Anne Barker and Hellena Souisa – Indonesia now has a 9.3 per cent death rate from coronavirus – the highest in the world – after the death toll rose by a quarter at the weekend, to 48 out of a total of 514 confirmed cases.
At first glance it is a stark figure, but the death rate is a crude percentage based on the number of known cases, as reported by the Indonesian Government.
Experts say the virus as a whole is widely expected to have a death rate of around 1 per cent and the explanation for the high death rate in Indonesia is that there are many more cases than have been detected, because they haven't been tested.
In fact, the rate of testing for COVID-19 – among the lowest in the world – means only a few thousand tests have been done for an entire population of 270 million people.
It is believed that once more testing kits become available – 150,000 arrived this morning from China – and the number tested for COVID-19 goes up dramatically across the country, the 9.3 per cent figure will come down.
Infectious diseases modelling from the Centre for Mathematical Modelling of Infectious Diseases (CMMID) in London already suggests the true number of positive cases in Indonesia is in the tens of thousands and could be as high as 250,000 – depending on the rate of transmission.
The vast majority appear to have gone undetected. Even a conservative estimate, according to the modelling, indicates there could already be 70,000 cases of coronavirus in the country.
CMMID, which has developed a tool that roughly predicts the likely number of cases in a country based on the known number of deaths, found in a study published this month that a single confirmed death in a country like Indonesia typically signals a true caseload in the tens of thousands.
"By the time a COVID-19 death is reported in a newly affected country, it is likely that there are already hundreds or thousands of cases in the population," the study said.
"This means containment through contact tracing will likely be very challenging, and alternative control/mitigation strategies should probably be considered."
Indonesia only confirmed its first two cases of coronavirus on March 2, after weeks of suspicion that the virus was already spreading rapidly and undetected across the archipelago.
But late last week only 1,727 tests had been done, representing one test for every 156,000 people. That's equivalent to about 160 tests in total for the entire Australian population.
The lack of testing means the COVID-19 virus has likely been spreading unchecked for months, during which time health authorities continually dismissed suggestions the virus was in Indonesia.
Health Minister Terawan Agus Putranto had attributed Indonesia's "luck" to prayer.
The CMMID modelling indicates there are likely 70,000 undetected cases already in Indonesia, and potentially as many as 250,000, depending on two key variables – the death rate and the rate of transmission, which measures how many people one person is likely to infect.
Recent studies suggest the transmission rate is somewhere between two and three.
The World Health Organisation (WHO) puts the death rate above 3 per cent, though many epidemiologists and virologists believe it could be below 1 per cent.
Taking, for example, a death rate of 1 per cent (one death per 100 cases) and a transmission rate of two (where one person infects two other people), the number of undetected cases in Indonesia today would be 70,848. And one death would indicate 1,476 cases in the community.
A transmission rate of three would translate to 251,424 cases, where one death would indicate there are 5,238 cases in the community. The true rate is likely somewhere between.
CMMID associate professor Stefan Flasche said that in the early stages the number of new coronavirus cases doubled roughly every seven days.
"One would expect that the roughly six reported deaths per day you see currently [in Indonesia] would rise to 12 deaths per day next week, and 24 per day the week after," he said.
"[That's] unless there are major efforts put in place that aim to curb the spread via for example, social distancing."
But Mr Flasche said it was difficult to predict how the spread of coronavirus would escalate in Indonesia in the coming weeks, without knowing what measures were put in place to contain the virus.
He also said, over time, immunity would slowly build up in the population and reduce the rate of spread.
When asked about the possibility of 1 million cases in Indonesia by the end of April, he said given the size of Indonesia's population it "may make a semi-reasonable, worst-case scenario".
And in a country of 270 million people – the fourth biggest nation in the world – even that number is unlikely to be anywhere near the peak.
As dire as the predictions are, other scientists support the broad numbers.
Professor Niall Ferguson from Imperial College in London agrees that one death corresponds to at least 1,000 cases in the community, assuming a death rate of about 1 per cent.
But he believes the spread of coronavirus may be doubling even more quickly.
"We think epidemics in the absence of control measures... probably double every five days or so. And only about one in 100 people infected die," Professor Ferguson said.
"The deaths we see today take 20 days or so from when you develop symptoms to when they die. So the deaths we see today correspond to the epidemic 20 days ago.
"And that gives you... the epidemic must have been about 10 times smaller then. You multiply that by a factor of 100 in cases to death. And you get about a 1,000-fold multiplier."
In fact, a lower death rate – of say 0.5, which some scientists support – would indicate an even higher number of cases in the community, because statistically it would mean one death per 200 cases instead of one in every 100.
Indonesia's health ministry in Jakarta last week forecast a possible 700,000 cases but put no timing on when that may occur.
Meanwhile, infectious disease modellers in Bandung have predicted "the epidemic peak will happen at the end of March and eventually end in mid-April".
In that mathematical model, which proved successful in predicting the beginning, end, and peak of the endemic SARS disease in Hong Kong in 2003, it was found that the South Korean Richard's Curve was the most suitable to be juxtaposed with COVID-19 reported case data in Indonesia.
"The projected number of cases will continue to increase to a maximum of around 8,000 cases, and an average daily increase of 600, until the end of March," the report said, citing researchers at the Centre for Mathematical Modelling and Simulation at the Bandung Institute of Technology.
The researchers compared Indonesia's rapidly rising curve of cases to those of other countries, and found it was similar to that in South Korea, which the report say has been "quite successful in preventing the NCOVID-19 pandemic from spreading".
"When based on data until March 14 through this modelling, the COVID-19 trend in Indonesia is similar to South Korea, [but] after the number of cases , the trend has become more similar to like in the United States," researcher at the Bandung Institute of Technology Centre of Mathematical Modelling and Simulation, Dr Nuning Nuraini, told the ABC.
Korea has also conducted mass-tests on its citizens, with a daily average test of 5,000 people out of every 1 million people.
However, the researchers acknowledged that the model they used was imperfect, and the actual number of cases could be far higher than was being reported.
Dr Nuraini pointed out that because their modelling is based on figures from March 14, the data was "still dynamic".
"[As a result], the shift in peak time and projected data accumulation can also change, especially after the number of patients becomes 369 cases."
Put simply, he said this could mean the peak would be longer than the initial projection.
"It can be imagined if these preventive measures are not taken seriously, then cases can multiply in tens, hundreds, thousands and even millions of sufferers," the researchers said.
Indonesia has so far taken limited steps to prevent the epidemic from worsening, having banned incoming travellers from the worst-hit nations, including Italy, China, Iran and South Korea.
But President Joko Widodo has ruled out the possibility of a lockdown in Jakarta, where the majority of cases and deaths are.
Nor is there any mandatory quarantine for travellers when they arrive, as there is in Singapore and now Australia.
There is no formal curfew in Jakarta, although national police have ordered the closure of mass religious events, as well as sports and entertainment activities.
Governor of Jakarta Anies Basweden has declared a state of emergency in the capital, and urged offices to shut, people to work or study from home, and a reduction in public transport.
However, even today, peak hour trains were packed with commuters and crowds still attended Friday's weekly prayers at mosques around the country.
The Government has designated more than 130 hospitals to treat coronavirus patients, while another 161 military and police hospitals have also been added.
And from today, the athletes' village from the 2018 Asian Games has been converted into an emergency hospital for at least 4,000 patients.
Mr Widodo has also ordered several million doses of the antimalarial drug chloroquine and the anti-viral Avigan, despite little or no evidence that they would work against coronavirus.
Given the lag between initial infection and death, it could be weeks before it becomes clear whether any of these measures will contain the spread of the virus.
In reality, Indonesia is likely to face a tsunami of cases and a health system that is very quickly overwhelmed.