Melbourne’s coronavirus second wave seemed impossible to defeat. Our modelling showed exactly...

Melbourne’s coronavirus second wave seemed impossible to defeat. Our modelling showed exactly how to do it


In August, our research team began working to help the Victorian Department of Health and Human Services model the likely outcome of various policies designed to control the progress of COVID-19 across our great state.

The idea was to devise an exit strategy that first got the numbers low, and by then keeping them low, enable Victorians to enjoy economic and social benefits now being withdrawn again across Europe and the US amid soaring case numbers and fresh lockdowns. We wanted a COVID-normal summer and especially a COVID-normal Christmas.

After modifying a model we had previously developed for Australia and New Zealand, we set to work. Many more ideas, inputs, iterations, tests, scenarios, assumptions, weekends, late nights, early mornings, bug fixes, and thousands of runs later, the results were ready. And on September 6, they were released to the public.

We estimated that with a prescribed set of restrictions, Victoria could achieve a 14-day reported average of less than five cases per day on or around October 26. In conjunction with already planned returns to schools and workplaces for earlier in October, this would enable further easing of restrictions and plot a pathway back to relative normality.

Victoria bravely rose to the challenge, stuck to the plan, and as it turned out, on the October 26, actual case numbers matched our predicted numbers pretty much exactly. Right on cue.

At the time we constructed the model in August the Victorian reported 14-day rolling average was only just dropping off 200 cases per day, and the possibility of achieving less than five was savagely challenged. The targets were described as “a guess”, “unachievable” and “impossible”. Our team was described by business representatives and academic colleagues as “inexperienced”, “cherry-picked” and “opinionated non-experts”. No obvious flaws, however, were found in the model itself — only its conclusions.


Three reasons our team’s approach worked

Why did some academic colleagues feel that making a target of us (and thereby unleashing a wave of both amateur and professional online trolls) was reasonable, but that a target of less than five cases per day by October 26 was not?

The answer to that question holds important messages for science and its role in society.

Here is what our team had going for us.

Firstly, our team was unencumbered by financial contracts. There was no financial nor other employee-employer relationship that influenced us to deliver DHHS a “preferred” solution. On such critical work, the arms-length (or at least 1.5m) freedom to agree and disagree with government in the collaborative process of model building is paramount.

Secondly, our team is multi-disciplinary. Core members include public health physicians, epidemiologists, psychologists, mathematicians, computer scientists, economists and geographers. In addition, our group model building approach also engages further input from an even broader range of experts on the ground. Everyone’s input is equally regarded and equally challenged.

As illustrated by the nature of the criticism our group has received, knowledge creation in the health sciences — and recognition of the authority of the knowledge creators — reflects the hierarchical nature of institutional medicine where challenges to authority are rare. Surely this is the antithesis of an enquiry-based culture upon which good science depends?

Lastly, the computational approach we used — agent-based modelling — does not restrict what is possible in the future to only that which has occurred in the past. Has any other country demolished a second wave before? Who cares? Let’s go for it.

Just like a football team that has never won a premiership but believes it could, we emphasise that if a strategy is set and a process is followed, we have a great chance of a win. In this circumstance, the team was Victoria, the process was what we all just did for one another, and the win is getting our lives back.

A photo of people sitting inside in a Melbourne cafe.
Now we’re enjoying the win — a gradual return to normality.(ABC News: James Hancock)

The goal was a COVID-free Christmas

Could it have been done differently? What if DHHS relied on the advice of our teams’ detractors? What if team Victoria had not bravely stuck to the game plan and instead listened to the loudest voices?

Sure, restrictions could have been lifted in mid-September. All kids could have gone back to school, movement could have been less restricted, and more businesses could have opened-up again.

But would the combined effects of these shortcuts have moved us any closer to our goal? Closer to a COVID-free Christmas and Summer? The chance once again to travel around Australia or even the world?

We don’t think so.

In fact, inputting those plans into our model suggests they would have sparked a fresh, elongated, health and economic disaster.

After six weeks of Stage 3 and a minor opportunity to reach Stage 2 in October, we reckon Stage 4 lockdowns would again have been re-imposed sometime between next week and late December, triggered by rolling averages between 1400 and 20,000 cases per fortnight.

Scientific progress should be driven by both theory and observation. Observation alone is not enough if all it leaves us with is an ability to describe the past.

Simulation models of the kind our team used allowed policymakers to think through the future and provide a basis for better-informed real-world decision making.

All of us — including world-leading scientists — have learned something important this year. We have learned the value of simulation models as an additional contribution to science, and to public health.

More importantly, we have also learned that Victorians are stubborn, brave, resilient and that by working together as a team we can achieve the incredible — even the impossible.

Dr Jason Thompson is a Australian Research Council DECRA Fellow at the Transport Health and Urban Design Research Lab, at The University of Melbourne. Professor Rod McClure is the Dean of the Faculty of Medicine and Health at the University of New England. Dr Thompson will appear on ABC Radio Melbourne this afternoon after 5pm.

Source link