US curve — the worst-case scenario, which is very unlikely — or South Korea’s curve — the best-case scenario. The next few days will determine the path we are taking.”

Ben Phillips, associate professor in ecology and evolutionary biology at the University of Melbourne, had developed an app to model the spread of coronavirus. He told Crikey forecasting the spread was difficult but important.

“The key question right now is how public health measures affect the rate of the virus … This week will be a massive decider,” he said.

Phillips stressed the importance of the government releasing data and modelling.

“It facilitates researchers being able to help … not releasing it limits the capacity for people to collaborate,” he said.