Sat. May 28th, 2022

Before Australia’s latest changes to the COVID test, it was basically straightforward to figure out when we would reach the top of cases.

We looked at the number of new daily cases, diagnosed via PCR. From there, we developed a number of other key indicators related to COVID proliferation, testing and hospitalization – each depending on these daily case numbers.

However, we have seen a huge increase in cases recently as people test positive using rapid antigen tests, especially as reporting their results to the state health authorities is now possible and becomes mandatory.

So it takes a few days before we can measure some key figures with any degree of accuracy. Only then will be able to say with confidence when we have hit the top and come down on the other side.

1. The number of new daily cases

Most people would now have seen an epidemic curve. It is a plot of the number of new cases of COVID-19 diagnosed each day. Here is the current epidemic curve for New South Wales.

A line graph showing the epidemic curve for cases of COVID-19 in NSW
Epidemic curve for NSW. Note the irregular case figures in recent days.(Delivered by: Adrian Esterman)

As for the date, states and territories use different cut-off times to define a 24-hour period. As the authorities conduct investigations, the date of some cases may change. So do we plot the daily announced case numbers, or the “true” case number after changes?

It sounds complicated, but even more complicated is trying to define a case.

Before rapid antigen tests became available to the public for home use, cases were diagnosed from positive PCR tests.

So due to huge queues at PCR test centers and many people, even those with symptoms who gave up and were not tested, our test system changed.

The National Cabinet agreed to remove the requirement for a PCR test to confirm a positive result of a rapid antigen test.

As most states and territories are moving in the direction of reporting both positive PCR tests and positive rapid antigen tests, we still need to smooth out the bumps in the data. Potentially someone could get both tests and be included twice!

The uncertainty in case numbers also affects other key parameters we use to monitor the current wave.

2. Reff

The effective reproduction rate (Reff) is a measure of how many other people on average infect each case. We want it to come below 1 to stop an outbreak. At its simplest, Reff’s case number is divided by the case number four days ago.

Since we currently have so many problems defining and counting case numbers, it will take a few days before we can consistently interpret Reff for each state and territory again.

Percentage of positive tests

This is the percentage of positive tests out of all COVID-19 tests taken. This is an important goal as it gives an indication of the amount of undiagnosed cases in the community.

The World Health Organization suggests that if it is below 5 percent, things are under control.

When the diagnosis was only by PCR test, we had good data on both the number of tests and the number of positive ones.

Now states and territories are going to report fast antigen test results, it’s not that straightforward.

Some jurisdictions like Queensland only ask you to report a positive result. This means that we no longer know how many samples were taken. SA Health encourages people to also report negative tests, which is a much better system.

Space to play or pause, M to mute, left and right arrows to search, up and down arrows for volume.

Play video.  Duration: 6 minutes 19 seconds

How to take a quick antigen test

4. Number of inpatients

As Australia opens up, we have been asked to pay more attention to COVID-19 hospital admissions instead of just the case numbers. But even that gets complicated.

It is clear that if someone tests positive for COVID-19 and then gets admitted to the hospital, they are an inpatient case. But what if they are admitted as a likely case?

And should the number of admissions include people managed in a hospital-at-home-type scheme? After all, they are still consuming hospital resources.

What if they were admitted for something else but subsequently diagnosed with COVID-19 in the hospital?

Even harder is trying to calculate rate of COVID-19 admission. This is the number of people in the hospital with COVID-19 divided by the number of people diagnosed. But you have to decide what time periods you are talking about, a completely different debate.

There are similar problems in measuring the number and frequency of people in intensive care.

How do these changes affect modeling?

NSW Health recently released modeling to look at what lies ahead.

With the current restrictions in place in NSW, the modeling shows a peak of 4,700 admissions, with 273 on intensive care in mid to late January.

A patient in the intensive care unit at Footscray Hospital, used with permission from the patient's family
As Australia opens up, we have been asked to pay more attention to COVID-19 admissions.(Delivered by: Penny Stephens / Western Health)

It is unclear whether changes in the test rules have been included in the modeling. However, it is understood that although the detection rate changes significantly, it does not affect any projection of when the peak will be reached so much.

Modeling will therefore probably still be reasonably accurate despite the changes in COVID testing. This is good news for other states and territories that rely on modeling results for planning.

Where to go from here?

A good start would be to have mandatory reporting of rapid antigen test results, both positive and negative. That way, we can calculate the percentage of positive tests again.

The United Kingdom has a good system. Once you have taken a quick antigen test there, scan a QR code on the package and report the test results as positive, negative or invalid to a central government database.

Importantly, let’s have one national body responsible for defining, collecting and reporting COVID-19 statistics. It could be the Australian Institute of Health and Welfare. Even better would be to have our own center for disease control, which people like myself have been calling for for a long time.

Chris Billington, from the University of Melbourne, contributed to the section on modeling.

Adrian Esterman is Professor of Biostatistics and Epidemiology at the University of South Australia. This piece first appeared on The Conversation.


Leave a Reply

Your email address will not be published.