If New Zealand hadn't bothered to lock down when COVID-19 arrived on our doorstep, it likely would have infected a third of us and killed more than 30,000 by the end of 2020, new modelling suggests.
Even just waiting a few weeks later to go to alert level 4 would have left hundreds dead, researchers at Te Pūnaha Matatini found, and made elimination almost impossible.
In contrast, shutting the border slightly earlier or later would have made little material difference, the modelling found - and locking down earlier might paradoxically have made elimination less likely.
In reality, New Zealand shut the borders on March 19, 2020, and went to alert level 3 on March 23 then level 4 on March 26. By May, we were out of lockdown and by early June, officially COVID-free. The death toll at the time was just 22.
The new modelling used real-life infection spread data and arrival dates of people infected with the virus to determine what effects changing the dates the Government took action might have had on the overall course of the pandemic.
Throughout March, as the outbreak worsened overseas, there were growing calls to take the unprecedented step to close New Zealand's border. From March 15, new arrivals were required to self-isolate a few days ahead of the border closure.
The modelling found bringing either of these measures ahead a few days had little effect on the outcome of the COVID-19 response.
That's because out of the 563 international cases that arrived before the start of managed isolation and quarantine (MIQ) and could have contributed to local transmission, just 14 percent of them - 78 - arrived before self-isolation was required of everyone entering the country. And of that 78, most - 52 - reported isolating immediately on arrival anyway.
Delaying the introduction of border measures by five days also had little effect, except to slightly reduce the likelihood of elimination.
In contrast, the alert level restrictions had a massive impact - so much so, that delaying level 4 by just five days would likely have resulted in hundreds more cases and 17 extra deaths.
Waiting 20 days longer would have ended up with a peak of about 500 cases a day - overwhelming the limited contact tracing systems at the time - and resulting in 11,500 confirmed cases and 200 deaths in seven weeks, at which point in reality we'd exited level 3. In this scenario, the chance of elimination was reduced by almost 90 percent, the researchers saying we'd probably have had to lock down a lot longer to get there.
But what if the Government moved us to alert level 4 five days earlier? Experience has shown going hard and going early is a recipe for success, but the Te Pūnaha Matatini team found an anomaly - while a five-day headstart would have likely reduced deaths by the end of level 3 to just 14, the modelling suggests we would have been slightly less likely to have achieved elimination.
"This counterintuitive result is due to the presence of an international case in the data that had an arrival date prior to the start of alert level 4 but a much later symptom onset date near the end of alert level 4," they said.
In reality, this person's most infectious period occurred under alert level 4 restrictions. But if the move to alert level 3 had come five days earlier, they would have been most infectious under alert level 3 where the modelling assumes a higher risk of transmission, "so this individual infects more people, on average".
"If this international case outlier is excluded from the data, the model predicts a very similar probability of elimination [to reality]."
While moving five days earlier would have saved lives under the model, the researchers say it would have been impractical in reality due to the "rapid escalation of the COVID-19 situation in mid-March", and "would have allowed less time to prepare for ongoing provision of essential services under alert level 4".
No alert levels?
What if the Government just decided to rely on the border, and not bother with alert level restrictions at all? The result shouldn't come as a surprise, after seeing what happened in countries that moved late like the UK and US - a "large uncontrolled outbreak" with "disastrous outcomes".
In this scenario, by April 27 - when in reality New Zealand moved to level 3 - we would have been recording more than 1100 cases a day, peaking at more than 47,000 by mid-June.
"By the end of the outbreak, around October 2020 on average, there could have been over 1.81 million reported cases in total and 31,905 deaths."
Unsurprisingly, the simulations showed there was a 0 percent chance of having eliminated the virus by June 18, compared to a 66 percent chance the model predicted for our actual course of action.
"This is an important result demonstrating that border measures alone would have been insufficient to prevent a serious outbreak from occurring and that stringent alert level 3/4 restrictions were necessary to have a chance of eliminating community transmission."
The outbreak last year involved the original strain of the SARS-CoV-2 virus, which causes COVID-19. It wasn't as infectious - the researchers calculating an R number of 1.8 before alert level 4 was implemented, 0.35 after (R being the number of people an infected person is likely to go on to infect), and R values of 0.95, 1.7 and 2.4 for alert levels 3, 2 and 1 respectively (estimated because actual case numbers were too low to calculate it).
The current outbreak - which the Government no longer has plans to try and eliminate - is much more infectious, with an R number above five in the absence of public health measures such as lockdowns and vaccines.
The research was published in journal Royal Society Open Science on Wednesday.