It was every state for itself as Australia’s state governments bickered over coronavirus border restrictions. Then, competition emerged over who was doing a better job of catching the virus in the community.
In a media conference on May 14, NSW Premier Gladys Berejiklian said she was aware of international reports that “NSW has the highest rate of testing in the world”. “It’s something we’re incredibly proud of, something we need to maintain,” she said.
Soon after, Victoria staked a similar claim.
On May 21, the state’s Health Minister, Jenny Mikakos, tweeted Victoria had “the highest per capita testing rate in Australia”.
Get Crikey FREE to your inbox every weekday morning with the Crikey Worm.
This was “one of the highest” rates in the world, she continued, adding a chart that suggested Victoria was second only to Iceland.
So, did NSW lead the world with its testing rate, and was Victoria’s rate among the highest?
Berejiklian’s claim is overblown, while Mikakos’s claim amounts to a fair call.
When the NSW premier made her claim (May 14), NSW ranked 22nd in the world for its cumulative testing rate, and 20th for its daily rate.
Victoria, meanwhile, was at 18th place in cumulative terms and had cracked the top three for daily tests.
Notably, NSW was counting the number of people tested, while Victoria was counting tests conducted.
Were both reporting the same thing, they may have placed roughly equal in cumulative terms.
In daily terms, though, Victoria had already doubled NSW’s test rate. It was soon leading the nation on both measures.
By the time Mikakos made her claim (May 21), Victoria’s total tests per thousand people was the 13th highest globally.
Though not quite the second place her tweet implied, this rate was certainly among the world’s highest. Victoria’s rate for daily tests was fourth highest, behind Bahrain, the United Arab Emirates and Lithuania.
But such comparisons come with serious caveats. Not all countries count the same things or report as regularly, and it cannot be guaranteed that different test types have been excluded.
Moreover, data is not available for every country.
What type of testing?
Mikakos said Victoria’s testing was about “seeing the extent of the virus’ transmission in the community”.
Broadly, there are two types of testing for SARS-CoV-2, the virus behind COVID-19.
Serology tests can show whether someone has been exposed to the virus based on the presence of antibodies in their blood following an immune response. But as the World Health Organisation explains, these tests are often possible only after a patient has recovered.
Viral tests, meanwhile, detect the virus itself. We have assessed the claims on the basis of so-called PCR tests, which look for the presence of the virus’s genetic material in a nasal or throat swab.
Picking a rate
Neither Berejiklian nor Mikakos specified whether they were referring to tests per day or cumulative tests to date. We have assessed their claims on both measures.
John Carlin, a professor of biostatistics at the University of Melbourne, said that daily rates were “the only way in which we can keep a handle on the intensity of effort that’s going into monitoring the current rate of spread of the disease”.
He said cumulative rates showed the effort countries had put in over the longer term, “but this will have been spread in very different ways over time”.
Catherine Bennett, Deakin University’s chair in epidemiology, said it made little sense to look at daily rates over time without also tracking policy changes.
Cumulative rates helped avoid short-term data fluctuations, she said, adding that while these rates would ideally be standardised to a common starting point (such as days since each country’s 100th case), the differences from not doing so diminished as the pandemic went on.
Ramon Shaban, University of Sydney’s clinical chair of infection prevention and disease control, said the important thing was to have “a continuing high rate of testing that goes on in jurisdictions over time”, and it was this, combined with falling community transmission, that showed whether a country was actively looking for cases.
Different places, different counts
Both claimants compared their rates of “testing” to those of other countries. However, not all countries count the same thing.
Some report the number of tests conducted, others the number of people tested. It is often not clear which method they are using.
Since people can be tested more than once but must have had at least one test to be counted, we have considered the number of people tested to reflect a jurisdiction’s minimum possible test rate.
In Australia, where hospitalised coronavirus patients must meet certain conditions including the return of two negative tests before being released from isolation, states report the number of tests conducted.
NSW was an exception. At the time of both claims, it counted the number of people tested, not the number of tests.
The state switched to counting tests on May 26, but unlike the other states and territories its counts still exclude follow-up tests for confirmed cases.
Sourcing the data
Separately, Data NSW also publishes daily test counts for the state.
Rates per thousand people were calculated using the latestAustralian Bureau of Statistics population estimates, for September 2019.
Victoria takes the lead
Mikakos claimed on May 21 that Victoria had among the world’s highest test rates.
The numbers show South Australia led the nation in cumulative testing per capita until the week of May 10, with NSW just behind.
Then, Victoria leapt ahead — overtaking NSW on May 12, then South Australia on May 14.
By May 21, Victoria had conducted a total of 57.1 tests per thousand people, followed by NSW (48.2) and South Australia (48.0).
Victoria’s daily test rate averaged 1.9 per thousand people, nearly twice that of the runner-up, NSW.
What about NSW?
Berejiklian made her claim on May 14. In a statement issued that same day, NSW Health announced the state had “one of the highest testing rates in the world”.
However, the numbers show both Victoria and South Australia already had higher cumulative rates than NSW. Importantly, NSW was at that stage still counting the number of people tested, not the number of tests.
When it switched to counting tests on May 26, its total rate jumped by 8% — well above the previous week’s 2% daily average increase.
Had the switch happened earlier, a similar boost on May 14 would have put NSW roughly level with Victoria at the top for cumulative tests.
However, it would have done little to dent Victoria’s lead in the daily stakes, with the southern state already conducting, on average, twice as many daily tests per person as NSW.
The big, messy picture
Who has the world’s highest test rate is a vexed question. There is no standardised reporting of COVID-19 test data, nor are numbers published by intergovernmental bodies such as the World Health Organisation.
In the absence of such a dataset, experts directed us to the websites Worldometer and Our World in Data.
Cited in Mikakos’s tweet, Our World in Data is run by Oxford University researchers and collates test figures from official government sources.
Detailed information on these sources makes clear that countries report their data inconsistently, and there is often limited information about what is being counted.
Illustrating the problem, the British government has been accused of including tests “posted out” in its count.
We have based our assessment of the claims of Berejiklian and Mikakos primarily on the website’s total rates and seven-day average test rates.
Where rates are marked as relating to “samples tested” or “swabs tested”, we take them to mean tests that have been conducted.
Widening the frame
Worldometer publishes COVID-19 test data for more than 100 jurisdictions not covered by Our World in Data, including many smaller territories (for example, Gibraltar, the Falkland Islands and the Isle of Man).
It compiles its numbers from both official and unofficial sources, and does not specify what is counted.
We have added four countries to the list compiled by Our World in Data using figures from official sources. These countries are Brunei, Cyprus, Kuwait and the United Arab Emirates (UAE).
Each has populations larger than 100,000 people and were reported as having conducted more than 40 tests per thousand people, according to a Worldometer snapshot that used estimates up to May 21.
Two other countries also met this criteria (Mauritius and Malta), but we were unable to find official test counts for them.
Australian states versus the world
Berejiklian suggested NSW was leading the world in her May 14 comments. But its cumulative rate of 41.1 tests per thousand people didn’t crack the top 20.
Meanwhile, Victoria took 18th place with 43.7 tests per thousand people, according to our analysis.
By May 21, when Mikakos made her claim, NSW (48.2) had risen one place to 21st place and Victoria (57.1) had jumped to 13th place.
Mikakos’s tweet from that time showed only Iceland ahead of Victoria.
In fact, 12 countries outranked Victoria, and rates for several of them were based on days-old data.
Five of those countries counted the number of people tested; or, it was not clear what was counted.
The daily rates
When Berejiklian made her claim, Victoria had already blazed ahead of NSW with its daily test rate.
Based on averages for the week to May 14, Victoria was conducting 2.4 tests per day per thousand people. That was the third-highest rate globally, behind only by Bahrain (3.8) and the UAE (3.6).
NSW took 20th place with 1.1 people per thousand tested each day.
By May 21, Victoria (1.9) had slipped one place, behind Bahrain (4.3), UAE (4.0) and Lithuania (2.3).
NSW (1.0) had gained four spots to 16th place.
Principal researcher: David Campbell