Covid infection numbers – what do they mean now?

5 minute read


Newer subvariants are coming, while testing and reporting rates have plummeted.


There has been chatter lately about early warning signs of an uptick in covid infections in Australia, but experts say the unreliability of the data makes it difficult to know what’s really happening.

As of 28 February, the number of new daily infections nationally – averaged over the previous seven days – was 2646 cases, up 1.1% from the previous week.

There’s considerable variation between the states and territories: infection rates in Tasmania were up nearly 9% on the previous week, in NSW they were up 2.3% and in Queensland they were up 5.1%, but in South Australia they were down by 5% and in the Northern Territory new infections were down by nearly 12%.

NSW Health’s latest report on covid surveillance notes that case numbers are low, but warns of a possible increase in the coming weeks as newer subvariants dominate. In NSW, there are several variants circulating, but the Omicron subvariants CH.1.1 and XBB – which carry mutations that make them both more transmissible and less susceptible to pre-existing immunity – are rising in prevalence.

Epidemiologist Professor Adrian Esterman says over the past three weeks, the effective reproduction number – the R0 – has been 1.01 or 1.02.

“That’s telling me that we’re starting to see a wave but it looks like it’s a very slow, gentle wave,” says Professor Esterman, professor of biostatistics at the University of South Australia in Adelaide.

But with testing and reporting of positive results now voluntary, it’s difficult to know just how accurate the daily new infection figures are.

Epidemiologist Professor Catherine Bennett estimates that fewer than half, and perhaps as few as 10% of covid infections, are actually being reported to health authorities.

“So [while] it’s showing us the trends, you just can’t believe the real numbers,” says Professor Bennett, chair in epidemiology at Deakin University in Melbourne.

Another possible indicator of what’s happening with infection rates is the percentage of PCR tests that are positive.

According to the World Health Organization, a test-positivity rate below 5% indicates that an outbreak is under control, but Professor Esterman says Australia’s test positivity rate currently stands closer to 10%.

“That tells us there’s an awful lot of covid out there that’s simply not being picked up,” he says.

Rapid antigen and PCR tests aren’t the only indicators of what covid is doing in the community.

Hospitalisations, intensive care admissions and deaths from covid have served as a measure of the prevalence and severity of covid, and they are more reliable than the “crude signal” of daily infection counts, says epidemiologist Associate Professor Hassan Vally, from Deakin University in Melbourne.

“You look at the hospitalisations and they’re low, and you look at the ICUs and ventilations and they’re all trending, and have been trending, the same way,” Professor Vally says.

On 28 February, just two deaths from covid were reported in Australia, down from the most recent peak of 54 as a rolling seven-day average at the start of this year.

“It’s consistent with what we would expect in a highly immune population,” Professor Vally says.

Another measure being used to track covid in the population is sewage monitoring. While the concentration of SARS-CoV-2 genes in sewage fluctuates in response to environmental conditions such as rain, the overall trend can give an indication of how much virus is in the community.

And according to NSW Health’s latest data, those concentrations have stabilised at “low levels” in recent weeks.

While all these measures can suggest trends, they’re far from an accurate portrait of covid prevalence.

In the United Kingdom, for example, there is a randomised population-level survey program that collects up to 120,000 blood samples and 227,300 swab test results every month from individuals aged two years and older across the UK to determine the prevalence of covid.

Such randomised sampling avoids the selection bias problem of voluntary testing, in which only symptomatic or concerned individuals go to the effort to test and report their results, and so gives a much more representative picture of infections across the entire community.

“We should have had that all the way through,” says Professor Bennett. She and others would like to see Australia implement similar surveillance to monitor for changes in the epidemiology of infection and disease.

“You’re looking for rapid changes in infection rates, rapid changes in severity or deaths associated with it, or a change in the epidemiology in terms of how it’s transmissible or whether particular sub-populations are becoming more vulnerable,” she says.

Professor Esterman is concerned that without an accurate measure of infections in the population, it’s difficult to predict or plan for the impact of long covid and other chronic conditions that could be precipitated by infection.

“Ten percent of them are going to end up with long term health problems, and at the moment we have very few services available for these people,” he says.

End of content

No more pages to load

Log In Register ×