Whether the world is in a second wave of the pandemic or still in the grip of the first is up for debate. But what is for sure is something remarkable has changed with more recent Covid-19 infections: fewer people appear to be dying.
The question is why, and if it will stay that way.
In April and May, Covid-19 was claiming around 300 deaths per day in the US, as many as 8 per cent of all those know to be infected. In early July, that was down to about 5 per cent.
It's not just a trend in the US. A study in Milan found a "dramatic drop" in mortality from 24 per cent of those hospitalised in March to just 2 per cent in May. In England, the death rate of people with Covid-19 in hospital has gone from 6 per cent to 1.5 per cent.
It sounds like good news. Certainly, scores of people who would have died if they had caught coronavirus earlier in the pandemic are still alive because they became infected more recently.
But an Australian epidemiologist has backed up warnings from top health officials that we shouldn't be too dazzled by the numbers. Total deaths are almost certainly about to rise, including here in Australia, and it's even possible we could be around the corner from a surge in new victims.
WHY FEWER PEOPLE DYING FROM COVID-19
On Thursday, the US reported its single biggest surge in new coronavirus cases with 65,000 infections. Yet, deaths are down 75 per cent since their April peak.
US President Donald Trump has ignored the soaring number of infections but revelled in the fewer number of deaths.
"99 per cent of [Covid-19 cases] are totally harmless," he wrote in a recent tweet, despite the fact the case fatality rate – the percentage of deaths to confirmed infections – in the US is currently about 6 per cent.
Nonetheless, deaths per known infections have indeed reduced. These are some of theories as to why that's happening.
One difference is clinical care. Healthcare workers have now had months to see what works best in tackling the disease.
"Before it was like we were stumbling in the dark. It feels a little better now," Dr Taison Bell, an infectious disease expert at the University of Virginia told The New York Times.
A really simple technique for severely ill patients, for instance, has been to put them in the prone position. This means laying them on their stomach which can improve lung function. In some cases, blood thinners have been of benefit while the steroid dexamethasone may help to calm out of control immune systems.
More vulnerable people are now better protected. Also, fewer people are catching Covid-19 in hospitals. As these people were ill already, they had a greater likelihood of dying.
VIRUS IS JUST AS VIRULENT
A general dip in cases in many countries has also meant hospitals aren't overrun, so they are better able to care for those who are sick.
However, Professor Peter Collignon, an infectious diseases expert at the Australian National University Medical School, said any hope the virus may have morphed into something less deadly was likely to be in vain.
"It looks like the death rate is falling but that's because we're doing more testing," he told news.com.au.
"It would be nice if Covid-19 was less virulent, but it's just that we've missed other cases."
In the early pandemic, when test kits were in short supply, only those who were already showing symptoms, and were therefore more ill, were tested. That meant there were more positive results and the case fatality rate (CFR) looked greater.
Now, people who feel well can get swabbed which is turning up lots of asymptomatic cases that can be treated early or shaken off with more ease, lowering the CFR.
Data from the US Covid Tracking Project (below) has backed up this trend. As testing has gone up massively in the US, positive results have gone down. That's even though a greater number of people overall are becoming infected.
Collignon has looked at mortality rates around the world particularly in South Korea and Australia, two countries that have handled the pandemic admirably so far, where 2.4 per cent of and 1.4 per cent of patients succumbed respectively. Both CFRs are relatively low – far below the US, for instance.
But it's Singapore, where case fatality is at just 0.1 per cent according to data from Johns Hopkins University, that may provide a bigger clue.
"In Singapore, the mortality rate looks really low because it's mostly in migrant workers who are young and healthy," he said.
In Florida, the average age of someone diagnosed with Covid-19 is now just 35 compared to 65 in March. As the age has come down, so has the mortality.
WHERE THE DEATH RATE WILL STABILISE
During most pandemics, the actual death rate sinks over time.
"The  swine flu pandemic had an original mortality rate of about 6 or 8 per cent in Mexico, which is why everyone thought it was so bad," Collignon said.
"Even in Australia the mortality was 3 per cent, but in reality it was more like 0.02 per cent when you looked at all the people that got it, so you have to be really careful when you judge mortality in the early part of illness."
Collignon said he expected the mortality rate in Australia to settle between 0.5 and 1 per cent.
"That's still between 10 and 20 times more lethal than seasonal influenzas."
Although, unlike President Trump's claim, that doesn't mean 99 per cent of people are going to come through unscathed.
"For every 100 people infected about one person will die but 10 will need hospitalisation."
For older people, the chance of death will be far higher. But younger people couldn't be blasé, he said.
"We should do we all we can to stop it spreading because even though 20 and 30-year-olds are less likely to die, if enough of them get it there will be deaths."
Top US infectious disease expert Dr Anthony Fauci is similarly worried.
"It's a false narrative to take comfort in a lower rate of death," he said on Tuesday pointing out that if Covid-19 spreads like wildfire, even a lower mortality rate still means lots of deaths.
"There's so many other things that are very dangerous and bad about this virus. Don't get yourself into false complacency," he told TV channel CNBC.
LAG IN DEATHS
Another fear is that the CFR figures are partly down because deaths haven't caught up to the recent rise in cases.
For victims, the average time from the onset of symptoms to dying was 14 days, according to the Covid Tracking Project.
But some younger patients who are ill may take longer to die, meaning these deaths are yet to come. Broadway actor Nick Cordero, who was just 41, went from diagnosis to death in an agonising 95 days.
In addition, a large uncontrollable spike in infections could overwhelm hospitals, meaning some patients are unable to get the care they need and die needlessly.
On Friday, Victoria's chief health officer Brett Sutton warned there would inevitably be more deaths in Melbourne.
"We will see an increase in hospitalisations and ICU cases and in deaths in the coming days because of the spike that we have seen in recent days."
As of Friday, there were 932 active cases in Victoria which now is Australia's worst-hit state. If the mortality rate is indeed 1 per cent, at least nine of these people could die. All eyes are on whether that will happen.
The more cases you have, the more overall deaths you get – that's the simple equation that isn't up for debate.