Steps out will be incremental, frustrating, even arbitrary

We have the magic numbers, the average number of new cases over 14-day periods that we will need to see to take the first steps out of stage four restrictions in Melbourne, or for regional Victoria to step down from stage three. Why 14 days? No one has explained this, but it makes this an even more conservative approach. So let’s look behind these numbers to see where they have come from, the modelling used to generate these thresholds and, even more importantly, the evidence behind the steps out.

Premier Daniel Andrews has unveiled the state’s path out of stage four restrictions.Credit:Luis Enrique Ascui

The government asked modellers to look at four different scenarios for average daily numbers to show what might happen after opening up: three, five, 10 and 25 cases a day. The simulations based on opening up with 25 cases still being notified on average per day suggest a six in 10 chance of a case resurgence before Christmas. Not good.

Statistical modelling can be a valuable decision-making tool, allowing you to test various scenarios and navigate the road ahead with some degree of foresight. But how useful it is depends on the insight we have into epidemic dynamics, how these are incorporated into the model and the accuracy of the parameters that underpin the model.

The model used here is based on international estimates for key parameters, including time to case reporting, infection periods, etc, and is reasonable enough. It builds in assumptions of compliance with mask wearing and how many people are essential workers, etc. The model then computes community transmission scenarios and spits out different forecasts depending on how the dice fall, say, whether that new case is an essential worker and still going to work or not. We then look at how the projected futures compare across simulation runs.

The model does not try to distinguish between those cases linked to known, contained clusters versus wider community transmission and in particular those "mystery cases". This is particularly important in Victoria as the chief health officer tells us that one third of new cases are in workers from health or aged care, and another third are their close contacts.

But the modelling itself is not the main issue here: the numbers sound quite sensible and not that different from the ideas being floated by experts these past few weeks. What is in question then is how they are tied to the actual steps, and what evidence has been used to compile these.

We have rich data from the second wave in Victoria. Most importantly it is our data, telling us how the virus has spread through our communities, and where our weak points are that will help us to map our way out of this wave and avoid the next. The Premier has said many times he is data driven, so I had assumed, perhaps naively, that this means comprehensive analysis of our own data first and foremost.

This data can tell us whether certain workplaces are more risky than others, how often people inadvertently attended work while infectious and passed the virus to their co-workers or clients, and how this varies across different types of industry or businesses. We can also examine how well various interventions have worked – did clusters associated with social gatherings drop away after the curfew, has the number of close contacts who test positive decreased since masks were introduced in regional Victoria?

The Premier has placed great stock in the modelling, so much so that it leaves no doubt in his mind about the road ahead. Yet the rest of the detail under the various steps is equally critical, and this comes down to the advice of the public health team. Naturally on Sunday we were all very keen to hear about what we have learnt from the Victorian data analyses, but there was nothing. I was left feeling extremely disappointed.

How can we not be using our own data to plan our road out? Or, if we are, then why was this not front and centre in the announcement? It was not even referred to when questions were asked from the press gallery.

Have we had outbreaks associated with gyms in Victoria, and are they really different from hairdressing salons when it comes down to risk? The chief health officer spoke of “personal attachment” when questioned on this, not the calculated risk based on our analysis of our 17,000 cases.

What really confuses me most is that the health department does not seem to have learnt the big lessons from our own experience. Stage three worked in the first wave, and our exit plan also worked then, though some still claim it didn’t despite the second wave being attributable to a new introduction of the virus. Stage three also turned around the second wave, and with masks started to bring case numbers down. Their own report says while slow, cases would halve each 49 days.

This is not about “pretending this is over because we want it to be”, or wanting to ease restrictions ahead of time, this is about using our data to inform tailored, precise, evidence-based responses. We may have missed that boat now for these next steps, but it is not too late for the government to analyse our second wave data so that we can know in more detail about our vulnerabilities as well as what worked, and can mount a strong, tailored prevention strategy going ahead.

We have done an amazing job turning this wave around. We are down to manageable low numbers of new cases and most are known and contained given they are linked to workplace clusters. Our steps out will be incremental, frustrating, and some may even seem arbitrary, but the road map does not tie us to these dates. If we can move through these steps faster than the road maps suggest, this might just help save our sanity, and our economy.

Catherine Bennett is chair in epidemiology at Deakin University.

Get our Morning & Evening Edition newsletters

The most important news, analysis and insights delivered to your inbox at the start and end of each day. Sign up to The Sydney Morning Herald’s newsletter here, to The Age’s newsletter here and Brisbane Times‘ here.

Most Viewed in National

Source: Read Full Article