The estimated likelihood of eliminating the SARS-CoV-2 pandemic in Australia and New Zealand under current public health policy settings
On April 10th, I wrote about issues we had observed in public policy discussions regarding the possibility of eliminating COVID-19 from Australia. We asked (because we were perplexed), “Why is elimination of COVID-19 in Australia apparently off the table?” and came up with a small set of not very many good reasons.
On the 10th I also provided a preview of some earlier-stage modelling of ours at the Transport Health and Urban Design Research Hub at the University of Melbourne, that suggested if we stick to the current plan, we might eliminate COVID-19 by mid to late June (June 19th as an ‘average’ date with the start of August being an upper (95%) confidence interval estimate). Our primary goal with this work was to show that a ‘pattern’ of elimination was at least possible in Australia and I’m glad to see that with each day it appears to be becoming more likely. In fact, we might almost achieve it ‘accidentally’.
In the 2 weeks since that post, we have seen 3 things happen. Firstly, numbers of new infections have continued to decline – Great! What we are doing is working.
Secondly, discussion of elimination is back! Which is also heartening because it is a real possibility and we can be proud that Australia and New Zealand really are leading the way, internationally.
But thirdly, as case numbers dwindle, we are also seeing increasing pressure to begin lifting restrictions, which will only get louder the lower numbers drop. Recent Google trends data in Australia confirms this (in a pretty ropey way) by showing that searches for ‘lift restrictions’ and ‘ease restrictions’ are on the way up.
So, 1) cases going down, 2) pressure to lift restrictions going up and 3) actual restrictions also being loosened. This is potentially cause for concern. The problem is, we have no idea how much of a concern it really is – and we possibly won’t know until it goes wrong. And then what? Put the restrictions back in place? This is the on-again, off again ‘Goldilocks’ strategy which would be extremely difficult for any Government to communicate and manage well.
To help answer this question, colleagues from Australia and New Zealand and I have drafted a paper with updated estimates from those previewed here on April 10th. Titled “The estimated likelihood of eliminating the SARS-CoV-2 pandemic in Australia and New Zealand under current public health policy settings” we attempt to show that despite our successes to date, Australia (and New Zealand) are still on a knife-edge when it comes to the elimination of COVID-19 from the population. We also show that dissolution of our current successful behavioural strategy over the 60 days post social distancing implementation could result in the dreaded ‘second-wave’ of infections, cause avoidable death and illness, and prolong the economic and social shutdown even longer. You can see a draft of the paper on Elsevier’s SSRN pre-print site here.
In this paper, we use an SEIR-inspired Agent-Based Model I have discussed previously under 2 scenarios. Scenario 1 locks-in the current social distancing policies both NZ and Australia put in place on March 26th and 28th, respectively, and doesn’t lift them until the virus is eliminated from the community (defined as no new cases and no current, non-recovered cases).
Scenario 2 implements the same policies on the same dates as above, but gradually eases off on the adherence to these policies in a non-linear way over 60 days, returning to baseline levels on May 26th/28th. Here’s a picture of the decay function we used – there are others that we and others also think could be realistic.
Here is a snapshot of our results.
Under Scenario 1 where no reduction of social distancing measures happens (e.g., no ‘easing restrictions’ during the 60 days), we estimate that the average date Australia and NZ might eliminate COVID-19 from the population is June 15th and June 17th, respectively, with a 95% confidence interval of between May 13th and July 30th for Australia and May 28th to August 7th for NZ.
Further, looking at thousands of trials of our model, we show that 90% of model runs acheived elimination on July 10th in Australia and 90% for NZ on July 14th.
So – 7 weeks from now, if we just sit tight and keep doing what we’re doing so well, we have a better than even chance of ridding COVID-19 from our shores. And 12 weeks from now, we have a 90% chance.
But what of Scenario 2 – where everyone starts to relax a bit; just ease off a little until the end of May, then – phew! Well, the answer is not good. Under this scenario, we estimate that Australia has only a 35% chance of elimination at all in the next 300 days, and New Zealand a similarly disappointing 36%. The pandemic drags on and on, well into next year and beyond. No social gatherings for you.
So what?
Our results highlight that in Australia and New Zealand, Government imposed public health responses and each country’s public adherence to these, has provided the potential to eliminate the COVID-19 infection from the community if these measures and levels of adherence are maintained.
However, reductions in adherence to physical distancing over the post-implementation period – even slightly, could lead to a resurgence in cases, and worse health, economic and social outcomes in the long-term – especially in the absence of effective and publicly accepted measures to improve contact tracing (e.g., contact tracing apps*) which we assess the benefits of here.
Now is not the time to ease off.
*At 6pm on Sunday 26th of April, the Australian Federal Government will make available a tracing app to augment existing state-based manual tracking and tracing. We have incorporated the uptake of this app and its effectiveness in boosting identification of cases into our work here.
A lot of people have also contacted me asking about the effect of schools re-opening – especially in NSW. You can see some example results of how this might work here.
HI. I know you are all busy but can I clarify if there has been any any modelling done for the return to school plan that was announced in NSW this week to understand how the plan may affect transmission and infection rates in kids and/or teachers?
Hi Natasha – we are currently working on a version that will incorporate school-aged return as far as possible. This is a somewhat tricky thing to do but we have a plan and will update the blog as soon as we have something reasonable.
What about our defence personnel? Are they able to enter and leave our shores as they wish? And millionaires seem to be getting special treatment too. Are air crews still exempt from quarantine? I don’t mean self quarantine, I mean the hotel approach.
Not sure I’m the best person to ask, Graeme.
There is a significant social, physical, mental and economic cost to your plan. With infectious diseases the virus will die out in your community when the R value is below 1. It is below 1 now meaning it is very unlikely to be able to spread given proper public health measures in identification and isolation.
For the rest of the world the R value has been about 3-3.25 so this means to get herd immunity (the community will be able to handle the virus) is approximately 70% of the population. No population is going to allow this to happen as health resources will bust and many many people will die. So the rest of the world will be also managing by spot cases and isolated
Until and if we get a vaccine
Hi Geoff – Then I’m not sure you’ve understood what I’ve written.
This is great Jase. Thank you very much.
Thanks for your blogs. Im particularly interested in elimination yet this doesn’t seem to be where state or federal governments are heading. It feels like they’ve thrown in the towel and are accepting that we will just have to live with the threat of the virus circulating ad infinitum. Now that the Covid-19 tracing app has been released, I’m wondering how this will contribute to the elimination models, or not.
Hi K – We have included a contact tracing app function in the model so we can see what sort of difference it might make. To cut a long story short, the tracing app in the absence of other measures will probably not work, whereas the other distancing measures will work even without the tracing app. So the purpose of the tracing app is really to enable faster opening up of activity in the hope that its benefits in reducing tracing time (and hence spread of and to unsuspecting people either in the incubation period or who are asymptomatic) off-set the increase in cases that a reduction in physical distancing might have – this is quite a leap of faith in my opinion. We know one strategy works. We don’t know the other one does. Why would we try to trade the one that works for the one we don’t know works? People will say that it’s not either / or but the message we are getting is that they are looking for a techno-fix in order to open up faster. This is just my current opinion – which might change!
According to your model, we can eliminated Covid 19 in Australia and New Zealand, which would be great. If we did that we would have freedom in those countries, but I assume that anyone entering Australia or New Zealand from another country would need to be quarantined for 14 days. That may be forever if no reliable vaccine is developed. Do you think that this would be accepted by the community and/or the government and what sort of economic impact would this have?
Hi Bron – Thanks for the question. My thoughts would be ‘yes’, especially given that anyone entering the country would still probably need to be quarantined, anyway, so there is really no difference in this regard in either scenario. A lot of people have mentioned to me that we should be heading toward herd immunity, but the current strategy of about 15 cases / day would take about 2500 years before we even reach 60% of infections so the alternative is not going to happen unless Government purposefully ramps up infection rates. The Grattan Institute has modelled the impact on GDP of this scenario if you search their blog.