While Australia was one of the earlier countries outside of China to record COVID-19 cases, by June we were congratulating ourselves, with apparent justification, for the success we had in flattening the curve.
We now know that we spoke too early and by late July we were detecting new cases in record numbers. Were there any signs in the readily available data that could have rung warning bells?
Source: For this and all the Charts below, the data is drawn from the daily updates provided by the Commonwealth Department of Health, Coronavirus (COVID-19) at a glance infographic collection (Available at: https://www.health.gov.au/resources/collections/coronavirus-covid-19-at-a-glance-infographic-collection)
Even as late as the last week in June, in contrast to what we were seeing elsewhere, our curve was looking encouragingly flat.
By then we had absorbed a new-found appreciation of logarithmic charts which also appeared more on the flat side relative to the shape of the charts recording the experience in many other countries.
Alongside the propagation of social distancing and the introduction of restrictions on gathering and social activities, we took pride in the extent of testing in Australia.
By early May we were averaging more than 20,000 tests a day and we kept ramping it up for another week peaking at an average of over 32,000 tests a day in the week ending 15 May.
Supported by an apparently flat curve, we took a spell from more testing and allowed the rate to ease, rise again, fall and then plateau over several weeks until we averaged slightly less than 24,000 in the week ending 12 June. Since then we again embarked on, and this time stuck with a strategy of more testing.
If a drop-off in testing was a failing, could we have known it at the time?
The national tally of new cases wouldn’t seem to suggest there was much to worry ourselves about. New cases remained pretty flat until the third week in June and by then we were still averaging only around 20 new cases a day.
There was still the chance that, as the President of the United States might point out, with less testing, we may have been picking up a smaller proportion of new cases.
However, at a national level, the rate of positive tests remained comfortably low until the third week in June in which the rate averaged 0.05%. This was well below the average rate of 0.086% recorded in the last week of April. The end of April rate was not exceeded until the last week in June and by that time the number of tests was on the rise again.
Were there better clues at the disaggregated level?
The two main players in Australia’s COVID-19 experience are the states of New South Wales and Victoria.
While the Victorian experience turned out very differently, for both states the number of new cases seemed to remain flat until the second half of June. And even then, Victoria averaged only 16.5 cases over the week ending 21 June. In the last week in June the average was a bit over 42 and it took off from there.
What of the rate of positive testing at a disaggregated level?
While, with the benefit of hindsight there may have been some earlier signs of the trouble ahead in a very close examination, still in the week ending 16 June the rate of positive testing in Victoria had not passed end-of April levels.
Were there more sensitive indicators?
One of the benefits of a federation – particularly one as homogenous as Australia’s – is the usefulness of comparative data. And here the story is more interesting.
The following chart shows the ratio of new cases in Victoria to new cases in NSW (7-day rolling averages). Victoria’s population is 82% the size of NSW’s and the dotted line (=0.82) is used demarcate between instances when the new detected cases per capita in Victoria is high or low relative to those picked up in NSW.
From very early in May, Victoria started to detect more cases than were being picked up in NSW. As these are 7-day rolling averages, this indicates that from late April/early May, Victoria started to identify greater numbers of cases per head of population than were being detected north of the Murray.
The same divergence appears in the rate of positive testing. From early in May Victoria was detecting more cases per test conducted than was the case in NSW.
Of course, hindsight is a wonderful thing. While neither the number of cases nor the rate of positive testing would have set alarm bells ringing, a simple comparison between the two states would have proved a more sensitive indicator.
Had there even been a tinkle of a warning at the start of May, some detailed examination of what was happening in Victoria would have been prudent.
At the very least, increased testing might have been on the cards. But that is not what happened and in fact, we went in the opposite direction.
The number of tests undertaken in Victoria peaked in the second week of May and was then wound back in fits and starts until well into the second week of June. It wasn’t until the end of June that it regained its earlier levels.
In NSW also, the foot was taken off the testing accelerator over much of this period.
But in the case of NSW, the early-warning signs were not pointing to the need for more tests.
In Victoria, by way of stark contrast, by late June when the number of Victorian tests was back at the sorts of levels seen in early May, the COVID cat was well and truly out of the bag and the second Australian wave was gathering pace.
