All these modifications require edits to parameters in the model, including, most importantly, the rate of transmission, or effective reproductive rate.
In an evolving outbreak where evidence on intervention effectiveness is still being built, this is a challenge, and models need ongoing adjustment and fine tuning, with real-time analysis of new daily data strengthening predictions.
So models aren’t going to tell us how many cases we will have on Sunday, or over the next week, but the more we understand the dynamics of this outbreak, the better we can see how interventions are expected to work.
If we understand this, and know where are case numbers are coming from, then we can anticipate what lies ahead; not the exact day or the exact number, but the trends to look for.
So what sits behind our stubborn daily new case numbers? It is the background community transmission rates that drive large clusters, and these in turn generate large numbers of cases quickly, and drive our daily case report numbers.
Large clusters in workplaces, for example, feed more cases back into the community before the cluster is identified and closed down. This amplification effect is what drives overall cases up even more quickly than widespread background community transmission, as one infectious person may infect a hundred or more – first their work colleagues, then all their close contacts, and so on.
When community transmission rates rise, so too does the likelihood of someone taking the virus to work. Large clusters in workplaces, public housing and residential aged care move from a risk to a reality.
We had the perfect storm in north-west Melbourne, with clusters quickly seeding other clusters, and a rapid escalation of the case numbers. This is now a Melbourne-wide outbreak, but remains concentrated in the north-west corridor where the second wave appears to have been seeded, showing how difficult it is to break this transmission cycle locally as well as more widely.
Restrictions suppress transmission by reducing the number of close contacts we each have, and the connections across households and into work or public places. The more restricted our movement, the less chance we have of getting the virus, and the smaller the risk we will share it if we already have it. This will be already happening, but we cannot see it yet because of the dynamics of testing and case reporting.
Our daily reports are dominated by the new cases linked to clusters for two reasons. First, clusters are where most new cases are likely to be generated. Second, we actively chase and test all close contacts of interrelated cases as part of cluster investigations and so are likely to have a more true count of case numbers.
Background community transmission case detection is likely to be less complete because it relies on people coming forward for testing, and we don’t test asymptomatic people unless they are close contacts of confirmed cases. In fact, we know we don’t have anything near complete capture because these are also where the mystery cases are found: people not linked to another confirmed case are proof we aren’t seeing all the links in the chain.
The good news is that restricting close contacts shuts down community transmission whether we can see it or not.
It is not surprising that we have yet to see a big dent in new case numbers, because there are still large numbers of cases being found that are linked to existing outbreaks, and these will mask any reductions in the background rate. But it is not the reduction of these community cases in themselves that will have the biggest impact, it is the flow-on effects that will be most important.
Suppression of community transmission will help close off existing outbreaks linked to workplaces and aged care sooner (fewer close contacts for each new case identified, so less people at risk). More importantly, pushing background prevalence down will reduce the chance of new clusters being seeded that would otherwise replace the number of cases still being generated within existing clusters (the status quo we were locked in over previous weeks).
Add these dynamics that reduce new infections together, and we will see rapid reduction in cases once we get on top of the last of the large outbreaks. We are not quite there yet, but watch the reports for news on existing clusters, and listen out for any news of new clusters, and you will have a better idea of how we are tracking beyond the daily numbers.
Catherine Bennett is chair in epidemiology at Deakin University.
Get our Morning & Evening Edition newsletters