Watch now on .

University of Melbourne’s Professor Tony Blakely in ‘The Age’ and .

A , posted online this week by researchers at the Australian National University and elsewhere, estimates 71,000 Australians had COVID-19 by mid-July 鈥� 60,000 more than official number of cases diagnosed by that stage.

The study involved testing 2,991 elective surgery patients in ten hospitals across four states, to see whether they had antibodies against SARS-CoV-2, the virus that causes COVID-19.

The study initially found 41 positive patients (1.4%), but then adjusted for the false positives that would arise due to the imperfect specificity of the antibody test, which the researchers estimate would produce 11 false positives for every 1,000 tests. This yielded an estimated prevalence of 0.28% 鈥� or eight 鈥渢rue鈥� positives from the 2,991 people sampled.

The researchers then extrapolated this estimate, including its uncertainty parameters, to the Australian population as a whole. They ultimately concluded the number of Australians with SARS-CoV-2 antibodies 鈥� and who have therefore presumably been infected with COVID-19 鈥� is somewhere between zero and 181,050, and most likely about 71,000.

This begs two main questions: should this alter our view on how best to contain the spread of COVID-19, and are there any limitations to the study that we should be aware of?

Let鈥檚 begin with the latter question. Here are four key things to consider when interpreting the results.

1. False positives

In countries with very low COVID-19 rates, such as Australia, the key requirement of an antibody test is to be highly specific 鈥� that is, to avoid false positives. This is even more important than being highly sensitive (avoiding false negatives).

The antibody test used in the new study reportedly has a specificity of 98.9%, and a sensitivity of 100%. This means, for every 1,000 tests, we can expect 11 false positives and no false negatives.

Imagine a place with high prevalence of the virus, such as New York City, where of people are estimated to have had COVID-19. A sample of 1,000 would, on average, contain 200 COVID-19 positive people, of whom the test would correctly identify all 200, with no false negatives. It would also find 11 people positive who were actually negative, giving an estimated prevalence of 211 out of 1,000, or 21.1% 鈥� which is close to the true figure.

Now imagine a sample of 1,000 Australians, with a COVID-19 prevalence of, say, 0.2%.

Just two people in this sample would correctly test positive, but again we would expect the test to deliver 11 false positives. This gives an estimated prevalence of 13 out of 1,000, or 1.3%, which is several times higher than the true figure.

Even if you revise your estimate to account for the expected false positives, as the authors did, we can see how hard it is to estimate low prevalences accurately. The small number of real cases is liable to be lost in the noise.

2. Sample size

A larger sample size could provide improved precision. The small sample size is why the study鈥檚 estimated range is so wide. In fact, it stretches all the way down to zero, even though we know there can鈥檛 possibly have been zero COVID-19 cases in Australia. But no matter the size of the study group, the false positive problem never really goes away as long as the prevalence is low.

3. Testing method

One solution would have been to retest the samples with currently available commercial with specificities of 99.9%. This would have offered a way to overcome the problem with false positives.

The suspicion that the sample included a substantial proportion of false positives is supported by the fact only one COVID-19 positive patient had contact with a known COVID-19 case, and none of those who tested positive had reported any COVID-19-like illness.

4. Extrapolation

There are also questions over how reliably the results can be generalised to the entire Australian public. The study involved people undergoing elective surgery, who may have had different risks of exposure to the virus.

It is hard to say from the available data whether any adjustment was made for variables such as age, sex and state of residence when extrapolating to the wider Australian population.

So what can we say for sure?

What can we determine from this study about the number of people exposed to COVID-19 in Australia? Unfortunately, without a much larger sample, wider sampling of the population, and a more reliable test, we know little more about the prevalence of exposure to COVID-19 than we already did.

This means it would be unwise to use these new findings to claim COVID-19 is any less dangerous or deadly than we thought.

Rather than take these estimates at face value, what we really need is more comprehensive testing of the prevalence of SARS-CoV-2 antibodies, including studies that track this prevalence over time.

In any case, Australia can consider itself fortunate to have low enough case numbers that the issues of false positives becomes a major caveat in interpreting studies such as this. Sadly, in many other places, false positives are buried in a landslide of genuine COVID-19 cases.

Editor鈥檚 note: Ian Cockburn, one of the study鈥檚 lead authors, told The Conversation false positives are indeed more likely to be a significant factor when trying to estimate low prevalence rates, but described the study as a 鈥渂est estimate鈥� based on two separate statistical analyses, which both arrived at the same result. He added the research team plans to use further statistical methods to check the study鈥檚 results before it is accepted for full publication.

He said the study sample 鈥渋s not a perfect cross-section鈥� of the population, and the ideal study size would be 6,000-10,000 people, but obtaining blood samples from the general population poses significant logistical and cost obstacles.

He added it can be difficult to verify commercial companies鈥� claims to have antibody tests with higher specificities, and that patients who register a false positive may also test false positive with another test if it works in the same way.

, Professor of Biostatistics, and , Scientia Professor and Head, Viral Hepatitis Clinical Research Program, Kirby Institute,

This article is republished from under a Creative Commons license. Read the .

From 11.59pm tomorrow (September 16), regional Victoria will take the out of COVID-19 restrictions, Premier Daniel Andrews announced today.

According to the revealed last weekend, the move to step 3 in regional Victoria could happen when the daily average number of cases for the previous 14 days was less than five, and there were zero community cases without a known source for 14 days.

Regional Victoria actually reached the 14-day moving average target on September 10, with , and by today the average had dropped to 3.6. They were just waiting to hit the mystery cases target.

Among the restrictions to be eased under step 3, up to ten people will be allowed to gather outdoors, and hospitality venues will be able to open again for sit-down service. Beauty salons and hairdressers will also reopen, and people who live in regional Victoria will be able to travel to other regional areas in the state.

Heading in the right direction

This easing of restrictions is undoubtedly good news for regional Victoria. But it鈥檚 also reason for people in Melbourne to be optimistic.

Certainly, regional Victoria and metropolitan Melbourne are quite different in terms of the epidemiology of their respective second waves. But the targets set out in the roadmap do appear to be achievable.

The 14-day moving average of daily case numbers for Melbourne is currently at , and needs to drop below 50 to reach the target for step 2.

Based on my modelling of the 14-day moving average, Melbourne could reach this target as early as Thursday.

This model takes the 14-day moving average for the last 30 days, and it assumes the continuing downward trend is exponentially decreasing. That is, it plots a slow downward curve that approaches zero.

The target to move to step 3 is a 14-day moving average of fewer than five cases per day statewide. A similar modelling strategy shows this is likely to occur on about October 22.

Save the date (or don鈥檛)

Notably, there was no date set for the move to step 3 in regional Victoria, which is very different to the set out for metropolitan Melbourne.

In Melbourne, the second step is not due to occur until September 28, and the third step not until October 26 鈥� and only then if case numbers have dropped below designated thresholds.

The key question for the Victorian government is whether to stick to this time frame, or allow for an earlier move to the second and third steps if Melbourne achieves the moving average targets ahead of time.

Read more:

In a recent on The Conversation, I called for a more nuanced approach to lifting restrictions. In other words, the government shouldn鈥檛 be too rigid with the roadmap.

The people of Victoria have been asked to make enormous sacrifices to get the outbreak under control, and it鈥檚 working. I believe the Victorian government should be willing to reward Melburnians by moving to the third step when they reach the relevant case threshold, regardless of the date.

What now?

It鈥檚 essential that over the next few weeks, Melbourne residents continue to stick to the restrictions, to help hit the targets as soon as possible.

The numbers of COVID-19 tests has dropped off in the past few days, averaging about tests a day. With the lower number of cases and fewer people with respiratory symptoms now winter is over, this is not really surprising. But it鈥檚 vital high rates of testing continue, with a focus on hotspot areas, such as the .

Read more:

As for regional Victoria, people there must also stick to the remaining restrictions. This will give them the best chance of moving towards step 4 and beyond.

, Professor of Biostatistics and Epidemiology,

This article is republished from under a Creative Commons license. Read the .

Expert epidemiology and biostatistics commentary from and 听颈苍 .

The Victorian government today announced the eagerly anticipated roadmap out of COVID-19 lockdown. It features several steps that reflect a much slower relaxing of restrictions than last time around.

While the government has provided a provisional time frame for the various steps, it is data, not dates, that will determine when restrictions are actually eased.

We applaud this strategy. The virus does not obey a timeline. Rather, we have to beat it down to a level at which easing of restrictions is safer.

What was announced?

Metropolitan Melbourne鈥檚 current stage 4 restrictions will be extended for two weeks, to September 27. But from 11:59pm on September 13, there will be a few key changes.

The nightly curfew will be shortened by one hour, and will be in place from 9pm to 5am. Also, two people or a single household can meet outdoors for two hours maximum, up from the previous one hour, for exercise or recreation.

For people living alone, and single parents with children under 18, there will be a 鈥渟ingle person bubble鈥� policy that allows them to designate one other person who can visit their home.

Regional Victoria is already faring better than Melbourne, and will have a faster timeline.

Premier Daniel Andrews wants to maximise the chance of getting to Christmas in something like stage 1, while minimising the chance of a third wave of infection that sends the state back into lockdown. This means staying in strict restrictions for longer, and easing out more gradually.

How did data influence the decision?

The Victorian government鈥檚 decision was based in part on the output of a model developed by researchers at the University of Melbourne and the University of New England. It simulates population movements in a simplified world, based on parameters that describe the spread of COVID-19 and people鈥檚 interactions with each other.

In the real world, these patterns are highly random. So the researchers ran the model 1,000 times, with thresholds for relaxing (or tightening) restrictions set at an average of 25, 10, and 5 cases per day on a fortnightly basis. The model could then report the probability, under a given set of policy settings, of having to lock Victoria down again before Christmas.

Opening up too soon is likely to cause a third wave. In simulations in which restrictions were eased once average daily cases dipped below 25 per day, there was a 62% likelihood of new lockdowns. But with restrictions retained until daily cases dropped below 5 daily cases, the lockdown likelihood was just 3%.

Viewed in that light, it is easy to see why the Andrews government opted to set strict criteria for lifting restrictions, knowing that short-term pain is better than the economic ravages of another lockdown in the long term.

What might hold Victoria back?

First, there鈥檚 the elephant in the room 鈥� the quality of Victorian contact tracing (especially in comparison to New South Wales). Living with the virus requires high-quality contact tracing. There鈥檚 no doubt contact tracing in Victoria has improved since June when our second wave started. There is therefore a real possibility that we may get the case numbers down faster, and hold off resurgences of case numbers more effectively or for longer than the modelling suggests.

Second, infection disease control in health care and aged care has not been up to scratch in Victoria (compared with, dare we say it again, New South Wales). These represent particularly dangerous settings. Older adults are much more likely to become severely ill with COVID-19, whereas health-care workers who become infected with the coronavirus risk infecting the most vulnerable and reduce health capacity when it is most needed.

And of course, health and aged care workers live in the community too, and if community restrictions are relaxed the virus will leak back out through family members and surge again. The Victorian government decided to deal with both community and health-care transmission simultaneously. We think that it is the right approach.

Is elimination still possible?

There were strong arguments for an explicit , requiring 鈥済oing hard鈥� for a six-week lockdown. Unfortunately, Victoria didn鈥檛 go hard early enough. The government waited three weeks, numbers got out of control, and then we went into stage 4. With the benefit of hindsight, it was a huge missed opportunity.

The Grattan Institute is also for an explicit elimination strategy and much longer hard lockdowns. It argues this will result in better economic outcomes in the long run. An by Australian National University researchers also supports the theory that elimination is better for both health and the economy in the long run (although this paper has not yet been peer-reviewed).

However, things have changed in the past two months. First, we are now closer to a vaccine, so in theory the long-term payoff for short-term pain will arrive sooner. Second, New Zealand (and Queensland) have taught us that elimination can be lost. Third, NSW has taught us you can live with the virus at low levels (so far). Fourth, the imminent border openings and hotspot strategy are not really consistent with the hard border controls needed to defend elimination in places that achieve it.

Andrews aptly termed the state鈥檚 strategy 鈥渁ggressive suppression鈥�. It may even achieve elimination, as the first wave effort so nearly did. We hope it does 鈥� but do not bank on it.

It鈥檚 in our hands now, both the government and citizens. With some good luck 鈥� and few would begrudge Victorians a little of that 鈥� the roadmap will pan out as planned.

, Research Fellow, Population Interventions Unit, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health, and , Research Fellow, Population Interventions Unit, Centre for Epidemiology and Biostatistics, Melbourne School of Population and Global Health,

This article is republished from under a Creative Commons license. Read the .

Security guards at have been widely blamed for Victoria鈥檚 current outbreak of COVID-19.

Reports have suggested they with people under quarantine, and did not properly follow instructions around infection control.

But late yesterday we heard the first positive case was in fact a at Rydges on Swanston, one of the hotels at the centre of the quarantine bungle. We don鈥檛 know how this person became infected, but there鈥檚 no suggestion it was a result of any improper behaviour.

This night manager has now become known as 鈥減atient zero鈥� in Victoria鈥檚 second wave of coronavirus infections. But what does this term actually mean?

The beginning of the chain of infections

The first case in a chain of infections is popularly called 鈥減atient zero鈥�. However, 鈥減atient zero鈥� is not a very .

In , we call the first case in an outbreak to come to the attention of investigators the 鈥渋ndex case鈥�. The actual individual who introduced the disease at the start of the outbreak is called the 鈥減rimary case鈥�.

According to these definitions, because the night manager was the first person recorded as being infected at the hotel (apart from the guests, who of course were already under quarantine), he or she would be the index case. However, the night manager was also the person who started the chain of infections, so he or she was also the primary case.

The one thing the night manager is not, however, is 鈥減atient zero鈥�. That expression should really be reserved for the first human ever to be infected with SARS-CoV-2 (the coronavirus that causes COVID-19).

Origins of patient zero

The expression 鈥溾�� originated from the HIV epidemic in the United States.

Reports emerged in early 1982 of sexual links between several gay men with AIDS in Los Angeles, and investigators from the Centers for Disease Control and Prevention (CDC) interviewed these men for the names of their sexual contacts.

The CDC gave each of the cases pseudonyms, and the person they eventually identified as the first to have the disease had a moniker beginning with the letter 鈥淥鈥�.

This was later mistakenly interpreted as a zero, and so we got the expression 鈥減atient zero鈥� for the first known case of a disease.

Why finding patient zero is important

It鈥檚 important for epidemiologists to find the first known case because it helps work out how the outbreak occurred, and gives us an idea of how to prevent further outbreaks in the future.

For example, scientists believe the COVID-19 pandemic started in the in Wuhan, China, in December 2019. If this proves to be correct (an international is underway to determine this), authorities may choose to close wet markets, or at least better regulate them to prevent future outbreaks.

Beyond 鈥減atient zero鈥� in the sense of the first ever case of a disease, it鈥檚 also important to find the first case in each particular outbreak.

In the case of the Rydges hotel night manager, this person would clearly have been infected by one of the hotel鈥檚 quarantined guests. Authorities now need to determine exactly how, where and when this person became infected, so they can tighten procedures to make sure this doesn鈥檛 happen again.

New Zealand is in a similar situation with its current COVID-19 . Until health authorities can work out who the primary case is, it will be very difficult to determine where the infection came from, and what actions they must take to ensure it鈥檚 not repeated.

Potentially, the primary case in this outbreak could have picked it up from a contaminated surface, a breakdown in quarantine regulations, or simply an asymptomatic person moving around in the community.

A political blame game

Unfortunately, finding out how Victoria鈥檚 second-wave outbreak started seems to have become a rather than a serious attempt to prevent it happening again.

The current finger-pointing is not only counterproductive 鈥� it could easily see the night manager designated as patient zero unfairly stigmatised, when that person is most likely blameless.

Richard McKay, a Cambridge academic who has written extensively on the concept of patient zero, captured the issue perfectly in an earlier :

Writing of a patient zero is a damaging red herring that distracts from constructive efforts to contain the epidemic. Let鈥檚 wash our hands of this toxic phrase.

, Professor of Biostatistics,

This article is republished from under a Creative Commons license. Read the .

It鈥檚 hard to recall a time when we didn鈥檛 nervously await the announcement of Victoria鈥檚 daily COVID-19 case numbers each morning.

It was certainly disconcerting when the state recorded more than 700 new cases on two occasions not long ago.

And likewise, now that we鈥檝e seen a few consecutive days of around 300-400 cases, it鈥檚 tempting to ask whether the peak of Victoria鈥檚 second wave is behind us.

The good news is, current daily case numbers do indicate we鈥檝e passed the peak of the second wave. But I would suggest we actually reached the peak at the end of July, and the reported case numbers are only now catching up.

Daily case numbers versus new infections

Before we can answer the question of whether Victoria has passed the peak of its second wave, we need to be clear about what we鈥檙e asking. Although it鈥檚 natural to focus on the reported case numbers because they鈥檙e highly visible, the outbreak鈥檚 progress is in fact driven by the number of new infections.

COVID-19 infections can take to be diagnosed and appear in the official case count. This is because an infected person must first pass through an incubation period (the time between becoming infected and symptoms presenting), and then be tested and wait for their result.

On average, the process takes , but it can vary substantially from person to person.

So today鈥檚 case numbers 鈥� rather than indicating new infections 鈥� actually reflect infections that occurred up to two weeks ago.

In other words, watching the case numbers doesn鈥檛 tell us the full story about the current spread of the virus. When asking whether we鈥檝e passed the peak, we really need to focus on the peak in daily infections.

That鈥檚 where data analytics come in

We don鈥檛 know how many new infections occur each day because infections remain hidden until symptoms develop or there鈥檚 some other reason for a person to get tested.

But we do have a good idea of how long it takes for someone to progress from infection to symptoms, and then from symptoms to diagnosis by a positive test.

By combining the observed case numbers with a mathematical model for the progress from infection to diagnosis, we can then reconstruct the pattern of past infections that would have led to the case numbers. This is an epidemiological analysis method called back-projection.

This analysis is an estimate, not an observation. But we can use it to explore whether there鈥檚 any evidence infection numbers have peaked, and at what point.

Looking back to the first wave

Earlier in the outbreak I used this approach to evaluate the effectiveness of the government鈥檚 control measures. In a study looking at across Australia, I showed that the timing of government restrictions matched almost exactly with the flattening and downturn of infection numbers.

This was despite the fact case numbers continued to rise after restrictions were introduced. In other words, the case numbers were hiding the good progress that was going on in the background.

Read more:

By clamping down early, we probably avoided tens of thousands of infections nationally. A recent study published in the estimated Victoria鈥檚 control measures averted between 9,000 and 37,000 cases in July.

Returning to Victoria鈥檚 second wave

We can use the same data analytics approach to explore the progress of the recent restrictions in controlling Victoria鈥檚 second wave.

My reconstruction of Victoria鈥檚 infection numbers during the second wave, shown below, illustrates an early rise in infections during June. This rise likely accelerated in the first half of July, when new infections would have been increasing at a substantially greater rate than was evident in the daily case numbers.

This lag in the case numbers makes it plausible the recent flattening of daily cases is being driven by a much more pronounced decrease in the underlying infection numbers. This is what the reconstructed infection numbers are suggesting in the graph, which shows a peak in late July.

Again, this is an estimate rather than an observation, and the very recent infection numbers have considerable uncertainty. (This is because we work backwards with this analysis, and very few of the most recent infections will have shown up yet in the case numbers.)

Room for optimism, but not complacency

The lower case numbers in recent days suggest we鈥檝e reached and passed the peak of Victoria鈥檚 second wave, and my analysis strengthens and supports this. It shows a peak and decline in new infections over the last couple of weeks.

If this is true there鈥檚 good reason to be optimistic the tough restrictions will drive the infection curve, and subsequently the case numbers, down even further.

But it鈥檚 sobering that my same analyses estimate Victoria has had about 2,000 more infections than case diagnoses. That鈥檚 an estimated 2,000 people who are infected but don鈥檛 yet know it.

So even if new infections have peaked, as we all hope, there鈥檚 plenty of potential for the curve to turn back up again if adherence to the restrictions wavers. Victorians have some reason to be optimistic that the peak has passed, but there鈥檚 no room for complacency.

, Professor of Biostatistics, NHMRC Clinical Trials Centre,

This article is republished from under a Creative Commons license. Read the .