How do you project something as large and variable as an Omicron outbreak? Jamie Morton talks to two leading Covid-19 modellers.
Another variant, another looming outbreak, another set of big, ugly numbers.
New modelling released today warned of nearly 2000 Omicron cases a day – that's 10 times the Delta peak – just six weeks into an Auckland outbreak.
Te Pūnaha Matatini's Professor Michael Plank, who has been preparing his own modelling, said the peak of a nationwide wave could certainly number in the thousands – and possibly top 10,000.
But such projections are mired in uncertainty, and not as straight-forward to calculate as with the original strain, or even the much more formidable Delta variant.
"We're still collecting and analysing data about Omicron, so the estimates that we have are much more uncertain," Plank said.
"And as we go further through the pandemic, there are more and more variables that make it hard to disentangle different effects."
Born to spread
The first factor to consider is Omicron itself.
We know that new variants are created as a virus makes "mistakes" - or mutations - while copying its own genome to replicate and spread.
If one of these mutations happens to offer some advantage, like more easily invading cells, then it becomes more likely that it increases in frequency.
In the Sars-CoV-2 virus that causes Covid-19, most of these mutations have tended to occur within the virus' distinctive "spike protein", which it uses to recognise and gain entry to our cells.
Such mutations have been troubling signatures of the variants that have come before – Alpha, Beta, Gamma and Delta among them – and some of these have aided the virus in spreading faster.
In Omicron, however, scientists quickly counted about 50 mutations, and more than 30 within its spike protein alone – about double that of Delta.
Not only did that signal the potential to infect more people, but also to develop more mechanisms with which to escape immunity, given most Covid-19 vaccines work by forming antibodies against the spike protein.
One recent study out of the University of Hong Kong indicated that, at just 24 hours after infection, Omicron replicated within the human bronchus around 70 times higher than the Delta variant and the original strain.
The good news was that Omicron replicated less efficiently – or more than 10 times less - in the human lung tissue than the original virus, which could partly explain why a given Omicron case was likely to be less severe than earlier Covid-19 infections.
In any case, Omicron's main trick was its ability to better escape immunity from vaccines and past infection – pushing the effectiveness of the Pfizer vaccine most of us have received to just 10 per cent, if it's been four months or more since we got jabbed.
In one Danish study in which researchers compared the spread of both Omicron and Delta in members of the same household, Omicron was estimated to be 2.7 to 3.7 times more infectious among vaccinated and boosted people.
The generation game
We've now heard much about Omicron's strikingly narrow incubation period – or the time from exposure to symptoms appearing.
That period has been around five days with early variants and about four days for Delta; in Omicron, it's as short as three.
The faster it takes an infected person to become contagious, the faster an outbreak spreads – something that has sobering implications for the test and trace system that has been a bedrock of New Zealand's response so far.
One study last month described an outbreak involving about 80 people at a restaurant in Norway.
Not only had nearly every infected person been vaccinated, they had also received a negative result from a rapid antigen test just two days before – meaning the virus had multiplied so quickly it had out-paced even the fastest way we have of detecting it.
"The incubation time is really important, but it's one of those things that's hard to observe directly," Plank said.
"What're more important for us, is actually the generation time – or the time between when someone gets infected, and when they spread it to someone else – because that determines how quickly the virus can move through a population."
This month, UK researchers reported Omicron's mean generation time could be just 1.5 to 3.2 days – giving it a 160 to 210 per cent transmission advantage over Delta, and partly explaining why it is rapidly replacing it around the globe.
Based on emerging data like this, Plank's fellow Te Pūnaha Matatini colleague Dr Dion O'Neale said we might infer that case numbers in a given outbreak could double every three days.
"We've seen that play out overseas, whether it's been in the UK, US or Australia," he said.
"Sometimes it's a bit shorter, sometimes it's longer – but most places are observing that doubling time to be in the order of three days."
That speedy growth rate, as seen in exponential outbreak curves in South Africa to Sydney, explains why modellers like Plank are picking an unprecedented wave here.
"If you start off with 10 cases, and you've had three doubling times of three days, you've gone from 10, to 20, to 40, just nine days later – so you don't need many weeks before you get up to 10,000," O'Neale said.
The fact at least two doubling times could elapse while a person developed symptoms, sought a test and got their positive result back illustrated why Omicron stands to quickly overwhelm our contact tracing services.
Assuming stray cases we had detected so far hadn't already started one, O'Neale said it was quite possible the looming outbreak could be discovered after it had had time to spread, as happened in August.
"But with Omicron, it can do a lot more spreading before you know what's happened."
Variants, vaccines and variables
Before Covid-19 arrived in New Zealand nearly two years ago, modellers tried to predict the potential scale of an outbreak using something called the basic reproduction number, or R0.
That measured the number of cases a single infected individual would cause if they were suddenly dropped into a population with no widespread immunity.
During the Delta crisis, we heard about a different value – the effective reproduction number, or Reff – because it measured how many secondary cases we might get in a population with mixed susceptibility.
If the "R" value was high, so was the danger: if it was below one, then the contagion could be snuffed out.
But now, O'Neale felt R didn't really apply either.
"I keep trying to convince people that R is a really dumb measure to use, because it's become a mix of pathogen, and interaction, and observation, all rolled into one," he said.
"The one thing that's both directly observable and relatively consistent around the world has been that doubling time."
O'Neale and colleagues recently constructed one of the most sophisticated models of its kind, which could simulate an outbreak right down to suburb level in real time – and even account for interventions like lockdowns.
This time, modellers were grappling with even more variables: namely the moving feast that was vaccination rates.
"Now, if you look at the variables that affect how many people are getting infected, you've got to look at the different number of doses that people have had," Plank said.
"A lot of people [903,464, or 54 per cent of the eligible population] have had three doses now – whereas with Delta, people had received just one or two doses."
Because booster shots could restore symptomatic protection against Omicron to as high as 70 per cent two weeks after the jab, the severity of a modelled outbreak could change much if we managed to get more time to roll it out.
Another obvious variable was the effect of child vaccinations – a roll-out that's so far covered just one in 10 of 476,294 eligible kids.
"The longer we can keep Omicron, the more people we can get boosted, the more kids we can get vaccinated ... all makes a big difference," Plank said.
While countries like the UK could offer useful data on vaccine effectiveness to plug into models, they had also had almost two years of exposure to circulating virus.
That's why Plank sees Australian states like Queensland and South Australia - both which have been reporting thousands of daily Omicron cases - as closer models for us.
"They have high vaccination rates like us, but they also weren't dealing with big Delta outbreaks like New South Wales and Victoria were, which complicates the picture a little bit."
Flattening the curve
There is also another important variable to consider: our own behaviour.
That includes the curve-flattening measures the Government put in place, but also how we responded ourselves.
"What we often see is that people do change their behaviour when you have big outbreaks, and that does reduce the Reff," Plank said.
"But what we're seeing now, more so than in the past, is that the Reff reduces as the virus runs out of susceptible hosts to infect."
A case in point is New South Wales, whose Omicron wave might now be subsiding, after washing over tens of thousands of people in little over a month.
Europe's main Omicron outbreak could also last a matter of months – the World Health Organisation predicts it could take just weeks for half the population to become infected – although scientists have also warned of re-infection risk.
In the modelling Counties-Manukau DHB put out this morning, a hypothetical surge that kicked off in Auckland on February 1 was projected to peak at about 1500 to 1800 a day by mid-March through April, before dropping away to 150 to 330 a day through May to September.
Over this period, the DHB modelled about 175 to 190 hospitalisations a month due to Omicron in March and April, falling to 45 to 70 throughout May to September.
O'Neale said a further behavioural factor is people returning to school and work.
"When everyone is on holiday in New Zealand, the interaction patterns are hugely different: and the number of people we meet is much smaller," he said.
"In modelling we did for Delta, we found that having kids back at school meant a bump in the number of kids being infected, which in turn brought an even bigger bump in the number of adults being infected.
"We can also see this quite clearly in flu tracking data each year."
Ultimately, Plank and O'Neale are clear on one point: that New Zealand likely can't avoid a large wave, be it bringing daily cases in the hundreds, thousands or tens of thousands.
But both are optimistic the combined impact of health measures and vaccination really does matter, even against a foe as quick and intimidating as Omicron.
"There's a lot of talk that Omicron is so transmissible that there's nothing we can do to stop it and therefore we shouldn't even try ... I don't agree with that at all," Plank said.
"Yes, Omicron spreads very quickly, and we're not going to be able to stop it altogether – but we can nevertheless, what we do can make a big difference to how many people get it."